Brain Meeting: Helen Blank

Please contact ion.fil.brainmeetings@ucl.ac.uk for a Zoom link.

Abstract: In this talk, I will present recent findings on how prior expectations shape the perception and neural representation of faces. Using a combination of computational modeling, fMRI, and eye-tracking, we investigated how expected and unexpected faces are processed in the brain and guide behavior. Participants viewed cued face images, including ambiguous morphs, and were faster to identify expected faces, showing a perceptual bias towards prior expectations. FMRI analyses revealed that expected face representations were sharpened in early face-processing regions, while unexpected faces triggered prediction error activity extending to high-level areas.

Further, we explored how the strength, or precision, of these expectations is encoded. When participants anticipated specific faces with varying probabilities based on scene cues, we found that high-level face-sensitive areas reflected the precision-weighted expectations, enhancing similarity to expected face patterns, while unexpected faces activated a prediction error network in regions such as the caudate and insula.

Additionally, eye-tracking data demonstrated that participants’ eye movements aligned with their expectations, with predictive saccades and longer fixations on anticipated features, indicating that both top-down and bottom-up processes drive the sampling of facial information. Together, these findings illustrate that prior expectations not only facilitate face recognition but also shape face representations by integrating graded, content-specific expectations across hierarchical brain networks and guiding sensory sampling.