Specific cortical and subcortical grey matter regions are associated with insomnia severity

There is an increasing awareness that sleep disturbances are a risk factor for dementia. Prior case-control studies suggested that brain grey matter (GM) changes involving cortical (i.e, prefrontal areas) and subcortical structures (i.e, putamen, thalamus) could be associated with insomnia status. However, it remains unclear whether there is a gradient association between these regions and the severity of insomnia in older adults who could be at risk for dementia. Since depressive symptoms and sleep apnea can both feature insomnia-related factors, can impact brain health and are frequently present in older populations, it is important to include them when studying insomnia. Therefore, our goal was to investigate GM changes associated with insomnia severity in a cohort of healthy older adults, taking into account the potential effect of depression and sleep apnea as well. We hypothesized that insomnia severity is correlated with 1) cortical regions responsible for regulation of sleep and emotion, such as the orbitofrontal cortex and, 2) subcortical regions, such as the putamen. 120 healthy subjects (age 74.8±5.7 years old, 55.7% female) were recruited from the Hillblom Healthy Aging Network at the Memory and Aging Center, UCSF. All participants were determined to be cognitively healthy following a neurological evaluation, neuropsychological assessment and informant interview. Participants had a 3T brain MRI and completed the Insomnia Severity Index (ISI), Geriatric Depression Scale (GDS) and Berlin Sleep Questionnaire (BA) to assess sleep apnea. Cortical thickness (CTh) and subcortical volumes were obtained by the CAT12 toolbox within SPM12. We studied the correlation of CTh and subcortical volumes with ISI using multiple regressions adjusted by age, sex, handedness and MRI scan type. Additional models adjusting by GDS and BA were also performed. ISI and GDS were predominantly mild (4.9±4.2 and 2.5±2.9, respectively) and BA was mostly low risk (80%). Higher ISI correlated with lower CTh of the right orbitofrontal, right superior and caudal middle frontal areas, right temporo-parietal junction and left anterior cingulate cortex (p <0.001, uncorrected FWE). When adjusting by GDS, right ventral orbitofrontal and temporo-parietal junction remained significant, and left insula became significant (p <0.001, uncorrected FWE). Conversely, BA showed no effect. The results were no longer significant following FWE multiple comparisons. Regarding subcortical areas, higher putamen volumes were associated with higher ISI (p