Visual Working Memory and Filtering Ability in Individuals with Autism Spectrum Disorder

Background:  Past investigations have reported impaired working memory performance in individuals with Autism Spectrum Disorder (ASD) (Geurts, de Vris, & van den Bergh, 2014). It remains unclear, however, to what extent the previously observed performance differences reflect a decreased memory capacity, difficulties with preventing irrelevant information from filling capacity (i.e., filtering ability), and/or disruption in focused attention.

Objectives:  The current study assessed the contribution of core processes (e.g. working memory capacity, attention, and visual filtering abilities) to visual working memory performance in adolescents and adults with and without ASD.

Methods:  Data from 48 subjects (24 with ASD; 24 without ASD) between 16 and 24 years old were analyzed. Estimated Full Scale IQ for all subjects was >80. Participants completed a computerized paradigm designed to systematically assess capacity, attention, and visual filtering abilities during visual working memory performance (Cowan et al., 2011; Mall et al., 2014). Subjects were shown visual arrays consisting of 2, 3, 4, 6, or 8 colored stimuli (circles and/or squares). After a short delay, memory for the color of one of the stimuli was probed (same or different color). Importantly, subjects were informed beforehand that the one of the shapes (e.g., circles; high frequency probes) was more likely to be probed compared to the other shape (e.g., squares; low frequency probes), allowing them to allocate their attention and filtering accordingly. Three different blocks of trials were presented: [1] trials with only one shape (100% high frequency probes), [2] trials with both shapes but only one shape probed (100% high frequency probes), and [3] trials with both shapes and both shapes probed (75% high frequency and 25% low frequency probes). Eye tracking data was simultaneously collected.

Results:  Results revealed comparable estimates of overall working memory capacity between groups (maximum mean capacity: ASD = 3.81 and non-ASD = 3.84), F(1,46) < 1, p = 0.44, h_P² = 0.01. However, performance for individuals with ASD was more impacted by increases in attention and visual filtering demands than individuals without ASD. Individuals with ASD also allocated their attention differently than non-ASD individuals and spent less time looking at relevant information, especially as the task increased in complexity (e.g., 25% versus 38% of time looking at relevant information for 8 stimuli), F(1,46) = 4.11, p = 0.01, h_P² = 0.08. The ASD group had more difficulty filtering distracting information in the most challenging condition (8 stimuli, high/low frequency), t(46) = 2.19, p = 0.03, Cohen’s d = 0.63. Specifically, individuals with ASD were less effective at allocating working memory capacity to the high versus low frequency stimuli in comparison to individuals without ASD.

Conclusions:  Findings suggest that visual working memory performance is similar between individuals with and without ASD when cognitive demands are low, but individuals with ASD are detrimentally affected when the cognitive load increases (increased attention and visual filtering demands). Given the complexity of our environments and need to filter visually distracting information, these findings may shed light on ASD-related difficulties in day-to-day functioning and provide a focus for intervention.