arXiv：cs.LG· Hubert Huang, Michelle McCleod, Brendan Ames, Evie Malaia·· 5 小时前AI 评分39
对比神经嵌入揭示对话角色之外的个体特质
Contrastive Neural Embeddings Reveal Individual Traits Beyond Conversational Role
AI 导读
研究将 CEBRA 应用于对话中双人 EEG 记录，分析其约束在 2D 球面上的嵌入。描述双人特征的标签解码显著高于随机水平：二分类 AQ 幅度为 0.77（多数基线 0.55），六类 |ΔAQ| 划分为 0.44（基线 0.25）。但冻结嵌入上置换标签得 p = 0.001，而每次置换下重训编码器得 p = 0.50，仅后者检验标签而非几何结构。
正文
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 $|\Delta$AQ$|$ 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.
| Subjects: | Neurons and Cognition (q-bio.NC); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.03410 [q-bio.NC] |
| (or arXiv:2610.03410v1 [q-bio.NC] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03410 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Evie Malaia [view email]
[v1]
Fri, 2 Oct 2026 15:00:02 UTC (3,938 KB)
来源：arXiv：cs.LG · arxiv.org