Guide for Scientists: Public Comment on OMB Proposed Rule for Federal Financial Assistance

Docket OMB-2026-0034 | Comment Deadline: July 13, 2026 | Published: May 30, 2026

Comment portal: https://www.regulations.gov/commenton/OMB-2026-0034-0001

Dear Colleague

On Friday, May 29, OMB published a 412-page proposed regulation in the Federal Register that codifies political appointee control over every federal science grant, government-wide, effective October 1, 2026. Peer review becomes advisory. Publication costs banned. Conferences require pre-approval. Active grants terminable without cause. The comment period closes July 13.

Your comment is not symbolic. It is the legal record courts will use to block the final rule. Silence in the record is evidence of acceptance.

This guide is a companion to my own comment on Docket OMB-2026-0034. The aim is to make it easy for any federally funded scientist to produce their own comment without starting from a blank page. Pick the sections and arguments that match your lab, your funding, and your discipline. Rewrite in your own voice. Volume of independent voices by July 13 is the goal, not uniformity.

1. What this is and why it is different from everything that came before

On May 29, 2026, the Office of Management and Budget published a proposed binding regulation that would rewrite 2 CFR, the government-wide rules governing all federal grants, cooperative agreements, and other forms of federal financial assistance. This is not a memo, not agency guidance, not an executive order. It is a proposed regulation under the Administrative Procedure Act, with a 45-day public comment period and a target effective date of October 1, 2026.

The distinction matters. Executive orders can be reversed by the next president. Agency guidance can be withdrawn. A finalized regulation under 2 CFR carries the force of law, applies to every federal science agency (NIH, NSF, DOE, NASA, DOD, ARPA-H), and can only be undone through a new rulemaking process or congressional action.

2025 was the test run. Every action the administration took against NIH in 2025 (the grant terminations, payline elimination, NOFO collapse, keyword screening, forward funding squeeze, workforce reductions) was implemented through memos, executive orders, and discretion. Each was individually reversible and litigable. Each was a test of whether anyone would push back hard enough to stop the next step.

The courts pushed back in some cases. The First Circuit struck down the indirect cost cap. Congress inserted protective riders in FY2026 appropriations. The system held, partially. The administration learned from those losses. OMB-2026-0034 takes every ad hoc policy that worked and codifies it into binding regulatory text, government-wide, on a single date. This is the deployment.

As Elizabeth Ginexi wrote in her provision-by-provision analysis (https://elizabethginexi.substack.com/p/summary-of-key-changes-in-ombs-proposed): "What OMB is proposing is not a reform of grants management. It is a complete political control apparatus layered over every stage of the federal science funding lifecycle."

2. Why your comment matters, and why silence is dangerous

Under the Administrative Procedure Act, OMB must collect all public comments and respond to substantive comments in the final rule. That response becomes part of the formal administrative record. If OMB finalizes the rule without adequately addressing substantive objections, the administrative record becomes the evidentiary basis for legal challenge under APA Section 706(2)(A) (the "arbitrary and capricious" standard).

This is not theoretical. The First Circuit Court of Appeals struck down the 15% indirect cost cap on January 5, 2026, concluding that NIH violated statutory law and its own regulatory procedures. The administrative record, including public comments, was central to that ruling. State attorneys general used the comment record to build their case. The same legal pathway applies here.

Do not wait for the courts. The indirect cost cap was announced in February 2025 and blocked within days, then struck down (First Circuit, January 5, 2026). It never took effect. Even so, the months of legal uncertainty forced universities to freeze hiring and spending. Labs lost staff. Training programs contracted. Courts can vacate a regulation. They cannot restore grants never funded, papers never published, students never admitted, or scientists who left the country. The comment period exists to prevent the rule from reaching that stage.

Under the APA, a sparse comment docket is evidence of tacit acceptance. An agency can cite low participation to justify finalizing without substantive revision. Every PI, trainee (or unemployed), citizen, taxpayer (you do not need to be in the USA or a citizen) who stays silent makes the legal challenge harder for those who do not.

As one NIH program officer put it: "The comments will be ignored by the administration. But they will not be ignored in the historical record. Silence will be inferred as support."

3. The legal attack surface

The legal vulnerability of this rule is unusually broad for a single rulemaking.

Statutory conflicts (the strongest claims). Section 200.205 directly contradicts 42 U.S.C. 289a, which mandates peer review by qualified experts as the basis for NIH funding decisions. The rule explicitly forbids appointees from deferring to those experts. Section 200.461 conflicts with the 2022 OSTP public access memorandum, which requires federally funded research to be freely available, while simultaneously making publication costs unallowable. These are not ambiguous tensions. They are direct contradictions between a proposed regulation and existing statutory and executive mandates.

Arbitrary and capricious (APA §706(2)(A)). The rule claims three objectives: transparency, regulatory clarity, and reduced burden. Every major provision does the opposite. Eliminating binding peer review reduces transparency. Adding conference pre-approval, E-Verify mandates, political alignment documentation, and publication cost workarounds increases burden. Defining compliance around an undefined "Gold Standard Science" reduces clarity. The First Circuit used exactly this internal-contradiction logic to strike down the indirect cost cap in January 2026. The preamble compounds the problem by citing Heritage Foundation reports and Manhattan Institute op-eds as its evidentiary basis rather than independent administrative or scientific assessments.

Void for vagueness. "Gold Standard Science" is required as a compliance benchmark but never operationally defined. "Anti-American values" is a disqualifying criterion with no measurable content. "National interest" exemptions from public posting requirements give the agency head unconstrained discretion. Each of these fails the basic due process requirement that regulated parties receive adequate notice of what conduct is expected.

Non-delegation and structural concerns. The rule centralizes grant authority in political appointees while stripping civil service protections (via Schedule F) from the career staff who historically exercised that authority. OMB is also claiming direct binding regulatory authority over all agencies through a 2 CFR restructure, raising questions about whether OMB has the statutory authority to override agency-specific enabling statutes (the PHS Act, the NSF Act, the DOE Organization Act) through a uniform grants regulation.

First Amendment concerns. Section 200.450 prohibits "issue advocacy" on federal funds. Section 200.421 restricts public communications. Combined with the content-based criteria in Section 200.205 (blocking awards touching on "the sex binary," immigration, or "anti-American values"), these provisions function as viewpoint-based restrictions on federally funded speech. Viewpoint discrimination triggers strict scrutiny, the highest bar in First Amendment law.

Impoundment Control Act. GAO already found NIH in violation (B-337203, August 2025). The terminate-at-will provision (§200.340) and the conference/publication restrictions create new mechanisms for effectively impounding appropriated funds without formally withholding them. If the rule makes it functionally impossible to spend appropriated money on research activities (publishing, conferences, journals, international collaboration), the effect is impoundment by regulation.

The aggregate problem. Any single provision might survive judicial review in isolation. The cumulative architecture is what makes the rule vulnerable. It simultaneously claims transparency while eliminating it, claims burden reduction while increasing it, overrides multiple statutory mandates, imposes viewpoint-based content restrictions, vests unreviewable discretion in political appointees, and relies on partisan advocacy documents as its evidentiary foundation. A court reviewing the rule holistically can see the pattern that each individual provision obscures.

The comment period is where you build the record on every one of these threads. Each substantive comment that identifies a specific statutory conflict or internal contradiction is a brick in the litigation wall.

4. The 19 provisions that affect your laboratory

The provisions below are organized by how they affect the daily work of a federally funded PI. Each includes the regulatory section number. When you write your comment, begin each point with the section number in brackets (e.g., [200.205]).

How your grants are funded

Political appointee pre-issuance review (Section 200.205). Senior political appointees must conduct a "pre-issuance review" of every discretionary grant before it is awarded. These appointees are explicitly forbidden from deferring to peer reviewers or routinely ratifying their recommendations. Awards must "demonstrably advance the President's policy priorities." The criteria include blocking awards touching on denial of "the sex binary in humans," illegal immigration, or anything deemed to "promote anti-American values."

The transparency framing sounds reasonable until you read that the criteria are whatever the appointee decides they are, with no obligation to follow the scores assigned by the scientists who actually reviewed the proposal.

For your comment: if you have served on an NIH study section, describe what peer review actually involves (conflicts of interest screening, structured scoring criteria, summary statements) and contrast it with what this provision substitutes.

Peer review reduced to advisory status (Section 200.205(d)). The rule states that peer review recommendations "remain advisory and are not ministerially ratified, routinely deferred to, or otherwise treated as de facto binding." This dismantles the post-WWII system in which independent expert peer review was the primary measure of scientific merit at NIH, NSF, DOE, NASA, and nearly every science agency. Before 2025, there were only two political appointees at NIH, and even these (the NIH and National Cancer Institute directors) were accomplished scientists. The proposed rule gives nonexpert appointees veto authority over the judgment of hundreds of study sections comprising thousands of domain experts.

For your comment: name the specific technical expertise required to evaluate proposals in your field. No political appointee possesses it.

"Gold Standard Science" as an undefined compliance test (Section 200.205). All grants must include benchmarks for compliance with "Gold Standard Science" (tied to Executive Order 14303). The standard is never defined in any concrete or measurable way. Agencies must prioritize institutions that have "demonstrated success in implementing Gold Standard Science." Because the standard has no operational definition, the administration retains unguided discretion over what qualifies.

Active grants terminable at any time without cause (Section 200.340). Active grants, including multi-year awards mid-project, can be terminated because a political appointee decides the work no longer aligns with agency priorities. No finding of misconduct, fraud, or noncompliance is required. No appeals process. Between 2012 and January 2025, NIH terminated fewer than six grants midstream in thirteen years. Since February 2025, the department has terminated more than 1,800 NIH grants (GAO, August 2025), with a later count of over 1,392 totaling $539 million in unliquidated obligations (CRS, April 2026). This provision converts that unprecedented pattern into permanent regulatory authority.

This is the provision that makes every other provision enforceable through fear. A PI who publishes findings inconvenient to the administration, collaborates with a foreign colleague, or studies a population the administration prefers not to acknowledge can lose funding mid-project with no recourse.

For your comment: if you hold an active NIH grant, describe what termination would mean concretely (staff laid off, patients withdrawn from studies, multi-year datasets abandoned, institutional debt from startup commitments).

Program design must align with presidential priorities (Section 200.202). Every new federal grant program must be designed with goals that explicitly align with administration policies. Science agencies must structure grant solicitations around the current administration's political agenda rather than scientific need, statutory mandate, or expert consensus.

Agency heads can exempt competitions from public notice (Section 200.204). A federal agency head can approve exemptions from the requirement to post funding opportunities on Grants.gov when "publicly announcing an opportunity would pose a risk to national security or is in the national interest of the United States." The phrase "national interest" has been used expansively by this administration and could justify conducting grant competitions outside public view.

How you do your science

Publication costs presumptively unallowable (Section 200.461). All journal publication costs, including article processing charges, open access fees, and page charges, are unallowable by default. Exceptions require a specific statutory mandate or case-by-case agency pre-approval. This directly conflicts with the 2022 OSTP memorandum requiring federally funded research to be publicly available. Open-access publication fees exist because the government required open access. Banning the mechanism while keeping the mandate is a defunding strategy dressed as fiscal responsibility.

For your comment: state the cost per paper in your field, the number of papers from your current award, and what happens to public access if you cannot pay those fees.

Conference attendance requires express pre-approval (Section 200.432). The current presumption that conference attendance related to grant work is a standard allowable cost is eliminated. Every conference must be pre-approved by the agency and written into the award terms. Conferences not anticipated when the award was issued cannot easily be added. The agency has full discretion to deny approval. The framing is fiscal accountability. The mechanism is information control.

For your comment: describe one collaboration, one corrected experimental direction, or one trainee career outcome that originated at a conference.

Journal subscriptions categorically unallowable (Section 200.454). Subscriptions to professional, academic, and technical journals are made categorically unallowable as direct costs on federal awards. Combined with Sections 200.461 and 200.432, the cumulative effect severs every mechanism by which federally funded scientists disseminate findings, receive critique, and access the work of peers.

Professional memberships require prior approval (Section 200.454). Professional society memberships are only allowable if they are "necessary to fulfill the award requirements" and receive prior written agency approval.

Public communications restricted (Section 200.421). All public relations costs are unallowable except those explicitly required by statute, restricting researchers from communicating findings to the public or press.

"Issue advocacy" prohibited (Section 200.450). Federal grant funds cannot be used for any messaging that promotes or opposes a "particular social, political, or public policy position unrelated to the statutory objectives" of the award. Given that the rule's preamble characterizes climate science, public health research, and equity research as "divisive ideologies," this prohibition could be deployed to bar researchers from speaking publicly about their own federally funded findings.

International collaboration

Broad prohibition on international collaboration (Section 200.220). Federal funds, including indirect costs, cannot support bilateral or multilateral collaboration with "covered foreign countries" or affiliated entities. The rule extends beyond China to all countries on broad sanctions lists, covering travel, research activities, technical assistance, and allocable indirect costs.

American science leads because the best scientists come here and collaborate here. Cutting off collaboration does not protect American science. It hands the competitive advantage to nations that are actively recruiting the collaborators we turn away.

For your comment: if you have international collaborators, co-authors, or trainees, describe what those relationships produce and what losing them would cost U.S. competitiveness in your field.

Domestic-first framework for research awards (Section 200.202(e)). Any international element in a federally funded R&D grant must be affirmatively justified on a case-by-case basis. Foreign entities cannot receive R&D awards at all except with written approval from a senior political appointee.

Your institution

DEI and related topics banned as grant conditions (Section 200.300). All federal award funds are prohibited from being used to "fund, promote, encourage, subsidize, or facilitate" DEI/DEIA policies or practices, "gender ideology," or gender transition assistance for individuals under 19.

E-Verify mandated for all grant recipients (Section 200.303). All recipients and subrecipients must enroll in and use E-Verify for every employee and contractor working on a federal award. Any Final Nonconfirmation must be reported to the federal agency.

Applicants can be denied based on organizational affiliations (Section 200.206). Risk factors agencies may use to deny applications are expanded to include membership in or affiliation with organizations that "advocate for the overthrow of the United States Government" or "undermine public safety or national security."

OMB gains direct binding authority over all agencies (2 CFR restructure). The rule restructures 2 CFR to make OMB guidance a directly binding regulation on all agencies. Individual agencies can no longer shield their communities from any of these changes through their own implementing rules.

OMB gains oversight of which institutions receive grants (new reporting provision). OMB can require agencies to submit reports identifying specific recipients of federal awards over any period, giving the White House direct leverage over which institutions receive funding.

5. The internal contradiction you should name in your comment

The rule names transparency as its first objective while eliminating every mechanism through which transparency operated. Replacing peer review with political appointee override eliminates the primary transparency mechanism. Making publication costs unallowable suppresses the mechanism by which findings enter the public record. Requiring conference pre-approval restricts venues where results are challenged and replicated. Making journal subscriptions unallowable limits access to the literature. Allowing grant termination without cause and without transparent criteria inverts accountability.

The Benton Institute noted that the preamble "cites Heritage Foundation reports, op-eds from the Manhattan Institute and City Journal, and Senate Republican committee reports as supporting authority" rather than independent scientific or administrative assessments.

Naming this contradiction in your comment matters because it creates the basis for an "arbitrary and capricious" challenge under APA Section 706(2)(A). If the final rule's provisions are internally inconsistent with its stated objectives, courts can vacate it.

6. How to write an effective comment

The comment form at regulations.gov allows 5,000 characters of text plus up to 20 file attachments (docx, pdf, pptx, xlsx, 10 MB each). You can submit as an Individual, Organization, or Anonymous.

What makes a legally substantive comment

Not all comments carry equal weight under the APA. OMB must respond to substantive comments but may dismiss form letters or generic expressions of opposition. A substantive comment does one or more of the following:

1. Identifies a specific statutory conflict. The PHS Act (42 U.S.C. 289a) requires that NIH funding decisions be based on peer review by qualified experts. Section 200.205 forbids political appointees from deferring to peer reviewers. The 2022 OSTP memorandum requires public access to federally funded research. Section 200.461 makes publication costs presumptively unallowable.

2. Provides concrete evidence of harm. State your active grant numbers, your IC, the specific provisions that affect your work. Name journals, conferences, costs, collaborators.

3. Demonstrates internal inconsistency. Show how specific provisions reduce the transparency the rule claims to advance.

4. Proposes an alternative. Courts give more weight to comments that suggest a workable alternative rather than simply opposing the rule.

Structure for maximum effect

Begin each point with the relevant section number in brackets, as the Federal Register notice instructs (e.g., "[200.205] The mandatory pre-issuance review by political appointees conflicts with..."). This ensures your comment is cataloged correctly in the docket.

A note on AI tools

Agency staff will almost certainly use AI summarization tools to cluster and characterize submissions. This is not a reason to avoid commenting. It is a reason to be deliberate about structure. If you use ChatGPT, Claude, or any other tool to help draft or refine your comment, ensure the language that matters most (the provision number, the statutory conflict, the concrete harm) appears prominently and repeatedly enough that no summarization algorithm can bury it.

However, if your comment reads like a form letter or copy-paste template, it will be treated as one. Under APA practice, agencies can batch identical or near-identical comments into a single entry. Even two sentences about how a specific provision affects your lab, your trainees, or your patients counts as a unique substantive comment OMB must address individually.

7. Modular draft arguments you can adapt

Pick the provisions that apply to your situation. Replace bracketed placeholders with your own specifics. Rewrite in your own voice.

On political appointee pre-issuance review (Section 200.205)

[200.205] The requirement that senior political appointees conduct pre-issuance review of every discretionary grant, with an explicit prohibition on deferring to peer reviewers, conflicts with the Public Health Service Act (42 U.S.C. 289a), which establishes peer review by qualified experts as the basis for NIH funding decisions. The United States built the world's preeminent biomedical research enterprise on merit-based peer review precisely because non-expert political override produces funding decisions optimized for political palatability rather than scientific impact. As a PI holding [number] active NIH grants in [your field], the effect of this provision is to make every funded project, including ongoing multi-year research, contingent on political alignment rather than scientific merit.

On peer review reduced to advisory status (Section 200.205(d))

[200.205(d)] Reducing peer review to explicitly advisory status, with political appointees forbidden from deferring to or routinely ratifying expert recommendations, reverses the foundational principle that has governed federal science funding since 1945. The proposed rule gives nonexpert appointees veto authority over the judgment of hundreds of study sections comprising thousands of domain experts. My field, [your field], requires [specific technical expertise] to evaluate proposals competently. No political appointee possesses this expertise. The provision will produce funding decisions that are scientifically arbitrary and legally vulnerable.

On publication costs (Section 200.461)

[200.461] Making publication costs presumptively unallowable directly conflicts with the August 2022 OSTP memorandum requiring that federally funded research be made publicly available at no cost to the reader. Article processing charges at journals in my field average [$X] per paper. My laboratory publishes approximately [Y] peer-reviewed papers per year. Under this provision, I would need [$Z] annually from non-federal sources to comply with the public access mandate while being prohibited from using federal funds to do so. The net effect is suppression of the scientific record.

On conference pre-approval (Section 200.432)

[200.432] In the past [time period], I presented federally funded research at [name conferences]. Under this provision, each of those presentations would have required express agency pre-approval written into my award terms at the time of funding. Conferences not anticipated when the award was issued cannot easily be added. Scientific opportunities, including invited talks, workshop participation, and emerging collaborations, cannot be predicted at the time of award. This provision gives political appointees gatekeeper authority over which scientific communities federally funded researchers can participate in.

On journal subscriptions (Section 200.454)

[200.454] Making journal subscriptions categorically unallowable as direct costs eliminates a standard mechanism by which investigators access the scientific literature foundational to their work. My research requires regular access to [name specific journals or fields]. Combined with Sections 200.461 and 200.432, the cumulative effect severs every mechanism by which federally funded scientists disseminate findings, receive critique, and access the work of peers.

On international collaboration (Section 200.220)

[200.220] My research involves active collaboration with [institution/country] on [scientific topic]. This collaboration is essential because [specific scientific rationale]. The blanket prohibition on using federal funds, including indirect costs, for collaboration with covered foreign countries would terminate this collaboration and [specific scientific consequence]. While legitimate national security concerns exist regarding certain foreign entities, this provision is sweeping enough to sever partnerships foundational to U.S. leadership in [your field].

