Abstract
Previous studies (using visual search, change blindness, and the attentional blink) have shown that when complex images such as faces are extremely familiar, they are encoded more quickly and require less attention to be processed than other less familiar stimuli. This suggests that repeated exposure to stimuli alters perceptual processing. Another body of work has shown that emotionally salient stimuli are also processed faster and require less attention than non-valenced stimuli. It in not clear, either for familiarity or emotional valence, how perceptual processing advantages are achieved. To address this question and to bring familiarity and emotionality together, we combined an associative learning task with visual perceptual tasks. Animal studies indicate that learned reward and loss value of stimuli have different effects on amygdala activity (Paz et al., 2006) that could potentially modulate activity in visual cortex via amygdalate-cortical pathways (Amaral & Price, 1984). Thus associative learning might modulate perceptual processing. To test this we exposed participants to face stimuli that were differentially associated with reward or loss in an effort to control the emotional value of faces whilst controlling for their perceptual exposure. Participants learned associations with computer-generated faces by playing a simple betting game. Faces were shown in pairs and participants had to choose between them in attempt to maximize payoff. Each face was associated with a probability (high or low) of gain, loss, or no financial outcome. (Different faces were used for different participants eliminating item effects.) After training, we quantified face perception by measuring the minimum stimulus onset asynchrony between a target face stimulus and a scrambled-face mask needed for criteria detection. We also measured detection of trained faces as the second target in an attentional blink task. The results indicate that reward and punishment learning have consequences for perceptual processing.