Lightnews — Scholar-powered news

Cogan Lab @coganlab.bsky.social · 1d

Stop by to say hello and see some great science!
#Sfn2025 #Neuroscience #neuroskyence

Cogan Lab @coganlab.bsky.social · 1d

Lastly (not least):

Wed. Nov 19 8am-12pm: 411.11 / MM10

Sensory-motor mechanisms for verbal working memory*

Postdoc Baishen Liang will be presenting his work on sensory-motor transformations for vWM
@gregoryhickok.bsky.social

*Also presenting at APAN

1 1

Cogan Lab @coganlab.bsky.social · 1d

Next:

Mon. Nov 17 8am-12pm: 173.10 / S11

Multimodal sensory-motor transformations for speech

@dukeengineering.bsky.social PhD Student Areti Majumdar will be presenting her work on multimodal sensory-motor transformations for speech

1

Cogan Lab @coganlab.bsky.social · 1d

Then:

Sun. Nov 16 1pm-5pm: 142.11 / LL17

Computational hierarchies of intrinsic neural timescales for speech perception and production

Former CRS @nicoleliddle.bsky.social (now at UCSD Cog Sci) will be presenting her work on intrinsic timescales and speech perception/production

1

Cogan Lab @coganlab.bsky.social · 1d

Next:

Sun. Nov 16 1pm-5pm: 142.06 / LL12

Hierarchical Speech Encoding in Non-Primary Auditory Regions*

Postdoc Nanlin Shi will be presenting his work on speech encoding in non-canonical areas

*Also presenting at APAN

1

Cogan Lab @coganlab.bsky.social · 1d

Then:

Sun. Nov 16 1pm-5pm: 142.05 / LL11

Verbal working memory is subserved by distributed network activity between temporal and frontal lobes

Former Neurosurgery Resident Daniel Sexton (now at @stanfordnsurg.bsky.social ) will be presenting his work on network decoding of verbal WM

1 1

Cogan Lab @coganlab.bsky.social · 1d

Next:

Sun. Nov 16 1pm-5pm: 137.10 / HH2

Intracranial EEG Correlates of Concurrent Demands on Cognitive Stability and Flexibility

Undergraduate Erin Burns and CNAP PhD Student Jim Zhang will present work from our lab and @tobiasegner.bsky.social Lab on cognitive control

1 1 1

Cogan Lab @coganlab.bsky.social · 1d

First up:

Sun. Nov 16 1pm-5pm: 126.20 / T11

Automated speech annotation achieves manual-level accuracy for neural speech decoding

@dukeengineering.bsky.social PhD Student Zac Spalding and Duke Kunshan undergrad Ahmed Hadwan will present work on validating automated speech alignment for BCI

1

Cogan Lab @coganlab.bsky.social · 1d

Coming to San Diego for SfN and/or APAN? Come check out the intracranial work from the lab (7 posters)! There's a bit of everything this year, so come say hello!
#Sfn2025 #Neuroscience #neuroskyence
@dukebrain.bsky.social @dukeneurosurgery.bsky.social @dukeengineering.bsky.social

1 3 9

Cogan Lab @coganlab.bsky.social · Sep 12

Come by tomorrow morning to hear about verbal working memory!

Cogan Lab @coganlab.bsky.social · Sep 10

Saturday Sept. 13 11am-12:30pm, Poster Session C

C54: Baishen Liang (Postdoctoral Associate) will be presenting his work on sensory-motor mechanisms for verbal working memory.

Hope to see you all there!

1 2

Cogan Lab @coganlab.bsky.social · Sep 12

Stop by this afternoon to see some intracranial speech decoding in the hippocampus and to say hello!

Cogan Lab @coganlab.bsky.social · Sep 10

Friday Sept 12 4:30pm-6:00pm, Poster Session B

B70: Yuchao Wang (Rotation CNAP PhD Student) will be presenting his work on auditory pseudoword decoding in the hippocampus.

1 3

Cogan Lab @coganlab.bsky.social · Sep 10

Saturday Sept. 13 11am-12:30pm, Poster Session C

C54: Baishen Liang (Postdoctoral Associate) will be presenting his work on sensory-motor mechanisms for verbal working memory.

Hope to see you all there!

Cogan Lab @coganlab.bsky.social · Sep 10

Friday Sept 12 4:30pm-6:00pm, Poster Session B

B70: Yuchao Wang (Rotation CNAP PhD Student) will be presenting his work on auditory pseudoword decoding in the hippocampus.