On grant termination without cause (Section 200.340)

[200.340] The authority to terminate active grants mid-award without a finding of misconduct or noncompliance creates an environment in which no PI can responsibly hire trainees, commit to longitudinal studies, or invest in infrastructure. I currently employ [number] trainees and staff on my active grants. Each of them made career decisions based on the reasonable expectation that funded awards would run their course. Between 2012 and January 2025, NIH terminated fewer than six grants midstream in thirteen years. Since February 2025, the department has terminated more than 1,800 NIH grants (GAO, August 2025), with a later count of over 1,392 totaling $539 million in unliquidated obligations (CRS, April 2026). The proposed rule would convert this unprecedented pattern into permanent regulatory authority.

On the training pipeline (applicable to multiple sections)

The combined effect of Sections 200.205, 200.340, 200.432, 200.454, and 200.461 falls disproportionately on trainees and early-career investigators. Graduate students and postdocs cannot attend conferences without pre-approval, cannot publish without non-federal funds, cannot access journals, and cannot trust that the grants supporting their positions will continue. In my laboratory, [specific example]. The 10-to-15-year training cycle from graduate admission to independent investigator means that damage to the pipeline now produces generational, not episodic, losses.

8. Model comment (text box version)

A complete model comment demonstrating how to structure a legally substantive APA comment is available for download alongside this guide. It references specific regulatory sections, identifies statutory conflicts, provides concrete evidence of harm from a working laboratory, and addresses the national security implications of the proposed provisions.

The downloadable version includes a 5,000-character text box version (for pasting directly into the regulations.gov form) and a 5-page PDF attachment expanding the arguments with a China competition analysis, void-for-vagueness challenge to "Gold Standard Science," and documentation of the preamble's reliance on partisan rather than independent sources. Use it as a structural template, not as text to copy. Independent voices in independent language are what build the administrative record.

9. What else you can do (parallel actions)

Submitting a comment is the single highest-leverage action available right now because it creates the legal record. The following actions are complementary.

Contact your House representative and both senators. Ask specifically for peer review protections and grant stability language in FY2027 LHHS appropriations. Markup timelines overlap with this comment period. Congress inserted protective riders in the FY2026 Consolidated Appropriations Act (Section 224, blocking the indirect cost cap; Section 240, capping multi-year obligations at FY2025 levels). The same mechanism can protect peer review. The Senate LHHS Subcommittee is chaired by Sen. Shelley Moore Capito (R-WV), Ranking Member Sen. Tammy Baldwin (D-WI). Arguments framed around medical progress, state economic impact, national security, and congressional intent are the arguments that move this specific committee.

Contact your state attorney general. State AGs have standing to challenge the final rule under the APA and led the successful indirect cost cap litigation. The Massachusetts AG led that challenge. AGs from states with major research universities (Massachusetts, California, New York, Pennsylvania, Maryland, North Carolina, Michigan, Illinois, Wisconsin, Texas) have direct standing because the rule threatens the economic base of their public university systems and academic medical centers.

Coordinate with your scientific society. Institutional comments from SfN, FASEB, BMES, AIMBE, AAU, APLU, AAMC, and COGR carry weight in the administrative record. Contact your society's government affairs office and ask whether they are preparing a comment. Your individual comment still matters because it provides specific evidence of harm that institutional letters cannot.

Forward this guide. The people most affected by these provisions are the least represented in policy feedback.

10. What to avoid

Do not submit identical comments. Form letters and copy-paste submissions are cataloged as a single comment. Rewrite in your own voice, even if you use the modular arguments above as a starting point.

Do not use DEI-flagged vocabulary. Terms such as "diversity," "equity," "inclusion," "underrepresented," "disparities" (in a workforce context), and "structural racism" are liabilities in any document entering the federal record under this administration. Translate equity arguments into capacity, stewardship, workforce sustainability, and merit-based language. The functional outcomes are identical. The risk profile is not.

Do not make the comment purely emotional or political. Courts give weight to comments that identify statutory conflicts, provide concrete evidence of harm, and propose alternatives. "This rule is terrible" is not a substantive comment. "Section 200.205 conflicts with 42 U.S.C. 289a because..." is.

Do not include personally identifiable information you do not want public. Comments and attachments are posted publicly on regulations.gov. Name, institution, and professional details strengthen your comment. Social Security numbers, home addresses, and personal health information should not be included.

Do not wait until July 13. The comment period may feel long. It is not. Coordinating institutional responses, drafting substantive comments, and building the record takes weeks. Start now.

11. A note to graduate students, postdocs, and early-career investigators

Your voice matters here, perhaps more than anyone's. You are the people most affected by the decisions this rule will encode, and you are the least represented in policy feedback. The comment form accepts anonymous submissions. You do not need to be a PI or have funding to comment.

The provisions hitting you hardest are specific and personal. Section 200.432 means you cannot attend a conference on your grant without express pre-approval. Section 200.461 means you cannot publish your work using grant funds. Section 200.454 means you cannot subscribe to the journals you need to read. Section 200.340 means the grant funding your position can be terminated at any time without cause. Say so in your comment. Name the conference you attended last year. Name the paper you published. Name the journal you need.

Public comment has changed NIH policy before. The BRAIN Initiative K99/R00 career transition awards exist in part because trainees and early-career scientists made the case that the pipeline was failing them. When NIH proposed the Grant Support Index in 2017 to redistribute funding, it was abandoned within a month after well-funded senior scientists mobilized while trainees stayed silent. The lesson is straightforward: the voices that show up in the record shape policy. Silence will be inferred as support.

12. Submission checklist

• [ ] Comment submitted at https://www.regulations.gov/commenton/OMB-2026-0034-0001

• [ ] Each point begins with the relevant section number in brackets (e.g., [200.205])

• [ ] At least one specific statutory conflict identified (42 U.S.C. 289a, OSTP 2022 memo, or other)

• [ ] At least one concrete example of harm from your own laboratory or institution

• [ ] Harm to taxpayers and citizens (including giving our technology lead to competition)

• [ ] 5,000-character limit for the text box (attach a PDF or Word document for longer comments)

• [ ] Attachments under 10 MB each, maximum 20 files

• [ ] No personally identifiable information you do not want public

• [ ] Submitted as Individual (recommended) or Organization

• [ ] Deadline: July 13, 2026

13. Sources

Primary documents

• Full proposed rule (Federal Register): https://www.federalregister.gov/documents/2026/05/29/2026-10817/regulation-for-federal-financial-assistance

• Comment submission portal: https://www.regulations.gov/commenton/OMB-2026-0034-0001

Expert analyses

• Elizabeth Ginexi, provision-by-provision analysis: https://elizabethginexi.substack.com/p/summary-of-key-changes-in-ombs-proposed

• Wiley Rein LLP legal analysis: https://www.wiley.law/alert-OMB-Proposes-Sweeping-Overhaul-of-Federal-Assistance-Regulations

• Benton Institute analysis (preamble sourcing from Heritage Foundation/Manhattan Institute): https://www.benton.org/blog/omb-proposes-changes-federal-grant-administration

• Mark Histed, "American Biomedical Science in 2026": https://donmoynihan.substack.com/p/american-biomedical-science-in-2026

• Mark Histed, "The Slow Dismantling of American Science" (Union of Concerned Scientists): https://blog.ucs.org/science-blogger/the-slow-dismantling-of-american-science-and-what-we-can-do-about-it/

News coverage

• Science/AAAS coverage: https://www.science.org/content/article/white-house-seeks-tighten-political-oversight-grantmaking

• Scientific American coverage: https://www.scientificamerican.com/article/white-house-proposes-new-rules-giving-political-appointees-final-say-on-research-grants/

• Inside Higher Ed coverage: https://www.insidehighered.com/news/government/science-research-policy/2026/05/29/omb-proposes-rules-establishing-political

• KFF Health News (Norton and Histed quotes, political appointee analysis): https://kffhealthnews.org/news/article/nih-grants-trump-political-appointees-agenda-alignment-peer-review/

Data and tracking

• Jeremy Berg, Bluesky (NIH funding data and policy analysis, 14,000+ followers): https://bsky.app/profile/jeremymberg.bsky.social

• Jeremy Berg, NIH Dashboard (grant obligation tracking): https://jeremymberg.github.io/jeremyberg.github.io/index.html

• United for Medical Research, 2026 Economic Impact Report: https://unitedformedicalresearch.org/annual-economic-report/

• UMR state toolkits: https://unitedformedicalresearch.org/nih-in-your-state/

Legal and legislative

• GAO decision B-337203 (Impoundment Control Act finding): https://www.gao.gov/products/b-337203

• Congressional Research Service, NIH Grants Policy (IF13131): https://crsreports.congress.gov

Companion guides

• Senate LHHS FY27 Outside Witness Testimony guide: https://bioniclab.org/news/senate-appropriations-open-comments

• NIH Strategic Plan RFI guide (NOT-OD-26-047): https://bioniclab.org/news/nih-rfi

Takashi D. Y. Kozai, Ph.D. Ernest E. Roth Professor of Bioengineering, University of Pittsburgh Director, B.I.O.N.I.C. Lab

Forward this guide to colleagues through LinkedIn, your lab website, email, or Signal. Volume of independent voices by July 13 is the goal.

Dear Colleague

OMB-2026-0034: The Permanent Architecture. 41 Days to Respond.

June 1, 2026

Colleagues,

TL;DR: On Friday, OMB published a 412-page proposed regulation that codifies political appointee control over every federal science grant, government-wide, effective October 1. Peer review becomes advisory. Publication costs banned. Conferences require pre-approval. Active grants terminable without cause. The comment period closes July 13. Your comment is not symbolic. It is the legal record courts will use to block the final rule. Silence in the record is evidence of acceptance.

Comment portal: https://www.regulations.gov/commenton/OMB-2026-0034-0001 Guide for writing an effective comment: https://www.bioniclab.org/news/OMB

2025 was the test run. This is the deployment.

Every action the administration took against NIH in 2025, the grant terminations, payline elimination, NOFO collapse, keyword screening, forward funding squeeze, workforce reductions, was implemented through memos, executive orders, and discretion. Each was individually reversible and litigable. Each was a test of whether anyone would push back hard enough to stop the next step.

The courts pushed back in some cases. The First Circuit struck down the indirect cost cap. Congress inserted protective riders in FY2026 appropriations. The system held, partially.

The administration learned from those losses. OMB-2026-0034 takes every ad hoc policy that worked and codifies it into binding regulatory text under 2 CFR, applicable to NIH, NSF, DOE, NASA, DOD, and ARPA-H simultaneously. A finalized regulation can only be undone through a new rulemaking, congressional action, or litigation. Executive orders can be reversed by the next president. This cannot, not easily, not quickly. And the rule goes further than anything attempted in 2025: it restructures who has the authority to make funding decisions in American science, permanently, by regulation, across every agency.

Do not wait for the courts.

The litigation will come. But the indirect cost cap was announced in February 2025 and blocked within days, then struck down (First Circuit, January 5, 2026). It never took effect. Even so, the months of legal uncertainty forced universities to freeze hiring and spending. Labs lost staff. Training programs contracted. Courts can vacate a regulation. They cannot restore grants never funded, papers never published, students never admitted, or scientists who left the country.

The comment period exists to prevent the rule from reaching that stage. Under the APA, OMB must respond to every substantive comment before finalizing. If the record shows the rule contradicts its own objectives, conflicts with statutes, or lacks rational basis, that record becomes the foundation for an injunction before October 1. Your comment is not a petition. It is a legal exhibit.

What it does

Full analysis by Elizabeth Ginexi (former NIH program officer): https://elizabethginexi.substack.com/p/summary-of-key-changes-in-ombs-proposed

The provisions that matter most:

• Section 200.205: Political appointees conduct pre-issuance review of every grant, explicitly forbidden from deferring to peer reviewers. Compliance with undefined "Gold Standard Science" required.

• Section 200.340: Active grants terminable at any time without cause, misconduct finding, or appeals process.

• Section 200.461: Publication costs presumptively unallowable, directly conflicting with the 2022 OSTP public access mandate.

• Section 200.432: Conference attendance requires express agency pre-approval written into award terms.

• Section 200.454: Journal subscriptions categorically unallowable. Professional memberships require prior approval.

• Section 200.220: International collaboration broadly prohibited, including allocable indirect costs.

They published it on a Friday afternoon, the slot reserved for announcements the administration does not want covered. The comment window closes July 13, mid-summer, after most faculty have left campus.

Why silence is dangerous

Under the APA, a sparse comment docket is evidence of tacit acceptance. An agency can cite low participation to justify finalizing without substantive revision. Every PI, trainee (or unemployed), citizen, taxpayer (you do not need to be in the USA or a citizen) who stays silent makes the legal challenge harder for the PIs who do not.

As one NIH program officer put it: "The comments will be ignored by the administration. But they will not be ignored in the historical record. Silence will be inferred as support."

What to do now

1. Submit a comment at regulations.gov (Docket OMB-2026-0034). Name one provision, one statutory conflict, one harm to your lab. That is a substantive comment OMB must address. Guide with modular arguments: https://www.bioniclab.org/news/OMB

2. Contact your House rep and both senators. Ask for peer review protections in FY2027 LHHS appropriations. Markup timelines overlap with this comment period.

3. Contact your state attorney general. State AGs have standing to challenge the final rule and led the successful indirect cost cap litigation.

4. Forward this letter to every colleague, postdoc, and graduate student in your department. The people losing conference funding, publication funding, and journal access first are the least represented in the record.

The post-WWII system in which the government funds science but does not direct it is not being reformed. It is being replaced, by regulation, on a 41-day clock, in a document published at the moment designed for no one to notice.

Please do not be silent.

TK

Full proposed rule: https://www.federalregister.gov/documents/2026/05/29/2026-10817/regulation-for-federal-financial-assistance

Docket OMB-2026-0034. Comment of Takashi D. Y. Kozai, Ph.D., Ernest E. Roth Professor of Bioengineering, University of Pittsburgh, in an individual capacity. I direct a laboratory studying why implanted brain-computer interfaces fail, and I was an invited witness to the 2026 National Security Commission on Emerging Biotechnology convening. The rule should be withdrawn. It contradicts the statutes it operates under, is internally contradictory under the APA, works against the national-security strategy Congress is enacting, and harms patients, trainees, and US competitiveness. Full analysis attached.

Federally funded discovery pays off on a multi-decade lag, after the discoverers retire. A 1966 NSF grant to a microbiologist at a Yellowstone hot spring seeded the enzyme behind PCR, now a market over $15 billion. No metric applied at award identifies the next PCR, so judging science by present alignment ([200.205] Gold Standard Science, an undefined benchmark failing Grayned) defunds the high-uncertainty work that yields the largest returns.

Peer review is imperfect but transparent, because scored critiques state why each proposal succeeds or fails, which replaced an era when awards turned on personal relationships. [200.205] Pre-issuance review by political appointees forbidden from deferring to reviewers, and peer review demoted to advisory, contradict 42 U.S.C. 289a, which makes expert peer review the statutory basis for NIH funding. Because the appointee need never state a reason, and the same authority lets an agency decline to fund and re-compete until a preferred applicant prevails, the rule reinstates unaccountable patronage under a merit label.

[200.340] Terminating active awards requires only a brief rationale, with no finding of cause and no appeal, making every other provision enforceable through fear. Since February 2025, HHS terminated over 1,800 NIH grants (GAO, 2025). Human BCI trials implant electrode arrays in the cortex, so mid-study termination strands volunteers with implanted devices and no funded team to explant them. Termination also raises net cost, because it draws litigation the public funds on both sides and, under the two-forum scheme of NIH v. American Public Health Association (2025), can end with the government paying the award plus the cost of the fight, plus shutdown, restart, and lost time-series data no verdict recovers.

[200.461][200.432][200.454][200.421][200.450] Unallowable publication costs contradict the 2022 OSTP public-access mandate, and the conference, journal, membership, and issue-advocacy limits are viewpoint-based speech restrictions that sever how funded scientists publish and read the literature, triggering strict scrutiny. The new conference and membership pre-approvals are also unworkable, because the department cut roughly 10,000 positions (about 1,200 at NIH) in layoffs a court found likely unlawful, leaving too few staff to grant them.

[200.220] The international-collaboration ban supplies the push competitors exploit. The 2025 chemistry Nobel laureate Omar Yaghi left UC Berkeley for Tsinghua in 2026, citing US funding cuts, and a CNN tally found at least 85 scientists left for China since early 2024. The ban is also vague enough to terminate an all-US project when a single vendor uses a foreign-made component, reaching work with no security concern (Grayned).

[2 CFR restructure] OMB claims binding authority over every agency through 31 U.S.C. 503, overriding agency enabling statutes over $1.1 trillion in assistance, which the major-questions doctrine says requires clear congressional authorization. [200.339(b)] The rule also permits agencies to aid private suits against grantees, asserting enforcement authority OMB lacks.

Contradiction with Congress. The FY2027 NDAA, on NSCEB recommendations, builds biotechnology national-security capacity, yet this rule makes the research feeding it terminable at will.

Arbitrary and capricious (APA 706(2)(A)). Every stated objective, transparency, clarity, and reduced burden, is contradicted by the provisions above, and the Regulatory Impact Analysis counts none of the termination, litigation, restart, or irrecoverable-data costs the rule imposes. Courts already voided the related NIH indirect cost cap as unlawful (First Circuit, January 2026).

Collateral harm. The R01 success rate fell from 21.6 percent (FY2023) to 13 percent (FY2025), and every NIH dollar returns about $2.57, so suppressing innovation raises long-run healthcare costs while taxpayers fund the failed legal defense.

Requested action. Withdraw the rule. If OMB proceeds, restore binding peer review and limit political review to compliance and security screening that cannot override scores, keep publication and conference costs allowable, require documented cause plus appeal for termination, protect active clinical trials absent a safety finding, narrow foreign restrictions to entity-specific determinations, and strike the private-suit provision.

Executive summary

OMB-2026-0034 should be withdrawn. The rule conditions every discretionary federal science grant on review by political appointees, forbids those appointees from deferring to expert peer review, and permits termination of active awards at will, which together replace the merit-based system that produced American scientific leadership with a system of political control. The rule is legally fragile on at least six independent grounds, it contradicts the biotechnology national-security strategy Congress is enacting in the FY2027 National Defense Authorization Act, and it inflicts concrete harm on patients, on the scientific workforce, and on United States competitiveness with China.

The core defect is a misunderstanding of how scientific value is created. Foundational research pays off decades after the work is done, usually after the people who did it have retired, so a rule that funds only what a nonexpert finds valuable today and cancels what a nonexpert later finds misaligned defunds exactly the high-uncertainty work that produced the polymerase chain reaction, the first CRISPR medicine, and the messenger-RNA vaccine platform. This comment establishes the submitter's standing, the compact that produced American science, the biological premise specific to implantable brain-computer interfaces, the specific research the rule would destroy, the legal deficiencies with controlling authority, the national-security contradiction, a provision-by-provision analysis, responses to the justifications OMB is likely to offer, the collateral harm to Americans who never held a grant, worked examples, and workable alternatives.

Each argument in this comment is a significant comment within the meaning of 5 U.S.C. 553, because each, if credited, would independently require OMB to withdraw or modify the rule, so OMB must respond to each on the record. The legal grounds in Section 5 are lettered and the provision objections in Section 7 are numbered P1 through P19 precisely so that each is separately addressable, and failure to respond to any one is an independent failure to consider an important aspect of the problem under Motor Vehicle Manufacturers Association v. State Farm (463 U.S. 29, 1983). A reviewing court may set aside the rule if OMB adopts a rationale this comment refutes, ignores evidence this comment supplies, or fails to address an alternative this comment proposes, so these objections are stated with the specificity that preserves them for review, and the comment closes with a summary reference table and a list of authorities.

The rule is also politically unsustainable, because its costs fall on people across the political spectrum. It cancels clinical trials that patients are enrolled in, strands volunteers who carry experimental implants, and sends Nobel laureates and their laboratories abroad. It wastes tax dollars already spent on terminated work, slows the treatments American families are waiting for, and raises the long-run cost of care. Opposition is correspondingly broad, spanning patient and disease-advocacy organizations, veterans who rely on federally developed devices, universities and hospitals in every state, industry that draws on federally funded discovery, and tens of thousands of public commenters. A measure that inflicts these harms on ordinary Americans, whatever its stated purpose, will not survive the scrutiny of the public or of a future administration once its effects are visible, which bears on both the reasoned basis of the action and its durability.

