Joined October 2015
- Tweets 2,657
- Following 2,039
- Followers 1,655
- Likes 1,383
285 Photos and videos
Pinned Tweet
27 Sep 2023
Grateful for this unexpected feature. I still have imposter syndrome days... I thought that was supposed to en be after tenure?
27 Sep 2023
Check out this wonderful feature about Associate Professor Takashi Kozai's journey through both biology and engineering fields, written by @elainevitone. news.engineering.pitt.edu/po…
2
2
23
6,301
🔊 Every intervention we've tried to extend BCI lifespan modifies the electrode or delivers a drug. Both work, both carry tradeoffs, new materials bring new failure modes, drugs go systemic w/ off-target effects.
We asked: what if we redirect the brain's immune response w/ sound?
1
1
139
🩺 Reduced encapsulation translated to improved chronic rec in rats. LIPUS applies non-invasively through the skull, so a person with an implanted BCI could receive it w/ no added surgery, as treatment or periodic maintenance, a delivery route electrode-level fixes cannot match.
1
21
🔗 Full mechanism, plus how 10-Hz electrical stimulation converges on the same microglial lever, in the latest Why BCIs Fail:
bioniclab.substack.com/s/why…
#WhyBCIsFail #BrainComputerInterface #Neuroscience #Neurotechnology #Ultrasound
1
24
🔬 Your tax dollars once paid a grad student to surgically remove a gland from a chicken's rear end.
In 1952, Bruce Glick at Ohio State started cutting the bursa of Fabricius out of chicks, an organ unexplained for 300 years. Nothing happened. It looked useless.
🧵1/6
1
136
📊 Everything builds on it. Monoclonal antibodies (a multi-billion-dollar drug class), CAR-T cancer therapy, vaccine design, bone marrow transplants, AIDS research.
Glick: 225 papers, 29 students trained, NIH/NSF/USDA funded, Golden Goose Award 2018.
1
1
26
🧪 Every vaccine, every antibody drug, every immune-deficiency diagnosis traces to a gland nobody understood in a chicken's posterior.
This is what foundational research looks like before anyone knows what it's for.
More 👇
bioniclab.substack.com/notes
#YourTaxDollarsAtWork #NIH #NSF
1
28
Define "Gold Standard Science" without publication, peer review, conferences, journal access, or international collaboration. You can't. Neither can OMB. The proposed 2 CFR rewrite (OMB-2026-0034) requires compliance with it anyway. 🧵
1
3
215
📅 OMB must respond to substantive comments. Name the provision, the legal theory, the contradiction. Deadline July 13, 2026.
Submit: regulations.gov/commenton/OM…
1
25
🔗 The full guide covers all 19 provisions, names the statutory conflicts (42 U.S.C. 289a, the 2022 OSTP memo), and gives modular arguments to adapt to your lab. Start from it, don't copy it.
bioniclab.org/news/OMB
#NIH #OMB #PeerReview #APA
98
The reframe is still human labor. Which question to ask, which absent measurement to run, which smoothed-over detail was the point. None of it is in the training data, because we haven't run the experiment yet.
The computer waits for someone to start the sentence.
1
4
See a more detailed analysis and a demo:
open.substack.com/pub/bionic…
4
This holds up better than the writers knew.
An AI minimizes error against its training data. A frontier research ? is the ultimate out-of-distribution input, so the most probable answer it returns is the field's conventional one. The consensus the unsolved problem already refutes
1
69
And you can't see it from outside. The model smooths the hard part into the comfortable popular ver.
In our work on why neural implants fail, the conventional answer was glial scarring for 20 yrs. Catchy, incomplete. A model trained on that lit hands it to you w/ total confidence
1
3