The Role of Integration in the Memory Deficits of ASC

Background: The study of memory functioning in ASC reveals a complex pattern of strengths and impairments, and many mixed findings. However recent studies have begun to converge on some recurring themes. Firstly, findings suggest individuals with ASC seem not to utilise semantic information to aid their recall in the same way as neurotypical individuals. This often results in impairment, for example when recalling a list of semantically related items (e.g. Toichi and Kamio, 2003; Renner, Klinger and Klinger, 2004; Tager-Flusberg, 1991; Minshew and Goldstein, 1992). Individuals with ASC also appear to be impaired in processing relations between items in memory (Gaigg, Gardiner and Bowler, 2008; Bowler, Gaigg and Gardiner, 2009). This could explain impairments seen in episodic memory and source memory, as recalling context requires relations between different elements of an event to be processed. Both the processing of inter-item relationships, and the use of semantic information to aid recall, requires integration of memory structures. Relational processing requires on-line integration of activated memory representations within working memory, both at the encoding and retrieval stages. Integration is also needed to enable semantic information to be used to aid encoding and retrieval processes; activated representations in working memory must be integrated with pre-existing long-term semantic memories. It is therefore possible that a more general impairment in integration of memories is underlying these difficulties seen in ASC. Objectives: This study aimed to investigate integration in ASC, both within working memory and between items in working memory and long-term memory. Task support was systematically varied to investigate whether any deficit was due to a problem in the self-initiation of integration or the actual integration process itself. Relationships between performance on these integration tasks, performance on other memory tasks and tests of executive function were then explored. Methods: The study compared the performance of a group of 16 children with high-functioning ASC to a group of 16 typically-developing children, matched on chronological age, verbal and non-verbal IQ. Firstly, the study investigated whether the mechanisms underlying basic integration of items within working memory were functioning correctly. Then, the ability to integrate activated items in working memory with existing long-term memory representations (both semantic and episodic) was explored. Integration was tested both with and without task support. Memory measures, including the free recall of related word lists, unrelated word lists and autobiographical memories, were taken to investigate the relationship between performance on memory tasks and integration abilities. Various tests of executive function were also carried out to assess their role in any impairment found. Results: At time of writing, data collection for an ASC-group was not complete. Conclusions: We predict that a deficit in memory integration may underlie some of the memory impairments seen in ASC.