1. Who is submitting this comment

I direct a neural engineering laboratory studying why implanted brain-computer interfaces fail inside the human brain. My group works across human, nonhuman primate, and rodent models with surgical and clinical collaborators who run human intracortical trials, producing the underlying science for devices that let paralyzed people move and feel through robotic limbs. I served as an invited expert witness to the National Security Commission on Emerging Biotechnology convening on implantable brain-computer interfaces on April 17, 2026, where a dedicated session addressed United States competition with China, and I contribute to the FDA-recognized implantable brain-computer interface collaborative community. The same work trains the American scientists and engineers who build the next generation of these devices. I declined a cofounder position at a prominent brain-computer interface company in October 2016 because the failure of these implants is a biological problem operating on timelines that venture capital will not fund, which in turn meant the biology had to be solved first in an academic setting supported by federal research dollars. My path required roughly ten thousand hours of deliberate training in biology and a comparable investment in engineering, because judging this work demands fluency in both, and I founded a biotechnology company as a graduate student, so I write as someone who has done the basic science, the engineering, and the translation. That combination is rare and takes more than a decade to build, which is precisely why a political appointee cannot substitute for the assembled expertise of a study section.

2. The compact that produced American science, and how scientific value is created

The United States built its scientific leadership on an arrangement established after the Second World War. In Science, The Endless Frontier (1945), Vannevar Bush argued that the federal government should fund basic research broadly while leaving scientists free to set its direction, because the practical returns of discovery cannot be forecast at the outset. That arrangement, under which government funds science but does not dictate its conclusions, produced the National Institutes of Health and National Science Foundation systems and the returns below. OMB-2026-0034 abandons it, substituting political direction of scientific conclusions for expert-directed inquiry.

Federally funded discovery pays off on a multi-decade lag, and the return is usually attributed long after the people who did the work have retired or died. In 1966, an NSF-funded microbiologist collected a heat-tolerant bacterium from a Yellowstone hot spring with no medical objective, and that organism supplied the heat-stable enzyme that made the polymerase chain reaction possible two decades later, a technique that now underlies a PCR market exceeding $15 billion and every COVID-19 PCR test. The study of unexplained bacterial immunity that became CRISPR produced the first approved CRISPR medicine, Casgevy, in December 2023, for roughly 16,000 Americans with sickle cell disease. Katalin Kariko's work on messenger RNA sat largely unfunded for close to thirty years before it powered two COVID-19 vaccines and a Nobel Prize.

These returns were invisible at the moment of funding, which exposes the central flaw in judging science by present political alignment. No metric applied at the time of award can identify which foundational project becomes the next PCR, because the value of discovery research is only legible in hindsight. A rule that funds what looks valuable today under an undefined Gold Standard Science benchmark [200.205], and cancels what a nonexpert later deems misaligned under terminate-at-will authority [200.340], therefore selects against the high-uncertainty work that produces the largest returns. It does so with no benchmark an awardee can meet and no path to appeal, so the defect is a due-process failure examined in Section 5, not merely unwise policy.

3. The scientific premise this rule ignores, that neural-interface failure is biological

Implanted brain-computer interfaces fail for biological reasons that engineering alone cannot fix, which is why sustained federal neuroscience funding is the precondition for the field rather than an optional supplement. Within seconds of insertion, the brain's resident immune cells extend processes toward the electrode and nearby blood vessels leak, disrupting the blood-brain barrier before the surgeon closes the incision. Two identical arrays placed by the same surgeon in the same region on the same day can perform markedly differently months later, driven by which vessels were grazed during a ten-second insertion. My laboratory was the first to document this insertion variability with two-photon imaging in living brain, and that finding informed the image-guided surgical methods used in current commercial systems.

Mechanical mismatch compounds the injury, because silicon electrodes are roughly six orders of magnitude stiffer than brain tissue, so ongoing micromotion from every heartbeat and breath drives repeated cellular strain that never resolves. My group pioneered carbon-fiber electrodes roughly 7 micrometers wide, published in Nature Materials, that reduce tissue displacement and chronic inflammation, yet even those probes trigger a biological cascade. Better materials reduce the injury but do not eliminate the biology, so the durable gains come from understanding the cellular response, work only sustained federal funding supports.

The failure is a cascade across cell systems on distinct timescales, not a single event. Within minutes, blood-brain barrier disruption admits serum proteins that activate microglia. Over days to weeks, reactive astrocytes form a sheath that electrically isolates the electrode from the neurons it must record. Across weeks to months, oligodendrocyte loss and demyelination degrade conduction of the signals the device measures, while neurovascular uncoupling starves the local tissue of oxygen. A further problem is computational rather than structural, because a neuron that still fires is not necessarily one that still communicates within its network, so recording a spike does not guarantee the circuit remains intact.

Myelin loss around an implant is a metabolic failure, not only a slowing of conduction. Oligodendrocytes wrap axons in myelin and also supply those axons with lactate as a metabolic substrate, so when oligodendrocytes are lost the underlying axons lose both their insulation and their energy supply, which drives them toward dysfunction and death. This reframes chronic recording decline as an energy-supply problem at the electrode, invisible to any evaluation that treats the device as a purely electrical object. Characterizing it requires expertise in oligodendrocyte metabolism that no political reviewer holds.

The neurovascular problem is the loss of active blood-flow control, not leakage alone. Mural cells, the pericytes and vascular smooth muscle wrapped around brain microvessels, adjust local blood flow to match neural demand, so they operate as a computational element of the circuit rather than passive plumbing. Insertion damages these cells and uncouples blood flow from activity, which starves active tissue and distorts the signals a device is placed to read. My group's work on glial and mural cell computation examines this layer directly, and it is the layer a nonexpert would overlook entirely.

Delivering current into cortex does not simply excite the nearest neurons. Stimulation preferentially recruits inhibitory interneurons and passing axons, so the net effect on a circuit depends on which cell types the current reaches and on the ongoing brain state at the moment of stimulation. The same electrode and the same current can therefore produce excitation in one state and suppression in another, which means fixed-parameter stimulation is neither reliably safe nor reliably effective. Our work on inhibitory recruitment and state-dependent cortical stimulation addresses this, and it explains why the six-week seizure finding matters, because unaccounted inhibitory dynamics accumulate into pathology over time.

Recording quality reflects the function of a network, not the health of one electrode's immediate neighborhood. Neurons communicate across cortical layers and between cortex and deeper structures such as the hippocampus, and that functional connectivity weakens with age. Because most people who receive neural implants are older, the aging brain's altered interlaminar and cortico-hippocampal connectivity changes what a device can decode, which our work on functional connectivity in the aging visual cortex and hippocampus quantifies. Evaluating a proposal in this area requires understanding aging network physiology, a further domain a single appointee cannot cover.

This premise has a direct legal consequence, because evaluating whether a proposal in this field is meritorious requires domain expertise in cortical electrophysiology, glial biology, and neurovascular coupling that no political appointee possesses. Appointee override therefore does not merely violate the peer-review statute examined in Section 5, it guarantees scientifically incompetent decisions. One safety finding shows the stakes, because half of the animals my group stimulated at the most common clinical frequency developed seizure-like activity only after six weeks. That delayed signal is invisible to any short-horizon study, and terminate-at-will [200.340] would make the work impossible to complete.

4. The specific research this rule would destroy

The abstract harms in this comment become concrete in specific lines of federally funded research my laboratory and its collaborators conduct. Each depends on sustained multi-year funding and on evaluation by reviewers who understand the underlying biology, so each is directly threatened by terminate-at-will authority [200.340] and by nonexpert pre-issuance review [200.205].

Glial encapsulation and the foreign-body response.

Every implanted cortical electrode is eventually walled off by reactive glia, which is the dominant reason recording quality declines over months to years. Within hours, microglia extend processes to the electrode surface, and over weeks reactive astrocytes deposit a dense sheath whose endfeet raise electrode impedance and push surviving neurons away from the recording site, so the device loses signal even where neurons remain alive. Interrupting that response is the difference between a device that works for a month and one that works for a decade, which determines whether the technology reaches patients at all. The six-week seizure finding in Section 3 shows why this work needs long timelines, because the critical signal did not appear until well after any short study would have closed.

Myelin and oligodendrocyte biology.

Oligodendrocytes, the cells that produce myelin and also supply axons with metabolic substrate, control long-term recording stability through mechanisms my group characterized. Manipulating a single RNA-binding protein in oligodendrocytes alters myelination and, in one direction, produces hypermyelination, which shows how tightly recording quality is tied to myelin state rather than to the electrode alone. Testing repurposed compounds produced the result that clemastine, an FDA-approved antihistamine that promotes remyelination, improves chronic electrode performance. The same oligodendrocyte biology reduced neuronal damage in an Alzheimer's disease knock-in model, so this line links neural-interface durability to the treatment of neurodegeneration, and a nonexpert reviewer would be unlikely to recognize why an antihistamine, or a myelin RNA-binding protein, belongs in a brain-electrode study.

Neurovascular and mural-cell dysfunction.

Insertion damages the microvasculature and the mural cells, pericytes and vascular smooth muscle, that actively match local blood flow to neural demand, which uncouples blood flow from activity and drives the per-electrode variability described in Section 3. Because mural cells participate in the circuit's computation rather than passively supplying it, their loss distorts the signals a device reads, not only the tissue's energy supply. The damage is set during a ten-second insertion and unfolds over weeks, so only chronic studies with intravital imaging can characterize it, and only reviewers who understand cerebrovascular physiology can judge the work.

Autophagy-lysosomal failure.

A newer line examines how the cellular waste-clearance machinery, the autophagy-lysosomal system, fails in the cells surrounding a chronic implant, because cells under sustained mechanical and inflammatory stress accumulate damaged components they cannot clear. The same failure appears in aging and in neurodegeneration, so strategies that restore lysosomal function bear on both implant longevity and disease treatment. This work has no near-term product and a payoff years away, which makes it exactly the high-uncertainty basic research present-alignment review [200.205] would defund first, and exactly the kind that produced the returns catalogued in Section 2.

Cross-species microstimulation safety and closed-loop control.

Restoring touch through a prosthetic limb requires delivering current into cortex, and doing so safely requires knowing the dose thresholds at which stimulation becomes harmful. The finding that half of animals developed seizure-like activity at the most common clinical frequency, only after six weeks, is a safety result with direct clinical and defense relevance that terminate-at-will [200.340] would make impossible to complete.

Sensory restoration through intracortical microstimulation.

Restoring the sense of touch requires delivering patterned current into the somatosensory cortex so a person using a robotic hand can feel contact and pressure, which closes the loop that makes dexterous control possible. This work depends on knowing safe and effective stimulation parameters over years rather than weeks, so it is precisely the long-horizon research terminate-at-will [200.340] would end. My cross-species framework for evaluating stimulation safety, developed with a bioethicist coauthor, exists to support this clinical translation, and it grew from the multi-site collaboration the conference restrictions in Section 7 would throttle.

Neural decoding and closed-loop control.

Converting cortical activity into intended movement requires decoding algorithms that stay stable as the underlying neural signals drift over months, an unsolved problem tied directly to the biological changes described in Section 3. Progress requires paired biological and computational work, because a decoder cannot compensate for signal loss it does not understand, so the engineering and the biology must advance together. Political review that funds one without the other, or cancels either midstream, breaks the coupling the problem requires.

Wireless and light-driven stimulation.

Current implanted systems depend on percutaneous connectors and wired hardware that create infection risk and limit use outside the laboratory, so a major research direction develops wireless and light-driven stimulation that removes the wired connection. One approach uses photovoltaic elements that convert light delivered through the skull into local stimulating current, eliminating both the wire and the battery, which lowers infection risk and enables use in daily life. This work spans materials, circuit design, and chronic biological testing, and its payoff is years away, which makes it the high-uncertainty engineering that present-alignment review [200.205] would deprioritize despite its clinical importance.

Aging and network-level connectivity.

Functional connectivity across cortical layers and between cortex and hippocampus weakens with age, and most implant recipients are older, so the aging network changes what a device can decode. Our work quantifying interlaminar and cortico-hippocampal connectivity in the aging visual cortex establishes that recording and decoding cannot be understood at the level of a single electrode, which is the level a nonexpert review defaults to. This bears directly on clinical translation, because a decoder validated in young tissue may fail in the older patients who ultimately receive the device.

Per-electrode variability and biomarker-guided model validation.

Two electrodes millimeters apart can fail by different biological mechanisms, so treating an array as uniform obscures the failure and misdirects the fix. My group's work reframes chronic performance around per-electrode variability and around biological markers that predict which mechanism dominates at each site, which lets a computational model be validated against measured biology rather than against an assumption. The same biomarker framework supports a mechanistic clinical-trial design in which per-electrode markers guide validation and define endpoints, rather than an active-comparator design that treats the array as uniform, which is how the field moves from anecdotal device performance to predictable outcomes. This biomarker-guided, mechanistic validation depends on the multi-year datasets terminate-at-will [200.340] would truncate before the biology is resolved.

Clinical neuromodulation as white-matter therapy.

Clinical neuromodulation depends on the same biology this rule threatens. Deep brain stimulation, established for movement disorders and under study for treatment-resistant depression and PTSD, works in part by remodeling white matter, so its therapeutic effect is a myelin and metabolic effect rather than a purely electrical one. Recent findings that stimulation engages mitochondrial and glucose metabolism support treating deep brain stimulation as a white-matter therapy, which means the myelin research this rule threatens underpins a clinical field that already helps patients. Cutting the basic myelin and metabolic research therefore reaches approved therapies and active trials, not only experimental recording arrays.

5. Legal deficiencies

OMB must respond to each significant comment in this section, and each lettered ground below is independently sufficient to require withdrawal or modification of the rule, so no single response disposes of the section. The grounds are stated separately, because leaving any one unanswered is visible on the administrative record and is the failure to consider an important aspect of the problem that State Farm makes reversible. Where a ground rests on evidence the preamble did not address, OMB must engage that evidence rather than restate its conclusion.

A. The rule contradicts the statutes it operates under.

Mandatory pre-issuance review by political appointees, forbidden from deferring to peer reviewers [200.205], contradicts 42 U.S.C. 289a, which makes peer review by qualified experts the statutory basis for NIH funding decisions. After Loper Bright Enterprises v. Raimondo (603 U.S. 369, 2024), courts no longer defer to an agency's reading of its own statute, so OMB cannot construe 289a to permit appointee override, and a reviewing court decides the question independently. Making publication costs unallowable [200.461] contradicts the August 2022 OSTP public-access memorandum, which as of this comment still requires federally funded results to be freely available, so the rule keeps that mandate in force while banning the mechanism that satisfies it. OSTP has announced it is in the process of repealing that memorandum, but the memorandum remains in effect, so the conflict is present rather than hypothetical. A subordinate regulation cannot displace a statute Congress enacted, and the loss of judicial deference makes that defect decisive rather than arguable. OMB will respond that 289a requires only that peer review occur and leaves the funding decision to the agency. That reading fails, because the statute makes review by qualified experts the basis for the decision, so forbidding the appointee from deferring to that review removes the statutory basis rather than supplementing it, and after Loper Bright the court, not OMB, decides whether 289a permits the override.

Peer review also supplies the reasoned basis judicial review requires, because a study section produces a written record of scored critiques stating why a proposal succeeds or fails, which is the articulated connection between facts and choice that State Farm demands. Replacing that record with appointee discretion that need state no reason removes the basis a court examines, so the override both exceeds 42 U.S.C. 289a and leaves the funding decision unreviewable in the way reasoned decision-making forbids, as P2 develops.

B. The rule is arbitrary and capricious under APA Section 706(2)(A).

Reasoned decision-making under Motor Vehicle Manufacturers Association v. State Farm (463 U.S. 29, 1983) requires an agency to examine the relevant data and to articulate a rational connection between the facts found and the choice made. The rule fails on its own terms, because each stated objective, transparency, clarity, and reduced burden, is contradicted by its provisions, as Section 7 documents. When an agency reverses long-standing policy it must show awareness that it is doing so and must account for serious reliance interests, a requirement the Court set in FCC v. Fox Television Stations (556 U.S. 502, 2009) and applied to reliance in Department of Homeland Security v. Regents of the University of California (591 U.S. 1, 2020). The rule ignores the reliance interests of institutions that hired staff, trainees who chose careers, and trial participants who accepted cortical implants, so it repeats the omission that invalidated the DACA rescission. The preamble also rests on policy and ideological rationale rather than an evidentiary record, which under Department of Commerce v. New York (588 U.S. 752, 2019) exposes a rule to reversal where the stated rationale does not match the real basis. Courts have already set aside a comparable NIH cost-cutting measure, because on January 5, 2026 the First Circuit struck the NIH indirect cost cap as a violation of an appropriations statute and the agency's own regulations, after the district court found the cap arbitrary and capricious. An independent analysis by a former NIH program officer reached the same conclusion, characterizing the rule as a system of political control layered across the funding lifecycle rather than a reform of grants management, which corroborates that the stated rationale does not match the design. OMB's own analysis ignores the costs termination imposes, because it does not account for the litigation the provision generates on both sides, the shutdown and restart of defunded laboratories, the loss of data that cannot be reconstructed, or the departure of trained researchers. It therefore fails to consider an important aspect of the problem in the sense State Farm forbids, as Section 7 documents at the termination provision.

C. Core compliance terms are void for vagueness.

Due process requires that a regulation give fair notice of what it demands and contain standards that prevent arbitrary enforcement, the test the Court stated in Grayned v. City of Rockford (408 U.S. 104, 1972). Gold Standard Science is required as a compliance benchmark [200.205] with no operational definition, and Section 2 explains why that gap cannot be cured, because the value of discovery research is only visible years after the award. Anti-American values operates as a disqualifying criterion [200.205], national interest triggers an exemption from public posting [200.204], and gender ideology governs an entire funding prohibition [200.300], each without measurable content. Section 7 analyzes these terms one by one, and each lets the agency approve or terminate an award for reasons it need never articulate, which is the standardless enforcement Grayned forbids. The rule is a binding regulation enforced by termination, not the internal guidance at issue in National Institutes of Health v. American Public Health Association, so the fair-notice requirement applies with full force rather than the relaxed expectation urged there for discretionary guidance.

D. The structure raises non-delegation, separation-of-powers, and major-questions problems.

OMB claims directly binding authority over every federal agency through a 2 CFR restructure, which raises whether OMB may override agency-specific enabling statutes, including the Public Health Service Act, the National Science Foundation Act, and the Department of Energy Organization Act, through one uniform grants regulation. Under the major-questions doctrine the Court applied in West Virginia v. Environmental Protection Agency (597 U.S. 697, 2022), an agency asserting authority of vast economic and political significance must identify clear congressional authorization, which the rule does not provide. The design compounds the problem, because it vests grant decisions in political appointees while Schedule F strips civil-service protection from the career experts who previously made them, so decision-makers lack domain expertise and expertise-holders lack tenure, an arrangement in tension with the intelligible-principle requirement of Whitman v. American Trucking Associations (531 U.S. 457, 2001). A reviewer will respond that OMB has issued government-wide grants guidance since 2013, so this is ordinary housekeeping. That response fails, because the rule converts guidance into binding regulation that overrides the enabling statutes of every science agency on a single date, which is a change in kind, and the economic and political significance of asserting that authority over roughly $1.1 trillion in annual assistance is precisely what triggers the clear-statement requirement West Virginia imposes.

