Low-dimensional signatures of neuronal activity associated with long-term operant conditioning in Aplysia

Operant conditioning is a form of learning in which a behavior is reinforced by reward. Operant conditioning has multiple temporal domains, ranging from short-term, lasting a few minutes, to long-term, persisting for at least 24 h. The extent to which short- and long-term operant conditioning memories rely on shared or separate neural mechanisms is poorly understood. Voltage-sensitive dye (VSD) imaging has been used previously to record the activity of a large number of neurons simultaneously in the buccal ganglion to measure changes in neuronal activity during short-term operant conditioning. We examined neuronal activity using VSD 24 h after operant conditioning and compared these results with those from short-term operant conditioning to assess the extent to which short-term and long-term operant conditioning share common neural correlates. Non-negative matrix factorization (NMF) isolated the temporal signature of neuronal activity. Similar to short-term operant conditioning, long-term operant conditioning resulted in an earlier recruitment of an NMF module that corresponded to the retraction phase of feeding behavior, which indicated that the temporal signatures of short- and long-term operant conditioning share similar features. In contrast to short-term operant conditioning, long-term operant conditioning engaged a larger population of retraction neurons in a region of the buccal ganglion containing sensory neurons. These findings suggest that a more extensive network is involved in long-term operant conditioning memory.