Depression following traumatic brain injury is characterised by dominant and recurring brain loops in self-referential areas

Depression is a major albeit neglected complication in patients with traumatic brain injury (TBI). Elucidating its neural correlates remains an important milestone with respect to understanding the disorder and helping with the rehabilitation process. Towards this direction, neuropsychological theories have proposed abnormal brain dynamics as the neural basis of depressive symptomatology. This observational study addressed the question of whether depression in TBI patients is related to abnormal brain dynamics using a sample of 81 TBI patients with depressive symptomatology. To explore brain dynamics we employed the Hidden Markov model that utilises resting-state fMRI data to identify the states that the brain visits sequentially during scanning. Spatial (highest activated regions) and temporal (occupancy, switching rate) characteristics of these states were used to analyse the networks involved and probe differences between depressed and non-depressed TBI patients. We found a significant positive association between depression score and the fractional occupancy and switching rate of two specific states that distinguished between depressed and non-depressed TBI patients. These states spanned default mode, subcortical and cerebellar regions while also forming a temporally coherent "metastate" that the depressed brain would recurrently visit. Depression in TBI patients is characterised by abnormal recruitment and repetitive sequencing between certain neural networks. These results point to the existence of a reinforced, self-referential circuitry that could provide the basis for targeted therapies during the recovery process.