E. The speech restrictions trigger strict scrutiny.

Government may not condition a federal benefit on the surrender of a constitutional right, the unconstitutional-conditions principle the Court applied in Agency for International Development v. Alliance for Open Society International (570 U.S. 205, 2013) when it struck a funding condition that compelled grantees to adopt the government's viewpoint. The prohibition on issue advocacy [200.450], the restriction on public communications [200.421], and the content-based criteria in [200.205] that block awards touching the sex binary, immigration, or anti-American values, together condition research funds on avoiding disfavored viewpoints. Legal Services Corporation v. Velazquez (531 U.S. 533, 2001) held that the government may not use funding conditions to distort the usual functioning of a medium of expression, and Rosenberger v. Rector and Visitors of the University of Virginia (515 U.S. 819, 1995) held viewpoint discrimination in a funding program presumptively unconstitutional. Viewpoint-based conditions trigger strict scrutiny, and the rule identifies no compelling interest and no narrow tailoring, so it fails that standard even accounting for the latitude in subsidy design recognized in National Endowment for the Arts v. Finley (524 U.S. 569, 1998). OMB will invoke Rust v. Sullivan (500 U.S. 173, 1991) for the proposition that the government may decline to subsidize speech it does not wish to fund. That principle does not reach these provisions, because Agency for International Development v. Alliance for Open Society International draws the line at conditions reaching a recipient's speech beyond the funded activity or compelling a viewpoint, and the issue-advocacy ban, the public-communications restriction, and the anti-American-values criterion condition funding on the viewpoint of speech outside the funded project itself, which is the unconstitutional side of that line.

F. The rule enables impoundment by regulation.

The Impoundment Control Act (2 U.S.C. 681 and following) requires the executive to spend funds Congress appropriates unless it follows the Act's rescission or deferral procedures, and the Supreme Court held in Train v. City of New York (420 U.S. 35, 1975) that the executive may not decline to spend appropriated money because it disfavors the program. GAO already found NIH in violation of the Act (B-337203, August 2025). Terminate-at-will [200.340], combined with the conference, publication, journal, and international-collaboration restrictions, makes appropriated research money functionally unspendable on research, which effects a withholding the Constitution assigns to Congress, accomplished by regulation rather than by open impoundment.

G. The combined structure is the decisive defect.

Any single provision might survive judicial review in isolation. The cumulative structure is what makes the rule legally fragile, because a court reviewing the whole can see the pattern that each individual provision obscures, a rule that claims transparency while eliminating it, claims burden reduction while increasing it, overrides multiple statutory mandates, imposes viewpoint-based content restrictions, vests unreviewable discretion in nonexperts, and rests on policy rather than an evidentiary record. Reviewing the provisions together, rather than one at a time, is how the rule's true design becomes visible. The rule is also one element of a coordinated design, because NIH's separate proposal to cap the number of research project grants held by one investigator and NSF's parallel replacement of its proposal and award guidance advance the same narrowing through different instruments, so the cumulative effect the aggregate inquiry captures extends across agencies acting in concert.

H. The rulemaking process is procedurally inadequate.

Notice-and-comment rulemaking under 5 U.S.C. 553 requires that the public receive enough information to comment meaningfully, and enough time to do so, in proportion to a rule's scope. This rule rewrites the government-wide framework for federal financial assistance across every science agency, yet the preamble omits the data a commenter would need to assess its effects, including any analysis of how the changes would alter grant outcomes, workforce retention, or scientific productivity. A comment period that pairs sweeping scope with a thin evidentiary record does not permit the informed participation the statute requires, which is an independent procedural defect apart from the substantive ones above. The mismatch is concrete here, because OMB allowed 45 days to comment on a 412-page rule rewriting 456 sections that govern roughly $1.1 trillion in annual assistance across 42 agencies, and it set the period to close before the fiscal-year deadline it chose for itself, so the schedule serves the agency's convenience rather than informed public participation. The record should reflect a request to extend the comment period in proportion to the rule's scope and to reissue the most consequential provisions as separate rules, because a denied procedural request of this kind is itself reviewable.

I. The standard of review requires vacatur, not remand.

Under APA Section 706(2), a reviewing court sets aside agency action that is arbitrary and capricious, contrary to constitutional right, in excess of statutory authority, or adopted without observance of required procedure. The defects above fall in each of those categories, so the rule is vulnerable on multiple independent grounds rather than one. Where a rule's defects are pervasive and go to its basic structure, as Section 5G explains, vacatur of the whole is the ordinary remedy, because severing individual provisions cannot cure a design that is unlawful as an integrated scheme. The government's likely response, that some provisions could be salvaged, does not save a rule whose central mechanism, appointee override of expert review, is itself unlawful. A severability clause, if OMB includes one, does not bind the court, because pre-issuance review, the terminate-at-will authority, and the speech and collaboration conditions are so interwoven with the remainder that the rule cannot operate coherently once they fall, so the proper remedy is vacatur of the rule as a whole rather than provision-by-provision trimming that would leave an incoherent instrument.

J. Enlisting federal enforcement against grantees and turning compliance error into litigation exposure exceeds OMB's authority.

Proposed 200.339(b) permits a federal agency to cooperate with private parties pursuing their own causes of action or remedies based on a recipient's failure to comply with law, regulation, or an award's terms and conditions, which reaches far beyond the misuse of funds. A laboratory like mine operates under dense terms and conditions, because invasive human and animal neural-interface work carries clinical-trial safety reporting, adverse-event timelines, other-support disclosure, animal-welfare assurances, and data-management obligations that generate constant filings. Because the provision facilitates private causes of action, including False Claims Act qui tam suits, a late adverse-event report in a chronic cortical-implant study could be pursued as a false claim with federal cooperation, so an ordinary documentation lapse becomes exposure to treble damages rather than a correctable administrative matter.

That exposure falls hardest on the high-compliance clinical research this comment defends, because the studies carrying the most reporting obligations carry the most fraud surface. OMB also lacks authority to direct how federal enforcement resources support private litigation, which no statute grants through a grants-management regulation, so the provision raises the separation-of-powers defect of Section 5D on a second front. The predictable response is that institutions decline the riskiest and most novel human studies to limit liability, which strips out the clinical work that turns neuroscience into treatments.

K. Converting guidance into binding regulation across 456 sections rests on authority OMB does not have.

The rule grounds its conversion of 2 CFR into binding regulation in 31 U.S.C. 503, a financial-management coordination provision, not a grant of power to displace the enabling statutes Congress wrote for each science agency. My own program is funded across agencies with distinct statutory missions and distinct implementing rules. Neural-interface research spans engineering, supported through the National Science Foundation and the Department of Defense, and biology, supported through the National Institutes of Health, so each agency now tailors its rules to the science it funds. Collapsing those frameworks into one self-executing regulation, effective on a single date without separate rulemaking, removes the tailoring that lets an interdisciplinary program satisfy an engineering sponsor and a biomedical sponsor at once. A provision addressed to financial-management coordination cannot bear that weight, so the conversion fails for want of statutory authority under the reasoning of Section 5D.

L. The preamble's evidentiary basis under the Administrative Procedure Act.

Reasoned decision-making requires that a rule rest on an evidentiary record adequate to support its choices, and a rule whose stated basis does not match its real basis is vulnerable to reversal, as Department of Commerce v. New York (588 U.S. 752, 2019) shows. The preamble to this rule relies substantially on advocacy publications rather than on data about grant performance, workforce outcomes, or scientific productivity.

Where the preamble offers support for its choices, it relies on policy and advocacy argument rather than on independent assessment of the peer-review system, the return on federal research, or the effect of the proposed changes on scientific output. An agency may consider policy arguments, but it may not substitute ideological sources for the factual record that State Farm requires, particularly when reversing a system that has operated since 1945.

The consequence is procedural and substantive, because a rule that rests on partisan sources rather than evidence invites the finding that its stated rationale is pretextual or unsupported. That finding is an independent ground for vacatur, and the rule forfeits the deference a well-documented record would earn. The record should reflect that the rule proceeded without the evidentiary basis its scope demands.

M. Chronology of the destabilization, 2025 to 2026.

The dated events below document the destabilization this rule would entrench, drawn from the sources listed under Selected Authorities.

February 2025. HHS begins large-scale termination of active NIH grants, departing from the pre-2025 practice of fewer than six midstream terminations across thirteen years.

2025. The R01 success rate falls to about 13 percent from 21.6 percent in FY2023, and R01 awards fall to 5,885 from 7,720 in FY2024.

Mid-2025. Seven Chinese central agencies issue national guidelines designating brain-computer interfaces a strategic industry, with milestones for 2027 and 2030.

August 2025. The Government Accountability Office finds NIH in violation of the Impoundment Control Act, documenting more than 1,800 grant terminations between February and June 2025 (B-337203).

Early 2024 to 2026. A CNN tally identifies at least 85 scientists leaving United States institutions for Chinese institutions, more than half in 2025, including an NIH neurobiologist.

January 5, 2026. The First Circuit affirms that the NIH indirect cost cap, blocked within weeks of its February 2025 announcement and never in force, is unlawful.

April 2026. The administration declines to seek Supreme Court review of the indirect cost decision, leaving the First Circuit ruling in force.

May 29, 2026. OMB publishes the proposed rule at issue, with a comment deadline of July 13, 2026.

July 2026. The 2025 chemistry Nobel laureate Omar Yaghi leaves the University of California, Berkeley, for Tsinghua University, citing United States funding cuts.

6. The rule works against the biotechnology national-security strategy Congress is enacting

Congress, acting on recommendations from the National Security Commission on Emerging Biotechnology, is building biotechnology national-security capacity in the FY2027 National Defense Authorization Act. Relevant provisions establish a dedicated biometrology research program at NIST to create the measurement standards that electrode-tissue interface stability and cortical microstimulation dose-response currently lack, a National Biotechnology Initiative with interagency coordination, critical-infrastructure protection for biotechnology and genetic-data systems, and a Treasury investment fund to close the capital gap that private markets leave in early-stage biotechnology. These provisions treat brain-computer interfaces and the broader biotechnology base as strategic assets requiring sustained federal investment. OMB-2026-0034 points in the opposite direction, making the federal research that feeds those assets politically contingent and terminable at will, which undercuts a strategy a coordinate branch is enacting for national-security reasons. The record below documents how far China's coordinated state program has advanced, and a rule that weakens the research base another arm of the government is deliberately strengthening cannot claim to serve the national interest it repeatedly invokes. A reviewer will respond that limiting collaboration with a strategic competitor is itself a security measure. That response mistakes where the advantage lies, because the openness and scale that draw the world's scientists to the United States are the source of the American lead, so a categorical ban that drives that talent and its methods to the competitor weakens the position it means to protect. Genuine risks are met instead by the targeted, entity-specific screening the fixes in Section 7 preserve.

United States competition with China in brain-computer interfaces

Brain-computer interfaces are dual-use, because the same neural decoding and closed-loop stimulation that restore movement to a paralyzed patient also underpin capabilities of defense interest. The National Security Commission on Emerging Biotechnology treated implantable interfaces as a strategic technology at its April 2026 convening for that reason. Leadership in this field depends on a sustained supply of basic neuroscience, trained scientists, and clinical infrastructure, none of which can be assembled quickly once lost.

China is pursuing that leadership through a coordinated state program. In July 2025, seven central government bodies, including the Ministry of Industry and Information Technology, the National Development and Reform Commission, the National Health Commission, and the National Medical Products Administration, jointly issued Implementation Opinions designating brain-computer interfaces a strategic future industry. The plan sets milestones to achieve breakthroughs in core interface technology, including electrodes, chips, and integrated systems at internationally leading levels, by 2027, and to build a safe and reliable industrial base with two to three globally influential companies and a full domestic supply chain by 2030. The effort builds on the China Brain Project, a fifteen-year national program launched in 2015.

State-backed clinical progress is already visible, because in 2025 the Chinese Institute for Brain Research and the company NeuCyber NeuroTech, working with Xuanwu Hospital in Beijing, reported a first cohort of patients implanted with a semi-invasive cortical device. More than 98 percent of recording channels remained functional after surgery, and paralyzed patients used the system to control computers and robotic arms. China also created a dedicated medical-insurance pricing category for brain-computer interface procedures in 2025, which subsidizes clinical adoption directly.

The United States posture under this rule points the other way. Terminate-at-will [200.340] makes the domestic research base unstable, the international-collaboration ban [200.220] relocates the scientists the taxpayer trained, and E-Verify and affiliation screening [200.303, 200.206] add the surveillance burden that competitor recruitment programs advertise against. A state-coordinated rival can out-fund and out-plan a venture-driven field, but it cannot quickly replicate the density of independent expertise the United States still holds, so a rule that drives that expertise abroad forfeits the one advantage that is hard to rebuild. The result is a self-inflicted transfer of a dual-use lead, with the defense and economic returns following the people who leave.

The dual-use stakes are specific, because neural decoding converts cortical activity into intended action, and closed-loop stimulation writes information back into the brain. The same methods that restore movement and touch to a paralyzed patient therefore bear on human performance, degraded-environment operation, and neural data as a category of sensitive information the FY2027 NDAA treats as critical infrastructure. Leadership in these methods therefore carries direct national-security weight, so ceding it through an unstable funding regime forfeits strategic ground, not only economic ground. The National Security Commission on Emerging Biotechnology reached the same conclusion when it treated implantable interfaces as a priority technology, so a rule that undercuts the supporting research base contradicts the government's own security judgment.

Competitor recruitment is organized to absorb the talent this rule displaces. China's programs, including the successor to its Thousand Talents effort and a new K visa for science and technology workers, offer stable multi-year funding and startup packages sized to move an established laboratory. France's Safe Place for Science and the European Union's Choose Europe for Science, the latter backed by roughly 500 million euros at launch and expanded since, target United States researchers explicitly. These programs compete on exactly the stability terminate-at-will removes, so each termination and each collaboration restriction functions as a recruiting advantage for a rival system.

7. Provision-by-provision analysis

The provisions below share a target the rule never names, the dense community of expert scientists and trainees that is the true source of American scientific advantage. That advantage comes from the concentration of skilled people who meet, argue, and share unfinished ideas, so a result from one laboratory is refined, extended, and applied by others within months rather than decades, and it comes from that concentration far more than from the size of the federal budget or the count of publications. This is the advantage against a rising competitor, and it is built from specific fragile behaviors, scientists meeting at conferences, forming collaborations across institutions and borders, training across disciplines by proximity to other fields, and sharing early ideas through a system of trust. China is building this density deliberately and is closing the gap, yet it does not hold the critical mass of world-class scientists or the open culture of early exchange the United States has accumulated over seventy years. Each provision below weakens that accumulated strength in a specific way, and each entry names the mechanism.

Two features make this community the decisive asset. Its returns are emergent and cannot be centrally planned, because the collaborations and cross-disciplinary insights that produce breakthroughs form from the bottom up when the right people meet, which is why concentrations of talent rather than directives have produced the largest advances. The community is also partly imported, because foreign-born scientists are 43 percent of the doctorate-level workforce and more than half of the STEM doctorate holders in defense-related industries, so the United States leads in part by attracting and keeping the world's best. Both features make the advantage fragile in one direction, because openness sustains it and restriction weakens it, and once a dense cluster disperses it reconstitutes elsewhere and compounds there rather than returning. Each departure removes not just one worker but that scientist and every collaboration that ran through them, so the loss reaches well beyond the individual.

Each entry follows a common structure, what the provision does, why the national evidence refutes it, how it reduces the expert community and strikes my own laboratory, and the fix. Some arguments recur, because a single mechanism such as termination at will drives several harms, and each recurrence explains its application to the specific provision at hand.

The nineteen provision objections below are numbered P1 through P19, and each is a significant comment OMB must answer individually, because each identifies a distinct provision, a distinct evidentiary omission the preamble did not address, and a distinct harm, so no single response disposes of them and failure to answer any one leaves that provision arbitrary as to the objection raised.

Each provision also carries a cost the taxpayer bears and a lead a competitor gains, and the entries below name both, because a rule presented as fiscal stewardship and national strength in operation wastes public money and cedes scientific ground abroad.

The taxpayer harm runs across many dimensions that Section 9 documents in turn. These include treatments delayed or never developed, clinical trials canceled and volunteers left carrying implanted cortical devices no funded team maintains, scientific and manufacturing jobs and the regional economies built on them, higher healthcare costs as therapies stall, weakened readiness for the next health crisis, forgone economic growth and the loss of a dual-use lead to a competitor, and a shrinking supply of new scientists that reaches children who would have been treated by discoveries never made. Each provision below feeds one or more of these, so the entries name the dimensions each provision most directly drives, while the fuller accounting sits in Section 9.

How grants are funded

P1. [200.205] Political pre-issuance review, no deference to peer review. This provision requires a senior political appointee at each agency to review every discretionary award before issuance, forbids routine deference to peer reviewers, and conditions each award on advancing the President's stated priorities, so the funding decision moves from the expert panel to a political officer who need not share the panel's field, methods, or standards. The shift is categorical rather than marginal, because the appointee may override any score for any reason the rule does not constrain, and the rule supplies no criteria the appointee must apply, so the operative question at the point of funding changes from whether the science is sound to whether the project pleases the reviewing official.

The national evidence refutes the premise that political selection improves outcomes. Expert-directed funding returns roughly $2.57 for every NIH dollar, stands behind nearly every new drug the FDA approved between 2010 and 2019, and produced the discoveries that built entire industries, while no comparable record shows political selection performing as well on any measure the government tracks. The strongest critique in the literature finds only that percentile scores discriminate poorly among proposals already judged fundable, which is an argument that expert ranking is imperfect at the margin, not an argument that a nonexpert should choose instead, so even read at its most skeptical the evidence places the appointee far below the panel. Replacing a system with a demonstrated return by one with no evidence base is arbitrary action under State Farm, a defect Section 5B develops.

The deeper damage is to the advantage, the dense community of expert scientists whose judgment, mobility, and willingness to collaborate constitute the actual American advantage. Political review selects for alignment rather than merit, and the scientists most able to leave, the ones with the strongest offers at home and abroad, are the first to exit a system that no longer rewards quality, so the community thins from the top where its value concentrates. A departure removes a scientist and every collaboration that ran through them, and the visible fact that funding now turns on politics deters the next cohort from entering American science at all, which shrinks the supply of scientists a decade forward. A rising competitor does not need to surpass the United States if the United States disperses its own talent, and political control of individual awards is a direct instrument of that dispersal, because it tells the world's best researchers that their careers here depend on a political officer rather than on the strength of their work.

In my laboratory, a single proposal asks whether glial encapsulation, cortical micromotion, or oligodendrocyte metabolism drives the failure of chronically implanted electrodes, a question a reviewer can score only by understanding cortical electrophysiology, glial cell biology, and neurovascular imaging together. A political appointee faced with that proposal can ratify the panel, which the rule forbids as routine deference, or substitute a criterion unrelated to merit, which produces an arbitrary result, so the most advanced work in the field, the interdisciplinary work that no single reviewer fully spans, becomes the work most exposed to a political veto. The predictable response is that investigators reshape proposals toward whatever they believe the reviewing official wants, redirecting effort from the science that is hard and valuable toward the science that is safe and legible, exactly the trade a competitor organized around directed priorities would welcome.

The justification the rule offers is accountability, but accountability for public money is already enforced through audit, financial reporting, progress review, and misconduct proceedings, none of which requires overriding the scientific merit of a proposal, so the provision adds political control without adding accountability those mechanisms do not already provide. The provision also conflicts with 42 U.S.C. 289a, which makes expert peer review the statutory basis for NIH funding, and after Loper Bright that statutory conflict is decisive because no judicial deference to the agency's reading can cure a direct contradiction of the enabling statute. Delayed and undeveloped treatments come first, because a proposal an appointee blocks on political grounds never matures into the therapy it promised, so the patients who would have received it keep managing their disease untreated. Global competitiveness weakens as the industries a vetoed discovery would have created form instead in countries that still fund on merit, relocating their future manufacturing and export base. Jobs and the regional economies around universities contract as the most mobile investigators exit a system that no longer rewards the strongest science, closing the laboratories that employ staff and train students. Healthcare costs stay high because a cure that never arrives cannot reduce the lifetime expense of managing the condition it would have ended, and that burden reaches children who grow up with diseases merit-funded research would by then have solved. Preparedness weakens in parallel, since a portfolio chosen for political fit rather than scientific promise leaves the country without the depth of knowledge the next health emergency will demand. The fix is to keep binding peer review as the basis for scientific merit and to confine any appointee review to defined, lawful statutory-compliance and national-security screening applied without displacing the panel's assessment of the science.

