Binding in Short-Term Memory across Modalities in Younger and Older Adults with Autism Spectrum Disorder

Background:

Adults with Autism Spectrum Disorder (ASD) show difficulties in long-term memory which are assumed to result from altered relational binding (Bowler et al., 2011). Relational binding is the process by which elements of an episode are related to form a coherent episodic representation and to enable flexible retrieval. There is some evidence that short-term memory can also be impaired in ASD, for example when asked to remember the temporal order of item presentation such as digits, words (Poirier et al., 2011) or locations in a grid (Bowler et al., 2016). Similar to ASD individuals TD older adults (TD OA) present an associative deficit (Naveh-Benjamin, 2000) and, therefore, they show lower memory especially in relational tasks compared to younger TD individuals, which was found for long-term (Rotblatt et al., 2015) and short-term memory (Lecouvey et al., 2015). The ageing analogy of ASD describes similarities in memory in ASD and TD OA (Bowler, 2007) in that the memory difficulties ASD individuals show at younger age are similar to that of TD OA. This was confirmed for order memory (Ring et al., 2016).

Objectives:

This study aimed to compare systematically short-term memory across different modalities to investigate whether it is impaired in ASD similarly to long-term memory when it depends on relational binding and to test which task is most difficult. In addition, the ageing analogy was tested. It was predicted that short-term memory in ASD would be reduced compared to TD adults and that age would have a significant negative effect on TD but not ASD memory.

Methods:

Fifty-three ASD and 54 typically developing (TD) participants matched on gender, chronological age (CA; M_CA = 43.19), and intelligence quotient (IQ; M_FIQ = 111) were tested with three tasks testing verbal (sequence of letters), visuo-spatial (sequence of crosses in the cells of a grid) and multimodal short-term memory span (sequence of letters in the cells of a grid).

Results:

TD participants had a significantly longer span compared to ASD individuals in all tasks (F = 12.24, p < .01, η_p² = .10). The multimodal task was hardest for both groups, followed by the visuo-spatial task (F = 47.97, p < .0001, η_p² = .31). ASD participants made more mistakes than TD participants (F = 6.37, p < .05, η_p² = .06). Whereas age had a significant negative effect on TD individuals’ memory (R² = .13, F(1,52) = 8.05, p = .006), short-term memory in younger and older ASD individuals did not differ significantly (R² = .01, F(1,51) = 0.53, p = .47).

Conclusions:

Similar to long-term memory, ASD individuals show difficulties in relational short-term memory that are similar across the verbal, visuo-spatial and multimodal domain. The analysis of the effect of age on short-term memory found confirmation for the ageing analogy. Age had a significant negative effect on TD but not ASD individuals’ relational short-term memory. ASD might be protective against the effects of age on memory.