Network Plasticity in Adaptive Filtering and Behavioral Habituation

The ability of organisms to seamlessly ignore familiar, inconsequential stimuli improves their selective attention and response to salient features of the environment. Here, I propose that this fundamental but unexplained phenomenon substantially derives from the ability of any pattern of neural excitation to create an enhanced inhibitory (or "negative") image of itself through target-specific scaling of inhibitory inputs onto active excitatory neurons. Familiar stimuli encounter strong negative images and are therefore less likely to be transmitted to higher brain centers. Integrating historical and recent observations, the negative-image model described here provides a mechanistic framework for understanding habituation, which is connected to ideas on dynamic predictive coding. In addition, it suggests insights for understanding autism spectrum disorders.