P2. [200.205] Peer review demoted to advisory. The rule reduces peer review to explicitly advisory input that does not bind the funding decision, reversing the principle that has governed federal science since 1945 and converting the considered judgment of assembled specialists into a suggestion a political officer may disregard. Peer review functions as a structured process rather than a single decision point, because a study section of specialists reads each application, meets to argue its strengths and weaknesses, and scores it against a cohort of competing proposals. Program officers then weigh fit to the agency's mission and an advisory council concurs before any award, so both expert and public oversight attach before money moves.

The reform peer review accomplished was transparency, because a study section records scored critiques stating why a proposal succeeds or fails, which replaced a system in which awards turned on a program officer's personal relationships rather than articulated merit. Demoting that review to advisory input an appointee may set aside restores the earlier problem in a sharper form, because the appointee need never state a reason a court or an applicant could test. That same authority also lets an agency decline to fund and re-compete until a preferred applicant prevails, so a determined official can iterate to a predetermined result. A rule that eliminates the mechanism producing articulated, reviewable reasons cannot be described as advancing the accountability and rigor its preamble claims, because it reintroduces the unaccountable patronage that scored peer review was built to end.

The refutation is that this process is itself a mechanism of scientific improvement, not merely of selection. When a panel argues a proposal, reviewers surface hidden assumptions, overlooked controls, and alternative approaches that an isolated author would never see, so the strongest applications are strengthened and the weakest exposed through exactly the expert exchange that defines the community. Demoting the review to advisory does not add a political check on top of this process, it removes the reason the process functions, because a judgment that does not bind is a judgment no one need heed.

The advantage weakens here in two distinct ways. Peer review runs almost entirely on unpaid expert labor, given freely on the understanding that the assessment determines outcomes, so a review a political appointee may override is a review serious scientists will decline to provide, and as experienced reviewers withdraw the government loses the very capacity it would need to evaluate proposals competently. The second loss is to collaboration, because study sections are where scientists from different institutions meet repeatedly over years, and those repeated encounters start the collaborations, cross-training, and early idea-sharing that later produce breakthroughs, so dismantling the authority of the panel dissolves one of the community's densest meeting points. China cannot yet convene panels of comparable depth across the range of fields American study sections cover, which is precisely the accumulated advantage this provision discards for no measured gain.

In my field, evaluating a cortical-stimulation proposal requires assembling reviewers in electrophysiology, glial biology, and neurovascular imaging, because no individual spans the whole question, and in my laboratory that assembled, arguing expertise is what distinguishes a fundable stimulation-safety study from one that would put animals or patients at risk. A political officer cannot reconstruct that distributed judgment, and a system that tells specialists their judgment is advisory will not long retain the specialists.

The justification that elected officials should govern unelected scientists conflates two functions, because Congress already sets national priorities and funding levels through appropriations and authorizing statutes, the legitimate channel for political direction, while this provision inserts political judgment into the technical evaluation of individual proposals, a function for which political officials have neither the competence nor a democratic mandate. The return on each research dollar falls, because when an official may override the expert panel, the same appropriation funds proposals a specialist would have ranked unfundable, so the public buys less science for the same money. Clinical trials become less safe, since a protocol reviewers would have flagged as hazardous can now proceed, putting volunteers at risk and forcing the expensive mid-study halts that follow. Treatments reach patients later as the loss of expert reviewers slows every subsequent funding cycle, and healthcare costs stay elevated because the therapies that would have displaced costly chronic care are among those the weaker selection fails to fund. The competitiveness effect is quieter but real, since a country that funds science by seniority rather than merit falls behind one that does not. The fix is to restore binding peer review of scientific merit and to confine political review to lawful, defined, non-scientific criteria applied after, and without displacing, the merit judgment.

History offers a direct test of what follows when political officials override expert judgment on scientific merit. In the Soviet Union under Stalin, the agronomist Trofim Lysenko rejected the genetics accepted by the scientific mainstream, won political backing, and was placed in charge of Soviet biology. Geneticists who would not renounce Mendelian genetics were then dismissed, imprisoned, or in some cases executed, and Nikolai Vavilov, a leading geneticist, died in prison in 1943. More than 3,000 biologists lost their positions, Lysenko's methods produced crop failures rather than the promised gains, and Soviet genetics, once among the world's strongest, fell decades behind and did not recover until the system that imposed the doctrine was gone. The mechanism was the one this rule installs, a political authority rather than the scientific community deciding which science is legitimate, so the precedent applies regardless of the field or the party in power.

P3. [200.205] The undefined Gold Standard Science benchmark. Every award must demonstrate compliance with a Gold Standard Science benchmark that Executive Order 14303 (May 23, 2025) introduced and the rule never defines in measurable terms, leaving each agency unguided discretion to decide what qualifies and each applicant no way to know in advance what will satisfy it. Due process requires fair notice and standards that constrain enforcement, the test set out in Grayned v. City of Rockford (408 U.S. 104, 1972), and a benchmark invoked but never defined supplies neither, so it operates less as a standard than as a discretionary label an official may attach or withhold at will.

The defect is structural rather than a matter of drafting, because the value of discovery research becomes legible only years after the award. The polymerase chain reaction began as curiosity about a bacterium in a hot spring, the first CRISPR therapy grew from study of bacterial immune systems, and the mRNA vaccine platform rested on three decades of work most reviewers thought unpromising, so no honest standard applied at the moment of funding could have identified any of them as gold-standard science. A benchmark that purports to sort fundable from unfundable work at the outset asks a question the history of science shows cannot be answered at the outset.

An undefined quality standard is especially damaging to this advantage, because the community's highest-value output is exactly the work that looks unpromising when funded, the cross-disciplinary and high-uncertainty projects that arise when a dense community lets one field's methods reach another field's problems. A nonexpert applying an undefined standard will systematically favor the safe, the incremental, and the legible over the boundary-crossing bet, so the rule selects against the emergent research that a competitor still organized around directed, near-term priorities cannot easily generate, surrendering the one kind of science hardest to copy. Investigators anticipating the standard will propose what looks defensible rather than what is most promising, which quietly lowers the ambition of the entire portfolio.

In my laboratory, a basic study of how oligodendrocytes supply metabolic support to axons has no product to point to at the time of funding, yet that biology later informs both the durability of neural interfaces and the treatment of neurodegenerative disease, so a nonexpert applying an undefined benchmark could reject it as insufficiently aligned with a stated priority. The work matured only because its funding did not depend on demonstrating a near-term application it could not yet show.

If the genuine aim were scientific rigor, the rule would adopt the measures that produce it, preregistration, funded replication, statistical review, and open data, none of which appears, so the benchmark cannot be defended as a rigor measure and functions instead as unbounded discretion. Crisis preparedness takes the sharpest hit, because the high-uncertainty basic work an undefined benchmark rejects is the kind that produced the messenger-RNA platform two decades before it enabled a pandemic vaccine, so refusing it now leaves the country without the tools the next emergency will require. Treatments are delayed across the board, since discoveries that become therapies fifteen years later never receive the early funding that begins the process. Global competitiveness and the jobs tied to it move abroad, because the industries that grow from such discoveries form wherever the foundational work was funded. Children inherit the loss most fully, since a benchmark applied today removes the future research whose payoff would have arrived in their lifetimes, and the economy forgoes the compounding returns basic research has historically produced at several dollars per dollar spent. The fix is to define any compliance standard operationally, with measurable criteria and adequate notice, before conditioning awards on it, or to remove the benchmark entirely.

P4. [200.340] Termination of active awards at will. The rule permits an agency to terminate an active award, including a multi-year award already in progress, whenever a political appointee determines the work no longer serves agency priorities or the national interest as it exists at the moment of termination, implementing Executive Order 14332. An agency need only recite a brief rationale, with no finding of cause, no performance trigger, and no appeal for discretionary terminations, which is why this comment describes the authority as termination at will. The scale of the change is measurable rather than speculative, because between 2012 and January 2025 NIH terminated fewer than six grants midstream across thirteen years, so termination answered genuine cause, while since February 2025 the department has terminated more than 1,800 NIH grants, so the authority now operates as a routine instrument of policy rather than a rare remedy.

Termination at will is the single most direct attack on this advantage, because the community depends on stability and trust and both collapse under revocable funding. The early idea-sharing that lets one laboratory's result be refined, extended, and applied by others depends on scientists believing their projects will run to completion, so a field living under constant threat of defunding becomes cautious and defensive, holding unfinished work back rather than sharing it, and the collaboration and idea-sharing that make the community productive gives way to self-protection. Instability also drives the mobile and the talented abroad, and every departure removes a scientist and their collaborations, while the threat alone, even unexercised, suppresses the ambitious long-horizon work that produces breakthroughs, because no rational investigator commits years to a project that can be cancelled for reasons unrelated to its merit.

This is also the enforcement engine behind every other provision in the rule. A researcher who can be defunded at any moment for any reason will avoid disfavored topics, decline foreign collaborations, moderate public statements, and steer clear of anything an official might dislike, whatever the other provisions technically permit, so terminate-at-will converts every ambiguous prohibition elsewhere in the rule into a binding constraint enforced by fear. A competitor need not match American science if American scientists are induced to police themselves.

In my laboratory, findings in chronic neural interfaces emerge over years, and a critical safety result, that half of the animals tested developed seizure-like activity at the most common clinical stimulation frequency, appeared only after six weeks of recording, so a mid-study termination would have ended the work before its most important signal ever surfaced, leaving a danger undiscovered. The gravest case reaches human trials, where terminating an award can strand a participant who carries an implanted cortical device with no funded team to monitor or explant it, an outcome that also ends recruitment for every future trial, because no one will accept an implant once it is known the supporting funding can vanish without cause.

The authority further enables impoundment by regulation, because appropriated research money rendered unspendable on research is money withheld from its congressional purpose, which the Impoundment Control Act and Train v. City of New York (420 U.S. 35, 1975) reserve to Congress, and the Government Accountability Office has already found NIH in violation of that Act. Canceled clinical trials are the most immediate harm, because terminating an award for convenience ends the study it funds regardless of the science. That cancellation strands the trial's participants, and as described above a discontinued cortical-implant study can leave a volunteer with an implanted device and no funded team responsible for it, a burden that then falls on the public. The taxpayer also loses the money already spent, so the provision discards sunk investment and produces nothing, manufacturing the waste it claims to prevent. Treatments are deferred as the therapies those trials would have delivered stall, keeping the associated healthcare costs on the public ledger, and research jobs and the local spending each laboratory supported vanish when the award ends. Once volunteers and institutions learn funding can end without cause, recruitment for future trials falls, so the harm compounds well beyond the trials directly terminated.

Termination multiplies taxpayer cost rather than reducing it, because a canceled award generates litigation the public funds on both sides. Grantees challenged the 2025 terminations, and a federal court found them arbitrary and capricious and vacated the underlying directives. The dispute then ran through the First Circuit and the Supreme Court, which in National Institutes of Health v. American Public Health Association (606 U.S. ___, 2025) split it into two forums, one for the guidance and one for reinstatement. That structure makes the government defend in two forums while grantees litigate in both, and where a termination is ultimately found unlawful the public pays the awarded amount plus the cost of the fight, so nothing is saved. Losses no verdict can restore are larger still, because a shuttered laboratory lays off staff, ends animal cohorts, and abandons cell lines, then rebuilds from nothing if funding returns. A terminated chronic study also loses time-series data that cannot be reconstructed, since months of aging or continuous cortical implantation cannot be re-run after the fact. Each interruption pushes trainees out of research and speeds the departure of senior scientists abroad, so the provision converts a claimed saving into a compounding public loss the Regulatory Impact Analysis does not count.

The fix is to require documented cause and a genuine appeals process for termination, and to shield active human-subjects trials from termination absent a specific safety finding.

P5. [200.202] Program design aligned to political priorities. Every new grant program must be designed to align with the administration's policy priorities, so agencies build their solicitations around the current political agenda rather than around scientific need, statutory mandate, or the consensus of the field, which embeds political direction before any individual award. The effect is that whole fields can be favored or starved before a single proposal is written, because the design of the funding opportunity, its topics, criteria, and eligibility, is set to political specification rather than to where the science is most promising.

The national record shows why this fails as a strategy, because the largest returns on federal research came from lines of work sustained across many administrations, precisely because their value was not visible on any single administration's agenda, so a program keyed to one year's political priorities cannot carry the decade-long inquiry that produces the results worth having. Directed, top-down program design is in fact the competitor's model, and adopting it surrenders the specific advantage the American system holds.

That advantage is the capacity to generate value where central planning cannot see it, in the unplanned collaborations and cross-field applications that emerge when a dense community of scientists is left free to follow the evidence. When programs are shaped to a political agenda, funding goes to the expected and the aligned, and the boundary-crossing work that a rising competitor cannot yet produce is exactly what goes unfunded, so the rule trades away the emergent, serendipitous research that is hardest to copy in exchange for a portfolio that looks orderly and produces less. A scientist who must anticipate the political shape of next year's programs will propose within them, which narrows the field's imagination to the current agenda.

In my laboratory, the myelin and metabolic research that now informs both neural-interface durability and the treatment of neurodegeneration would not have survived had each year's funding tracked that year's political priorities, because for years its eventual value was invisible to any near-term agenda and legible only to the specialists in the field. The work existed because program design followed scientific need rather than political direction. Whole fields are starved at once, because shaping programs to political priorities before any proposal exists defunds entire lines of inquiry, so the loss is structural rather than project-by-project. Global competitiveness and the economy bear the largest share, since the industries, exports, and growth a sustained field generates migrate to whichever country keeps funding it. Treatments that require a decade of continuous work are lost, because a field defunded on a one-year political cycle cannot carry a therapy to completion. Children face the later cost, inheriting unsolved diseases whose solutions depended on fields abandoned before they were born, and the jobs a mature field would have created never appear. The fix is to require that program design follow authorizing statutes and demonstrated scientific need, with political priorities expressed through the budget Congress sets rather than through the design of individual programs, which keeps democratic direction in its proper channel without distorting the science.

P6. [200.204] National-interest exemption from public posting. An agency head may exempt a funding competition from public posting on Grants.gov whenever public announcement is deemed contrary to the national interest, a phrase the rule leaves entirely undefined, so a competition can be run outside public view where it cannot be scrutinized by outside observers, contested by excluded applicants, or even discovered by laboratories that would have competed. The undefined trigger fails the same fair-notice test as the Gold Standard benchmark, because the agency can invoke it without articulating any reason a court could review, and it directly undercuts the transparency the preamble claims as a justification for the entire rule.

Open competition is part of what keeps the community dense, because a publicly posted opportunity lets new entrants, less-established institutions, and unexpected interdisciplinary teams compete for support on the strength of their ideas, while a hidden competition narrows the field to those already inside a favored circle who happen to learn of it. Restricting who can even see an opportunity shrinks the pool from which collaborations and cross-institutional teams form, so the provision thins the community quietly at its point of entry, before any proposal is written or any team assembled.

The mechanism also invites directed funding to predetermined recipients, which corrupts the merit competition that produces quality, because an award steered to a favored institution is an award not won on the science. In my field, a quietly posted competition could channel neural-interface funding toward a chosen recipient while excluding laboratories with stronger records and more relevant expertise, and no excluded laboratory would ever know it had been passed over, so the loss would be invisible and therefore uncorrectable.

The likely justification is that some competitions involve sensitive national-security matters, which is real but narrow and does not support an open-ended exemption triggered by an undefined national-interest standard. Public money is misallocated, because a competition run outside public view can be steered to a predetermined recipient, so funds go to a favored institution rather than the strongest proposal, and the science bought is weaker at the same price. Treatments slow as a result, since work that would have advanced fastest under the best team instead goes to a chosen one, delaying whatever therapies the stronger proposal would have produced. The waste is self-perpetuating, because no excluded laboratory learns it was passed over, so the correction open competition normally provides never operates and the public overpays in every hidden round. Competitiveness declines as the open contest that identifies the strongest science, and keeps the country ahead, narrows. The fix is to permit non-posting only under a narrow, documented national-security standard subject to review, so that genuine sensitivity is accommodated without creating a general power to fund in secret.

How science is done and disseminated

P7. [200.461] Publication costs made unallowable. Journal publication costs, including article-processing charges and the open-access fees that make an article free to read, become unallowable by default, which places the rule in direct conflict with the August 2022 OSTP public-access memorandum that, as of this comment, still requires the results of federally funded research to be freely available, so the government would simultaneously mandate open access and forbid paying for it. The fiscal saving is trivial against the cost of the research that produced the result, so the provision cannot be defended as a meaningful economy and instead operates as a constraint on how, and whether, findings are disseminated.

The provision also cedes ground in the competition for talent, because publication is where reputations form and where recruiters find the researchers they pursue. Competitor nations have treated that visibility as a strategic asset. China paid cash bonuses for publications for roughly three decades, lifting its indexed output from about 120,000 papers in 2009 to about 450,000 in 2019, before curbing the bonuses in 2020 to check quantity-over-quality incentives. Making United States investigators pay open-access and page charges from personal or departmental funds, or forgo publishing, lowers the visibility that retains domestic scientists and attracts foreign ones, compounding the relocation Section 6 documents.

Publication is the mechanism by which the community shares and compounds knowledge, because a result becomes useful to the field only when others can read it, test it, and build on it, and the peer-reviewed record is the shared memory on which cumulative progress depends. Suppressing publication support slows the sharing of results through the community, and by pricing out open access specifically it pushes American findings behind paywalls while competitor systems that fund open publication distribute theirs freely, handing away part of the reach and influence that accompany scientific leadership.

Open publication is also a recruitment and retention instrument, because the visibility of American science is part of what draws the world's best to work here, and a system that makes it harder for its scientists to publish openly becomes a less attractive place to build a reputation. The compounding harm joins this provision to the conference and subscription restrictions, since together they sever the three channels, publishing, meeting, and reading, through which a scientist participates in the field at all, so the rule does not merely raise a cost, it isolates the investigator from the community.

In my laboratory, the failure-mode findings that now inform device design, why an electrode fails in the first seconds after insertion, how myelin loss degrades a recording over weeks, reach clinicians and other engineers only through open publication, so pricing that out would slow the diffusion of exactly the biology that lets the field avoid repeating known mistakes. A finding that cannot be read cannot prevent the next avoidable device failure.

The likely justification is cost discipline, but the cost is negligible and the informational loss is large, so the trade is irrational on its own terms, and Section 5A notes the additional conflict with the government's standing public-access obligation. Treatments that build on a prior result are delayed, because a finding others cannot read cannot be built upon, so the next therapy in a chain of discoveries starts late or never begins. The public pays twice, having funded the research and then being denied even the ability to read what it bought, while the article charges the provision bars would have made those results freely available. Competitiveness shifts abroad as competitor systems that publish openly gain the reach and citation that move a field's center, and its future commercial activity, out of the country. Preparedness suffers too, since in a fast-moving emergency the results that guide a response spread through publication, and a walled-off literature slows that diffusion when speed matters most. The fix is to keep publication and open-access costs allowable, so the government can meet the public-access requirement it has already imposed on itself.

P8. [200.432] Conference attendance requires pre-approval. Conference attendance loses its standing as a routine allowable cost and requires express agency pre-approval written into the award, and because scientific opportunities such as invited talks, panel invitations, and the presentation of new results cannot be predicted at the time an award is made, this hands political officials control over which scientific communities a funded researcher may join and when. The cost of conference travel is a small fraction of any research grant, so the provision is not a fiscal measure, its operative effect is control over scientific association.

The requirement is also unworkable given the workforce that would administer it. That structure layers conference-by-conference and membership-by-membership pre-approval on top of the pre-issuance review of every discretionary award, at the same time the department eliminated roughly 10,000 positions, including about 1,200 at NIH, in cuts a federal court found likely unlawful and arbitrary and capricious. Fewer staff cannot clear more approvals, so the approval step operates as denial by default or as delay that arrives after the invited talk or the annual meeting has passed. A rule that both expands mandatory approvals and removes the people who would grant them increases burden and forecloses participation at once, which contradicts the burden-reduction the rule claims and supplies an independent arbitrary-and-capricious defect.

This provision strikes this advantage at its most direct point, because conferences are where the community physically exists. They are the rooms in which scientists meet, hear results months or years before those results are published, argue them into sharper form, and form the collaborations that carry a finding from one laboratory into the work of many others. The refinement-expansion-application cycle that the American system runs on begins in these encounters, when a question asked after a talk becomes a joint project, or a method demonstrated in one field is carried by a listener into another, and none of this can be scheduled, directed, or reproduced by an official deciding in advance which meetings serve the administration's priorities.

