Fig. 3: mPFC-NAc and vHip-NAc are differentially sensitive to unrewarded outcomes.

A Venn diagram representing the relationship between the mutual information and conditional entropy that exists between observed mPFC-NAc and vHip-NAc signals. Conditional entropy is a measure of the additional unique information contributed by a second signal given fully knowledge of a first signal. B Conditional entropy in mPFC-NAc is reduced on rewarded relative to unrewarded trials regardless of previous outcome (male n = 11, female n = 10; U→U vs U→R: Z = 2.8644, p = 0.0083; R→U vs R→R: Z = 3.5185, p = 0.0009) indicating that less unique information is carried in mPFC-NAc after reward. C Conditional entropy in vHip-NAc is increased on rewarded relative to unrewarded trials only when the prior outcome was unrewarded (male n = 11, female n = 10; U→U vs U→R: Z = −3.8585, p = 0.0002) indicating that more unique information is carried in vHip-NAc when reward follows nonreward. Comparison of activity at ITI end on currently rewarded or unrewarded trials considering prior outcome history up to three trials back shows that (D) mPFC-NAc activity is suppressed on every currently rewarded trial indicating that mPFC-NAc consistently encodes current outcome via relative suppression regardless of outcome history (male n = 11, female n = 10). In contrast, (E) vHip-NAc activity is suppressed on currently rewarded trials except when current reward is preceded by two (Z = 1.2310, p = 0.8606) or three (Z = 0.8398, p = 0.9834) prior consecutive rewards indicating that vHip-NAc ceases to encode current outcome via relative suppression after consistent reward (male n = 12, female n = 10). See Supplementary Table 1 for all comparisons. Comparisons were performed using a two-sided Z-test and Sidak’s method to adjust for multiple comparisons. Error bars represent SEM around the estimated mean. Source data are provided as a Source Data file. **p < 0.01, ***p < 0.001, ****p < 0.0001.