Neuromodulation Increases Decision Speed in Dynamic Environments

Neuromodulation is considered a key factor in learning and memory in biological neural networks. Recent computational models of modulated plasticity have shown increased learning capabilities also in artificial neural networks. In this study, a rewardbased dynamic scenario is employed to test networks evolved with modulatory dynamics. The analysis on the networks shows that neuromodulation does not only allow for better learning, but accelerates the computation in decision processes. This appears to derive from topological features in modulatory networks displaying more direct sensorymotor connections, whereas non-modulatory networks require longer pathways for signal processing. This computational advantage in increased decision speed could contribute to unveil the fundamental role of neuromodulation in neural computation.