Arousal and Emotion Recognition in Music Among Youth with Autism Spectrum Disorders

Background:  Many studies have investigated the so-called amygdala theory of autism (Baron-Cohen et al., 2000) with evidence signifying that atypical amygdala function likely underlies high levels of anxiety in autism spectrum disorders (ASD) (Amaral, Bauman, & Schumann, 2003).  The amygdala has been shown to be involved in the processing of emotions in music, specifically in fearful and sad emotions (Gosselin, Peretz, Johnsen, & Adolphs, 2007).  However, Quintin, Bhatara, Poissant, Fombonne, & Levitin (2011) found that high-functioning autism spectrum disorder (ASD) adolescents did not differ from typical (TYP) adolescents in the accuracy of their behavioral ratings related to recognizing emotions in music.  

Objectives:  We extended the Quintin et al. study in two important ways: we measured physiological response during stimulus presentation, and we included an additional, younger participant group. We hypothesized that individuals with ASD (the ASD group) would show more arousal during scary music, relative to age-and IQ-matched typical controls (the TYP group).

Methods:  There was a bimodal age distribution among participants.  The younger group consisted of 50 participants ages 8-11 (M=9.92, 25 ASD).  The older group consisted of 50 participants ages 15-18 (M=16.8, 25 ASD).  The task consisted of participants listening to 20 second musical clips in randomized order, including five of each emotion (sad, happy and scary).  After each clip participants assessed the emotion and intensity.  Skin conductance response (SCR) was collected during the task with amplitude being analyzed according to the area under the curve following each stimulus.

Results:  Behavioral measures of accuracy showed few reliable between-group differences. However, repeated measures ANOVA of the SCR showed that for the younger child group, there was a significant main effect for Emotion Condition, with the Scared condition showing higher arousal than the other conditions in both groups.  There was no main effect for diagnostic group, but there was a significant Group x Condition interaction, indicating a substantially greater response to the scared condition in the ASD group.  For the older teen group, there was a significant main effect for diagnostic group, as the mean SCR response for the ASD group (3.02μS) was significantly less compared to that of the TYP group (5.99μS) despite similar baseline arousal values.  There was no main effect for condition and no significant Group x Condition interaction in the older sample.  The ASD group but not the TYP group showed a main effect for age with younger ASD children more responsive overall than the older cohort. 

Conclusions:  Schumann et al. (2004) suggested that in ASD the amygdala is larger and perhaps overactive in younger children, with growth slower compared to controls in adolescence.  These data may reflect a functional consequence of this atypical growth pattern.  Future research regarding emotion recognition in ASD should attend to probable developmental changes. Different psychophysiological responsiveness in the face of similar behavioral reports suggests atypical integration of cognitive and affective cues in ASD; the study of mechanisms underlying observable behavior may be useful for understanding known strengths and weaknesses in behavioral performance.