Skip to main content
. 2016 Jan 27;36(4):1173–1184. doi: 10.1523/JNEUROSCI.2985-15.2016

Figure 8.

Figure 8.

Articulation-aligned clusters. Articulation-aligned information flows (IFs) were analyzed with k-means clustering to identify network-level transitions. Each cluster is shown in a different shade of gray. Top, Arrows indicate time integrals of IFs for the two cluster timeframes shown. Color and width of arrows are proportional to the magnitude of the normalized integrated information. States A–D are shown in Figure 7 and described in the analysis time locked to stimulus onset; State D and E are similar, though derived from distinct concatenations of the time series. State E is from 600–450 ms prior to articulation, and State F from 400 ms to onset of articulation. Bottom, BGA time traces are overlaid on top of shaded clusters (periods of overlap are not depicted). The clusters were derived from all IFs values, not just those that achieved significance.