Laura Gwilliams
@lauragwilliams.bsky.social
Language processing - Neuroscience - Machine Learning - Assistant Professor at Stanford University - She/Her - 🏳️🌈
This dynamic code serves the computational function of avoiding destructive interference between neighbouring inputs, and provides an implicit "time code" for the relative position of items in a sequence!
8/8
8/8
October 22, 2025 at 5:21 AM
This dynamic code serves the computational function of avoiding destructive interference between neighbouring inputs, and provides an implicit "time code" for the relative position of items in a sequence!
8/8
8/8
Finally, we find that each level of the hierarchy is encoded in a dynamic ensemble of neural patterns, and neural activity unfolds faster for more sensory features like parts of speech sounds, and unfolds more slowly for more symbolic features like word meaning and sentence structure.
7/8
7/8
October 22, 2025 at 5:21 AM
Finally, we find that each level of the hierarchy is encoded in a dynamic ensemble of neural patterns, and neural activity unfolds faster for more sensory features like parts of speech sounds, and unfolds more slowly for more symbolic features like word meaning and sentence structure.
7/8
7/8
When we group the features into 6 hierarchical levels, we find that the hierarchy is processed in a remarkably parallel manner, highly overlapping in time, with long-lived neural responses. Also, higher order features are decodable earlier than lower order features - a "reverse hierarchy".
6/8
6/8
October 22, 2025 at 5:21 AM
When we group the features into 6 hierarchical levels, we find that the hierarchy is processed in a remarkably parallel manner, highly overlapping in time, with long-lived neural responses. Also, higher order features are decodable earlier than lower order features - a "reverse hierarchy".
6/8
6/8
We confirm that a rich hierarchy of 50 language features can be decoded from neural activity.
5/8
5/8
October 22, 2025 at 5:21 AM
We confirm that a rich hierarchy of 50 language features can be decoded from neural activity.
5/8
5/8
21 participants listened to short stories while we recorded neural activity with MEG. We annotated those stories for a comprehensive set of language features, from the parts of individual speech sounds, to the meaning of words, and the structure of sentences.
4/8
4/8
October 22, 2025 at 5:21 AM
21 participants listened to short stories while we recorded neural activity with MEG. We annotated those stories for a comprehensive set of language features, from the parts of individual speech sounds, to the meaning of words, and the structure of sentences.
4/8
4/8
We tested the hypothesis that each level of the hierarchy is encoded dynamically, where different neural patterns are activated in temporal sequence. We call this "Hierarchical Dynamic Coding".
3/8
3/8
October 22, 2025 at 5:21 AM
We tested the hypothesis that each level of the hierarchy is encoded dynamically, where different neural patterns are activated in temporal sequence. We call this "Hierarchical Dynamic Coding".
3/8
3/8
To process speech, the brain generates a rich hierarchy of representation, from sound to meaning.
2/8
2/8
October 22, 2025 at 5:21 AM
To process speech, the brain generates a rich hierarchy of representation, from sound to meaning.
2/8
2/8
looking forward to seeing everyone at #CCN2025! here's a snapshot of the work from my lab that we'll be presenting on speech neuroscience 🧠 ✨
August 10, 2025 at 6:09 PM
looking forward to seeing everyone at #CCN2025! here's a snapshot of the work from my lab that we'll be presenting on speech neuroscience 🧠 ✨
[3] @irmakergin.bsky.social designed a method of collecting behavioural readouts of comprehension during natural continuous listening! her scanner-compatible slider device, with millisecond resolution, captures meaningful dynamic fluctuations in understanding!
link: osf.io/preprints/ps...
link: osf.io/preprints/ps...
December 21, 2024 at 12:47 AM
[3] @irmakergin.bsky.social designed a method of collecting behavioural readouts of comprehension during natural continuous listening! her scanner-compatible slider device, with millisecond resolution, captures meaningful dynamic fluctuations in understanding!
link: osf.io/preprints/ps...
link: osf.io/preprints/ps...
[2] Ellie Abrams designed clever acoustic stimuli - spectrally non-overlapping, but evoke the same pitch percept. we find shared neural patterns of pitch, abstracted from the sensory input! AND, tonal context modulates temporal dynamics!
link: www.biorxiv.org/content/10.1...
link: www.biorxiv.org/content/10.1...
December 21, 2024 at 12:47 AM
[2] Ellie Abrams designed clever acoustic stimuli - spectrally non-overlapping, but evoke the same pitch percept. we find shared neural patterns of pitch, abstracted from the sensory input! AND, tonal context modulates temporal dynamics!
link: www.biorxiv.org/content/10.1...
link: www.biorxiv.org/content/10.1...
[1] @kriesjill.bsky.social finds similar dynamic neural coding of speech sounds for healthy older adults (~70yo) and age-matched individuals with aphasia. BUT healthy adults process phonetic features sig more robustly in the face of lexical ambiguity!
link: www.biorxiv.org/content/10.1...
link: www.biorxiv.org/content/10.1...
