0463 Preliminary Findings from COGMUSE-AUT: Adaptive Cognitive Training for Autistic Adults with Insomnia

Abstract Introduction Autistic adults are an underserved population who experience a high prevalence of insomnia (~80%) and cognitive challenges. Limited access to trained providers for behavioral insomnia treatments, adverse side effects of sleep medications, and the lack of medications to improve cognition has prompted research on alternative behavioral interventions such as cognitive training. Our prior work has shown benefits of multimodal cognitive training to cognition and associated functioning, including sleep, in older adults; but this has not been tested in autistic adults. This study optimized a novel adaptive cognitive training program (COGMUSE-AUT, patent pending) for autistic adults with insomnia and tested its preliminary impact on cognition, pre-sleep arousal and insomnia symptoms. Methods Autistic adults (N=6, Mean Age=30.8, SD=12.2; 3 men/3 women) with insomnia [DSM-5 criteria plus either Insomnia Severity Index (ISI) of 10+ or baseline sleep diary confirmation of 6+nights/14 >30 mins of sleep onset latency or wake after sleep onset] completed 6 weeks of COGMUSE-AUT at home 3x/week for 60mins/day. COGMUSE-AUT (patent pending) is a computerized multimodal cognitive training targeting processing speed, spatial attention, perceptual motor skill, working memory and inhibition. Participants completed assessments of insomnia (Insomnia Severity Index), pre-sleep arousal (Pre-sleep Arousal Scale-cognitive/somatic subscores), and objective cognition (NIH Toolbox-Cognitive Battery) during a 14-day baseline and post-intervention period. Paired t-tests examined outcome change across timepoints. Results COGMUSE-AUT led to statistically significant improvements in delayed verbal memory (Rey Auditory Verbal Learning Test-Delayed; p=.03, hedges g=.45, moderate effect) and trending improvements in attention and processing speed (Oral Symbol Digit Test, p=.059, hedges g=.16, small effect), and inhibition (Flanker Task, p=.085, hedges g=.11, small effect). Statistically significant reductions were also observed in pre-sleep cognitive arousal (p=.02, hedges g=.43, moderate effect), trouble falling asleep (p=.006, hedges g=.94, large effect), and trouble staying asleep (p=.04, hedges g=.62, moderate effect). Conclusion Preliminary findings indicate that among autistic adults with insomnia, COGMUSE-AUT shows near-transfer benefits to non-trained cognition and far-transfer effects on pre-sleep cognitive arousal, sleep initiation and sleep maintenance. Larger randomized controlled trials with extended follow-up and objective assessment of arousal and sleep mechanisms underlying these transfer effects are warranted. Support (if any) American Academy of Sleep Medicine Foundation (325-BS-23, PI: Curtis).