Developmental biobehavioral markers of working memory impairment and psychopathology risk

• In Event: Early signals of risk for psychopathology: Developmental approaches informed by the Research Domain Criteria

Abstract

Introduction: Working memory is a critical top-down control process that develops across childhood and adolescence. It is implicated as a transdiagnostic mechanism of psychopathology (Abramovitch, Short, & Schweiger, 2021; Huang-Pollock, Shapiro, Galloway-Long, & Weigard, 2017) and highlighted within the RDoC Matrix. Within attention-deficit/hyperactivity disorder (ADHD), a majority of children experience working memory impairment. Yet vast within-group heterogeneity exists, and some children’s working memory appears to normalize with age (Fair, Bathula, Nikolas, & Nigg, 2012; Karalunas et al., 2017). Initial evidence suggests that distinct patterns of working memory development may be related to persistence/remission of ADHD symptoms and onset of commonly co-occurring problems, such as depression (Karalunas et al., 2021). It remains unclear, however, how normalization of working memory performance is related to changes in neural networks that support top-down cognition or how well trajectories of working memory development predict clinical outcomes.

Study population: 437 children (nADHD= 297) enrolled in a longitudinal study between the ages of 7-11-years-old and were followed annually for up to 12 years. 84% of the sample identified as white/non-Hispanic. Median income range was $75-$100,000. 37% of children were prescribed stimulant medication at baseline.

Methods: At all years, parents and teachers completed comprehensive clinical assessment. Children completed a cognitive battery, including multiple tests of working memory. Children were classified into groups based on working memory development in childhood (Years 1-3 of the study) using latent class growth models (LCGA). A subset of 265 children (nADHD =162) were invited to complete an add-on electroencephalogram (EEG) and cognitive testing visit at one time point between Years 5-8 of the larger study. At the add-on visit, children completed a whole-report version change detection task (Adam, Mance, Fukuda, & Vogel, 2014) to measure working memory and attention while 32-channel EEG was recorded. Whole-report task accuracy and rate of attention lapses were compared between LCGA groups using standard ANOVAs. Time-frequency analyses of EEG focused on differences in posterior alpha and frontal midline theta activity prior to stimulus presentation and during memory retention.

Results: LCGA using data from Years 1-3 of the study identified three patterns of working memory development in the ADHD group: “Unimpaired,” “Impaired, Recovering,” and “Persistently Impaired.” During follow-up (Years 5-8), groups continued to perform as would be expected based on these trajectories (Figure 1). The “Unimpaired” and “Impaired, Recovering” groups were indistinguishable from their non-ADHD peers on measures of working memory accuracy and time-on-task effects. In contrast, EEG revealed persistent differences in brain functioning in the “Impaired, Recovering” group. Whereas the non-ADHD and “Unimpaired” ADHD groups showed expected changes in alpha power prior to stimulus presentation, these changes were not observed in either the “Impaired, Recovering” or “Persistently Impaired” groups (Figure 2). Findings demonstrate a lack of normalization in neural processes supporting working memory and attention, despite normalized performance. Normalization of working memory predicted greater ADHD symptom recovery but not decreased risk for developing depression or suicidal ideation. Results highlight the importance of multi-modal measurement and individual differences in developmental trajectories for linking RDoC constructs to clinical phenomenon.

Authors

• Sarah Karalunas, Purdue University, West Lafayette
  Presenting Author

• Jason Dude

• Hanna C Gustafsson, Oregon Health & Science University (OHSU)

• Hannah Morton

• Joel T Nigg, Oregon Health & Science University (OHSU)