Figure 7. Contextual processing of information.

A) Sketch of numerical experiments. The wave train injected at the left input of the neuronal chain conveys raw information (stimulus to be processed, the same for all experiments). The informative train interacts with a periodic train injected at the right input, which sets the context for the wave-processing. The context features are the spatial period and length of the contextual train. After trains’ interaction the processed information can be readout from the right end of the chain. B) Four examples of different contexts (trains with 3 and 9 waves spaced by 30 and 100 neurons) and the corresponding outputs (processed trains). Red trains are different, whereas black train are the same, which suggests divergence/convergence of the contextualization. C) Influence of the context on the information extracted from the raw message. Main graph: the length of the output train linearly depends on the length of the contextual train and its period modulates the sensitivity. Inset: relative entropy of the output has higher variability for contextual train with shorter periods (horizontal displacements of symbols serve for visualization only).