1 1

Cogan Lab @coganlab.bsky.social · Sep 10

Coming to DC for SNL later this week?

Come check out our posters on speech decoding and verbal working memory using intracranial recordings!

@snlmtg.bsky.social
#SNL2025

1 3 8

Cogan Lab @coganlab.bsky.social · Sep 9

❔3️⃣: In Figs. 4 and 5, do you obtain similar results if you operate directly on the spike trains instead of on the PCA-reduced spike trains? Why is PCA necessary first?

Thank you to the authors for your work!
cc: Alexis Arnaudon, Mauricio Barahona, Pierre Vandergheynst

Cogan Lab @coganlab.bsky.social · Sep 9

If separate animals were treated as separate manifolds with an embedding-agnostic MARBLE, would you still expect an informative latent space to be learned without any need for post-hoc alignment?

1 1

Cogan Lab @coganlab.bsky.social · Sep 9

❔2️⃣: It seems that a linear transformation between MARBLE representations of different animals was necessary because the same information is present in the latent space but not necessarily with the same ordering... (con't)

1

Cogan Lab @coganlab.bsky.social · Sep 9

❔1️⃣: It is stated that non-neighbors (both within and across manifolds) are negative samples (mapped far) during the contrastive learning step. Does treating non-neighbors within and across manifolds as similarly “distant” lead to less interpretability of larger distances in latent space?

1

Cogan Lab @coganlab.bsky.social · Sep 9

🤍3️⃣: The comparisons to state-of-the-art latent dynamical systems models are great for properly contextualizing the performance of MARBLE.

1

Cogan Lab @coganlab.bsky.social · Sep 9

🤍1️⃣: The initial proximity graph is a clever way to define distance and neighborhoods between inputs that can be used for downstream training.
🤍2️⃣: The rotation invariance is important and likely useful for extracting shared latent representations from systems with minor differences.

1

Cogan Lab @coganlab.bsky.social · Sep 9

They find that MARBLE successfully decomposes complex dynamical activity from spike trains into informative and easily decodable latent representations. This 🧵 explores our thoughts (🤍 & ❔). www.nature.com/articles/s41...

MARBLE: interpretable representations of neural population dynamics using geometric deep learning - Nature Methods

MARBLE uses geometric deep learning to map dynamics such as neural activity into a latent representation, which can then be used to decode the neural activity or compare it across systems.

www.nature.com

1

Cogan Lab @coganlab.bsky.social · Sep 9

Last week, Zac Spalding (@zspald.bsky.social‬, 4th year PhD student, @dukeubme.bsky.social‬) presented Adam Gosztolai, Robert Peach, and colleagues’ 2025 paper on MARBLE, a method for finding interpretable latent representations of neural dynamics.

MARBLE: interpretable representations of neural population dynamics using geometric deep learning - Nature Methods

MARBLE uses geometric deep learning to map dynamics such as neural activity into a latent representation, which can then be used to decode the neural activity or compare it across systems.

www.nature.com

1 1 2

Reposted by Cogan Lab

Duke Institute for Brain Sciences @dukebrain.bsky.social · Aug 27

Exciting work by PhD student Zac Spalding @zspald.bsky.social and multiple labs across departments at Duke! A combination of human electrophysiology, modeling/decoding, and cognitive concepts to improve BCI for speech in patients. 👏 🧠📈

Cogan Lab @coganlab.bsky.social · Aug 25

We’re happy to present @zspald.bsky.social 's work on shared neural representations of speech production across individuals! We find that patient-specific data can be aligned to a shared space that preserves speech information, enabling cross-patient speech BCIs.
www.biorxiv.org/content/10.1...

Aligned neural space for speech production allows for improved decoding in micro-ECOG patients for BCI

5 8

Cogan Lab @coganlab.bsky.social · Aug 26

From Cogan Lab Journal Club with @zspald.bsky.social
these decomposition acronyms are getting out of hand!

1 4

Cogan Lab @coganlab.bsky.social · Aug 25

We’re happy to present @zspald.bsky.social 's work on shared neural representations of speech production across individuals! We find that patient-specific data can be aligned to a shared space that preserves speech information, enabling cross-patient speech BCIs.
www.biorxiv.org/content/10.1...

6 12