Relational Memory Processes in Adults with Autism Spectrum Disorder

Background: Previous literature suggests that individuals with Autism Spectrum Disorder (ASD) show mostly intact performance in item memory (Boucher et al., 2005). However, they show particular difficulties with relational memory (Gaigg et al., 2008; Poirier et al., 2011) which is defined as the processing of relations between items and their contexts in order to enable flexible retrieval (Bowler, Gaigg & Lind, 2011) which are functions thought to be supported by the hippocampus (Opitz, 2010). Other theories suggest a problem with the processing of complex information and executive functions, thus implicating the prefrontal cortex (Minshew & Goldstein, 2001). Few studies compare item and relational memory directly and often the item test does not have the same level of complexity/ difficulty as the relational test. Moreover, most previous work has been done using verbal stimulus material or nameable pictures but few have used abstract shapes.

Objectives: In this study we were seeking to compare item memory with different forms of relational memory. Most importantly we were using the same format for all 4 tasks to make the complexity of the item and relational memory tasks more comparable. Stimuli were abstract shapes presented in black and white. All tasks had a working memory component in that items were presented in several shape triplets followed by the test, but only one task involved working memory load in the testing phase as well, in order to investigate the above mentioned theories more specifically.

Methods: Eighteen adults with ASD and 18 IQ, age and gender matched TD individuals were tested with a paradigm adapted from Konkel et al. (2008). In the study phase, which was the same for all 4 tasks, participants were presented with a series of shape triplets. The shapes of a triplet appeared one after the other in specific screen locations (e.g. top right). During test trials, participants received either a repeated or a manipulated set of previously studied triplets. In the item test participants were asked to indicate if they had seen an item triplet before when they were presented either with the same 3 items as in the study phase (repeated) or with 1 old and 2 new items (manipulated). In the relational tasksno new items were presented keeping either the locations, order or item-combination the same (repeated) or mixing them up (manipulated). Measures taken were percentage of correct answers, hit rate, false alarm rate and hits minus false alarms.

Results: Contrary to our predictions there were significant differences between groups in all measures (except hit rates). Inspection of effect size shows the biggest effect for the location task, followed by the order test, the item test and the item-combination task.

Conclusions: Results suggest item memory impairments when items have to be manipulated. They provide support for both theories and will be discussed in relation to the role working memory processes might play during the encoding and retrieval of item and relational information in ASD.