The cross-training that dense communities provide is also a conference phenomenon, because a graduate student or early-career scientist learns the shape of adjacent fields by standing in the room with the people who lead them, and that exposure is how the interdisciplinary researcher is made. Gating attendance behind political pre-approval severs those meetings, isolates trainees from the communitys that determine their careers, and slows the early idea-sharing that a rising competitor, still lacking the same density and frequency of world-leading gatherings, cannot yet match. The advantage the United States holds comes from the dense connection among its scientists rather than from any individual superiority, and this provision attacks that connection directly.

In my laboratory, the cross-species microstimulation framework I now lead grew directly out of a multi-site collaboration that formed at a scientific conference, a collaboration this provision would have throttled at inception, and every conference a trainee in my group cannot attend is a set of future collaborations that never begins and a body of adjacent knowledge they never absorb. The loss compounds over a career, because each missed meeting forecloses the collaborations and cross-training it would have produced, so it far exceeds the value of the trip itself.

The likely justification is travel-cost control, but conference costs are negligible against a grant and the provision's real function is to control where and with whom scientists gather, which no budgetary rationale supports. The next scientific generation absorbs the first cost, because a trainee shut out of the meetings where collaborations form and adjacent methods are learned loses the professional connections that determine a career, so the public forfeits part of the workforce its training dollars built. Discovery slows for everyone, since the early exchange at conferences that refines and accelerates results is severed, pushing back the point at which any of that work becomes a treatment. Competitiveness declines as those collaborations reform abroad, at gatherings a competitor still hosts, and children inherit the delay because the therapies a faster, better-connected field would have produced arrive later or not at all in their lifetimes. The fix is to restore conference attendance related to the funded work as a standard allowable cost, without case-by-case political approval.

P9. [200.454] Journal subscriptions made unallowable. Subscriptions to professional and technical journals become categorically unallowable as direct costs, so an investigator cannot charge to an award the literature access required to know the current state of the field, and together with the publication and conference restrictions this removes the third of the three channels, reading, through which a scientist stays connected to the work of others. The saving is trivial and the provision is again a constraint on participation rather than a genuine economy.

Access to the literature is access to the community's accumulated knowledge, because the published record is what lets a scientist begin from the field's most advanced current work rather than rediscover what is already known, and cutting that access forces isolated, redundant work of exactly the kind a dense and well-read community never wastes effort on. The provision therefore degrades both the quality of the individual scientist and the productivity of the community, because a field in which investigators cannot read each other's results loses the coordination that makes distributed effort cumulative.

The harm compounds with the publication restriction in a particularly self-defeating way, because the rule discourages paying to publish open-access articles and simultaneously discourages paying to subscribe, so it obstructs both ends of the same exchange and leaves the investigator less able to contribute to the record and less able to read it. A competitor whose scientists retain full literature access gains a relative advantage that costs the United States almost nothing to have avoided.

In my laboratory, keeping current with the glial-biology and neurovascular literature is a precondition for designing sound experiments, not a discretionary expense, because an experiment built on an outdated understanding of the underlying biology fails before it begins and wastes the public funds that supported it. Research dollars are wasted at the design stage, because an investigator cut off from the current literature builds experiments on an outdated understanding, so public money funds work that duplicates a known result or fails for want of a published finding. Treatments are delayed later, since redundant or misdirected experiments consume the time and funding a correctly informed study would have spent moving toward an application patients would see. The inefficiency compounds across a field, so the economy loses the productivity current knowledge would have bought, and competitiveness slips as researchers elsewhere who retain full access advance while ours repeat settled questions. The fix is to restore journal subscriptions necessary to the funded work as an allowable cost.

P10. [200.454] Society memberships require pre-approval. Professional society memberships are allowable only with prior written agency approval and only where deemed necessary to award requirements, which adds a political approval step to routine professional participation and, more consequentially, lets the agency screen and curate the professional affiliations of the researchers it funds. Membership is the ordinary route by which scientists reach standards bodies, annual meetings, working groups, and professional connections, so a requirement to approve it inserts political oversight into the everyday conduct of professional life.

The approval step also fails on its own terms, because the reduced workforce described at P8 cannot process membership requests any more than conference requests, so a cost that was routinely allowable becomes contingent on an approval that may never arrive.

Scientific societies are the institutional form of the community, the durable structures through which scientists convene across institutions, set the technical standards their fields depend on, and sustain the communities that produce and evaluate new work. Conditioning membership on political approval therefore reaches directly into how the advantage is organized, because a researcher whose society membership must be approved is a researcher whose associations the government may shape, and the free formation of professional community that drives collaboration and standard-setting gives way to a curated list of politically acceptable affiliations.

The provision also deters participation at the margin, because an investigator uncertain whether a given membership will be approved, or whether it might later be questioned, will simply not join, and each such non-participation is a connection the collaboration does not form. The aggregate effect is a quieter, less connected professional community, which is the opposite of the concentration of connected expertise that constitutes the American advantage.

In my field, society membership is the route to the consensus-standards work on which the safe evaluation of new neural devices depends, so restricting it slows the development of the very standards regulators and clinicians need to bring these devices to patients responsibly. Patients and abandoned device pathways bear the clearest cost, because the consensus-standards work professional societies host is what allows a new medical device to be evaluated and approved, so cutting membership slows those standards and leaves devices stuck in the laboratory rather than reaching the patients who need them. Treatments are delayed for the same reason, held up by an approval bottleneck the provision creates rather than by any limit of the science. Healthcare costs stay higher because the cheaper or better devices those standards would have cleared never displace the existing ones, and competitiveness weakens as the American standards effort those societies lead slows while other countries set the specifications that shape global markets. The fix is to restore necessary professional memberships as an allowable cost without case-by-case political approval.

P11. [200.421] Public communication of findings restricted. Public-relations and public-communication costs become unallowable except where a statute affirmatively requires them, which restricts a funded researcher's ability to communicate findings to the public or the press, and together with the issue-advocacy prohibition it narrows the scientist's freedom to explain publicly funded results to the taxpayers who paid for them. Because the restriction attaches a funding condition to protected speech about the work, it raises the First Amendment problem Section 5E develops.

Communication is the step at which the community's output becomes public value, because research reaches the clinicians, engineers, companies, and policymakers who apply it largely through explanation aimed outside the academy, so restricting that explanation slows translation at the point where discovery would otherwise become practice. A finding that is never explained to the people positioned to use it is a finding whose social return is deferred or lost.

The restriction also weakens the standing of science at the moment that standing most needs defending, because a field that cannot explain its work to the public cannot sustain public support for its funding, and a defunded field disperses, taking its collaborations with it. Silencing scientific communication therefore feeds the same dispersal the rule's other provisions drive, and it does so while a competitor communicates its own scientific achievements aggressively to build prestige and attract talent.

In my laboratory, explaining to the public why implanted electrodes fail, and what that failure biology means for the patients who might one day receive these devices, is part of the public value of federally funded work, and a restriction on that explanation serves no purpose the rule articulates. The historical cost of penalizing public communication is on record, because Carl Sagan was denied tenure and later blocked from the National Academy of Sciences in part because colleagues treated his public engagement as evidence he was not a serious scientist, and the resulting shortfall in public understanding is part of why science now lacks the public constituency it needs to defend its funding. Treatments reach the field later, because the clinicians, engineers, and companies that turn research into practice often first learn of a result through its public explanation, so restricting that explanation slows translation to patients. The broader economy of research suffers as the public understanding that sustains funding declines, weakening support for the work that produces future therapies and the jobs around them. Competitiveness declines because a competitor that promotes its science aggressively gains the prestige and talent that visibility attracts, and preparedness weakens where public communication is exactly how findings reach the practitioners who must act in an emergency. The fix is to permit reasonable communication of the results of funded research.

P12. [200.450] Issue-advocacy prohibition on funded speech. Federal funds may not support messaging that takes a position on any social, political, or public-policy matter unrelated to the award's statutory objectives, and because the preamble characterizes research on climate, public health, and equity as divisive, the provision creates a real risk that investigators could be barred from discussing their own findings whenever those findings touch a contested area. Conditioning research funds on the avoidance of disfavored viewpoints is viewpoint discrimination, which triggers strict scrutiny the rule cannot satisfy under Agency for International Development v. Alliance for Open Society International (570 U.S. 205, 2013), as Section 5E explains.

A scientific conclusion is not advocacy, and the community's value depends on scientists being free to state plainly what their results show, because a field that must self-censor findings to protect its funding produces a compromised and untrustworthy record that others cannot safely build on. When the government can recast an empirical finding as prohibited advocacy, investigators suppress inconvenient results, and the candid early exchange through which a scientific community corrects and refines itself is replaced by caution, which degrades the quality that makes the community worth having.

The deterrent effect is broad because the boundary is undefined, so a rational investigator avoids not only clear advocacy but any statement that might later be characterized as advocacy, which sweeps in ordinary scientific communication about results with policy relevance. A competitor need only wait while American science quiets itself on entire categories of consequential questions, ceding both the findings and the public conversation around them.

In my field, stating that a particular stimulation parameter is unsafe, or that a specific patient population carrying a device is at risk, is a scientific conclusion that a broad reading of this provision could treat as prohibited advocacy, even though patient safety depends on saying it clearly and publicly. Patient safety is the first casualty, because a conclusion that a stimulation parameter is dangerous or that a device population is at risk could be recast as prohibited advocacy, so the finding that would protect patients goes unspoken. The public inherits a compromised research record, since results deterred as advocacy are absent from the evidence base clinicians, regulators, and future researchers rely on, which degrades every treatment later built on that record. Children are specifically exposed where a suppressed finding concerns pediatric safety, and the economy loses the trust in federally funded results that lets industry and regulators act on them. The fix is to remove content-based conditions on funded speech and to limit any restriction to genuine lobbying rules of general application, which address the real concern without silencing science.

International collaboration

P13. [200.220] Blanket ban on covered foreign collaboration. Federal funds, including allocable indirect costs, may not support collaboration with covered foreign countries or affiliated entities, and the covered set reaches well beyond China to broad sanctions and screening lists, while the prohibited activity reaches travel, joint research, data-sharing, and the overhead that supports them, so the ban is categorical rather than targeted at demonstrated threats. The rule treats international collaboration as a vulnerability to be eliminated, when the evidence shows it is a principal mechanism by which the United States builds and maintains its scientific lead.

The definition is also vague enough to reach ordinary domestic research, because the covered set turns on foreign involvement without a threshold that separates a genuine collaboration from an incidental supply-chain link. A project whose personnel and partner companies are all United States based can still fall within the ban if one vendor incorporates a foreign-made component, so the provision exposes work with no security concern to termination on the same footing as a real risk. No rule that cannot distinguish a Chinese co-investigator from a foreign transistor in a microscope satisfies the notice requirement of Grayned that Section 5C develops, and it deters the routine international sourcing that modern instrumentation depends on.

The workforce data alone refute the premise. Foreign-born workers are 19 percent of the United States STEM workforce and 43 percent of doctorate-level scientists and engineers, more than half of the STEM doctorate holders in defense-related industries are foreign-born, and roughly 71 percent of the foreign nationals who earn science doctorates here remain five years later, with about 65 percent still present at ten years. American science leads in part because it attracts and retains the world's best, so a rule that cuts collaboration and signals hostility to foreign-born researchers does not protect the enterprise, it attacks the mechanism that built it.

A large part of the advantage is imported, assembled by drawing top talent to work and collaborate here, and the openness that draws them is the same openness this ban forecloses. International collaboration brings talent in rather than losing it, because a foreign partnership brings methods, datasets, instruments, and people into the American research community, and severing it removes those additions while telling every prospective arrival that the United States has become a less reliable place to build a career. The damage compounds over time, because agglomeration, once disrupted, does not return, it relocates, so the talent and collaborations the ban expels reconstitute in whichever countries keep their doors open and strengthen those systems instead of this one.

The competitor context makes the miscalculation stark. China issued coordinated national guidelines in 2025 designating brain-computer interfaces a strategic industry with milestones for 2027 and 2030, backed the field with an 11.6-billion-yuan brain-science fund, and is building scientific density deliberately. The United States, currently ahead, proposes to expel its own advantage. This loss is already visible at the top of the field, because the 2025 chemistry Nobel laureate Omar Yaghi moved from Berkeley to Tsinghua in 2026 amid United States funding cuts, and a CNN tally counted at least 85 scientists leaving United States institutions for China since early 2024, and every such departure removes a scientist along with the collaborations that ran through them.

The ban also damages ties to allied democracies, whose laboratories are central to neural engineering and whose partnership is a strategic asset, because a blanket prohibition and its associated screening deter exactly the alliance-based collaboration the United States should be deepening, so the rule trades a real cost to allied science for a marginal gain against an adversary that funds its own program regardless. In my laboratory, key methods and datasets in neural engineering are developed jointly across borders, and the foreign-born colleagues who make up nearly half the doctorate-level workforce are integral to the work, so a categorical ban would sever productive collaborations and cut the people who do the science, without a corresponding security benefit. Jobs and the trained workforce are lost first, because foreign-born scientists are a large share of the researchers the public educated, and driving them out shutters the laboratories that hire technicians and sustain local economies. Defense and competitiveness suffer together, since the same methods underlie dual-use capability and future industries, so the ban surrenders strategic and commercial ground to the competitor that receives the departing talent. Treatments are lost with the people who would have developed them, and preparedness declines as the country sheds the expert depth an emergency response draws on. The economy forgoes the discoveries, companies, and tax base those researchers would have generated had they stayed. The fix is to replace the blanket ban with targeted, entity-specific screening grounded in evidence of actual risk, which addresses genuine threats precisely while preserving the openness that is the true source of American advantage.

P14. [200.202(e)] Case-by-case approval of any international element. Any international element in a research award must be affirmatively justified case by case, and foreign entities cannot receive research awards except with senior political-appointee approval, so ordinary international exchange becomes slow and politically contingent even where no plausible security concern exists. Because the stay rate for temporary-visa science doctorates is high, roughly 71 percent at five years and 65 percent at ten, the talent this friction deters is disproportionately talent that would otherwise have remained and contributed to the United States.

Friction is itself a filter, because the world's best scientists have options, and a system that makes routine collaboration and recruitment slow, uncertain, and dependent on political approval loses them at the margin to systems that do not impose the same cost. Every added approval step is a reason for a prospective collaborator, postdoctoral researcher, or faculty recruit to build a career elsewhere, and each such loss removes a potential recruit before they ever join the community, so the provision thins the advantage quietly by suppressing the new arrivals that build it rather than by any single dramatic exclusion.

The requirement also signals unreliability to allied institutions, because a partner who must clear political approval for every joint element is a partner whose participation cannot be counted on, which accelerates the realignment of scientific collaboration toward other hubs that competitors are actively cultivating. The signal compounds the direct effect, since reputations for openness or for obstruction both persist and shape where the next generation chooses to work.

In my field, cross-border method development is routine and often essential, so a requirement of per-element political approval would stall standard scientific exchange and steadily push collaborations toward partners in more open systems, degrading the American research community one deterred partnership at a time. The harm is the talent that never arrives, because making every international element slow and politically contingent deters the recruits and collaborators the country competes to attract, so their future jobs, trainees, and discoveries accrue elsewhere. Competitiveness falls as United States projects are cut off from the global collaborations that accelerate results, isolating domestic work from advances it could have shared. Treatments slow because those partnerships are often how a method or a patient cohort reaches a domestic laboratory, and preparedness weakens as international scientific ties, which matter most in a cross-border health crisis, are made harder to sustain. The fix is to narrow review to entity-specific determinations grounded in evidence and to remove the blanket approval requirement for collaboration with allied and non-covered institutions.

The institution and its people

P15. [200.300] Prohibitions on diversity and gender-related activity. Award funds may not fund, promote, or facilitate diversity policies, gender ideology as the rule defines it, or gender-transition assistance for individuals under 19, and the operative terms are broad and undefined, so institutions face compliance exposure across ordinary operations, student support, training activities, and clinical care that have no connection to the aims of any research award. Facing an undefined prohibition enforced by termination at will, institutions will predictably over-comply to protect their funding, extending the rule's practical reach far beyond its text and failing the fair-notice requirement of Grayned.

The advantage is built by continually widening the pool from which talent is drawn and trained, because scientific density is a function of how many able people a system recruits, develops, and retains, so anything that narrows recruitment or discourages the mentoring of newcomers shrinks the community over a generation. An undefined prohibition that sweeps in ordinary training, outreach, and student-support activities deters exactly the broadening of participation that keeps the American supply of scientists larger and deeper than a competitor's, and it does so at a moment when that competitor is investing heavily to expand its own talent base.

The deterrent effect is the mechanism of harm, because the vagueness of the terms means a risk-averse institution will curtail lawful and productive activities rather than risk a termination it cannot appeal, so the loss is not confined to whatever the drafters intended to prohibit. A supply of scientists narrowed by precaution produces fewer scientists years later, which is a direct subtraction from the community's future density.

In my field, routine training and mentoring activities that help develop the next generation of neural engineers could be swept in by an undefined term, deterring the recruitment and development that build the workforce a decade out. Children are directly affected, both because the provision reaches pediatric care and because a supply of scientists narrowed by precautionary compliance produces fewer of the scientists whose future discoveries would have treated them. The trained workforce shrinks as institutions curtail lawful recruitment and mentoring to protect funding, so the public trains fewer researchers per dollar spent. Treatments are delayed because a smaller and less varied next generation makes fewer discoveries, and competitiveness slips as a competitor expanding its own talent base gains exactly the margin the country gives up. The fix is to define any such term narrowly and against measurable conduct, or to remove the prohibition, so that lawful workforce development is not deterred by uncertainty.

P16. [200.303] Mandatory E-Verify for all award personnel. Every recipient and subrecipient must enroll in E-Verify for all employees and contractors working on a federal award and must report any final nonconfirmation, which imposes a new administrative and surveillance burden on every funded institution in the country and directly contradicts the burden-reduction the rule claims as an objective, a contradiction Section 5B treats as an independent arbitrary-and-capricious defect. The requirement is not calibrated to any demonstrated problem in federally funded research and falls on a workforce that is disproportionately foreign-born.

Because foreign-born scientists are 43 percent of the doctorate-level workforce and more than half of the STEM doctorate holders in defense-related industries, any added employment friction lands hardest on the imported half of the advantage, the very people the United States competes globally to attract and keep. A mandatory verification-and-reporting regime, layered on top of the immigration processes these researchers already navigate, treats them as a compliance risk rather than an asset, and it does so at the moment competitor nations are offering stable, well-funded positions expressly designed to draw them away.

The harm compounds with termination at will, because an institution operating under both a surveillance mandate and revocable funding becomes a markedly less secure place for a foreign-born scientist to build a career, and each researcher deterred at recruitment or displaced mid-career is a scientist the community loses and cannot easily replace, since a mid-career specialist embodies years of accumulated training and many established collaborations. The reporting requirement also creates a mechanism by which routine administrative outcomes become grounds for federal attention, deepening the instability the rule's critics identify.

In my laboratory, which depends on foreign-born scientists who make up roughly 43% of the doctorate-level research workforce nationally, added employment friction directly threatens both the team and the continuity that multi-year neural-interface work requires, because a project that loses a key specialist mid-course may lose years of momentum. The trained workforce is deterred, because the provision falls hardest on the foreign-born researchers the country competes to keep, so it pushes out the people who produce treatments. Institutional costs rise as every funded recipient must build and run the screening system, an administrative overhead the public ultimately pays out of the same appropriations, diverting money from research to paperwork. Treatments are delayed as the projects deterred or displaced specialists would have carried stall, and each mid-career researcher lost is years of public training investment forfeited, so the economy loses both the sunk investment and the future output. The fix is to remove the government-wide E-Verify mandate, or at minimum to limit it to what existing law already requires, so the rule does not impose a uniform burden that falls hardest on the talent the country is trying to retain.

P17. [200.206] Affiliation-based denial of applications. Agencies may deny applications based on an applicant's membership in or affiliation with organizations deemed to undermine national security or public safety, which shifts the basis of decision from the merit of the proposal to the associations of the applicant and compounds the First Amendment concerns Section 5E develops. Because the disqualifying categories are undefined, the provision lets an agency exclude a disfavored applicant without articulating a reason a court could test, so it operates as unreviewable discretion over who may receive federal research support.

