Hypothalamic volumes predict sleep dysfunction in genetic frontotemporal dementia

Abstract Background Sleep dysfunction is common in neurodegenerative disorders, however, its neural correlates, remain poorly characterized in genetic frontotemporal dementia (FTD). Atrophy in two hypothalamic nuclei, the suprachiasmatic nucleus and the lateral hypothalamic area, important for sleep regulation, may be related to this dysfunction. Thus, we examined changes in cerebral and hypothalamic structure across the lifespan in genetic FTD and their relations to measures of sleep dysfunction. Method Data was retrieved from the Genetic Frontotemporal Dementia Initiative (GENFI). T1‐weighted structural MRI images and scores on the Cambridge Behavioural Inventory‐Revised (CBI‐R) sleep subscale were obtained from subjects with mutations causative of FTD (n = 491, scan number = 1029) and healthy controls (n = 321, scan number = 739). MRI images were processed for cortical thickness using CIVET 2.1 and hypothalamic volumes using a deep learning segmentation algorithm (Billot et al., NeuroImage 2020). Using linear mixed‐effects models, we examined changes in sleep dysfunction, vertex‐wise differences in cortical thickness, and volumetric changes in hypothalamic regions in mutation carriers compared to controls. Further, using linear mixed‐effects models, we examined associations between cortical and hypothalamic atrophy and changes in the CBI‐R sleep subscale while controlling for age, sex, scanning site, and disease severity based on the MMSE. Result Mutation carriers showed greater sleep dysfunction across the lifespan, and this increased closer to the predicted onset of symptoms, compared to controls (p < 0.01), with MAPT carriers having greater dysfunction overall (figure 1). All mutation carriers showed patterns of cortical thinning (figure 2) commensurate with the literature (p < 0.05, FDR corrected). Further, cortical thinning in frontal and parietal regions were associated with greater sleep disturbance in C9orf72 and GRN mutation carriers (p < 0.05, FDR corrected) (figure 3). Lastly, MAPT mutation carriers showed consistently significant hypothalamic volume loss across the lifespan (figure 4) (p < 0.01) and reduced hypothalamic volumes were related to increased sleep dysfunction (p < 0.05) (Figure 5). Conclusion These findings suggest that while cortical thinning in C9orf72 and GRN carriers non‐specifically correlate with increased sleep dysfunction, the increased sleep dysfunction observed in MAPT carriers may be attributable to increased hypothalamic atrophy.