Altered Dynamic Network Stability in Remitted Late Life Depression Associated with Depression Recurrence.

Background: Late-life depression (LLD) is associated with negative outcomes including high rates of recurrence and cognitive decline. However, the neurobiological changes influencing such outcomes in LLD are not well understood. Disequilibrium in large-scale brain networks may contribute to LLD-related cognitive decline.

Methods: Never-depressed older adults and participants in early remission from LLD were recruited as part of the REMBRANDT study. At study entry, participants completed a resting-state fMRI scan and neuropsychological testing and were subsequently monitored over two years for depression recurrence. Using a previously described algorithm, recurring whole-brain states of spatial co-activation were identified by k-means consensus clustering. Co-occurring network state properties from never-depressed participants (n = 40) were then compared to LLD participants who remained in remission (n = 50) or experienced depression recurrence (n = 33).

Results: A three-network solution overlapping anatomically with the Default Mode Network, Cognitive Control Network, and Anterior Salience Network best explained recurring network states. Compared with never-depressed older adults, participants who remitted from LLD exhibited decreased network resilience and altered transitions between networks. Stability of specific networks were associated with baseline clinical and neuropsychological markers in never-depressed and sustained remission participants but were blunted for participants who experienced depression recurrence.

Conclusions: Collectively, these data suggest that LLD alters dynamic network stability lasting into remission. Furthermore, stability of specific networks states is associated with clinical and neuropsychological markers which may predict the likelihood of a recurrent episode of LLD.