Memory Deficits for Faces and Non-Social Stimuli in Children with ASD

Background: The ability to remember faces is important for normal social communication. Despite mounting research on face processing deficits in children with autism spectrum disorder (ASD), little is known about their face memory deficit, particularly regarding its domain-specificity and the impact of memory demand. Moreover, the relation between the face memory deficit and affect processing in ASD has not been thoroughly researched, although the emotion recognition deficit is a key diagnostic criterion. Hence, a systematic investigation of the face memory deficit and its relation with affect processing in ASD is desired for improving diagnosis and treatment of affected individuals.

Objectives:  This study aimed to investigate (a) whether children with ASD show a specific memory deficit for faces compared to non-social objects, (b) whether memory demand (i.e., time delay) affects memory performance in children with ASD, and (c) whether memory for faces is associated with affect processing.

Methods:  Preliminary analysis on a sample of 8 children with ASD and 10 typically-developing (TD) controls (age range=8-12 yrs, M=10.72 yrs) is presented below. Memory for faces and non-social stimuli was assessed using (a) Memory for Faces and Memory for Designs subtests (both immediate and delayed) from the Developmental Neuropsychological Assessment Battery (NEPSY-II), and Faces and Dot Locations subtests (both short and long delay) from the Children’s Memory Scale (CMS). Affect processing was assessed using the Affect Recognition subtest from NEPSY-II. Standardized scores of all these assessments were used.

Results:  A 2x2x2 ANOVA with Domain (Face vs. Non-social), Delay (Short vs. Long), and Group (ASD vs. TD) on the NEPSY-II Memory for Faces and Memory for Designs subtests revealed a significant effect of group, F(1,16)=15.13, p<0.001, indicating overall lower memory ability in children with ASD (M=9.47) compared to TD children (M=12.63). No other main effects or interactions were significant. This result was replicated on the CMS Faces and Dot Locations Memory subtests; a 2x2x2 ANOVA with Domain (Face vs. Non-social), Delay (Immediate vs. Long), and Group (ASD vs. TD), F(1,16)=6.71, p<0.02, showed inferior memory performance in ASD children regardless of domain and time delay. Further correlation analysis found a significant correlation between long delayed face memory and affect recognition scores on NEPSY-II across all individuals, r(16)=0.647, p<0.004, and a significant relation was only observed in TD, r(8)=0.774, p<0.009 but not ASD, r(6)=0.058, p=0.891.

Conclusions: Although further analysis with a larger sample size is needed to corroborate these findings, current preliminary analysis revealed impaired memory for faces in children with ASD compared to their TD peers. However, this deficit was not specific to faces, and memory demand had no effect on its severity. Rather, children with ASD appear to have a general memory deficit across social and non-social domains. Furthermore, face memory seems to associate with affect processing in TD, but this relationship is likely to be disrupted in children with ASD.