Memory Deficits, Communicative Ability, and Self-Monitoring Limit Utility of Existing Forensic Interviewing Methods for Youths with High Functioning Autism Spectrum Disorder

Background: Individuals with Autism Spectrum Disorder (ASD) exhibit deficits in autobiographical memory, and, in legal application eyewitness memory, compared to typically developing individuals (TD), particularly on free recall tasks (e.g., “Tell me what you remember about X”). Such free recall accounts in ASD are reportedly sparse, lack coherence, and include irrelevant or idiosyncratic details. In eyewitness memory, free recall accounts are the preferred statements used by legal investigators. Interview protocols, derived from TD memory abilities, which aim to elicit accurate accounts, may be ineffective as they rely on cognitive and social abilities that may be diminished in ASD (e.g., narrative ability, pragmatic understanding, self-monitoring, and use of organizational frameworks at encoding/retrieval).

Objectives: This study compared memory for details (events, people, actions, and setting/object) in youths with and without high functioning ASD (HFASD) as well as compared the efficacy of two forensic interview protocols for use with ASD samples.

Methods: Forty-eight youths [age: 13.96(2.47)] experienced a semi-distressing event involving four confederates. Half of the participants had a diagnosis of HFASD (n = 24). Two to three weeks later, they returned for a memory interview, half receiving the Cognitive Interview and other half the 10-Step (shortened NICHD derivative version) protocols to aid in retrieving details. Parents rated youths’ language (pragmatic difficulties) and executive functioning (self-monitoring ability). Free recall details were scored as general theme events, or more specific people-, action-, or setting/object-related.

Results: Controlling for age, working memory, and attempt duration, a 2 (Diagnosis: TD vs. HFASD) x 2 (Interview Protocol: CI vs. 10) ANCOVA revealed deficits for the HFASD group on proportion of events remembered, F (1, 41) = 8.37, p < .01, η_ρ² = .17. Planned comparisons revealed these differences for the CI groups only, F (1,41) = 6.44, p = .02, η_ρ²= .14. For detail categories, a 2(Diagnosis) x 2 (Interview Protocol) x 3 (Detail: people, actions, objects/setting) repeated measures ANCOVA, revealed diagnostic group differences, F (1,41) = 21.61, p < .001, η_ρ² = .35 as well as a within Detail effect for both groups, F (1, 41) = 26.41, p < .001, η_ρ² = .39. Both groups demonstrated the lowest and greatest accuracy for details related to actions and people, respectively, F’s (2,40) ≥ 84.70, p’s < .001, η_ρ² = .81-.83. Regression models predicting detail accuracy revealed that pragmatic ability (β ≤ -.41, p <.01) and self-monitoring X Interview Protocol (β ≥ .39, p <.05) predicted a substantial portion of the variance in the proportion scores for actions (R² = .58, F = 11.78, p <.001) and setting/objects actions (R² = .51, F = 8.78, p <.001).

Conclusions: While deficits were observed, HFASD group did exhibit similar memory trends for type of detail reported relative to TD peers, and remembered most events. Lower accuracy was associated with pragmatic difficulty and receiving the CI while having low levels of self-monitoring. Moreover, the HFASD group did not seem to benefit from either protocol’s strategies. Results are discussed with legal implications in terms of modifying cognitive demands of existing protocols.