In this abstract, we proposed an optimum window-size in a sliding-window approach for dynamic functional connectivity analysis. The proposed window-size was derived from the instantaneous period and energy of each intrinsic mode functions (IMF) obtained from empirical mode decomposition. IMFs track local periodic changes of non-stationary time series and therefore can capture subtle temporal variations. Using dynamic functional connectivity matrix computed with the proposed window-size as features, a higher accuracy was obtained in classifying cognitively impaired fighters from cognitively normal ones; and a larger behavioral variance was found in HCP data.
This abstract and the presentation materials are available to members only; a login is required.