Perceptual difficulty modulates the direction of information flow in familiar face recognition

Hamid Karimi-Rouzbahani, Farzad Ramezani, Alexandra Woolgar, Anina N. Rich, Masoud Ghodrati

Abstract

Humans are fast and accurate when they recognize familiar faces. Previous neurophysiological studies have shown enhanced representations for the dichotomy of familiar vs. unfamiliar faces. As familiarity is a spectrum, however, any neural correlate should reflect graded representations for more vs. less familiar faces along the spectrum. By systematically varying familiarity across stimuli, we show a neural familiarity spectrum using electroencephalography. We then evaluated the spatiotemporal dynamics of familiar face recognition across the brain. Specifically, we developed a novel informational connectivity method to test whether peri-frontal brain areas contribute to familiar face recognition. Results showed that feed-forward flow dominates for the most familiar faces and top-down flow was only dominant when sensory evidence was insufficient to support face recognition. These results demonstrate that perceptual difficulty and the level of familiarity influence the neural representation of familiar faces and the degree to which peri-frontal neural networks contribute to familiar face recognition.

Introduction

Traditional neuroscience often treats face recognition as a binary state: a face is either familiar or unfamiliar. However, human experience suggests a gradient of familiarity, where we recognize ourselves, close friends, and celebrities with varying degrees of efficiency. This study investigates the neural correlates of this familiarity spectrum and explores how the brain's internal communication—specifically the balance between bottom-up sensory processing and top-down cognitive feedback—adapts to the clarity of visual evidence.

Methods

Researchers recorded brain activity from 18 participants using Electroencephalography (EEG). To test the limits of recognition, they employed a dynamic noise masking technique where the phase coherence (clarity) of face images was varied from 22% (high noise) to 55% (low noise). The noise pattern was updated every 16.7ms, requiring the brain to integrate visual information over time.

Experimental paradigm and general behavioral performance. Panel A illustrates the stimuli at varying coherence levels and the trial structure. Panels B and C show the overall effect of coherence on accuracy and reaction time.

A novel informational connectivity method based on Representational Similarity Analysis (RSA) was developed to track the flow of identity-specific information between peri-occipital (sensory) and peri-frontal (cognitive) electrode clusters.

Results

The Familiarity Spectrum

Behavioral and neural data confirmed a graded representation of familiarity. Participants identified "Self" faces most rapidly and accurately, followed by "Personally Familiar" (friends/family) and "Famous" faces. Neural decoding of EEG signals mirrored this behavioral hierarchy, with more familiar faces producing more distinct neural signatures.

Behavioral performance across the familiarity spectrum. Panels D and E show accuracy and reaction time for Self, Personally Familiar, and Famous faces across coherence levels.

Dynamic Information Flow

The study revealed a significant shift in brain communication strategy based on perceptual difficulty and familiarity. At high coherence (55%), information flow was predominantly feed-forward, moving from sensory to frontal areas. Conversely, at low coherence (22%), top-down flow from peri-frontal areas became dominant. This suggests that when sensory input is ambiguous, the brain recruits higher-order cognitive regions to provide feedback based on internal templates. Furthermore, feed-forward flow was found to dominate for the most familiar faces, while top-down feedback was primarily recruited when sensory evidence was insufficient to support recognition.

Discussion

These findings indicate that the brain does not process all familiar faces identically. Instead, it dynamically adjusts its processing strategy based on both the level of familiarity and the quality of the visual signal. While the fast feed-forward sweep is sufficient for clear images, difficult perceptual conditions trigger a top-down recruitment of frontal networks. Future work utilizing fMRI-EEG fusion could provide better spatial localization of these feedback loops, potentially offering insights into the mechanisms underlying face-processing disorders like prosopagnosia.

Summary updated

Abstract

Abstract Humans are fast and accurate when they recognize familiar faces. Previous neurophysiological studies have shown enhanced representations for the dichotomy of familiar vs. unfamiliar faces. As familiarity is a spectrum, however, any neural correlate should reflect graded representations for more vs. less familiar faces along the spectrum. By systematically varying familiarity across stimuli, we show a neural familiarity spectrum using electroencephalography. We then evaluated the spatiotemporal dynamics of familiar face recognition across the brain. Specifically, we developed a novel informational connectivity method to test whether peri-frontal brain areas contribute to familiar face recognition. Results showed that feed-forward flow dominates for the most familiar faces and top-down flow was only dominant when sensory evidence was insufficient to support face recognition. These results demonstrate that perceptual difficulty and the level of familiarity influence the neural representation of familiar faces and the degree to which peri-frontal neural networks contribute to familiar face recognition.
BibTeX
@article{karimirouzbahani2021perceptual,
  author = {Hamid Karimi-Rouzbahani and Farzad Ramezani and Alexandra Woolgar and Anina N. Rich and Masoud Ghodrati},
  title = {{Perceptual difficulty modulates the direction of information flow in familiar face recognition}},
  journal = {NeuroImage},
  year = {2021},
  doi = {10.1016/j.neuroimage.2021.117896},
}
Figure