Fig. 2.
Results of the self-paced experiment. A) Matrix representation of the network; the ordering of the 78 AAL regions is overlaid. Note that the values in the matrix are the ICA derived mixing coefficients. B) 3D representation of the same network, thresholded for visualisation. Lines show connections, with thicker lines indicating stronger connections. Circles represent the summed magnitude of connectivity between that region and the rest of the brain. C) Time evolution of the network during the self-paced finger movement task, averaged across trials in all subjects (black line). Time represents the position of the centre of the 6 s window, relative to the button press at t = 0 s. The grey shaded region represents the null distribution based on a hypothesis that the response is not time locked to the button press. D) Sign-flip analysis, again showing the mean response across all 15 subjects (black line). The grey shaded area represents the null distribution based on a null hypothesis that the modulation is driven by a small number of subjects. Significance (pcorrected<0.05) is attributed if the black line appears outside the null distribution in both C and D. Note that the network clearly represents the primary somatosensory, motor and supplementary motor regions and demonstrates significant modulation with the task. (An interactive version of this Figure can be found at http://nottingham.ac.uk/~ppzgo/ica_nets.).