Supplement 1:Detailed description of behavioral tasks.
Probability Discounting Task: Each trial (300 trials/session) started with a blue rectangle at the center of the screen. Once the subject correctly touched the blue stimulus, 2 different stimuli were displayed: a red circle and a green plus sign. Touching of the red circle led to a guaranteed, low reward (0.016 ml/kg, 100% of responses), whereas responding on the green plus sign led to a 5 times higher reward (0.080 ml/kg). Subjects were trained to criterion (choice of high reward 80% for 3 consecutive sessions), after which the probability of receiving the high reward was reduced from 100% to 10% in 5 consecutive blocks of 60 trials (100%, 75%, 50%, 25%, and 10%) during each session. Each subject completed 10 sessions. Stimuli disappeared from the screen once the animal had responded to it or after 5 sec without response. Responses outside the stimuli were defined as errors and resulted in the start of the next trial without reward. The inter-trial interval (ITI) was 2 sec. The reduction of the choice of the stimulus associated with the high reward at each probability level indicated the discounting of probabilistic rewards. In addition, the RT immediately after rewarded and non-rewarded trials were calculated as an indication of increased error monitoring1, 2.
Delay Discounting Task: Each trial (200 trials/session) started with a blue rectangle at the center of the screen. Once the subject correctly touched the blue rectangleat the center of the screen, 2 different stimuli were displayed: A black sun stimulus yielding a fixed reward (0.070 ml/kg) that was only available after a delay, and a red arrow which yielded an immediate, but variable reward. The delay for the fixed reward was increased from 0-10 sec on different sessions. Each subject completed 4-5 sessions at 0, 1, 2, 5, and 10 sec delays. The immediate reward magnitude was started at (0.016ml/kg: 20% of delayed reward) at the beginning of each session and adjusted by 15% based on the response on the previous trial (after selecting the immediate variable reward, the reward magnitude was decreased 15%, whereas after fixed reward selection, the variable reward was increased 15%)3. By means of forced choice trials, which occurred after 3 consecutive selections of the same stimulus (changing the variable reward magnitude by -39% or +52%), subjects were forced to sample the other stimulus. The response on forced trials did not change the value of the variable reward. The rate of progress through a session was made independent of choice behavior by adjusting the ITI which was minimally 3 sec. The plateau value of the average amount of the variable reward over the last 150 trials of each session was calculated to reflect the discounting of delayed rewards.
Reversal Learning Task: Each trial (200 trials/session) started with a blue rectangle at the center of the screen and once the animal correctly touched the blue stimulus, 2 different abstract images were presented (from a collection of 400 stimuli). One image was associated with a large reward (0.070ml/kg), whereas the other stimulus was associated with a small reward (0.014ml/kg). The criterion for successful discrimination (DS) was defined as the selection of 9 correct stimuli (associated with the high reward) out of blocks of 10 stimuli. Once criterion was met, reward contingencies were reversed. Upon successful achievement of a reversal (9 out of 10 stimuli correct), a new pair of stimuli were presented and the animal proceeded with the next DS. The number of reversals completed during each of 10 sessions and the number of errors made following reward contingency reversal were analyzed as an indication of performance.
Delayed Match to Sample (DMS): Each trial (200 trials/session) started with a sample stimulus of an abstract image (from a collection of 400 stimuli), which the subject was required to hold for 1 sec. Subsequently, after a (randomized) delay of 0, 10, 20, or 40 sec, two stimuli, one of which matched the sample stimulus, were presented side by side. The subject received a reward (0.075 ml/kg) for correctly touching the stimulus matching the sample stimulus within 10 sec. The inter-trial interval was 2 sec and incorrect responses or lack of response resulted in a(n additional) 5 sec timeout. Accuracy was defined as the ratio of correctly matched stimuli to the total number of responses.
Continuous Performance Task: The continuous performance task consisted of 600 trials. Subjects had to respond within 1 sec to 2 out of 4 different visual stimuli presented sequentially (2 sec ITI) in random order to receive a water reward. These stimuli were associated with a large (0.15ml/kg) or a small (0.03ml/kg) reward which occurred in 10 and 30% of trials, respectively. The remaining two stimuli, which each occurred on 30% of trials, were not associated with any reward and touching them led to a 3 sec timeout in addition to the regular ITI.
References:
1.Rabbitt, P.M. Errors and error correction in choice-response tasks. Journal of experimental psychology 71, 264-272 (1966).
2.Ridderinkhof, K.R., Ullsperger, M., Crone, E.A. & Nieuwenhuis, S. The role of the medial frontal cortex in cognitive control. Science 306, 443-447 (2004).
3.Richards, J.B., Mitchell, S.H., de Wit, H. & Seiden, L.S. Determination of discount functions in rats with an adjusting-amount procedure. Journal of the Experimental Analysis of Behavior 67, 353-366 (1997).
Figure S1: The expected value of reward as a function of reward probability.
Over the course of the probability discounting task, the probability of receiving reward for selecting the stimulus associated with the large reward was reduced from 100% to 10% (60 trials at each probability level per session). As a result the expected value for the stimulus associated with the large reward decreased during the session whereas the expected value for the guaranteed, small reward remained constant. Based on the 5-fold difference in reward size between the stimuli, it became advantageous only at the 10% reward probability to select the small, guaranteed reward.