Ben Hayden
@benhayden.bsky.social
Professor of Neurosurgery, Baylor College of Medicine
Finally, in 3-person conversations we found the brain uses the same subspace rotation principles for binding meaning to different speaking partners, with greater rotation between self and other than between specific others.
September 22, 2025 at 2:15 PM
Finally, in 3-person conversations we found the brain uses the same subspace rotation principles for binding meaning to different speaking partners, with greater rotation between self and other than between specific others.
If I mention my hand, Im not talking about your hand. But if I mention the moon, it’s probably the same moon you mentioned. Semantic-identity binding varies by word. And we found that subspace rotation angle does so too. Body parts were most speaker-specific; verbs and function words were least.
September 22, 2025 at 2:15 PM
If I mention my hand, Im not talking about your hand. But if I mention the moon, it’s probably the same moon you mentioned. Semantic-identity binding varies by word. And we found that subspace rotation angle does so too. Body parts were most speaker-specific; verbs and function words were least.
Technical aside: Cross-speaker semantic tuning is more orthogonalized than within-speaker half-split estimates of tuning, which controls for lot of thing, like imperfect embeddings, differential fit quality for speaking and listening, and, of course, neural variability, which is very high.
September 22, 2025 at 2:15 PM
Technical aside: Cross-speaker semantic tuning is more orthogonalized than within-speaker half-split estimates of tuning, which controls for lot of thing, like imperfect embeddings, differential fit quality for speaking and listening, and, of course, neural variability, which is very high.
Here’s where manifolds save the day. If the brain can build cross-speaker vectorial semantic representations very carefully so that they live in an intermediate space between fully orthogonal and fully collinear, you can generalize and differentiate at the same time. www.cell.com/cell/fulltex...
September 22, 2025 at 2:15 PM
Here’s where manifolds save the day. If the brain can build cross-speaker vectorial semantic representations very carefully so that they live in an intermediate space between fully orthogonal and fully collinear, you can generalize and differentiate at the same time. www.cell.com/cell/fulltex...
Not only that, but we find overlapping semantic tuning at the single neuron level. A neuron that responds to hearing the word “dog” will tend to have a stronger response for speaking the word “dog.”
September 22, 2025 at 2:15 PM
Not only that, but we find overlapping semantic tuning at the single neuron level. A neuron that responds to hearing the word “dog” will tend to have a stronger response for speaking the word “dog.”
We converted all the spoken words to semantic embeddings using BERT (with speaker tokens) and regressed firing against embeddings to get semantic tuning curves for each neuron. Not surprisingly, neurons have maximal encoding around the time of speaking and a few hundred ms after words heard.
September 22, 2025 at 2:15 PM
We converted all the spoken words to semantic embeddings using BERT (with speaker tokens) and regressed firing against embeddings to get semantic tuning curves for each neuron. Not surprisingly, neurons have maximal encoding around the time of speaking and a few hundred ms after words heard.
We used fancy microphones to record participants having conversations while recordings hippocampal neurons. We transcribed and diarized every word spoken and heard.
Automated methods aren’t accurate enough, so our stellar team did it, painstakingly, by hand, using Praat.
Automated methods aren’t accurate enough, so our stellar team did it, painstakingly, by hand, using Praat.
September 22, 2025 at 2:15 PM
We used fancy microphones to record participants having conversations while recordings hippocampal neurons. We transcribed and diarized every word spoken and heard.
Automated methods aren’t accurate enough, so our stellar team did it, painstakingly, by hand, using Praat.
Automated methods aren’t accurate enough, so our stellar team did it, painstakingly, by hand, using Praat.
Love this paper. I just want to highlight this sentence from the abstract:
September 8, 2025 at 1:45 PM
Love this paper. I just want to highlight this sentence from the abstract:
Total chaos.
Utter insanity.
Utter insanity.
September 4, 2025 at 12:27 PM
Total chaos.
Utter insanity.
Utter insanity.
If you had predicted this 20 years ago they would have laughed at you and called you crazy.
(And laughed at you again when you predicted the mouse would be the most important organism in behavioral neuroscience)
(And laughed at you again when you predicted the mouse would be the most important organism in behavioral neuroscience)
September 4, 2025 at 11:20 AM
If you had predicted this 20 years ago they would have laughed at you and called you crazy.
(And laughed at you again when you predicted the mouse would be the most important organism in behavioral neuroscience)
(And laughed at you again when you predicted the mouse would be the most important organism in behavioral neuroscience)
It uses vague, unsubstantiated concerns that we will "miss out on innovations" and then offers its methods as a monitoring solution to a problem it pretends to identify but does not.