December 21, 2024 at 12:47 AM
[1] @kriesjill.bsky.social finds similar dynamic neural coding of speech sounds for healthy older adults (~70yo) and age-matched individuals with aphasia. BUT healthy adults process phonetic features sig more robustly in the face of lexical ambiguity!
link: www.biorxiv.org/content/10.1...
link: www.biorxiv.org/content/10.1...
looking forward to discussing the relationship between human language processing and LLMs at APS with @alexanderhuth.bsky.social and @neuranna.bsky.social ! 🧠🤖
December 17, 2024 at 5:52 PM
looking forward to discussing the relationship between human language processing and LLMs at APS with @alexanderhuth.bsky.social and @neuranna.bsky.social ! 🧠🤖
Holiday party time!! 🤩 @rdhawkins.bsky.social @coryshain.bsky.social
December 7, 2024 at 4:25 PM
Holiday party time!! 🤩 @rdhawkins.bsky.social @coryshain.bsky.social
GLySN Lab photo wall!!! 😍😍😍😍 (thank you @kriesjill.bsky.social for your vision and execution!!!!)
November 26, 2024 at 4:20 PM
GLySN Lab photo wall!!! 😍😍😍😍 (thank you @kriesjill.bsky.social for your vision and execution!!!!)
excited for this menu of GLySN Lab research at the neurobiology of language conference! 🧠
October 23, 2024 at 6:29 AM
excited for this menu of GLySN Lab research at the neurobiology of language conference! 🧠
Finally, we find that each level of the hierarchy is encoded in a dynamic ensemble of neural patterns, and neural activity unfolds faster for more sensory features like parts of speech sounds, and unfolds more slowly for more symbolic features like word meaning and sentence structure.
7/8
7/8
April 20, 2024 at 6:45 PM
Finally, we find that each level of the hierarchy is encoded in a dynamic ensemble of neural patterns, and neural activity unfolds faster for more sensory features like parts of speech sounds, and unfolds more slowly for more symbolic features like word meaning and sentence structure.
7/8
7/8
When we group the features into 6 hierarchical levels, we find the hierarchy overlaps a lot in time. Also, the higher order features are decodable earlier than lower order features. Thanks to prediction, the meaning of a word is decodable before the person hears it!
6/8
6/8
April 20, 2024 at 6:44 PM
When we group the features into 6 hierarchical levels, we find the hierarchy overlaps a lot in time. Also, the higher order features are decodable earlier than lower order features. Thanks to prediction, the meaning of a word is decodable before the person hears it!
6/8
6/8
We confirm that a rich hierarchy of 50 language features can be decoded from neural activity.
5/8
5/8
April 20, 2024 at 6:44 PM
We confirm that a rich hierarchy of 50 language features can be decoded from neural activity.
5/8
5/8
21 participants listened to short stories while we recorded neural activity with MEG. We annotated those stories for a comprehensive set of language features, from the parts of individual speech sounds, to the meaning of words, and the structure of sentences.
4/8
4/8
April 20, 2024 at 6:43 PM
21 participants listened to short stories while we recorded neural activity with MEG. We annotated those stories for a comprehensive set of language features, from the parts of individual speech sounds, to the meaning of words, and the structure of sentences.
4/8
4/8
Here we tested the hypothesis that each level of the hierarchy is encoded in a dynamic ensemble of neural patterns, which are traversed at a speed commensurate with the feature’s position in the hierarchy-- we call this "Hierarchical Dynamic Coding".
3/8
3/8
April 20, 2024 at 6:43 PM
Here we tested the hypothesis that each level of the hierarchy is encoded in a dynamic ensemble of neural patterns, which are traversed at a speed commensurate with the feature’s position in the hierarchy-- we call this "Hierarchical Dynamic Coding".
3/8
3/8
To process speech, the brain generates a rich hierarchy of representation, from sound to meaning.
2/8
2/8
April 20, 2024 at 6:42 PM
To process speech, the brain generates a rich hierarchy of representation, from sound to meaning.
2/8
2/8
At each location, most neurons responded to a particular type of speech sound (e.g., consonants vs vowels). This explains the activity we see at the brain surface using techniques like electrocorticography. [4/6]
December 13, 2023 at 4:52 PM
At each location, most neurons responded to a particular type of speech sound (e.g., consonants vs vowels). This explains the activity we see at the brain surface using techniques like electrocorticography. [4/6]
High-density Neuropixels probes were used to map the superior temporal gyrus, a critical area for speech in the human brain. This allowed us to record from hundreds of neurons across all layers of the cortex. [3/6]
December 13, 2023 at 4:51 PM
High-density Neuropixels probes were used to map the superior temporal gyrus, a critical area for speech in the human brain. This allowed us to record from hundreds of neurons across all layers of the cortex. [3/6]
Gwilliams Lab of Speech Neuroscience 1.0!!! First lab photo, very proud 🙌
October 28, 2023 at 10:16 AM
Gwilliams Lab of Speech Neuroscience 1.0!!! First lab photo, very proud 🙌
Merci, Marseille! A wonderful reunion with the neurobiology of language conference
October 27, 2023 at 12:27 PM
Merci, Marseille! A wonderful reunion with the neurobiology of language conference