Can you recognize the faces below amongst the noise? Maybe, if top-down frontal-to-occipital mechanisms play a major role in improving incoming sensory information. A recently published paper in NeuroImage by Hamid Karimi-Rouzbahani and colleagues found that when recognising familiar faces, feedback mechanisms dominated when the noise in the image was high, sending "category-relevant" information to sensory areas. When the stimulus was clear or highly familiar (e.g., one's own face), however, the direction of information flow was reversed (i.e., predominantly feed-forward). In order to derive these findings, Karimi-Rouzbahani et al. developed a novel "informational" connectivity analysis which targets and tracks the flow of specific aspects of information across the brain in EEG. In addition, building on recent studies which showed enhanced representations for familiar over unfamiliar faces (categorically defined), they showed that there is a gradient of face familiarity decodable from EEG activity corresponding to the increasing behavioural advantage for more over less familiar faces. The data and codes for their study are freely available at https://osf.io/s3j4z/ and the paper at https://doi.org/10.1016/j.neuroimage.2021.117896
