Figure 6.

Models to account for a decrease in the context specificity with progression of training. The model is based on one of the two models we proposed to account for a decrease in the sensitivity of the CR to US devaluation with extension of training (Mizunami et al. 2019). The other model is not described here for simplicity. The basic assumptions of the model are that (1) it represents neural circuits in the lobe of the mushroom body (MB), consisting of “CS” neurons that encode CS signals (Kenyon cells), “CR” neurons for which activation leads to the CR (output neurons of the MB lobe) and “OA2” neurons that code US value signals (octopamine neurons projecting to the MB); (2) two kinds of memory traces are formed by standard training, one being the strengthening of synaptic connection from “CS” to “OA2” neurons, which represents the CS-US (S-S) connection, and the other being that of synaptic connection from “CS” neurons to “CR” neurons, which represents the CS-CR (S-R) connection; (3) simultaneous activation of “CS” neurons and “OA2” neurons is needed for activating “CR” neurons (AND gate) and hence for producing the CR after standard training; and (4) activation of “OA2” neurons is inhibited when the animal has been satiated with water US. With extended training the CR becomes insensitive to US devaluation due to strengthening of the CS-CR connection. This implies that the activation of “OA2” neurons is no longer required for the CR to be executed. In A, it is assumed that the context signals control or gate (AND gate) activation of “OA2” neurons mediated by activation of “CS” neurons. In B, it is assumed that the context signals control or gate activation of “CS” neurons by presentation of the CS after standard training, but the control is lost after extended training. Notice that the models are hypothetical and await future validation in physiological studies.