Fatigue-induced changes in brain nonlinearity inferred by nonparametric and differential equation models of fMRI

In this paper we investigated the nonlinear sources of neurophysiologic changes in the motor cortex due to peripheral fatigue using functional magnetic resonance imaging (fMRI). The changes in nonlinearity was tracked from pre- to post-fatigue resting state using patterns of singularities in the complex plane. A significant decrease in nonlinearity with an accompanying decrease in determinism was observed post fatigue. The possible contributing factors for such a change include decreased excitability of cortical neurons and/or changes in neurovascular coupling. The dynamical changes of nonlinearity during the task showed a decrease in nonlinearity in the initial phase of the task followed by a partial recovery that continued into the post-fatigue resting state. This observation not only proves the hypothesis that cortical effects of fatigue extend beyond the fatiguing task, but also strengthens the possibility of changes in the nonlinearity of neurovascular coupling as a result of fatigue. Subsequently the source of nonlinearity was characterized using a polynomial differential equation model. We found that the nonlinearity is primarily driven by a square relation and to some extent by a logarithmic relation. Also, the magnitude of the coefficients contributing to these nonlinearities decrease post-fatigue confirming the earlier observation.