Screening scientists by association rather than conduct attacks the community at the level of who is permitted to belong to it, because the collaborations, consortia, and professional societies that constitute the advantage are precisely the affiliations a broad and undefined standard could later recast as disqualifying. A researcher who must weigh whether an ordinary professional membership might one day be treated as a security disqualification will limit the associations that produce collaboration, so the provision deters the collaboration and idea-sharing on which the community's productivity depends.

The undefined standard also invites selective and viewpoint-based enforcement, because an agency can characterize the affiliations of applicants it wishes to exclude as threatening while ignoring identical affiliations of applicants it favors, which corrodes the merit competition that produces scientific quality. The harm is both to the excluded individual and to the field, which loses whatever that individual and their collaborations would have contributed.

In my field, an investigator's membership in a scientific or professional society could be recast as a disqualifying affiliation under a sufficiently broad reading, deterring the participation that consensus-standards work and collaborative research require. Public funds buy exclusion rather than merit, because screening applicants by association applies a filter unrelated to whether the science is sound, so the money is spent narrowing the pool instead of funding the best work. Treatments slow as the collaborations and society memberships such screening deters are the ones that generate discoveries, so cutting them delays the research that becomes therapy. Competitiveness falls as productive partnerships move to systems without the filter, and the economy loses the output of the collaborative scientific activity the provision suppresses. The fix is to remove affiliation-based denial and to rely on conduct-based, evidence-tested national-security review that addresses genuine risk without screening scientists by association.

P18. [2 CFR restructure and 31 U.S.C. 503] Binding OMB authority across 456 sections. The rule converts 2 CFR from guidance into binding regulation by deleting the longstanding statement that publication in the Code of Federal Regulations does not change the nature of guidance, and it replaces guidance with regulation across 456 sections in a single action effective government-wide on one date, without separate rulemaking by the agencies that will administer it. OMB rests this on 31 U.S.C. 503, a financial-management coordination provision, rather than on any statute granting it power to displace the enabling acts Congress wrote for each science agency, which raises the major-questions and non-delegation problems Section 5D develops.

The agency-level tailoring the rule removes is what allows the most collaborative, cross-trained research to be funded at all, because interdisciplinary work spans agencies with different missions and different rules, and each agency's ability to adapt its terms to the science it supports is what accommodates a project that fits no single mold. Collapsing every agency into one uniform, self-executing regulation forces the boundary-crossing researcher, who is the characteristic product of a dense community, to satisfy a rigid framework built for none of their fields, which pushes science back toward narrow single-agency silos and away from exactly the cross-disciplinary work a competitor cannot yet reproduce.

Centralizing this authority in a single office also concentrates the power to reshape the entire federal research enterprise in one place, removing the distributed, statute-specific judgment that has let different agencies serve different scientific communities in different ways. Uniformity imposed from the center is not neutral, because it privileges the kinds of work that fit the common template and disadvantages the interdisciplinary and the unconventional, which are the community's highest-value output.

My own program is funded across agencies with distinct statutory missions and distinct implementing rules, because neural-interface research spans engineering supported through NSF and the Department of Defense and biology supported through NIH. Each agency currently tailors its requirements to the science it funds, so collapsing those frameworks into one regulation would remove the tailoring that lets a single interdisciplinary program satisfy an engineering sponsor and a biomedical sponsor at once. Compliance costs rise for every recipient, because collapsing agency-specific rules into one framework adds overhead the public funds out of the same dollars, so less of each award reaches research. Treatments that cross disciplines are lost, since the provision forces interdisciplinary work into single-agency silos, which is where the highest-value boundary-crossing discoveries occur. Competitiveness declines as the country cedes the cross-field advantage a competitor cannot yet match, and the economy forgoes the industries that interdisciplinary therapies and technologies would have built. The fix is to retain the guidance status of 2 CFR Subtitle A and to preserve agency-level implementation under each agency's authorizing statute, so that coordination does not become the displacement of the statutes Congress enacted.

P19. [reporting provision] Recipient-level reporting to OMB. OMB may require agencies to report the specific recipients of federal awards over any period it designates, which gives the Executive Office of the President direct visibility into, and control over, which institutions receive federal research funding, a function historically insulated from centralized political control. Recipient-level oversight of this kind invites political targeting of disfavored universities, because the information exists precisely to enable intervention, and combined with termination at will it converts a reporting requirement into a mechanism for conditioning institutional funding on political considerations.

Universities are the physical centers of the community, the places where scientists and trainees concentrate and where the collaborations, seminars, and cross-training that constitute the advantage occur, so a tool that lets the government identify and pressure specific institutions threatens the community at its foundations. An institution that comes to believe its funding turns on the political acceptability of its faculty's findings will steer away from controversial or inconvenient lines of work, and the independence that lets a dense community pursue evidence wherever it leads gives way to institutional caution.

The deterrent effect operates through anticipation, because administrators, faculty, and recruits adjust their behavior to protect the institution's funding long before any explicit intervention occurs, so the mere existence of targeted recipient tracking reshapes what universities are willing to support. A set of institutions optimizing for political safety produces less of the ambitious, boundary-crossing science that is the American advantage.

In my field, a university whose faculty publish findings inconvenient to a given administration could be identified and pressured through its research funding, deterring both the independence research depends on and the willingness of its scientists to share results early and openly, which is the behavior the advantage is built on. Public oversight is misused, because targeted institutional tracking spends the government's monitoring capacity on political pressure rather than genuine transparency, so the resource is diverted from its purpose. Treatments are lost as the universities that sustain the research workforce steer toward political safety, declining the ambitious work that yields therapies. Competitiveness falls as institutions choose caution over discovery, and the economy and its jobs suffer because the campuses being pressured are also the largest research employers in their regions, whose retreat from bold science reaches the communities built around them. The fix is to limit recipient reporting to aggregate, statutorily grounded transparency rather than targeted institutional tracking that can be turned to political ends.

8. Responses to the justifications OMB is likely to offer

A final rule must respond to the justifications the preamble and its supporting documents advance, and each justification below is mapped to the provisions and legal grounds that refute it. OMB cannot adopt any of these rationales without engaging the specific refutation, and adopting a justification this comment has already answered, without addressing the answer, is itself arbitrary action under State Farm.

Accountability.

Peer review already supplies accountability, because independent experts evaluate every proposal against the state of the field and rank it on the record, so replacing that expert judgment with political control inverts accountability rather than strengthening it, substituting the preference of a single appointee for the assessment of the scientific community. The deeper harm is to transparency, because scored peer review states the reason for each decision, whereas an appointee who may cancel and re-compete until a preferred applicant prevails need never state one, which returns the unaccountable patronage peer review replaced, as P2 explains. Accountability for spending is separately enforced through audit, reporting, and misconduct proceedings, none of which requires overriding scientific merit. This rationale is refuted by P1, P2, and P3 and by Section 5A, so adopting it obligates OMB to explain why those existing mechanisms do not already provide the accountability claimed.

Stewardship of taxpayer funds.

The return record in Section 2 shows that expert-directed funding is the best available stewardship, returning roughly $2.57 for every NIH dollar and roughly $1 trillion on the Human Genome Project's $3.8 billion investment. Terminate-at-will is the opposite of stewardship, because canceling a multi-year project midstream forfeits the sunk investment and produces no result, so genuine stewardship maximizes the return on funds already committed by letting sound projects finish. Beyond the sunk investment, termination raises net cost rather than lowering it, because it draws litigation the public funds on both sides and can end with the government paying the award plus the cost of the fight, alongside shutdown, restart, and irrecoverable-data losses no verdict restores, as P4 develops. This rationale is refuted by P4 and by Section 5F, so adopting it obligates OMB to explain how canceling funded work midstream conserves rather than wastes public money.

National security.

Section 6 shows that the rule weakens the security posture Congress is building, because a blanket international-collaboration ban relocates the scientists the taxpayer trained and terminate-at-will destabilizes the domestic base, while surveillance and affiliation screening give competitor recruitment programs exactly the instability they advertise against. Security is served by targeted, evidence-based screening of specific entities and individuals, not by conditions that drive talent to competitors. This rationale is refuted by P13, P14, P16, and P17 and by Section 6, so adopting it obligates OMB to explain how handing a dual-use lead to a state-coordinated rival advances national security.

Democratic accountability over unelected scientists.

Elected officials already set scientific priorities and budgets through appropriations and authorizing statutes, which is the constitutional channel for democratic control of science funding, so the rule's insertion of appointee judgment into individual merit decisions is a different and illegitimate act, because it politicizes the evaluation of technical quality rather than the setting of national priorities. This rationale is refuted by P1, P2, and P5 and by Section 5D, so adopting it obligates OMB to identify the authority that lets political officers override expert assessments of scientific validity.

Reducing administrative burden.

The rule increases burden rather than reducing it, because it adds mandatory E-Verify enrollment, conference pre-approval, membership pre-approval, and documentation of political alignment across the award lifecycle, so an agency cannot claim burden reduction while imposing new procedural requirements at every stage. This rationale is refuted by P8, P10, P16, and P1 and by Section 5B, which treats the internal contradiction as an independent arbitrary-and-capricious defect, so adopting it obligates OMB to reconcile the claim with the new requirements the rule imposes.

Scientific rigor.

The rule invokes Gold Standard Science as though it advanced rigor, yet it adopts none of the measures that improve rigor and reproducibility, because rigor is strengthened by preregistration, by funding replication and negative results, by statistical review of study designs, and by open sharing of data and code, none of which the rule requires or supports. Substituting political judgment for expert review and cutting the publication and data-sharing costs that rigor depends on moves in the opposite direction. This rationale is refuted by P3, P1, and P2 and by Section 5C, so adopting it obligates OMB to explain how dismantling peer review and defunding open data advance the rigor it claims.

Transparency and preventing waste and fraud.

The rule presents itself as advancing transparency and preventing waste, yet it permits competitions to run outside public view, directs recipient-level reporting that enables targeting rather than genuine transparency, and cancels multi-year projects midstream, which wastes the sunk public investment. Transparency is served by public posting and aggregate reporting, and waste is prevented by letting sound projects finish, so this rationale is refuted by P4, P6, and P19 and by Section 5G, and adopting it obligates OMB to reconcile the claim with provisions that reduce transparency and increase waste.

9. Collateral harm to Americans who never held a grant

Training of new scientists and a lost generation.

The training cycle from graduate admission to independent investigator runs 10 to 15 years, so damage now produces generational rather than episodic loss. The former NIH director Francis Collins has warned that the United States risks losing an entire generation of young scientific talent, and the data already show contraction, with the R01 success rate falling from 21.6 percent in FY2023 to 13 percent in FY2025 and R01 awards falling from 7,720 in FY2024 to 5,885 in FY2025. My laboratory's central discoveries, that chronic neural-interface failure is driven by oligodendrocyte and myelin loss, were made by American graduate students and postdocs whose training taxpayers already funded and who have since moved into independent United States research positions. Terminate-at-will [200.340] plus isolation from journals, conferences, and collaborators pushes these people out of science, forfeiting the public investment already spent to create them.

Loss of the best United States-trained scientists to China and other competitors.

The loss of scientific talent is measurable and now reaches the most senior figures. In July 2026 the 2025 Nobel laureate in chemistry, Omar Yaghi, left the University of California, Berkeley, after nearly five decades in the United States to lead an artificial-intelligence materials institute at Tsinghua University, and he tied the decision in part to United States funding cuts. Below the Nobel tier, a CNN tally identified at least 85 established and rising scientists who left United States institutions for Chinese research institutions full-time since early 2024, more than half of them in 2025, including a National Institutes of Health neurobiologist.

These departures are enabled by the conditions this rule intensifies. A March 2025 Nature survey found more than 75 percent of United States-based scientists considering leaving the country, rising to nearly 80 percent among early-career researchers. China's research-and-development spending, meanwhile, reached roughly 96 percent of the United States level in purchasing-power-parity terms in 2023, up from 72 percent a decade earlier, and OECD estimates it surpassed the United States in 2024. The durable United States advantage in brain-computer interfaces rests on the density of trained scientists and cross-institutional expertise that converts basic neuroscience into devices, not on publication count, and cutting the training of scientists does not stop the science, it relocates the people the taxpayer trained, so the next generation of neural-interface breakthroughs becomes likelier in Beijing than in Pittsburgh.

Allied research relationships.

The international-collaboration ban [200.220] reaches beyond adversaries, because its breadth and case-by-case approval requirement deters partnerships with allied democracies whose laboratories are central to neural engineering, including institutions in the United Kingdom, the European Union, Japan, Canada, and Australia. Severing those relationships weakens the alliances that multiply United States research capacity while doing little to stop a determined adversary that funds its own program. The ban therefore trades a real cost to allied science for a small security gain, which inverts the tradeoff a targeted restriction would achieve.

Clinical trials and patients left with implanted devices.

Human brain-computer interface trials implant electrode arrays directly into the cortex, and participants volunteer for multi-year studies on the reasonable expectation that funded work will run its course. Terminate-at-will [200.340] can end such a trial mid-study with no misconduct finding and no appeal, leaving volunteers carrying implanted intracortical devices with no funded clinical team to monitor, maintain, or surgically remove them. Explantation is itself a neurosurgical procedure that carries risk, so abandonment is an ethical and human-subjects catastrophe that also exposes universities and the government to liability. It further destroys future recruitment, because no one will volunteer for a cortical implant once funding is known to be revocable at will, which would end the human trials on which the field depends.

Patients and veterans waiting on these treatments.

The people who lose most are patients who never see the trial reach them. Brain-computer interfaces are being developed to restore movement and communication after spinal-cord injury, stroke, and amyotrophic lateral sclerosis, and to restore touch through prosthetic limbs, a capability with direct relevance to wounded veterans and to the Department of Defense and Veterans Affairs research that has long funded this field. Terminating that research mid-course does not save the cost of care, it defers the treatments that would reduce lifelong disability, so the harm lands on paralyzed patients, injured service members, and the families who care for them.

Children and the future of treatment.

The treatments this research builds toward include control of drug-resistant epilepsy and therapy for dementia, because the same myelin biology that governs implant failure also governs neurodegeneration. My group showed that protecting oligodendrocytes reduces neuronal damage and attenuates disease progression in an Alzheimer's model. Pediatric and developmental applications depend on the standing research capacity this rule dismantles, so children who would benefit from treatments a decade away lose them before the work can mature.

Jobs, regional economies, and healthcare inflation.

Federal research is among the highest-return public investments available. Every dollar of NIH funding returns roughly $2.57 in economic activity (United for Medical Research), NIH is about 0.7 percent of the federal budget, and federally funded research is traceable in nearly every new drug the FDA approved between 2010 and 2019. The Human Genome Project returned an estimated $1 trillion on a federal investment of about $3.8 billion, and roughly $250,000 in federal funding for a mid-century insect-sterilization program has saved the United States livestock industry billions since. NIH funding reaches every state and employs technicians, nurses, coordinators, and domestic manufacturers, so the near-term effect of cutting it is lost American jobs across the country, not only in coastal universities. Because medical innovation lowers the long-run cost of treating disease, suppressing it keeps expensive chronic care in place and raises healthcare inflation while the research spending relocates to competitors.

Collapse of the peer-review system that oversight depends on.

Tens of thousands of scientists review proposals and manuscripts without pay because the system is trusted to measure merit, and that volunteer labor is what makes federal oversight of scientific quality possible at all. Reducing peer review to advisory status [200.205] removes the reason experts donate that time, because a review that an appointee may disregard carries no weight. As experienced reviewers withdraw, the government loses the very capacity it would need to evaluate proposals competently, which deepens rather than cures the problem the rule claims to address.

Regional research institutions and later private-sector research.

NIH and NSF funding sustains research at public and private institutions in every state, many of which sustain regional economies and train the local technical workforce, so instability hits these institutions harder than the largest and best-endowed universities. Private industry depends on that academic base for trained workers and for the foundational discoveries it does not fund itself, so cutting the base raises costs and slows development across the private sector. The harm therefore reaches companies and communities that never received a federal grant.

Crisis preparedness.

Rapid response to the last pandemic rested on decades of federally funded basic research. The messenger-RNA vaccine platform depended on work that sat largely unfunded for close to thirty years, and population-scale COVID-19 testing depended on PCR, itself descended from a curiosity-driven NSF grant. Political control plus terminate-at-will destroys the standing capacity a nation needs before the next biological or technological threat, because preparedness cannot be surged on demand and has to be maintained continuously.

Litigation costs paid by taxpayers.

The rule's legal fragility guarantees litigation, and the Department of Justice will defend it with taxpayer money. The indirect cost cap was proposed, litigated, and struck down (First Circuit, January 5, 2026), so taxpayers paid twice, once for the disruption the attempted change caused before and during the litigation and again for the failed legal defense. This rule repeats that pattern at government-wide scale across NIH, NSF, DOE, NASA, DOD, and ARPA-H simultaneously, which multiplies both the litigation exposure and the interim damage.

State and local economies.

NIH and NSF awards support jobs in every state, and the effect concentrates in the research institutions that sustain regional economies. Each award funds not only investigators but technicians, nurses, study coordinators, animal-care staff, and domestic suppliers of reagents and instruments, so a terminated grant eliminates a chain of local employment rather than a single position. Because every NIH dollar returns roughly $2.57 in economic activity (United for Medical Research), the near-term effect of destabilizing this funding is measurable job loss across congressional districts that host universities and academic medical centers, most of which sit outside the largest coastal metros. The harm therefore lands on communities and workers with no direct connection to any research question.

Loss of measurement and standards infrastructure.

Progress in neural interfaces is limited by the absence of agreed standards for electrode-tissue interface stability and cortical stimulation dose-response, which is why the FY2027 NDAA directs NIST to build a biometrology program, as Section 6 notes. Developing those standards requires a stable base of federally funded academic research generating comparable data across laboratories, because standards are built from accumulated measurements rather than declared by fiat. Terminate-at-will [200.340] and the collaboration restrictions fragment that base, so the rule undercuts the measurement infrastructure the same government is trying to create. The consequence is slower and less reliable evaluation of every device that follows, which raises cost and risk across the field.

Credit ratings and the cost of public borrowing.

A major credit-rating agency has already assessed the rule as credit negative for institutions dependent on competitive federal funding, in a June 2026 analysis reported in the municipal-bond press. Because roughly 36 cents of every state dollar comes from federal grants, and because universities and academic medical centers borrow against expected research revenue, a rule that makes that revenue terminable at will [200.340] raises borrowing costs for hospitals, universities, and the state and local governments tied to them, which reaches taxpayers who never held a grant. This is an independent, non-scientific harm that OMB must weigh, and the reliance interests it implicates are the kind State Farm requires the agency to address, so leaving it unexamined is a further arbitrary-and-capricious defect.

The device-approval pathway depends on this research.

Implantable brain-computer interfaces reach patients through the FDA device pathway, and that pathway depends on the federally funded research this rule would destabilize. A novel implantable interface is a significant-risk device, so it requires an Investigational Device Exemption before human testing and, for market authorization, a Premarket Approval supported by clinical and preclinical safety data. The preclinical safety data behind these submissions come substantially from federally funded academic work, because the multi-year animal studies that establish stimulation dose thresholds, chronic biocompatibility, and failure modes are not studies private developers fund themselves.

A six-week safety signal my group recorded, seizure-like activity in half of animals stimulated at the most common clinical frequency, is exactly the kind of result that informs a device submission. I contribute to the FDA-recognized implantable brain-computer interface collaborative community, which develops consensus standards and methods for evaluating these devices, and that standards work presupposes a stable academic base generating the underlying data. Cutting that base does not only slow discovery, it removes the evidentiary foundation on which the FDA approves the devices, so the rule reaches the regulatory pathway itself. The measurement standards the FY2027 NDAA directs NIST to develop, addressed in Section 6, are the standards this research base is needed to inform, which ties the device pathway to the national-security strategy Congress is enacting.

Clinical populations and unmet need.

The treatments this research builds toward address conditions affecting millions of Americans, and the scale of that unmet need is quantifiable, with the figures below drawn from standard public-health estimates.

