Neural correlates of fatigue after traumatic brain injury.

Fatigue is the main cause of disability after traumatic brain injury and has negative impact on social, physical and cognitive functions, participation in daily activities, and ability to work. Since the neural underpinnings are largely unknown, few causal treatments are currently available. This study therefore aimed to investigate the neural correlates of subjective fatigue after traumatic brain injury, controlling for differences in cognitive performance, motor performance and subjective psychological covariates such as depression, anxiety and apathy. Seventeen chronic traumatic brain injury patients (10 with and seven without fatigue) and 11 age, sex, and education-matched healthy controls participated in the study. The dependent variable, overall fatigue, was quantified as the sum of the subscales of the multivariate fatigue inventory. Subjective psychological covariates were extracted from appropriate questionnaires. Brain activation during a two-back task and functional connectivity at rest were reconstructed from high-density EEG. Cortical excitability was quantified from motor evoked potentials induced by transcranial magnetic stimulation over the primary motor cortex. Cognitive performance was assessed with a two-back task as well as with a comprehensive neuropsychological test battery. Motor performance was quantified with Jamar dynamometer. Beside the between-group differences in most fatigue subscales resulting from the group attribution, participants also differed in subjective memory functions, depression, anxiety and apathy. Conversely, objective neuropsychological performance was similar across groups in most domains, except for alertness and divided attention (P ≤ 0.039). At the neural level, we observed no difference in corticospinal excitability, but a significant disruption of global resting-state alpha-band functional connectivity between cortical midline structures and the rest of the brain in patients with fatigue (P = 0.006). Furthermore, individuals with fatigue exhibited reduced signs of overall brain activation compared with healthy controls throughout the cognitive task (P = 0.032) without time-on-task effect. In a multivariate regression model, resting-state functional connectivity (P = 0.013) and subjective psychological questionnaire scores (P < 0.0001) were independent predictors of fatigue. In conclusion, our results suggest that disrupted network interactions are the primary independent neural predictor of fatigue. This may serve as a new target for therapy.