Neurophysiological evidence for crossmodal (face-name) person-identity representation in the human left ventral temporal cortex

Abstract

Putting a name to a face is a highly common activity in our daily life that greatly enriches social interactions. Although this specific person-identity association becomes automatic with learning, it remains difficult and can easily be disrupted in normal circumstances or neurological conditions. To shed light on the neural basis of this important and yet poorly understood association between different input modalities in the human brain, we designed a crossmodal frequency-tagging paradigm coupled to brain activity recording via scalp and intracerebral electroencephalography. In Experiment 1, 12 participants were presented with variable pictures of faces and written names of a single famous identity at a 4-Hz frequency rate while performing an orthogonal task. Every 7 items, another famous identity appeared, either as a face or a name. Robust electrophysiological responses were found exactly at the frequency of identity change (i.e., 4 Hz / 7 = 0.571 Hz), suggesting a crossmodal neural response to person identity. In Experiment 2 with twenty participants, two control conditions with periodic changes of identity for faces or names only were added to estimate the contribution of unimodal neural activity to the putative crossmodal face-name responses. About 30% of the response occurring at the frequency of crossmodal identity change over the left occipito-temporal cortex could not be accounted for by the linear sum of unimodal responses. Finally, intracerebral recordings in the left ventral anterior temporal lobe (ATL) in 7 epileptic patients tested with this paradigm revealed a small number of "pure" crossmodal responses, i.e., with no response to changes of identity for faces or names only. Altogether, these observations provide evidence for integration of verbal and nonverbal person identityspecific information in the human brain, highlighting the contribution of the left ventral ATL in the automatic retrieval of face-name identity associations.

A robust neural familiar face recognition response in a dynamic (periodic) stream of unfamiliar faces

“…It also opens new perspectives for investigating the functional neural networks of FFR by combining this frequency-tagging approach with fMRI (e.g., Gao et al, 2018) and intracerebral human recordings (e.g., Jonas et al, 2016). Besides, this approach could be extended in future studies to investigate biological markers for personally familiar face recognition, the effect of face learning, and multi-modality person recognition (e.g., names, voices, and the interactions of the modalities) (e.g., Volfart et al, 2020). It could also be applied in forensic and clinical settings to assess one's ability to recognize human faces, including super-recognizers, and people with developmental difficulties at face recognition.…”

Measuring self-similarity in empirical signals to understand musical beat perception

“…"-* ) is the standard deviation across the neighboring bins. This approach has been widely used in the frequency-tagging literature, as it offers a straightforward way to statistically test whether the response stands out significantly from the noise in the recorded signal (Liu-Shuang et al, 2014;Jonas et al, 2016;Lochy et al, 2018;Volfart et al, 2020;Hagen et al, 2021).…”

Identifying conceptual neural responses to symbolic numerals

“…The copyright holder for this preprint (which this version posted July 4, 2023. ; https://doi.org/10.1101/2023.07.04.547627 doi: bioRxiv preprint and/or numbers: Cohen, 2009;Lochy & Schiltz, 2019;Volfart et al, 2020;Marlair, Crollen & Lochy, 2022). Taken together, the limited variability within symbolic number exemplars, and the high variability across number exemplars, leaves responses highly susceptible to stimulus confounds.…”