
Cancer neurotechnologist @Imperialbrains
London
Joined May 2011
- Tweets304
- Following1.6K
- Followers226
- Likes902
Romeo RACZ retweeted
Cancer research needs neuroscience and neuroscientists | Nature Neuroscience nature.com/articles/s41593-0…
Romeo RACZ retweeted
I have added some visualization to the #Brain effective connectivity toolbox:
1. A directed graph + tractography visualization
github.com/alecrimi/effconnp…
2. Dynamic visualization of connected regions according to fMRI/EEG time series, but with issues:
github.com/alecrimi/effconnp…
In @elife: Reproducibility of in vivo electrophysiological measurements in mice doi.org/10.7554/eLife.100840…
1
100
Romeo RACZ retweeted
Suthana et al. studied brain recordings from implanted electrodes in freely moving humans. They show that neural dynamics encoded actual and imagined routes similarly. @SuthanaLab @m_seeber
nature.com/articles/s41562-0…
Romeo RACZ retweeted
1/ Delighted to share the published version of this work from brilliant #CancerNeuroscience post-doc @TaraBarronPhD detailing GABAergic neuron-to-glioma synapses in #DIPG/#DMG. 🧵
Open access link:
nature.com/articles/s41586-0…
Romeo RACZ retweeted
You will never know how much work this reel took!!
I think if I had known how hard it was when I started, I wouldn’t have started at all 😂
Scrolling through a brain MRI is like driving down the road. There are signposts & check marks along the way.
These are some of the anatomic landmarks I check as I scroll through a brain MRI
These help you to orient yourself to the anatomy and help you to remember to look for pathology in small places.
Of course, these are not all the landmarks—it’s just a start.
I want to show more signposts that I stop by when reading a brain MRI—so follow along—I think all the extra work will be worth it!!
Romeo RACZ retweeted
How the brain gets rid of its chemical waste through the glymphatics during sleep, via blood vessel oscillations, regulated by norepinephrine. Impeded by Ambien.
cell.com/cell/abstract/S0092… @CellCellPress
science.org/content/article/… @ScienceMagazine
Romeo RACZ retweeted
Physics Photo Of the Day:
This is an actual picture of a molecule taken by IBM, created by using a technique called Atomic Force Microscopy (AFM).
Romeo RACZ retweeted
The synchronization of energy patterns across the brain may be the basis for the sense of self
nature.com/articles/s41586-0…
@realJoshSiegel @ndosenbach
Romeo RACZ retweeted
What is or isn't a "BCI"? (Brain Computer Interface)
We argue in @natBME today that if the tech stimulates or records brain activity AND does computation, it's a BCI.
This definition would align with the popular concept of BCI, but we still need a way to discuss different types of BCI separately.
We argue that classifying BCI based on their applications e.g. "movement BCI" or "therapeutic BCI" would allow regulators, scientists, physicians, and patients to discuss specific BCI classes narrowly.
Just in time for your holiday conversation starter - our gift to you!! 🎁
nature.com/articles/s41551-0…
Thanks to a great group of co-authors!
@SumnerLN @Matt_R_Angle @tgconstandinou @JonAMichaels @neuroamyo @chethan @andpru @crozSciTech @nishalpshah @MaryamShanechi @SameerShethMD @SergeyStavisky @EricMTrautmann @xiechong and may others not on X! 🙏
#BCI #BrainComputerInterface
Romeo RACZ retweeted
"Why does the brain need billions of neurons to process 10 bits/second? Why can we only think about one thing at a time?"
by @jieyusz and @mameister4 @Caltech @NeuroCellPress cell.com/neuron/fulltext/S08…
Romeo RACZ retweeted
My best/most research was never ever previously proposed in any grant. Actual research outcomes are weakly correlated to proposed research. So we need less emphasis on proposed research and more on recent accomplishments…
I can cite numerous examples of basic research leading to important therapeutic developments over the last 60 years. I can also point to enormous inefficiencies in the allocation & use of tax dollars for research. It behooves all researchers to think about how to improve the system, not just complain.
This is amazing. Hydrogel expansion enables molecularly informed reconstruction of brain tissue at synaptic resolution with light microscopy.
biorxiv.org/content/10.1101/…
42
Romeo RACZ retweeted
Imagine regaining touch after decades without sensation—thanks to advanced neurotechnology. For Scott Imbrie, this is now reality.
