Do Models Hear Like Us? Probing the Representational Alignment of Audio LLMs and Naturalistic EEG
This work addresses the problem of understanding neural alignment in audio models for researchers in AI and neuroscience, but it is incremental as it builds on existing probing methods with new datasets and metrics.
The study investigated whether the internal representations of 12 open-source Audio LLMs align with human EEG signals during naturalistic listening, revealing rank-dependence splits, spatio-temporal alignment patterns like peaks in the 250-500 ms window, and affective dissociation where negative prosody reduces similarity but enhances covariance.
Audio Large Language Models (Audio LLMs) have demonstrated strong capabilities in integrating speech perception with language understanding. However, whether their internal representations align with human neural dynamics during naturalistic listening remains largely unexplored. In this work, we systematically examine layer-wise representational alignment between 12 open-source Audio LLMs and Electroencephalogram (EEG) signals across 2 datasets. Specifically, we employ 8 similarity metrics, such as Spearman-based Representational Similarity Analysis (RSA), to characterize within-sentence representational geometry. Our analysis reveals 3 key findings: (1) we observe a rank-dependence split, in which model rankings vary substantially across different similarity metrics; (2) we identify spatio-temporal alignment patterns characterized by depth-dependent alignment peaks and a pronounced increase in RSA within the 250-500 ms time window, consistent with N400-related neural dynamics; (3) we find an affective dissociation whereby negative prosody, identified using a proposed Tri-modal Neighborhood Consistency (TNC) criterion, reduces geometric similarity while enhancing covariance-based dependence. These findings provide new neurobiological insights into the representational mechanisms of Audio LLMs.