Dimensional Autistic Traits Predict Susceptibility to False Memory: Sex Differences in Source Monitoring and Gist Construction

Background:  Fuzzy trace theory (FTT) has been used to account for autistic individuals’ often exceptional verbatim memory, and to suggest that the same autistic cognitive style may confer resistance to false memory.

Objectives:  To assess the relationship between dimensional variations in autistic traits and false memory.

Methods:  The Deese-Roediger-McDermott (DRM) word-list false-memory paradigm was used in combination with Autism Spectrum Quotient (AQ) subscales and two performance-based measures of autistic traits – the 'Reading the Mind in the Eyes' Test (RMET) and the Embedded Figures Test (EFT) – to assay effects of individual and sex differences in empathising and systemising traits on authentic and false memory in 89 (45 female) typically developing young adults.

Results:  A Sex x AQ(Social) interaction predicted correct recalls, F(1, 74) = 5.19, p = .026, and the output positions thereof, F(1, 74) = 4.13, p = .046: in females high AQ(Social) predicted fewer (r = -.288) and earlier (r = -.173) correct recalls. Likewise, in females high RMET predicted more correct recalls, F(1, 75) = 5.23, p = .025, r= 0.281. No such effects were present in males. Sex x AQ(Social) x AQ(Detail) influenced number of critical lure false recalls, F(1, 74) = 5.92 p = .017: among males with high social skills (low AQ(Social)), attention to detail (high AQ(Detail)) increased the number of false recalls, but in males with low social skills, attention to detail decreased the number of falsely recalled critical lures. Females exhibited the opposite pattern, high social skills yielding a negative and low social skills a positive relationship between attention to detail and number of falsely recalled critical lures. Again in females, faster disembedding on the EFT (Sex x EFT F(1, 75) = 6.98, p = .01, r = 0.429) and lower RMET score (Sex x RMET F(1, 75) = 8.97, p = .004, r = 0.308) both were associated with fewer critical lure false recalls.

Conclusions:  The restriction of semantic gist (reduction in critical lure recalls) associated with increased AQ(Detail) in those females with low AQ(Social) suggests that in females the main determinant of false recall is systemising. Whereas heightened RMET in males only reduced false memory, implying a source-monitoring effect, females’ RMET increased both correct and false memory, associating heightened empathising skill with heightened gist construction. These relationships suggest that semantic and theory-of-mind aspects of females’ empathising abilities are more closely related than males': Empathising deficits may impair semantic association in females, but source monitoring in males. Implications include the identification of traits predicting individual differences in false-memory susceptibility (a rarity in false-memory research), greater understanding of cognitive mechanisms underlying the uniquely autistic style of memory, and appreciation of the potential clinical importance of false-memory susceptibility in persons with autism spectrum conditions: This interpretation predicts that in males with Broader Autism Phenotype or in general high levels of autistic traits, confusion of endogenous beliefs with exogenous observations – particularly where impaired social perception creates a vacuum of true observations – might in part underlie deficits in the formation and maintenance of social relationships.