On Nonlinear Analysis of Brain Dynamics for Revealing Cognitive Processes

Neurophysical processes that give rise to cognitive and behavioral phenomenon, such as social evaluation of contexts or the interplay between bodily sensation (whether exteroceptive or interoceptive) and the brain, are composed of systems of large scale interactions and self-interacting parts that are not immediately separable, thus demand assessments that exploit nonlinear and complex system methods. After introducing neuroimaging modalities and nonlinearities that sustain them, we analyze the state of the art of nonlinear and complex systems analysis that is commonly employed to study brain signals, indicating potential for improvements. We then derive a nonparametric, nonlinear, probabilistic state-space method to extract coupling measure from time series with promise for strong specificity and sensitivity, showing its application in both stationary processes or nonstationary processes, e.g. task-invoked responses in neuroimaging data. We apply nonlinear analysis to uncover nonlinear spectral effects of stress processing in hemodynamic signals and their implications in body-brain interactions, while also investigating effects of cortical connectivity given social contextual cues such as nutritional claims on value conditioning behaviors for food items.