This dataset contains EEG and behavioural measures taken from participants observing images of faces and cars, with different expectations. Probabilistic information associated with a high and low expertise category (faces and cars respectively) was manipulated, and behavioural and EEG data was collected for conditions with high and low face expectation. Raw data is provided for 57 participants. The data was collected to study how neural signals of expectation effects on face processing relate to a measure of face expertise (the Cambridge Face Memory task - long; also provided in the dataset alongside with the Cambridge Cars Memory Test).