B.I.O.N.I.C. Lab

Bio-Integrating Optoelectronics, Neural Interfaces, and Computation

The cellular and vascular biology of brain-machine interfaces and neuromodulation, with bidirectional translation between mechanism and device, biology and engineering, bench and clinic.

Directed by Takashi D.Y. Kozai, Ernest E. Roth Professor of Bioengineering, University of Pittsburgh.

Kozai TDY*, et al. Solving the Problem of Inception: A cross-species perspective on strategies for a mechanistic refinement of intracortical microstimulation. Journal of Neural Engineering. 2026.

What We Do

We study how cortical tissue responds to chronic perturbation. We use implanted neural interfaces and multi-modal neuromodulation as instruments for questions in adult cortical plasticity, glial and vascular biology, and the cellular failure modes shared by brain-machine interfaces and neurological disease (MS, AD, TBI, stroke).

For the engineering and translational scientist. Brain-machine interfaces fail in tissue. Materials science alone does not explain or fix the failure. We pair parametric in vivo perturbation (ICMS, ultrasound, optogenetics, chemogenetics) with longitudinal two-photon imaging and chronic electrophysiology to characterize the device-tissue interface across rodent, NHP, and human preparations. We collaborate directly with the Pitt human BCI program and with clinical neuromodulation partners. [ Engineering and Translation ]

For the neuroscience and glioscience student. Chronic implantation, combined with the standard methods of contemporary neuroscience (optogenetics, chemogenetics, transgenics, pharmacology), gives experimental access to adult cortical glial and vascular plasticity that other in vivo preparations do not offer. We use this access to ask whether myelin remodels on behavioral timescales, what mural cells do under perturbation, how microglia maintain homeostasis under chronic input, and how oligodendrocyte progenitors respond to injury and activity. [ Basic Biology ]

For the disease-focused quantitative scientist. Glial and vascular failure modes around implanted devices share cellular mechanisms with multiple sclerosis, Alzheimer's disease, traumatic brain injury, and stroke. Parametric stimulation in healthy adult cortex provides controlled longitudinal access to four processes (neuroinflammation, myelin remodeling, neurovascular uncoupling, and oligodendrocyte vulnerability). Disease preparations cannot provide this level of experimental control. [ Disease Connections ]

Research Architecture, Two Coupled Axes

Our research runs along two axes. Both axes share methods, trainees, and scientific motivation.

Axis 1, Tissue-Device Biophysics. How does the brain respond to implanted devices and to exogenous perturbation (electrical, ultrasonic, optical, chemogenetic, pharmacological) at the cellular and vascular level? Axis 1 characterizes the biological cascades that determine device performance over chronic timescales (weeks to months). Axis 1 output is mechanism that informs device design, stimulation parameter selection, and clinical deployment.

Axis 2, Glial and Vascular Neurocomputation. What computational roles do astrocytes, oligodendrocytes, microglia, and the cerebrovasculature play in healthy brain function, learning, and disease? Axis 2 maps these contributions using several readouts (two-photon imaging, chronic electrophysiology, electrochemistry, and computational modeling). Axis 2 output is mechanistic accounts of cortical function that go beyond the neuron-centric view of neural circuits.

Observations from Axis 1 (how implants and parametric stimulation perturb glial and vascular systems) generate hypotheses that Axis 2 tests in basic biology terms. Findings from Axis 2 specify what Axis 1 must engineer to advance the device-tissue interface and its clinical applications. Most papers from the lab contribute to both axes. Trainees therefore develop the experimental and reasoning skills to operate across both. (philosophy).

(see details)

Who Will and Will Not Thrive Here

We recruit students and postdocs from a range of backgrounds (biology, engineering, physics, applied math, computational science). Applicants should want to work on basic glial and vascular biology, contemporary neuroscience, and translational neural engineering. Applicants should want to develop competence across multiple in vivo perturbation and readout modalities.

We are not the right fit for trainees whose main interest is descriptive immunohistochemistry of foreign body response, terminal histology as a primary endpoint, or device fabrication without the underlying tissue biology. Descriptive characterization of glial scarring is no longer a primary question in this field.

For complementary interests within Pitt, students focused on electrode and device design should consider Dr. Tracy Cui's lab. Students focused on human BCI clinical research should consider Drs. Jennifer Collinger and Robert Gaunt. The B.I.O.N.I.C. Lab partners with these programs. Our lab is the right home for trainees whose questions live in the cellular and vascular biology that determines whether devices and human BCIs succeed or fail.

Trainee Outcomes

Every graduate student trained in the lab has secured external fellowship or scholarship funding during their PhD, a 100% rate sustained across or scored highly competitively with applications pending. Fellowships have included the NSF Graduate Research Fellowship, NIH F31 individual predoctoral awards, NIH F99/K00 transition awards, ARCS Foundation Scholarships, and Pitt-internal awards (including the Bevier Dissertation Fellowship). Postdoctoral trainees have secured NIH F32, K99/R00, and foundation postdoctoral awards at 100%.

We train for both academic and industry careers. Alumni have placed in academic faculty positions, in industry research positions at neural interface and medical device companies, and in policy and translational roles. Each trainee chooses the track. Our core skill set prepares trainees for both tracks without forcing early commitment. That skill set includes four elements (mechanism at the engineering-biology interface, multi-modal neuromodulation, longitudinal in vivo readout, and computational analysis).

How We Work

Our two axes bidirectional translational program describe what we study.

Work runs in two directions. Forward, mechanism from Axis 2 informs device design, stimulation parameters, and clinical deployment in Axis 1, through industry and clinical collaborations. Reverse, the chronic devices and parametric stimulation of Axis 1 are used as instruments for the basic biology of Axis 2.

Rodent preclinical rodent preparation carry the mechanistic work. Conditional knockouts, transgenic reporters, chronic two-photon windows, and parametric stimulation at proper sample sizes give the experimental access needed for mechanistic claims. Trainees extend rodent-derived findings into NHP and human collaborations through the Pitt human BCI program and other clinical partners as projects mature

The Toolkit

We pair a broad set of in vivo perturbation methods with longitudinal readouts. Perturbation modalities include intracortical microstimulation, focused ultrasound with microbubbles, low-intensity pulsed ultrasound, optogenetics, DREADDs, conditional and transgenic genetic tools, and pharmacology. Readouts include two-photon and multiphoton imaging across weeks to months, chronic single-unit and multi-unit electrophysiology, behavioral measures, computational and biophysical modeling, and quantitative immunohistochemistry (terminal validation only). Trainees develop competence in 2 to 3 perturbation modalities and 2 to 3 readout modalities, chosen by scientific question rather than prior specialization.

Beyond the Lab

Industry translation. Dr. Kozai founded Fontis Biotechnology as a graduate student and declined a co-founder position at Neuralink in 2016 on scientific grounds. The lab maintains active engagement with the neural interface industry, and trainees benefit from internships, introductions, and post-graduation placement support.

Public engagement and science policy. The lab contributes to national science policy through Senate Labor-HHS appropriations testimony, the National Security Commission on Emerging Biotechnology, and engagement with NIH, NSF, FDA, and ARPA-H. Dr. Kozai writes the "Why BCIs Fail" series on LinkedIn and Substack.

Institutional role. Dr. Kozai is a lead faculty organizer of UP NExT, an Associate Editor at the Journal of Neural Engineering, and a member of committees serving three deans and the Senior Vice Chancellor for Research.

Page updated

Google Sites

Report abuse