Extended Data Fig. 6: Anatomical gradient in a different subset of probability selective neurons.

Here we test for the anatomical gradient across those neurons that were probability selective at choice, but did not meet the criteria for RPEs. This analysis is to test the robustness of the gradient result by assessing whether it replicates in an independent set of neurons. The neuron optimism of these neurons is measured using the reversal point. We present this data here to supplement the gradient analyses in Fig. 1, but note that it is less clear what the predictions of distributional RL are for these non-RPE neurons, and so it is unclear exactly what the reversal point means in these neurons. Nevertheless we present this result to demonstrate the gradient of the reversal point measure replicates in an independent set of neurons (R = 0.28, P = 0.013, by Pearson correlation).