Self-Referential Processing in Autism: Does Valence Matter?

Background:  A strong sense of self incorporates both affective self-attributions, and social-cognitive processes. Individuals with high-functioning autism (HFA) endorse a less positive view of themselves than typically-developing individuals, and do not preferentially remember self-relevant over other-relevant information. These differences represent atypical organization of both affective and cognitive components of the self in individuals with autism. However, the associations between the affective encoding and later memory for personally-relevant information, and dissociable correlates of each have not been examined on the self-referenced memory task.

Objectives:  The goals of the current study were to examine 1) group differences in endorsement of positively- versus negatively-valenced information about self and others on the self-referenced memory task, 2) associations between endorsement and memory for self- and other-referenced information and 3) affective and socio-cognitive correlates of the endorsement and memory phase of the task in children with HFA and a matched comparison sample (COM). 

Methods:  Children and adolescents with HFA (N=76, 65 males, Mage=12.51, SD=2.58) and a comparison sample (N=72, 52 males, Mage=13.28, SD=2.15) completed a self-referenced memory paradigm, where they judged positive and negative trait adjectives with reference to themselves, and a familiar other person (i.e., Harry Potter). After a delay, participants were asked whether they recognized the adjectives from a list among distractor adjectives. An index of Self-Positivity Bias (SPB) was calculated as the difference in endorsement between self-referenced adjectives of positive and negative valence. In the recognition phase, self- and other-referenced memory, as well as the difference between the two (i.e., Preferential Self-referenced Memory) were computed. Self-reported Internalizing problems were evaluated by the Behavior Assessment Scale for Children (BASC). A Theory of Mind (ToM) composite was created by standardizing and averaging performance on two tasks: Reading the Mind in the Eyes and Strange Stories. 

Results:  Controlling for gender and VCI, HFA participants displayed a lower Self-Positivity Biases and reduced Preferential Self-referenced Memory, than COM participants, F(1, 145) = 6.98, p=.009, η²_partial=.05, and F(1, 145) = 21.27, p<.001, η²_partial=.13, respectively. The Self-Positivity Bias was predictive of Preferential Self-referenced Memory in the COM group, r(71)=.23, p=.049, but not the HFA group, r(75)=-.06, p>.05 (Figure 1). Self-reported internalizing problems were inversely related to Self-Positivity Bias (p’s<.05; Figure 2), but not preferential self-referenced memory in both groups. In contrast, the ToM composite was associated with memory performance in the self- and other-referential conditions, but not the difference score in the HFA (p’s<.001), but not COM group (p’s>.05).

Conclusions:  Typically-developing children demonstrated a coherent sense of self, where positive affective self-understanding supported enhanced socio-cognitive processing. However, the same was not true in children with HFA. Children with HFA may demonstrate a more dissociated sense of self, with positive self-perceptions linked to lower levels of internalizing problems, while recognition of self- and other-referenced adjectives supports social cognition. Two avenues of intervention could target the dissociated sense of self for children with HFA. Certain individuals may benefit from interventions that target negative self to reduce internalizing problems, while others need training of socio-cognitive skills to improve awareness of themselves and others.