Dependency of the activity in BLA and VTA on reward magnitude. Panel (A) shows the maximal firing rate in BLA around CS-onset and reward delivery during the first and last trial of conditioning, for different reward magnitudes. For each value of the reward magnitude, the CS1-US1 association is presented 15 times, and the maximal activity in BLA around CS-onset (between 900 and 1100 ms after the start of each trial) and reward delivery (between 3900 and 4100 ms after the start of the trial) is recorded. The experiment is repeated 10 times (without different initial values), and the mean (solid line) and standard deviation (colored area) of these measurements are plotted. The blue dotted line shows the maximal activity at CS-onset during the first trial, which does not depend on reward magnitude, as no learning has taken place yet. The red dotted line shows the maximal activity at reward delivery during the first trial, which is proportional to the reward magnitude because of learning in the LH → BLA projection during the sensitization phase. For the last trial of conditioning, the blue and red solid lines show the dependency on reward magnitude of the maximal activity in BLA at CS onset and reward delivery, respectively. While the US-related response is proportional to the reward, the CS-related activity only appears for reward magnitudes bigger than 0.1. Panel (B) shows the dependency on reward magnitude of the VTA bursts in the same conditions (blue dotted = CS onset at trial 1, red dotted = US delivery at trial 1, blue solid = CS onset at trial 15, red solid = US delivery at trial 15). While there are no CS-related bursts during trial 1, the US-related burst is proportional to reward magnitude. A similar relationship can be observed for the CS-related burst at the end of learning. However, the US-related burst after learning shows a different pattern: small rewards (magnitude smaller than 0.4) elicit burst proportionally to their magnitude, but bigger rewards elicit strongly attenuated bursts, showing that the cancelation of US-related bursts is dependent on reward magnitude.