Spinal cord injury. About 300,000 Americans live with spinal cord injury, with roughly 18,000 new injuries each year, and about 46,000 are veterans. Brain-computer interfaces are being developed to restore movement and communication for this population.

Stroke. Roughly 795,000 strokes occur in the United States each year, many leaving lasting motor or speech impairment that neural interfaces aim to address.

Amyotrophic lateral sclerosis. About 5,000 Americans are diagnosed with ALS each year and roughly 30,000 live with it, a population for whom communication interfaces can restore the ability to speak.

Drug-resistant epilepsy. About 3.4 million Americans have epilepsy, and roughly one third do not respond to medication, so closed-loop neural devices offer a route to seizure control that drugs cannot provide.

These are the people a mid-study termination abandons and the treatments a defunded research program delays. The lifelong cost of untreated disability dwarfs the cost of the research designed to reduce it, so cutting the research raises the long-run burden on patients, families, and public programs.

10. Worked examples of discovery chains under this rule

One discovery chain from my own laboratory shows concretely how the rule would have prevented results now entering clinical use. Federal funding supported basic imaging of how brain tissue responds to an implanted electrode, work with no near-term product and an uncertain outcome. That imaging revealed that damage begins within seconds of insertion and varies with which blood vessels are grazed, a finding that informed the image-guided surgical placement used in current commercial systems. Separate federal funding supported the counterintuitive idea that an oligodendrocyte, a cell that makes myelin rather than a cell that records signals, controls long-term recording quality.

Testing that idea produced the result that clemastine, an FDA-approved antihistamine repurposed to promote remyelination, improves chronic electrode performance, and the same biology reduced neuronal damage in an Alzheimer's model. Under Gold Standard Science review [200.205], a nonexpert would likely have scored basic tissue imaging and a myelin hypothesis as insufficiently aligned with a present priority, and terminate-at-will [200.340] would have ended the multi-year work before the clinical payoff emerged. The chain shows the pattern in Section 2 in concrete form, value invisible at funding, realized years later, and destroyed by present-alignment review.

11. Workable alternatives

Courts give more weight to comments that propose a workable alternative than to comments that only object. The following alternatives preserve legitimate oversight and national-security screening while curing the statutory and constitutional defects above:

1.  Retain binding peer review as the basis for merit decisions, and limit political review to statutory-compliance and national-security screening that does not override or substitute for scientific scores.

2.  Preserve publication, journal, and conference costs as allowable, so the government can continue to satisfy its own public-access mandate and investigators can access and contribute to the literature.

3.  Replace termination without cause with termination for documented cause plus an appeals process, and expressly protect active human-subjects trials from mid-study termination absent a safety finding.

4.  Define any compliance standard operationally, with measurable criteria and adequate notice, before it is used to condition awards, recognizing that the value of foundational research is only visible years after funding.

5.  Narrow international restrictions to entity-specific, evidence-based determinations rather than blanket country bans that sever partnerships foundational to United States leadership.

6.  Replace blanket public-posting exemptions and recipient-level reporting with narrow, documented national-security carve-outs, so transparency remains the default the rule claims to serve.

7.  Align the rule with the biotechnology national-security priorities Congress is enacting in the FY2027 NDAA, rather than making the underlying research base terminable at will.

12. Requested action

Withdraw the proposed rule, and if OMB proceeds, respond on the record to each statutory conflict identified above and adopt the alternatives in Section 11. The post-World War II compact in which the government funds science but does not direct its conclusions produced American leadership in medicine and technology, and it produced that leadership precisely because it funded work whose value could not be predicted at the outset. This rule replaces that compact by regulation, on a compressed timeline, and the record should reflect that a working scientist identified the specific statutory conflicts, the specific harms to patients, veterans, children, and taxpayers, and the specific fixes before the rule was finalized. If OMB does not withdraw the rule, it should at a minimum extend the comment period in proportion to the rule's scope and reissue the pre-issuance-review, termination, speech, and foreign-collaboration provisions as separate rules, each with its own record, so the public can comment on each in the detail its consequences require.

Additional traced examples at the national level.

The chains summarized in Section 2 are traced here in full. Each case follows a federally funded project from an origin that no present-alignment reviewer would have scored as valuable to a return measured decades later, which is the pattern Gold Standard Science review [200.205] and terminate-at-will [200.340] cannot accommodate.

The intracortical electrode array and restored speech.

Federal support from NIH and the Defense Advanced Research Projects Agency in the late 1980s funded Richard Normann's development of the intracortical microelectrode array, a bed of silicon needles for recording from cortex, with no clinical product in view. The array sat in academic laboratories for roughly two decades before the first human intracortical trials began in 2004, and the same design later allowed a participant with amyotrophic lateral sclerosis to produce attempted speech decoded at about 62 words per minute in 2023. The chain from a basic recording device to restored communication ran more than thirty years and crossed engineering, neuroscience, and clinical medicine, so a Gold Standard Science benchmark applied at the outset [200.205] could not have scored it and terminate-at-will [200.340] applied at any stage would have ended it. This example sits in the exact field this comment addresses, so the risk the rule creates is demonstrated rather than hypothetical.

Polymerase chain reaction and Thermus aquaticus.

An NSF grant in the 1960s supported Thomas Brock's study of microbial life in the hot springs of Yellowstone National Park, work with no medical or commercial objective. Brock and Hudson Freeze discovered Thermus aquaticus in 1966, a bacterium that thrives near boiling temperatures, and characterized the heat-stable enzymes that let it survive. Two decades later, that heat-stable polymerase became the core of the polymerase chain reaction, because the reaction requires an enzyme that survives repeated heating and cooling. Kary Mullis received the 1993 Nobel Prize in Chemistry for the technique, which now underlies a global market exceeding $15 billion, every COVID-19 PCR test, forensic DNA analysis, and routine clinical diagnostics. No reviewer in 1966 could have scored hot-spring microbiology as aligned with a diagnostic priority, because the diagnostic application did not yet exist, so the return depended precisely on funding work whose value could not be stated in advance.

CRISPR gene editing and Casgevy.

Federally funded curiosity about an odd pattern of repeated DNA sequences in bacteria, first noted in the late 1980s and pursued through the 2000s, revealed a bacterial immune system that cuts viral DNA at programmed sites. Basic research on that system, supported across NIH and NSF, established that it could be directed to edit DNA at chosen locations, which opened genome editing as a field. In December 2023, the FDA approved Casgevy, the first CRISPR-based medicine, for sickle cell disease, a condition affecting roughly 100,000 Americans. The path from bacterial oddity to approved therapy ran more than three decades and crossed microbiology, structural biology, and clinical medicine, so terminate-at-will [200.340] applied at any stage would have ended the chain before the therapy existed.

Messenger RNA and the COVID-19 vaccines.

Katalin Kariko pursued messenger RNA as a therapeutic tool through the 1990s and 2000s against sustained skepticism, and her funding was repeatedly denied as the work was judged unpromising. A central advance, chemically modifying mRNA so the immune system tolerates it, was published in 2005 and drew little immediate attention. That foundation enabled the two mRNA COVID-19 vaccines deployed at scale in 2020 and 2021, and Kariko shared the 2023 Nobel Prize in Physiology or Medicine. A project a nonexpert would have defunded as misaligned, judged by the standards of its moment, became the basis of a national pandemic response fifteen years later, which is the exact outcome present-alignment review forecloses.

Screwworm eradication and the livestock economy.

Federal entomology funding in the mid-twentieth century, on the order of a few hundred thousand dollars, supported Edward Knipling and Raymond Bushland's work on the sterile insect technique against the New World screwworm, a parasite that consumes living livestock. Releasing sterilized male flies collapsed the pest population without pesticides, eradicating the screwworm from the United States. The Department of Agriculture estimates the program saves the livestock industry on the order of $1 billion or more each year, a return of thousands to one on the original federal investment. A reviewer asked to fund research on insect reproduction would have struggled to connect it to any priority, which again illustrates the lag between discovery and measurable return.

Four cases do not prove that every project returns value, and most do not. The point is narrower and decisive, because the projects that produced the largest returns were indistinguishable at the time of funding from projects that returned little, so no benchmark applied at the award can separate them. A system that funds only what looks valuable today therefore forfeits the returns that define the record, which is why the merit judgment must rest with experts who can assess scientific quality rather than present alignment.

13. Anticipated responses to this comment, and why each fails

OMB and its reviewers, including any automated system used to draft a response, will raise a predictable set of objections to this comment. Each is stated below in the form it will take and answered on the record, so that proceeding requires rebutting the answer rather than restating the objection, and so that a reviewing court can see the objections were anticipated and met. The comment period affords a single opportunity to build this record, so each objection is answered here in advance, and any answer OMB does not squarely rebut in the final rule stands as an unaddressed significant comment under 5 U.S.C. 553 and Motor Vehicle Manufacturers Association v. State Farm (463 U.S. 29, 1983).

A reviewer may respond that the harms described here are speculative. They are documented rather than predicted, because the relevant events have already occurred. These include more than 1,800 grant terminations since early 2025, a decline in the National Institutes of Health R01 funding rate, the departure of the 2025 chemistry Nobel laureate and other senior researchers to Chinese universities, and a federal court's decision striking a related funding change, so the pattern this rule accelerates is on the record.

It will be said that peer review is preserved and only made advisory. Advisory review a political appointee may disregard is not merit review in any operative sense, because the notice forbids the appointee from deferring to the panel, so where scientific judgment and political priority conflict the panel carries no weight, which is the arrangement 42 U.S.C. 289a was written to prevent.

The rule will be defended as adding oversight rather than banning good science. Several provisions remove capabilities rather than review them, because they make publication, conference attendance, professional-society membership, and collaboration with named foreign entities unallowable outright, so the effect is prohibition of ordinary scientific practice, not added scrutiny of it.

OMB may point to its Regulatory Impact Analysis as proof it weighed costs. That analysis states its own benefits are not readily quantifiable and does not estimate the losses this rule imposes, among them canceled trials, forgone treatments, departed researchers, and ceded competitiveness, so it does not reflect the reasoned consideration of costs and alternatives that Motor Vehicle Manufacturers Association v. State Farm (463 U.S. 29, 1983) requires.

The objections will be characterized as policy disagreements within executive discretion. The defects identified here are legal rather than matters of policy, because Loper Bright Enterprises v. Raimondo (603 U.S. 369, 2024) ended deference to an agency reading of its own authority, and these objections rest on statutory conflict, viewpoint discrimination, non-delegation, and procedural inadequacy, none of which a policy preference can cure.

A response may assert that any defective provision is severable, leaving the rest intact. The central mechanism, appointee override of expert review, is itself unlawful and runs through the entire award lifecycle, so the instrument cannot operate coherently once it and the terminate-at-will, speech, and collaboration provisions fall, which is why vacatur of the whole is the remedy under Section 5I regardless of any severability clause.

The foreign-collaboration limits will be defended as protecting national security. The density and openness of the American research system is itself the security advantage, because it is what draws and holds the world's scientists, so a categorical ban that drives that talent and its dual-use methods to a competitor weakens security rather than protecting it, as Section 6 develops against the strategy Congress is enacting.

Opposition from universities and scientists will be discounted as self-interest. The harms set out here fall on patients awaiting treatment, volunteers carrying experimental devices, taxpayers whose funds are wasted, and national competitiveness, not on the commenter, and the opposition on the record reaches well beyond grant recipients to patient organizations, industry, and tens of thousands of members of the public, so the objections stand apart from any interest a grantee holds.

The rule will be described as applying only to new awards after October 1, so that no ongoing trial is terminated. In fact the termination and related authorities attach both to new awards and to new funding actions on existing multi-year awards, so a continuing grant is exposed at each yearly increment, and the prospect of cancellation without cause deters trial enrollment and long-horizon work immediately, before any single award ends.

OMB may argue that national-interest exemptions and agency-head waivers preserve flexibility, so posting and collaboration are not truly restricted. A discretionary waiver held by a political appointee is not equivalent to a rule of general applicability, because the delay and uncertainty of seeking one deter the activity, and lodging the waiver with an appointee reproduces the pre-issuance-review defect of Section 5 rather than curing it.

It may be said that making peer review advisory reflects the President's authority over the executive branch. Section 289a of Title 42 is a statutory command that peer review occur, and Article II authority does not displace a statute, because the spending power belongs to Congress, so an executive preference cannot override the review Congress required, which is the separation-of-powers defect in Section 5D.

The foreign-collaboration prohibition will be compared to ordinary export controls. Targeted, entity-specific export controls already exist and rest on evidence, whereas a categorical and undefined government-wide ban differs in kind, because it sweeps in benign collaboration with no security nexus, the overbreadth the fix to Section 7 would correct by narrowing to entity-specific determinations.

The Gold Standard benchmark will be cast as a requirement of scientific rigor, which this comment claims to support. Rigor is already the object of expert peer review, so a second and undefined standard applied by nonexperts adds none, and the notice ties the benchmark to presidential priorities, which reveals that it operates as an alignment filter rather than a quality standard and fails the fair-notice test of Section 5C.

The DEI and gender provisions will be justified as enforcing existing civil-rights law. Existing law already prohibits discrimination, so a provision enforced by termination against undefined terms reaches lawful recruitment, mentoring, and pediatric care that add to the trained workforce taxpayers fund, and the vagueness rather than any anti-discrimination purpose is the defect examined in Section 5C.

The historical examples will be dismissed as cherry-picked winners that ignore the many projects returning nothing. This comment states plainly that most projects do not return measurable value, so the argument does not depend on the winners being typical. The decisive point is that those winners were indistinguishable from the rest at the time of award, so no benchmark applied at award can separate them, which is a limit on present-alignment review rather than a claim that all basic research pays off.

A reviewer may argue that this comment cannot both defend open international collaboration and treat China as a competitor. The advantage defended here is the openness and density that draw the world's scientists to the United States, so remaining the most attractive place to do science is how the country stays ahead of a competitor, and genuine threats are met by the targeted, evidence-based screening the fixes preserve, whereas a categorical ban forfeits the advantage while accomplishing little a targeted measure could not.

The rule will be defended as fiscal stewardship that saves taxpayer money, when it does the opposite where the money is largest. Terminating awards generates litigation the public funds on both sides, and under the two-forum scheme of National Institutes of Health v. American Public Health Association a wrongful termination can end with the government paying the terminated amount plus the cost of the fight. Those payments come on top of the shutdown, restart, lost-data, and departed-talent costs no verdict recovers, none of which the Regulatory Impact Analysis counts, as Section 7 details at the termination provision.

A common response holds that research worth doing will attract private funding, so public support is unnecessary. Private capital cannot substitute for it on the work that matters most, because the returns from basic research arrive on a multi-decade lag and cannot be captured by the firm that pays for them, since competitors read the same findings. Venture funds run on a seven-to-ten-year horizon, and an individual investor expects a return within a working lifetime. Neither will finance the thirty-year path that modified mRNA traveled before the COVID vaccines, or the decades PCR needed after a 1966 Yellowstone grant. Large pharmaceutical companies face the tightest horizon of all, because quarterly-earnings pressure, executive compensation that vests over a few years, and shareholders saving for or drawing down retirement all reward near-term returns. These firms therefore concentrate on late-stage development and acquire discoveries already de-risked by public science, rather than funding the basic research that precedes a drug candidate. A twenty-year patent clock that starts at filing reinforces the pattern, because a discovery needing fifteen years of pre-commercial work would reach the market with little exclusivity left, so a rational firm lets public money carry those years and enters afterward. The record matches the economics, because every example in Section 10 reached the market only after public funding carried it through the pre-commercial decades no private backer would underwrite, the market failure the 1945 compact was built to correct.

14. Paths to reversal for Congress, the courts, and a future administration

If OMB finalizes the rule despite this record, the record is built to support reversal through each channel open to the actors who can undo it. The paths below are stated so that a litigant, a legislator, or a later administration can act on this comment directly.

For the courts and litigants.

The grounds most likely to support vacatur are developed in Section 5. The rule conflicts with the enabling statutes Congress wrote for the science agencies, judged without deference after Loper Bright Enterprises v. Raimondo (603 U.S. 369, 2024). Its stated aims are contradicted by its own provisions, which makes it arbitrary and pretextual under Motor Vehicle Manufacturers Association v. State Farm (463 U.S. 29, 1983) and Department of Commerce v. New York (588 U.S. 752, 2019). Conditions the rule places on funded speech are viewpoint-based, triggering Agency for International Development v. Alliance for Open Society International (570 U.S. 205, 2013) and Legal Services Corporation v. Velazquez (531 U.S. 533, 2001). OMB's assertion of government-wide authority raises delegation and major-questions problems under West Virginia v. Environmental Protection Agency (597 U.S. 697, 2022), on a comment period too short for a 412-page rule under 5 U.S.C. 553. A federal court has already vacated a related funding change on internal-contradiction grounds, which supplies a direct model, and because these defects reach the rule's basic structure the remedy is vacatur of the whole rather than provision-by-provision severance.

For Congress.

Congress can disapprove the final rule through a joint resolution under the Congressional Review Act, which if enacted voids the rule and bars a substantially similar one without new authorization. Appropriations offer a second route, as Congress already used by barring the indirect-cost change in the enacted fiscal year 2026 appropriations, and the authorizing committees that oversee the science agencies offer a third. Members of both parties have raised concerns about the direction of these changes, so the record here supplies the specific, agency-level harms a disapproval resolution or an appropriations rider can cite.

For a future administration.

Because the rule is a regulation rather than a statute, a later administration can rescind it through the same notice-and-comment process that created it. This comment and the public record establish the evidentiary basis a rescission would rely on, including the documented harms to patients, trials, the workforce, and competitiveness. A rescission resting on that record would supply exactly the reasoned analysis this rule lacks.

The political durability of the rule is as weak as its reasoned basis, because a measure that cancels the trials patients are enrolled in, strands volunteers who carry experimental implants, wastes funds already spent, and speeds the departure of American science to a competitor is not one a future administration will defend once its effects are visible. A rule can be legally vulnerable and politically survivable, or the reverse, yet this one is both, which is why the record should preserve, for whoever later moves to undo it, the specific statutory conflicts, the specific harms to Americans, and the specific fixes set out above.

Selected authorities

Statutes and executive mandates

42 U.S.C. 289a (peer review as the statutory basis for NIH funding). Impoundment Control Act, 2 U.S.C. 681 and following. OSTP public-access memorandum (August 2022). 31 U.S.C. 503 (the financial-management provision OMB invokes to convert guidance into binding regulation). False Claims Act, 31 U.S.C. 3729 and following. Executive Order 14332 (August 7, 2025, termination for convenience).

Cases

Motor Vehicle Manufacturers Association v. State Farm, 463 U.S. 29 (1983). FCC v. Fox Television Stations, 556 U.S. 502 (2009). Department of Homeland Security v. Regents of the University of California, 591 U.S. 1 (2020). Department of Commerce v. New York, 588 U.S. 752 (2019). Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024). West Virginia v. Environmental Protection Agency, 597 U.S. 697 (2022). Whitman v. American Trucking Associations, 531 U.S. 457 (2001). Grayned v. City of Rockford, 408 U.S. 104 (1972). Agency for International Development v. Alliance for Open Society International, 570 U.S. 205 (2013). Legal Services Corporation v. Velazquez, 531 U.S. 533 (2001). Rosenberger v. Rector and Visitors of the University of Virginia, 515 U.S. 819 (1995). FCC v. League of Women Voters, 468 U.S. 364 (1984). National Endowment for the Arts v. Finley, 524 U.S. 569 (1998). Train v. City of New York, 420 U.S. 35 (1975). United States Court of Appeals for the First Circuit, indirect-cost-cap decision (January 5, 2026).

Reports and data

Government Accountability Office decision B-337203 (August 2025). Congressional Research Service, NIH grant-termination count (April 2026). United for Medical Research, NIH economic-return estimate. OECD, gross domestic expenditure on research and development (2023). Nature, survey of United States scientists considering departure (2025). CNN, tally of scientist departures to China (2025 to 2026). Seven central ministries of China (Ministry of Industry and Information Technology and others), Implementation Opinions on the brain-computer interface industry (July 2025).

Respectfully submitted, Takashi D. Y. Kozai, Ph.D., Ernest E. Roth Professor of Bioengineering, University of Pittsburgh.