Neural Noise- During auditory narratives, EEG and physiological signals show temporal correlations across subjects (intersubject correlation, ISC), which increase when attention is directed to the story.1
- Identifying what drives these shared responses is clinically relevant for interpreting EEG ISC in unresponsive patients.1 1
Weekly enrichment (2026-07-20)
- The preprint (bioRxiv, posted 20 Feb 2026, doi 10.64898/2026.02.19.706583) is by Emilia Fló, Álvaro Cabana, Juan Valle-Lisboa, Damian Cruse, Jens Madsen, Lucas C. Parra and Jacobo Sitt, spanning the Paris Brain Institute (ICM/Sorbonne), UdelaR Montevideo, the University of Birmingham and City College of New York.2
- The study dissected whether EEG ISC during auditory narratives is driven by low-level acoustic features (envelope, spectrogram) versus higher-level linguistic features (word onset, word surprisal), using two groups of healthy participants across passive, attentive and distracted listening conditions.2
- Contributions of each feature were estimated with temporal response functions (TRFs), while ISC itself was computed using correlated component analysis (CorrCA / CCA), a standard method for isolating shared spatiotemporal components across subjects.2
- Acoustic-feature TRFs explained a larger fraction of EEG variance than linguistic-feature TRFs and were the main contributors to ISC, indicating that shared stimulus acoustics dominate the synchronization.2
- The attention-related increase in ISC was driven by all features, but word surprisal affected ISC only during active story engagement, with timing and scalp topography consistent with genuine language processing.2
- Linear TRF responses accounted for only a small portion of total ISC, implying that ISC is largely generated by nonlinear responses to the narratives, a caveat for interpreting linear-model markers.2
- The authors propose combining ISC and TRFs as markers of language processing in patients with disorders of consciousness (DOC) and offer practical implementation recommendations; competing interests were declared as none.2
- This builds on prior work showing DOC patients have significantly lower narrative-speech ISC than healthy controls and, unlike controls, fail to differentiate forward from time-reversed speech, so ISC can index clinical diagnosis at the bedside where motor-based assessment fails.3