July 9, 2025 at 9:47 PM
It uses vague, unsubstantiated concerns that we will "miss out on innovations" and then offers its methods as a monitoring solution to a problem it pretends to identify but does not.
In fact it's a good thing because they should spend their time doing science, not writing it up.
This paper sounds the alarm over hallucinations and mistakes, but, despite lots of data, doesn't show that those are on the rise in LLM-assisted papers.
That's sensationalism.
This paper sounds the alarm over hallucinations and mistakes, but, despite lots of data, doesn't show that those are on the rise in LLM-assisted papers.
That's sensationalism.
July 9, 2025 at 9:45 PM
In fact it's a good thing because they should spend their time doing science, not writing it up.
This paper sounds the alarm over hallucinations and mistakes, but, despite lots of data, doesn't show that those are on the rise in LLM-assisted papers.
That's sensationalism.
This paper sounds the alarm over hallucinations and mistakes, but, despite lots of data, doesn't show that those are on the rise in LLM-assisted papers.
That's sensationalism.
Finally, we find direct links between contextualization and next-word prediction. For example, hippocampus neurons encode the three most likely upcoming words, with strength corresponding to their likelihood.
June 24, 2025 at 5:56 PM
Finally, we find direct links between contextualization and next-word prediction. For example, hippocampus neurons encode the three most likely upcoming words, with strength corresponding to their likelihood.
To implement contextualization, you need positional encoding - a map of where words are. We find that too. Both linear encoding (not too surprising) but also, superimposed, a sinusoidal (periodic) encoding, which was proposed in Vaswani et al.
June 24, 2025 at 5:54 PM
To implement contextualization, you need positional encoding - a map of where words are. We find that too. Both linear encoding (not too surprising) but also, superimposed, a sinusoidal (periodic) encoding, which was proposed in Vaswani et al.
This pattern explains contextualization for many types of contextualizing pairs, including, for example, adjectives and nouns that they modify.
But not all! For example, verbs contextualize objects, but we don't see evidence that subjects contextualize their verbs!
But not all! For example, verbs contextualize objects, but we don't see evidence that subjects contextualize their verbs!
June 24, 2025 at 5:52 PM
This pattern explains contextualization for many types of contextualizing pairs, including, for example, adjectives and nouns that they modify.
But not all! For example, verbs contextualize objects, but we don't see evidence that subjects contextualize their verbs!
But not all! For example, verbs contextualize objects, but we don't see evidence that subjects contextualize their verbs!
Here's is the key figure from the paper, where the inferred APGs look really similar, even though they are generated from entirely unrelated data (and three more examples):
June 24, 2025 at 5:51 PM
Here's is the key figure from the paper, where the inferred APGs look really similar, even though they are generated from entirely unrelated data (and three more examples):
We developed a method for inferring the attentional weightings (attention pattern grids) for each words and then found that the empirical weightings from brain data and from LLMs look very similar.
June 24, 2025 at 5:49 PM
We developed a method for inferring the attentional weightings (attention pattern grids) for each words and then found that the empirical weightings from brain data and from LLMs look very similar.
Pronouns refer to their antecedents; adjectives specify nouns; nouns modify verbs, and so on. Context matters. The meaning of “pool” differs if a few words before it you see “billiards” or “backstroke”.
June 24, 2025 at 5:44 PM
Pronouns refer to their antecedents; adjectives specify nouns; nouns modify verbs, and so on. Context matters. The meaning of “pool” differs if a few words before it you see “billiards” or “backstroke”.
We happen to be very excited about this one in the lab right now. X: a stimulus parameter. Y: a neural correlate.
April 18, 2025 at 12:38 PM
We happen to be very excited about this one in the lab right now. X: a stimulus parameter. Y: a neural correlate.
Wow it’s been a really long time since he released any new music.
April 17, 2025 at 12:05 PM
Wow it’s been a really long time since he released any new music.
Strong disagree.
I would contend people cannot meaningfully answer these questions without comparing their own experience to what they imagine other people experience.
I would contend people cannot meaningfully answer these questions without comparing their own experience to what they imagine other people experience.
April 15, 2025 at 12:20 AM
Strong disagree.
I would contend people cannot meaningfully answer these questions without comparing their own experience to what they imagine other people experience.
I would contend people cannot meaningfully answer these questions without comparing their own experience to what they imagine other people experience.
The older I get the more I interpret the "metacogntive deficit" theory as just meaning the same as "aphantasics are misdescribing their subjective experience" or (to be glib) "aphantasics are wrong"
April 14, 2025 at 10:29 AM
The older I get the more I interpret the "metacogntive deficit" theory as just meaning the same as "aphantasics are misdescribing their subjective experience" or (to be glib) "aphantasics are wrong"