Date of Award

5-2026

Document Type

Thesis

Degree Name

Master of Arts in Psychological Science

Department

Psychology

First Reader/Committee Chair

Brunet, Nicolas

Abstract

Faces are among the most socially significant stimuli humans encounter, requiring recognition not just at the category level but at the level of individual identity. While the fusiform face area has been established as a dedicated region for face processing, whether it encodes distinct representations for different individuals or simply detects the presence of a face remains an open question. Evidence for identity-specific coding in the brain has come primarily from invasive single-neuron recordings in epilepsy patients, leaving open whether the same principles govern face processing in healthy populations. The present study addressed this gap using scalp electroencephalography to examine whether the N170 event-related potential, an early face-sensitive brain response, is modulated by identity variation in multi-face displays.

Two experiments were conducted, each presenting participants with 2x2 grids of faces that systematically varied in the degree of facial redundancy and identity variation. In Experiment 1, 36 participants viewed displays of unfamiliar faces across three conditions: a single large face, four identical copies of the same face, and four different individuals. In Experiment 2, 34 participants viewed displays of celebrity faces across three conditions: four identical photos of the same celebrity, four different photos of the same celebrity, and four photos of four different celebrities. This second experiment introduced a critical middle condition that allowed the study to dissociate image-level variability from identity-level variability. ERP components were measured across occipital, left temporal, and right temporal electrode clusters, with the N170 serving as the primary component of interest and the P100 and P200 examined as secondary components.

In Experiment 1, the N170 was significantly larger for displays of four different individuals than for four identical copies of the same face at occipital sites, providing evidence that identity variation modulates early face processing above and beyond what face count alone can explain. This finding is consistent with sparse coding principles, which predict that distinct identities should activate partially separate neural populations. In Experiment 2, the N170 was sensitive to image-level variability, responding more strongly to different photographs of the same celebrity than to identical repeated images, but did not distinguish between different images of the same identity and images of entirely different individuals. The P200 showed a similar pattern in Experiment 2, responding to image-level variability but not to identity-level differences. Control analyses confirmed that none of the observed ERP effects were attributable to low-level visual differences between conditions such as luminance, contrast, color saturation, or spatial frequency content.

These findings offer partial support for identity-sensitive coding at the level of the N170 and suggest that detecting such coding non-invasively may require conditions that more directly engage identity processing, such as familiar faces or explicit identity judgment tasks. By establishing a non-invasive baseline for early identity-sensitive face processing in healthy participants, this study lays the groundwork for future research into populations where face processing is disrupted, including individuals with prosopagnosia, autism spectrum disorder, and schizophrenia.

Keywords: Fusiform Face Area, N170, Event-Related Potentials, Multiple Faces, Sparse Coding

Download

  Contact Author

Share

COinS