The University of Southampton
University of Southampton Institutional Repository

Contrastive neural embeddings reveal individual traits beyond conversational role

Contrastive neural embeddings reveal individual traits beyond conversational role
Contrastive neural embeddings reveal individual traits beyond conversational role
Contrastive representation learning is increasingly used to recover low-dimensional structure from neural recordings, but its output is typically validated by decoding accuracy rather than by the geometry of the manifold it produces. We apply CEBRA to EEG recorded from dyads in conversation, and analyze the resulting embedding, which training constrains to the 2D sphere. Labels describing the dyads, including the absolute difference between partners' autism-quotient scores, decode well above chance (0.77 against a 0.55 majority baseline for binary AQ magnitude; 0.44 against 0.25 for the six-class |ΔAQ| partition). However, the two permutation controls have notable differences in results: permuting labels over a frozen embedding yields p = 0.001, whereas retraining the encoder under each permutation yields p = 0.50. Only the latter tests the label rather than the geometry. Consistent with this, spherical mixture structure and per-class dispersion track identity rather than autism trait differences in dyads; frequency-band and non-oscillatory activity ablation controls do not change the results. However, participant-level model does separate from its identity-aware null (p = 0.0099) while speaker-versus-listener role analysis performs at chance in the same embedding, indicating a manifold organized by individual -- and, in contrast with current neurolinguistics models, almost invariant to speaking vs. listening. Based on these results, we suggest that retraining-based nulls should be the default for grouped-data contrastive embeddings.
Huang, Hubert
fbb9c453-33a9-4988-8b17-d8f9db41446a
Mcleod, Michelle
6103710b-3805-4a96-a93c-3d5c4847ca77
Ames, Brendan
8ca36119-6cf2-495b-9cf0-983c976e12f7
Malaia, Evie
7d1b42ff-6599-4b36-9c96-ed66d4490e77
Huang, Hubert
fbb9c453-33a9-4988-8b17-d8f9db41446a
Mcleod, Michelle
6103710b-3805-4a96-a93c-3d5c4847ca77
Ames, Brendan
8ca36119-6cf2-495b-9cf0-983c976e12f7
Malaia, Evie
7d1b42ff-6599-4b36-9c96-ed66d4490e77

Huang, Hubert, Mcleod, Michelle, Ames, Brendan and Malaia, Evie (2026) Contrastive neural embeddings reveal individual traits beyond conversational role. NeurIPS 2026: NEUREPS 2026 - Symmetry and Geometry in Neural Representations, Australia, Sydney, Australia. (In Press)

Record type: Conference or Workshop Item (Paper)

Abstract

Contrastive representation learning is increasingly used to recover low-dimensional structure from neural recordings, but its output is typically validated by decoding accuracy rather than by the geometry of the manifold it produces. We apply CEBRA to EEG recorded from dyads in conversation, and analyze the resulting embedding, which training constrains to the 2D sphere. Labels describing the dyads, including the absolute difference between partners' autism-quotient scores, decode well above chance (0.77 against a 0.55 majority baseline for binary AQ magnitude; 0.44 against 0.25 for the six-class |ΔAQ| partition). However, the two permutation controls have notable differences in results: permuting labels over a frozen embedding yields p = 0.001, whereas retraining the encoder under each permutation yields p = 0.50. Only the latter tests the label rather than the geometry. Consistent with this, spherical mixture structure and per-class dispersion track identity rather than autism trait differences in dyads; frequency-band and non-oscillatory activity ablation controls do not change the results. However, participant-level model does separate from its identity-aware null (p = 0.0099) while speaker-versus-listener role analysis performs at chance in the same embedding, indicating a manifold organized by individual -- and, in contrast with current neurolinguistics models, almost invariant to speaking vs. listening. Based on these results, we suggest that retraining-based nulls should be the default for grouped-data contrastive embeddings.

Text
2610.03410v1
Download (3MB)

More information

Submitted date: 24 August 2026
Accepted/In Press date: 30 September 2026
Venue - Dates: NeurIPS 2026: NEUREPS 2026 - Symmetry and Geometry in Neural Representations, Australia, Sydney, Australia, 2026-12-11

Identifiers

Local EPrints ID: 515201
URI: https://eprints.soton.ac.uk/id/eprint/515201
PURE UUID: 02f5c9e3-438e-4cec-bb14-2cf4e543979e
ORCID for Brendan Ames: ORCID iD orcid.org/0000-0003-0810-7956

Catalogue record

Date deposited: 07 Oct 2026 16:52
Last modified: 08 Oct 2026 02:45

Export record

Contributors

Author: Hubert Huang
Author: Michelle Mcleod
Author: Brendan Ames ORCID iD
Author: Evie Malaia

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of https://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×