Abstract
Animals must respond selectively to specific combinations of salient environmental stimuli in order to survive in complex environments. A task with these features, biconditional discrimination, requires responses to select pairs of stimuli that are opposite to responses to those stimuli in another combination. We investigate the characteristics of synaptic plasticity and network connectivity needed to produce stimulus-pair neural responses within randomly connected model networks of spiking neurons trained in biconditional discrimination. Using reward-based plasticity for synapses from the random associative network onto a winner-takes-all decision-making network representing perceptual decision-making, we find that reliably correct decision making requires upstream neurons with strong stimulus-pair selectivity. By chance, selective neurons were present in initial networks; appropriate plasticity mechanisms improved task performance by enhancing the initial diversity of responses. We find long-term potentiation of inhibition to be the most beneficial plasticity rule by suppressing weak responses to produce reliably correct decisions across an extensive range of networks.
Learning to associate relevant stimuli in our environment is important for survival. For example, identification of an object, such as an edible fruit, may require us to recognize a unique combination of features – color, shape, size – each of which is present in other, perhaps inedible, objects. Thus, how the brain associates distinct stimuli to produce specific responses to particular combinations of stimuli is of fundamental importance in neuroscience. We aim to address this question using computational models of initially non-functional, randomly connected networks of spiking neurons, which are modified by correlation-based learning rules identified experimentally. Correlation-based learning rules use the spikes of neurons to change connection strength between neurons. Correlation-based learning rules can enhance stimulus-pair representations that arise naturally in random networks. Altering the strength of inhibitory-to-excitatory connections alone was the most beneficial change, generating high stimulus-pair selectivity and reliably correct decisions across the widest range of networks. Surprisingly, changing connections between excitatory cells alone often impaired stimulus-pair selectivity, leading to unreliable decisions. However, such impairment was ameliorated or reversed by changes in the inhibitory-to-excitatory connections of those networks. Our findings demonstrate that initial heterogeneity and correlation-based changes of inhibitory synaptic strength can help generate stable network responses to stimulus-pairs.