Neural Correlates of Affective Priming in ASD

Background: People take less time to evaluate a word as emotionally positive or negative if it is preceded by another word of the same emotional valence. This effect is called affective priming and is thought to reflect early bottom-up influence  of emotional valence of the first (prime) word on the recognition of the second (target) word and on response preparation. There is some evidence that individuals with ASD process emotional stimuli differently, however, it is not clear whether this difference occurs at an early or later stage.

Objectives: To investigate behavioral and neural correlates of affective priming in ASD using reaction times (RT) and EEG.

Methods: Our aim is to recruit 20 participants with ASD and 20 matched healthy controls. Participants read prime-target word pairs on the screen and responded as fast as possible whether the target was emotionally positive or negative. Time interval between the onset of the prime and the target was 300 ms. In addition, we measured Autism Quotient (AQ) and Empathy Quotient (EQ) in all participants.

Results: Preliminary analysis (7 ASD and 14 controls) revealed that that priming effect for positive target words significantly correlated with AQ (Pearson's r=0.46, p<.05) and EQ (r=-0.47, p<.05). Priming effect for negative target words showed a marginal inverse correlation with AQ only (r=-0.39, p=.08). In other words, participants with ASD were significantly speeded up by affectively congruent positive primes, but less so by congruent negative primes. In the EEG data, target words preceded by an affectively incongruent compared to congruent prime elicited an increased Late Positive Component (LPC) amplitude in the time window between 500-700 ms. The magnitude of the LPC effect positively correlated with AQ score of the participants. Looking at positive and negative words separately, we found that AQ score correlated significantly with the LPC effect for negative target words (r=0.7, p<.01), but not for positive words.

Conclusions: Participants with ASD seem to inefficiently process negative valence specifically. In participants with ASD, but also in controls with high AQ scores, presenting an emotionally negative prime did not lead to faster evaluation of negative targets, suggesting that it was either ignored or not processed in time. LPC amplitude which is associated with increased processing effort and deeper stimulus encoding was also found to correlate with AQ scores for negative targets only. While in participants with low AQ negative targets elicited high LPC amplitudes irrespective of the priming condition, in participants with high AQ congruent negative prime-target pairs elicited much lower LPC, suggesting that it was not efficiently encoded. We conclude that processing of negative emotion is impaired in ASD, while processing of positive emotion is relatively unaffected.