FIGURE 3.

The final detected microstates with different topographical clustering methods. (A) The identified microstates by Case 1, Case 2, Case 3, and Case 4. All are bottom-up-based data-driven approaches. (B) The identified microstates by Milz et al. (2016) and Case 5. Milz et al.’s (2016) results could be considered as standards in the existing EEG microstate studies, in which the underlying neural mechanisms have been verified with functional magnetic resonance imaging studies. Case 5’s results could be more capable of reflecting the specific brain states under tasks, as the task information is used as a guide in the detection procedure (top-down-based task information-guided approach). A good microstate detection performance should share a similar topography distribution property with the standard templates and the task information-guided templates.