Poster E104, Monday, March 26, 2:30-4:30 pm, Exhibit Hall C
Naturalistic decision-making dynamics in spiking neuron circuits
John C. Ksander1, Donald B. Katz1, Paul Miller1; 1Brandeis University
Animal foraging provides an ethological paradigm for studying fundamental human decision-making. Foraging behavior can be distilled as a series of basic decisions between two choices: consume the immediately available food (i.e. “stay”), or seek an alternate food source (i.e. “leave”). The current study provides a novel account of how the brain may implement such decision-making with bistable spiking-network models. Animal behavior in a taste preference task was simulated with pools of exponential leaky integrate-and-fire neurons. The within-pool and between-pool neuron connections produced network spiking activity that abruptly transitioned between two discrete states. The network’s active state indicated the animal’s behavior, with one state representing “stay” and the other “leave”. Simulations with this model reproduced two key aspects of foraging dynamics. Behaviorally, increasing the relative palatability of one stimulus caused a preference for the high-palatability stimulus. Neurally, the network dynamics demonstrated sudden changes in cortical spiking that have been shown to predict consumption behaviors. Further simulations evaluated whether animals only “stay” given a sufficiently hedonic stimulus, or if “leaving” requires aversive stimuli. Increasing within-pool inhibition caused rapid state transitions without palatable stimuli, reflecting a behavioral disposition for seeking other stimuli. Alternatively, increasing cross-pool inhibition produced a network which rarely transitioned without unpalatable stimuli, reflecting a behavioral aversion to leaving stimuli. These implementations yielded distinct network transition dynamics, providing empirically testable predictions for these accounts. Together, these results show realistic behavior in a taste preference task, produced by spiking-network models that provide plausible ways the brain may implement such naturalistic decision-making.
Topic Area: THINKING: Decision making