The lab operates a deliberately bidirectional research program. This page develops the bidirectional logic in detail and explains how it shapes project design, training, and the lab's relationship to the broader neural interface and translational neuroscience fields.
In the forward direction, cellular and vascular mechanism informs device design, stimulation parameter selection, and clinical deployment. This is the standard direction in engineering-portfolio research, and the lab's forward-translation outputs include the carbon-fiber microthread electrode design that informed industry developments in ultrasmall electrodes, the parametric framework for ICMS that resolves longstanding questions on perceptual fading through neurovascular and metabolic mechanisms, the pharmacological repositioning of clemastine as an oligodendrocyte-targeted intervention for chronic recording stability, and the demonstration that low-intensity pulsed ultrasound modulates microglial activation as a non-invasive perturbation modality.
Forward translation requires sustained engagement with industry partners, clinical collaborators, and the regulatory environment in which devices ultimately deploy. The lab's industry network, anchored by the Fontis founding history, the Neuralink co-founder offer, and ongoing collaborations with neural interface and medical device companies, supports this engagement.
In the reverse direction, the chronic devices and parametric stimulation of Axis 1 serve as instruments for the basic biology of Axis 2. This is the direction that distinguishes the lab from labs operating in only one direction, and it produces basic neuroscience contributions that are credible to the glial and vascular biology community as well as to the engineering community.
Reverse translation outputs from the lab include the first in vivo demonstrations of activity-dependent oligodendrocyte function in adult cortex on behavioral timescales (FusOLcKO experiments, Wellman et al., JNE 2026), the longitudinal characterization of mural cell dynamics under chronic perturbation (Wellman et al., Biomaterials 2025), the resolution of microglial surveillance dynamics during sustained microstimulation (Preszler et al., JNE 2026), and the identification of metabolic and oxygen consumption constraints as gating factors for sustained neural firing under stimulation.
Most papers from the lab advance both directions simultaneously. The Wellman 2026 FusOLcKO paper is the cleanest illustration. In the reverse direction, it is the first in vivo demonstration that genetic perturbation of oligodendrocyte cholesterol biosynthesis preserves CA1 single-unit fidelity and stabilizes network dynamics over 16 weeks, a basic biology contribution to the field of activity-dependent myelin plasticity. In the forward direction, it identifies myelin plasticity as a previously uncharacterized determinant of chronic recording stability, with implications for stimulation protocol design and chronic BCI longevity. The same dataset advances both directions, and the trainees on the project develop skills in both.
The bidirectional logic operates at the project level, the publication level, and the cross-species level. The Solving the Problem of Inception perspective (Kozai et al., JNE 2026), with 16 co-authors across bioengineering, neurobiology, computational neuroscience, neuroethics, philosophy, ophthalmology, rehabilitation medicine, and English, formalizes a Parallel Translation framework, in which human psychophysical reports, rodent mechanistic studies, NHP experiments, and computational modeling run simultaneously and inform each other iteratively, rather than sequentially. The lab's research program is structured to operate in this Parallel Translation mode, with rodent mechanism, NHP and human collaborations, and computational modeling running concurrently rather than in a fixed pipeline.
Trainees in the lab learn to operate this dual structure, which is the skill set that distinguishes them in both academic and industry hiring contexts. Forward translation alone produces engineers, reverse translation alone produces basic biologists, the lab produces scientists who do both, which is what the contemporary neural engineering and translational neuroscience job markets actually want. The 100 percent fellowship rate among graduate trainees and the alumni placement record across academic, industry, and policy roles are downstream evidence of this training approach.