Interpreting EEG and MEG signal modulation in response to facial features: the influence of top-down task demands on visual processing strategies

Frei, Luisa Sophie (2012) Interpreting EEG and MEG signal modulation in response to facial features: the influence of top-down task demands on visual processing strategies. PhD thesis, University of Glasgow.

Abstract

The visual processing of faces is a fast and efficient feat that our visual system usually accomplishes many times a day. The N170 (an Event-Related Potential) and the M170 (an Event-Related Magnetic Field) are thought to be prominent markers of the face perception process in the ventral stream of visual processing that occur ~ 170 ms after stimulus onset. The question of whether face processing at the time window of the N170 and M170 is automatically driven by bottom-up visual processing only, or whether it is also modulated by top-down control, is still debated in the literature. However, it is known from research on general visual processing, that top-down control can be exerted much earlier along the visual processing stream than the N170 and M170 take place. I conducted two studies, each consisting of two face categorization tasks. In order to examine the influence of top-down control on the processing of faces, I changed the task demands from one task to the next, while presenting the same set of face stimuli. In the first study, I recorded participants’ EEG signal in response to faces while they performed both a Gender task and an Expression task on a set of expressive face stimuli. Analyses using Bubbles (Gosselin & Schyns, 2001) and Classification Image techniques revealed significant task modulations of the N170 ERPs (peaks and amplitudes) and the peak latency of maximum information sensitivity to key facial features. However, task demands did not change the information processing during the N170 with respect to behaviourally diagnostic information. Rather, the N170 seemed to integrate gender and expression diagnostic information equally in both tasks. In the second study, participants completed the same behavioural tasks as in the first study (Gender and Expression), but this time their MEG signal was recorded in order to allow for precise source localisation. After determining the active sources during the M170 time window, a Mutual Information analysis in connection with Bubbles was used to examine voxel sensitivity to both the task-relevant and the task-irrelevant face category. When a face category was relevant for the task, sensitivity to it was usually higher and peaked in different voxels than sensitivity to the task-irrelevant face category. In addition, voxels predictive of categorization accuracy were shown to be sensitive to task-relevant, behaviourally diagnostic facial features only. I conclude that facial feature integration during both N170 and M170 is subject to top-down control. The results are discussed against the background of known face processing models and current research findings on visual processing.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: cognitive neuroscience, EEG, MEG, Bubbles, face perception, vision, mutual information, classification images, reverse correlation, task effects, top-down, bottom-up
• Subjects: B Philosophy. Psychology. Religion > BF Psychology
• Colleges/Schools: College of Science and Engineering > School of Psychology
• Supervisor's Name: Schyns, Prof. Philippe G. and Gross, Prof. Joachim
• Date of Award: 2012
• Depositing User: Dr Luisa Sophie Frei
• Unique ID: glathesis:2012-3025
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 23 Nov 2011
• Last Modified: 14 Mar 2023 09:20
• Thesis DOI: 10.5525/gla.thesis.3025
• URI: https://theses.gla.ac.uk/id/eprint/3025