Figure 1

Measurement of autonomic response to reward cues. (a) In our task, participants were required to fixate an illuminated disc, and heard a recorded voice indicating how much reward could be won by making a speeded eye movement. Three reward levels were used: 0p; 10p; and 50p, randomized across 216 trials. After a 1,200–1,600 ms delay, a saccade had to be made to the second of two other discs that illuminated, one slightly later than the other. (b) Correct saccades were those that went directly to the target, whereas on error trials an initial saccade was made to the distractor. Percentages indicate the range of proportion of correct and error trials over all participants. (c) Reward was numerically displayed at the target, based on speed, and scaled up by the amount on offer on that trial. The value fell off exponentially with increasing response time (measured from distractor onset until gaze arrived at the target), with adaptive time constants that maintained a constant average rate of reward. (d) The pupil diameter was recorded at 1,000 Hz from the time of the cue for 1,800 ms. The pretrial baseline was subtracted, and average traces are shown over all trials for each group of participants. The shaded area indicates the standard error within subjects. No difference between groups was seen for the overall pupillary traces, as visible in the zoomed inset. (e) For the healthy controls, the pupil trace was averaged over trials of each reward level. Note that visually, there was no difference between reward levels. The three traces indicate that the pupil was more dilated after hearing a “50p” incentive, compared with “0p” or “10p” incentives (1p≈2 US cents).