Circuit principles of memory-based behavioral choice (360G-Wellcome-205050_Z_16_Z)

We aim to elucidate the circuit mechanisms underlying three key computations essential for memory-based behavioral choice: 1) updating valences attached to sensory cues, when actual and expected outcomes differ; 2) computing the “value” for each action, based on multiple, conflicting cues; and 3) selecting one action and suppressing other physically incompatible competing actions. One obstacle to progress in this field has been the problem of identifying underlying circuits with synaptic resolution, and causally relating structural motifs to their proposed function. Both insects and vertebrates have evolved cerebellar-like higher-order parallel-fiber systems specialized in forming large numbers of associative memories and in guiding memory-behavioral choice. However, no synapse-resolution wiring diagram of any such system has been available to guide analysis and inspire understanding.    We have recently mapped the synaptic-resolution wiring diagram of one such system, the insect mushroom body, in Drosophila larva, which reveals multiple novel circuit motifs and provides clues about learning and decision-making models and their neuronal implementation. An exquisite genetic toolkit available in this model system allows selective manipulation of individual neuron types to establish causal relationships between their activity and behavior. We are now in a unique position to causally relate the identified structural motifs to their function.