Neural Correlates of Relational Memory In Autism

Background:

While many studies have observed relatively typical recognition memory for objects in individuals with ASD, less is known about memory for object-context relations. Dual-process models of recognition memory (e.g. Aggleton & Brown, 1999) propose two mechanisms by which object-context pairs can be remembered: (i) a ‘feeling of familiarity’ that acts on blended object-context representations; (ii) hippocampus-dependent recollection of the relations between objects and the contexts in which they appear, thought to facilitate more ‘flexible’ memory retrieval (Eichenbaum, 1997). Several authors have proposed that individuals with high-functioning ASD have impairments in recollection/relational memory, but relatively intact familiarity-based recognition (e.g. Joseph et al., 2005; Gaigg et al., 2008; Boucher et al., 2008). Although neurotypical adults predominantly use recollective memory to remember object-context associations, individuals with ASD may show atypical reliance on less flexible familiarity-based mechanisms. In the present study we used event-related brain potentials (ERPs) to examine this proposal, because familiarity and recollection have been associated with different components of the ERP response to a recognized stimulus (e.g. Curran et al., 2006).

Objectives:

To assess whether recognition memory for object-background pairs relies on atypical neural mechanisms in adults with high-functioning ASD.

Methods:  

Participants were 15 high-functioning adults who met gold-standard diagnostic criteria for ASD, and 15 age- and IQ-matched neurotypical controls (NT group). Participants were asked to remember a series of object-background pairs.  At test, participants saw familiar backgrounds paired with either i) the same object as during learning (Same); ii) an object previously seen on a different background (Rearranged) or iii) a novel object (Novel).  Participants were asked whether the object-background pairing was one they had seen before. During the task, brain activity was continuously measured with high-density electroencephalography.  Initial analyses explored i) accuracy and reaction time of behavioral responses and ii) ERPs during correct responses to test stimuli.

Results:

Relative to the NT group, the ASD group found it significantly more difficult to correctly identify Same and Rearranged but not Novel stimuli, suggesting difficulty remembering object-background relations.  Preliminary analyses of brain activity focused on two ERP components previously associated with memory: an early frontal component associated with familiarity-based recognition, and a later parietal component associated with recollection-based recognition.  As expected, in the NT group neural responses in the Same and Rearranged conditions significantly differed during the recollection-associated parietal component; there was no significant difference between Same and Rearranged conditions over this component for the ASD group.  In contrast, the familiarity-associated frontal component significantly differed between Same and Rearranged conditions for the ASD group but not the NT group. These findings support the hypothesis that recognition of object-background pairs relies on atypical neural mechanisms in adults with high-functioning ASD.

Conclusions: Preliminary findings are consistent with the proposal that participants with ASD find it more difficult to remember object-background relations than age- and IQ-matched neurotypical controls, and suggest that participants with ASD may rely on less efficient familiarity-based retrieval mechanisms to compensate for deficits in recollective recognition.  Further work will explore how these differences relate to reduced cognitive flexibility in autism.