Figure 3. Outcome encoding.
(a) The mean response difference between winning and losing (at t=0 ms) in self-experienced trials in the AMY, rmPFC and rACC (n=125, 95 and 138, respectively). Top panels: percentage of neurons with values outside of the 95-percentile of values recorded during the 1 s immediately before outcome (bold lines show percentages which are higher than expected based on a binomial test with α=0.01). Middle panels: the mean (+/− s.e.m.) response difference values (arrowheads and horizontal lines mark the upper end of the 95% c.i. of 10,000 bootstrapped means during the 1 s before outcome). Bottom panels: time points where the mean response difference values were significantly different from those before outcome (P<0.01/3, t-test, nAMY=125, nrmPFC=95, nrACC=138). (b) Pearson correlation coefficients between the individually plotted mean response difference values from a, bold traces highlight a significant correlation between the two variables (P<0.01/3, nAMY=125, nrmPFC=95, nrACC=138). (c) The same analysis as in b in the rACC but with values averaged over the time window highlighted there in orange. Every data point represents the mean response difference between winning and losing in a single rACC neuron, in all three comparisons we found a significant Pearson correlation (p (S-E versus Obs)<10−4<0.01, r(S-E versus Obs)=0.425; p(S-E versus Slot)<10−4<0.01, r(S-E versus Slot)=0.812; and p(Obs versus Slot)<10−4<0.01, r (Obs versus Slot)=0.311, n=138). (d) Waveform of all spikes recorded from an rACC neuron (n=10,828, mean +/− s.e.m in red, middle 95% of values in black, insert: interspike interval frequency plot). (e) For the same neuron as in d raster plots and peristimulus time histograms of wins (top) and losses (bottom) in self-experienced (left), observed (middle), and slot machine trials (right). This example neuron not only responds differentially to winning and losing in all three trial types, but notably does so inversely for observed outcomes compared with self-experienced and slot machine outcomes.