Verbal and Spatial Working Memories in Children with High-Functioning Autism Spectrum Disorder, and Their Relationships with Symptom Severity and Adaptive Function

Background: Whether children with autism spectrum disorder present a deficit in working memory, a component of Executive Function (EF), remains controversial in past literatures. In Baddeley’s model (1986), working memory is depicted as two slave systems, the phonological loop and the visuospatial sketchpad, under a central executive system. Recently, Williams et al. (2005) reported spatial working memory (SWM) impairment in 8- to16-year-olds with high-functioning autism spectrum disorder (HFASD); while their verbal working memory (VWM) was found unaffected. The present study further examines if similar patterns also exist in younger children with HFASD, and if psychopathology of ASD is associated with the patterns of working memory.

Objectives: We aim to (1) examine the verbal and spatial working memories in young children with HFASD, to (2) explore working memory challenges encountered in their daily activities, and to (3) examine the correlation between the performance of two working memories and symptom severity. 

Methods: We recruited 30 children aged 7–10 years with HFASD, and 27 control participants (TD) with matched Age, Verbal, Performance, and Full Scale IQ. Diagnosis was verified with ADI-R and ADOS. All HFASD children met diagnostic criteria for ASD in DSM-5. The VWM was measured by Digital Symbol subtest (including forward and backward digit span) from WISC-III, and SWM was measured by Spatial Memory subtest from Leiter-R. The BRIEF (Behavior Rating Inventory of Executive Functions) was completed by parents for daily life executive performance of their children. Parents were also interviewed by using the VABS-II (Vineland Adaptive Behavior Scales, 2^nd Edition) to assess the adaptive function of their children.

Results: The results showed that HFASD shows statistically significant scores in all subscales of BRIEF compared to TD. The most salient disparity was found in working memory subscale (F(1,55) = 41.512, p = .000). Moreover, we found a significant difference between the HFASD and TD in SWM, whereas HFASD performed poorly (F(1,55) = 7.628, p = .008) compared to TD. Nevertheless, no difference was found in forward- or backward-VWM. A strong correlation between SWM and backward-VWM was found (r = .652, p < .000). Although no correlation was found in working memories and autistic symptom severity, we found HFASD with higher performance on SWM had better performance IQ (F(1,28) = 6.230, p = .019) as well as higher socialization domain in adaptive function (F(1,28) = 4.860, p = .036).  

Conclusions: The present study suggests parents’ reports of working memory deficit in their autistic children was the most challenging issue among all other EF components, which is different from past studies. When IQ and age is matched, children with HFASD show only spatial working memory deficit. It indicates a dissociation between verbal and spatial working memory in HFASD, supporting the conclusion made by Williams et al. (2005). Although the deficit may not be explained by the psychopathology in ASD, it is associated with decreased socialization domain in adaptive function. Possible influencing factors and the implications for intervention to the spatial working memory in HFASD children are also discussed in details.