Figure 6.
Population activity. Average population firing for the entire population of SNc (top) and LC neurons (bottom). All recorded neurons were included, whether they showed task related activity or not. We plotted the average activity around two events of a trial: cue onset (left) and correct bar release (right). Around cue onset (left), we separated 1st from non-first trials in a schedule. The horizontal bars indicate the standard epoch used for the analysis, “pre-cue” (P, from 400 ms before the cue to cue onset); “cue” (C, from 0 to 500 ms after cue onset); “wait” (W, from 0 to 500 ms after the wait signal); “go” (G, from 250 to 0 ms before the bar release); “blue” (B, from 0 to 250 after the blue point); and “outcome” (O, from 10 to 260 ms after trial outcome). The first three epochs are time-locked to each other, then there is a variable period at the end of the wait, followed by the go, blue, and outcome that are also almost completely time-locked (there is a 250 ms of jitter in outcome timing). Around bar release (right), we separated rewarded and unrewarded trials. Each trace was generated by averaging firing rates across all recorded neurons in the corresponding subset of trials. The broken lines represent the error (SEM) envelop around the mean. The baseline rates were so similar in general that no normalization of firing rates across neurons was used. On average, the firing of both SNc neurons was greater in first trials around cue onset. Around bar release, the average firing was greater in rewarded trials for SNc neurons but greater in unrewarded trials for the population of LC neurons.