FIGURE 2.

Extraction and tracking of whole-brain functional networks at different spatial and temporal scales using the whole-brain model. (A) We simulate neuronal time series at different spatial scales (from 100 to 400 regions). We then create different bin sizes of the time series (using bins from 10 to 3,000 ms), the bin size corresponds to the temporal scale. The binned time series are binarized using a point process paradigm, resulting in an event matrix. (B) We extract whole-brain functional networks using independent component analysis, resulting in a network matrix (see ribbon plot). These networks are tracked over time by projecting the event matrix onto the networks, resulting in an activity matrix (not displayed). (C) The richness of the switching between functional networks is estimated by calculating the entropy of their switching probability. The entropy is compared across spatial and temporal scales.