Decoding Task and Stimulus Representations in Face-responsive Cortex Decoding Task and Stimulus from Faces

• Published in: Cognitive neuropsychology Vol. 33; no. 7-8; pp. 362 - 377
• Main Authors: Kliemann, Dorit, Jacoby, Nir, Anzellotti, Stefano, Saxe, Rebecca R.
• Format: Journal Article
• Language: English
• Published: 01-01-2016
• ISSN: 0264-3294; 1464-0627
• DOI: 10.1080/02643294.2016.1256873

Faces provide rich social information about others’ stable traits (e.g., age) and fleeting states of mind (e.g., emotional expression). While some of these facial aspects may be processed automatically, observers can also deliberately attend to some features while ignoring others. It remains unclear how internal goals (e.g., task context) influence the representational geometry of variable and stable facial aspects in face-responsive cortex. We investigated neural response patterns related to decoding i) the intention to attend to a facial aspect before its perception, ii) the attended aspect of a face and iii) stimulus properties. We measured neural responses while subjects watched videos of dynamic positive and negative expressions, and judged the age or the expression’s valence. Split-half multivoxel pattern analyses (MVPA) showed that (i) the intention to attend to a specific aspect of a face can be decoded from left fronto-lateral, but not face-responsive regions; (ii) during face perception, the attend aspect (age vs emotion) could be robustly decoded from almost all face-responsive regions; and (iii) a stimulus property (valence), was represented in right posterior superior temporal sulcus and medial prefrontal cortices. The effect of deliberately shifting the focus of attention on representations suggest a powerful influence of top-down signals on cortical representation of social information, varying across cortical regions, likely reflecting neural flexibility to optimally integrate internal goals and dynamic perceptual input.