A new @Nature article shares his journey using a robotic arm powered by @UChicago and our tech. Sensory feedback has reignited Scott’s connection to the world, enhancing his quality of life.
Huge thanks to the visionary researchers and to Scott, whose dedication drives this tech forward. The future of neural interfaces is about reviving the sensory richness of human experience.
Read more here: nature.com/articles/d41586-0……
#Neurotech #BCI #Prosthetics #Innovation
Romeo RACZ retweeted
A PhD position is open in my lab for those interested in studying the neuronal basis of visual perception - utilising virtual reality for mice, optogenetics, and in vivo imaging methods. See the following link for further information: findaphd.com/phds/project/as…
Romeo RACZ retweeted
📢 Opportunity Alert! We have two open PhD positions in our lab at @PDN_Cambridge starting in Autumn 2025. Exceptional candidates may have the opportunity to join earlier as RAs and receive PhD funding if scholarship applications are unsuccessful.
Advert: keshavarzilab.com/blog/phd-p…
Romeo RACZ retweeted
Come for the videos - stay
for the science
If you are into circuits, x-rays, olfaction, connectomics - you will not be disappointed
Hi @SfNtweets,
Neurons of the same olfactory circuit in the mammalian brain share firing properties while showing individual traits.
More about it today PM, poster E37 (PSTR341.09).
@YuxinZhang97 @APacureanu @AndreasTSchafer
@TheCrick @esrfsynchrotron @psich_en
#SfN24
I scored 8.99 BPS (55 NTPM) on @neuralink's cursor control benchmark: neuralink.com/webgrid/
2
93
Romeo RACZ retweeted
People still doubt it? NeuroAI is really the best approach to improving AI.
The brain is the current state of the art in Intelligence, by a wide margin. The difference is truly substantial, both in terms of performance, and in terms of underlying implementation,
ChatGPT's "neurons" are still missing tons of important details compared to our brain's neurons. In the same time, despite the impressive feats of today's AI, I still don't have a robot to do the simplest things like tie my shoes. It is very unlikely that these two facts are unrelated to each other. Fortunately, the foundational operation of AI and the brain can be very compatible. Both are connectionist models (i.e. networks of connected nodes), meaning we can port any mechanism of the brain we understand and find useful, directly to AI models. This makes neuroscience an enormous pool of excellent candidate mechanisms to improve AI. So, bridging the gap between the two implementations (biological and artificial) is our best shot and our fastest path to bridging the gap in performance too.
Your gut reaction that this is easier said than done is correct. But the reason for the difficulty should be pinpointed and can be surpassed.
The difficulty is in that Neuro-AI is not merely multidisciplinary, but truly interdisciplinary. For example, studying the literature on, say, neuronal ion channels does not suffice. One has to also ask what computations arise from this biophysics, then understand the open problems in AI applications, and how people have tried to solve those, to ultimately find a conceptual hole that can be filled by a model of the ion dynamics.
This is hard work that not everyone can afford to do. Worse, this degree of synthesis is probably impossible, even for a multidisciplinary team, unless it is guided by people with experience in all these disciplines simultaneously. So, what does it take? It takes training people in all these fields, i.e. joint Neuro-AI academic programs, long-term venture capital investing in interdisciplinary deep tech, big-tech companies directing their "moonshot" investments specifically to neuroAI projects, and so on.
Honestly, among the hypothetical paths to AGI, this is the most systematic, least speculative one, and thus the AI investment with the highest Expected Value.
I am not the only one to think so.
"to accelerate progress in AI, we must invest in fundamental research in NeuroAI." Signed, a team of superstar researchers, 2023. How incredibly refreshing this was to read!
OTOH, I would emphasize the role of private venture over state-directed funding. Even if I find it a great cause to fund, I disagree with the notion that the government or I have the moral authority to decide what taxpayers do with their own money, or that centralized interventions are better decided or executed than voluntary investment. The latter seeks the highest efficiency in capital allocation, and rationally balances the long-term and short-term horizons.
On my end, this is precisely what I am currently doing. Investing my time and money in bringing NeuroAI to the mainstream. It is time. In a shorter timeframe than what people might think.



























