Amnesia in Dissociative Identity Disorder: A Meta-analytic Assessment

Abstract

Individuals with dissociative identity disorder (DID) often report an inability to retrieve information associated with other identities, termed inter-identity amnesia (IIA). Research investigating IIA has begun to amass and interest surrounds whether objective deficits in retrieval mechanisms necessarily underlie the experience of IIA, or whether aberrant metacognitive processes may instead be implicated. The latter conclusion stems from experimental findings that certain information reported as irretrievable by an individual with DID may actually be accessible through indirect measures of memory. This paper conducted a systematic literature review to qualitatively ascertain findings related to this body of work. Additionally, meta-analyses were conducted to quantitatively assess the qualitative findings of the empirical literature. The question underpinning the review was: from the cognitive empirical literature, are DID patients’ clinical reports of retrieval failure across identities substantiated by controlled measures of memory? More broadly, the review aimed to investigate whether the cognitive empirical research on IIA in DID is robust and comprehensive enough to provide clarity on the nature of IIA.

Nineteen empirical and four case studies informed the systematic review. Here, IIA could not be wholly substantiated by controlled memory tasks; instead, qualitative conclusions suggested a degree of transfer of information across identities reporting amnesia. The meta-analyses comprised 12 of the included studies. Data were compared 1) between identities that reported amnesia for each other within individual DID patients and, 2) between patients in their amnesic identity and healthy controls who were instructed to simulate amnesia. Comparable data included the participants’ proportion of correct results in tasks of memory accessibility using recall, recognition, and cognitive priming tasks, as well as the reaction time data from such tasks. Thus, four separate analyses were conducted. Here, two of the four analyses, comparing both the performance and reaction times of DID participants’ amnesic identities to that of simulators instructed to feign amnesia (i.e., between-subjects design), provided statistical support for the accessibility of at least some memory representations across dissociative identities, despite self-reports of amnesia. The remaining two analyses however, which compared memory performance and reaction times within DID individuals across dissociative identities (i.e., within-subjects design) failed to provide support for such a notion. Instead, support was found for an interpretation of amnesia across identity states, akin to that reported clinically by the patients (i.e., IIA). Importantly, closer examination drew attention to methodological considerations that may limit definitive conclusions about the nature of IIA drawn from the present studies. These issues include substantial heterogeneity between subjects’ scores which is masked by group statistics, a small and homogenous cumulative sample utilised across studies, limited research teams, and minimal domains of memory assessed. The paper urges a nuanced understanding of the phenomenon of IIA in light of current findings.