Multivariate Pattern Analysis of Event-related Potentials Predicts the Subjective Relevance of Everyday Objects

Potentially decision-relevant stimuli have been proposed to undergo immediate semantic processing. The current study investigated whether information regarding the general desirability ('Wanting') of visually presented 'everyday' objects was rapidly and automatically processed. Participants completed a foreground task while their electroencephalogram (EEG) was recorded, and task-irrelevant images were presented in the background. Following this, participants rated the images with regards to Wanting and the potentially related attributes of Relevance, Familiarity, Aesthetic Pleasantness and Time Reference. Multivariate pattern classification was used to predict the ratings from patterns of EEG data. Prediction of Wanting and Relevance was possible between 100 and 150ms following stimulus presentation. The other dimensions could not be predicted. Wanting and Relevance ratings were highly correlated and displayed similar feature weight maps. The current results suggest that the general desirability and subjective relevance of everyday objects is rapidly and automatically processed for a wide range of visual stimuli.