TK (Takashi DY) Kozai

@bioniclab.bsky.social

Ernest E Roth Professor. Dept Bioengineering, University of Pittsburgh. BionicLab.org Neural Engineering; Neural Interfaces: Neurostimulation; Neurocomputation

When I interviewed at Michigan for their Biomedical Engineering PhD, the chair told me I was an experiment. He meant it descriptively. The program had only ever accepted engineering majors. I had a biology degree from Colorado. 🧵

9 lessons from Bethesda Declaration organizers on how to rescue US science: www.ovid.com/jnls/dartmou... 1. Scientists can no longer safely claim the privilege of being outside of, or above, politics 2. Inaction can be riskier than action. 3. Progress requires work. 4. Do not wait for someone else.

Nine lessons from Bethesda Declaration organizers... : Dartmouth Journal of Healthcare Delivery Science

Abstract was not provided for this article.

ovid.com

Monday is the deadline for public comment! The Heritage Foundation has been mobilizing its members to leave comments in favor of the rule change, so every unique comment matters (don’t use a form letter or copy/paste—similar comments are batched and treated as one).

Bethany Brookshire@beebrookshire.bsky.social · 4w ago

There's an OMB rule change that would require political oversight of scientific grants, and could pull any that don't "reflect presidential priorities" Another country has tried this before! It was the USSR and it was called Lysenkoism. By @thsaey.bsky.social www.sciencenews.org/article/omb-...

🧠 Memory held even where the plaques never cleared. In an aged Alzheimer's model, boosting oligodendrocytes kept spatial working memory at wild-type levels (56.7±2.5% vs 39.5±2.1%, P=0.0004), without lowering amyloid. 🧵

One current pulse drives a single cortical column in a mouse. The same waveform in a paralyzed human BCI user yields a different percept, or none. That cross-species gap is the problem BRIDGE-STIM 2027 is built to close. 🧵

Bild

Absolutely this. Bell Labs was very productive when I was a kid. But they, as a monopoly, basically had the ability to tax the entire US population. Moreover, while they published or otherwise disseminated some of their findings, they kept some just for A T & T.

Brent Simmons@thrashy.bsky.social · last mo.

It's worth pointing out that one of the few private entities to ever do significant amounts of basic research was Bell Labs, a subsidiary of a telecom monopoly at the height of its monopolistic abuses. Basic research is a costly endeavor, and few entities outside gov. can afford it.

🧬In 2001, a member of the National Academy of Sciences told my 400-person lecture to watch Jennifer Doudna, because she'd probably win a Nobel. What he didn't say: taxpayers were already paying for the work that would get her there. 🧵

Bild

🧠 NSF just funded BRIDGE-STIM 2027 (Award #2624028, CBET). May 17-19, Pittsburgh. The workshop assembles experimentalists running intracortical microstimulation (ICMS) in mice, NHPs, and humans, addressing the cross-species translation gap stalling sensorimotor BCI restoration.

Bild

📊 NIH NOT-OD-26-086 proposes capping RPGs per PI at 2, 3, or 4. The cap counts grants rather than dollars controlled, so an MPI on a five-site U19 spends a full slot for a fraction of a budget while a sole PI spends one slot for an entire R01.

⚡ A brain implant delivers a microsecond pulse to cortex, and the textbook answer is that the neuron fires. The biology turns out to be far richer, which matters for every patient feeling through a prosthetic hand, suppressing a seizure, or treating depression by stimulation.

Bild

Roughly half of intracortical microelectrodes lose recording capability within a year of implantation. The B.I.O.N.I.C. Lab at Pitt is hiring a postdoc to attack the mechanism on two parallel grant-funded threads. 🧵 (1/7)

🔬 Your tax dollars once paid a physicist whose own colleagues told him he was wasting the university's money. In 1951, Charles Townes sat on a D.C. park bench and worked out, on the back of an envelope, how to force molecules to release coherent radiation. 🧵 1/5

Bild

In 1944, Vannevar Bush handed FDR "Science: The Endless Frontier," arguing federal investment in basic research was a national security requirement. Eighty-two years later, the FY27 NDAA biotech amendments make the same case. 🧵 (1/8)

🌊 High school biology. Teacher plays a clip. Frame one, ATP synthase, the protein that makes ATP from a proton gradient in every cell. Frame two, a water turbine. Nearly identical structures, 14 orders of magnitude apart, identical mechanism. Biology was an engineering system. 🧵

🛩 Every commercial pilot trains on simulators. None earns a license through simulators alone. The simulator is not the airplane. On July 10, 2025, NIH announced future funding opportunities will no longer be issued exclusively for animal models. The shift is now active policy.🧵

🧮 We built a computational model that predicted which electrode channels would fail in human BCI participants before they failed. The mouse data isn't misleading. The biology of failure is the same in humans, and now we can map it.

Bild

🧬 In 1995, Katalin Karikó was demoted at UPenn because her grants weren't landing and her papers weren't cited. The mRNA chemistry she developed over the next 20 years (modifying uridine to pseudouridine) became the COVID-19 vaccines and the 2023 Nobel Prize.

🧪 I came up under peer review. The people who built it did not. Senior colleagues told me what it replaced: strong proposals from capable researchers going unfunded, not for weak science, but because the path to money ran through personal proximity to the program officer who held it.

🔧 As a kid I was confused and frustrated by school. Japanese and English jumbled together, home unstable. What I could control was small. So I built things: rubber bands, paperclips, then sheet metal and hardwood. Mechanical logic was the one place the world made sense. 🧵