Figure 6. Effects of lesioning/silencing areas on the activity and number of attractors.
Silencing occurs throughout the full trial for each area indicated here. (A) Number of active areas in the example attractor as a function of the number of (randomly selected) silenced areas. (B) The activity of the areas which remain as part of the attractor decreases with the number of silenced areas. (C) The number of active WM areas decreases faster when areas are incrementally and simultaneously silenced in reverse hierarchical order. (D) When considering all accessible attractors for a given network (G = 0.48, Jmax = 0.42), silencing areas at the top of the hierarchy has a higher impact on the number of surviving attractors than silencing bottom or middle areas. (E) Numerical exploration of the percentage of surviving attractors for silencing areas in different lobes. (F) Silencing areas at the center of the ‘bowtie hub’ has a strong impact on WM (adapted from Markov et al., 2013). (G) Numerical impact of silencing areas in the center and sides of the bowtie on the number of surviving attractors. For panels (E) and (F), areas color-coded in blue/red have the least/most impact when silenced, respectively.