Exploring the Relationship between Face-Specific Activity and Conscious Face Perception

Introduction

Face perception is a fundamental cognitive process that enables humans to effectively identify faces in the environment, thus facilitating better social interactions. Extensive research has identified a specific region in the ventral visual cortex of the brain that shows significant increased activity in response to facial stimuli. However, it remains unclear whether this face-specific sensitivity is directly related to face perception (e.g., subjective perception). Comparing neural activity under different perceptual states can directly demonstrate the role of these signals in conscious face perception.

The main purpose of this study is to investigate whether face-specific activity in the ventral visual cortex is associated with conscious face perception. To this end, researchers used human intracranial electroencephalography (electrocorticography, ECoG) technology, which provides direct measurements of the brain’s electrophysiological responses with high spatial and temporal resolution. Through this method, researchers were able to compare neural activity when subjects “saw” faces (“seen” condition) and when they “did not see” faces (“unseen” condition), thereby exploring the association between these activities and face perception performance.

Authors and Publication Background

This paper was co-authored by Wenlu Li, Dan Cao, Jin Li, Tianzi Jiang, and other researchers. They are affiliated with the Brainnetome Center at the Institute of Automation, Chinese Academy of Sciences; School of Artificial Intelligence, University of Chinese Academy of Sciences; School of Psychology, Capital Normal University; Zhejiang Laboratory; and Xiangbrain Health Research Institute, Yongzhou Central Hospital. This paper was published in the journal “Neurosci. Bull.” on November 25, 2024.

Research Process

Data Source and Recording

The human brain data used in this study came from a public repository (https://searchworks.stanford.edu/view/zk881ps0522), which includes ECoG recording data provided by epilepsy patients who volunteered to participate. These patients participated in the experiment during clinical monitoring at Harborview Hospital in Seattle. The equipment used in the study included Neuroscan’s synamps2 amplifier and the BCI2000 EEG stimulation and acquisition program. Subjects viewed pictures of faces and houses while maintaining fixation through a visual task.

Experimental Design

The experiment included two tasks: “faces_basic” and “faces_noisy”. In the “faces_basic” task, subjects viewed grayscale images of faces and houses displayed in random order (each image shown for 400 milliseconds) and reported a simple target (inverted house). In the “faces_noisy” task, subjects performed a face detection task, with images noise-added using a “phase perturbation” method to test the perceptual threshold of the images.

Statistical Analysis and Data Preprocessing

Significant differences between experimental conditions were analyzed using t-tests, analysis of covariance (ANCOVA), and non-parametric cluster-based permutation tests. ECoG data preprocessing included denoising, re-referencing, baseline correction, and time-frequency decomposition using Morlet wavelets to extract broadband gamma activity (BGA) power spectrum values for statistical analysis.

Electrode Localization and Face-Selective Site Definition

Electrode localization within the ventral visual cortex was performed using the database and BrainNet View toolbox. Electrodes were defined as “face-selective” if BGA for face images was significantly higher than baseline and house images 100 to 300 milliseconds post-stimulus, with 8 face-selective sites included in subsequent analyses.

Results

Behavioral Results

Subjects’ face detection accuracy under different noise levels showed a Logistic (S-shaped) change, and an S-shaped function was used to fit the behavioral measurement curve to determine the perceptual threshold of 43.8%. The behavioral measurement curve showed a sharp change within this range, indicating a transition in perception.

Correlation of Face-Specific Activity

By comparing BGA under “seen” and “unseen” conditions at different noise levels, results showed that face-specific BGA was significantly higher in the “seen” condition than in the “unseen” condition. Further, using a nearest neighbor classification algorithm predicted whether faces were seen with accuracy exceeding random guessing and permutation classification accuracy. Directional information transfer analysis indicated significantly increased information transfer in the face perception state, supporting the role of the ventral visual cortex in face perception.

Correlation between Face-Specific Activity and Perceptual Performance

Analysis of partial correlations between BGA and face perception performance (reaction time and detection accuracy) showed a significant positive correlation between BGA peak delay and subject response delay, as well as a significant positive correlation between BGA amplitude and face detection accuracy. Both showed significant trends varying with face detection accuracy under different noise levels.

Discussion

The research results indicate that face-specific activity in the ventral visual cortex is associated with human conscious face perception. This finding provides support for understanding the neurophysiological mechanisms of face perception: analysis of ECoG data revealed changes in brain activity when seeing faces. In particular, the increase in BGA and reliable neural responses reflect the formation of high-level abstract face representations in the ventral visual cortex. This provides direction for future exploration of face-specific activity.

Although the current study only explored the results of face perception, future research could further investigate the mechanisms of “seen” versus “unseen” images across different object categories. Additionally, this study emphasizes the key role of higher-order visual cortex in processing visual information, laying the foundation for a comprehensive understanding of the neural mechanisms of face perception.

Conclusion

Through precise experimental design and rigorous data analysis, this study successfully revealed the close connection between face-specific activity in the ventral visual cortex and conscious face perception, providing new insights into understanding the neurophysiological mechanisms of face perception.