Gain modulation of probabilistic selection without synaptic relearning

Adaptation of behavior requires the brain to change goals in a changing environment. Synaptic learning has shown its effectiveness in changing the probability of selecting actions based on their outcome. In the extreme case, it is vital not to repeat an action to a given goal that led to harmful punishment. The present model proposes a simple neural mechanism of gain modulation that makes possible immediate changes in the probability of selecting a goal after punishment of variable intensity. The results show how gain modulation determines the type of elementary navigation process within the state space of a network of neuronal populations of excitatory neurons regulated by inhibition. Immediately after punishment, the system can avoid the punished populations by going back or jumping to unpunished populations. This does not require particular credit assignment at the 'choice' population (the branching node) but only modulation of the gain of units active at the time of punishment (at the end of the punished branch). In this way, gain modulation encodes memories of past experiences that change behavior without modification of synaptic efficacies. This neuronal non-synaptic learning mechanism does not require statistical relearning. It helps the system not repeat harmful choices that may lead to further punishments. Thus, such a neuronal learning mechanism can complement synaptic plasticity.