Explicit Probabilistic Incentive Learning Tasks (EPILT)
In this task participants are asked to simultaneously learn discriminations for four pairs of stimuli.
In two of the pairs, the choice of the optimal stimulus is probabilistically associated with the receipt of money, and the choice of the non-optimal stimulus is associated with no reward ("Win or Not Win" or gain approach).
In the other two pairs, the choice of the optimal stimulus result is probabilistically associated with no loss of money, while the choice of the non-optimal stimulus is probabilistically associated with the loss of money ("Not Lose or Lose" or loss avoidance).
Stimuli are color images of landscapes or other types of nature scenes appearing on a white background, one pair at a time.
On "Win or Not Win" trials, if the optimal item is selected, participants see an image of a nickel coupled with the word "Win!"
If the non-optimal item is selected, they see "Not a winner, Try again!"
On "Not Lose or Lose" trials, the optimal response receives the feedback "Keep your money!"
If the non-optimal item is selected, participants see an image of a nickel with a red line through it, coupled with the word "Lose!".
The optimal response is reinforced on 90% of trials in one pair and on 80% of trials in the other pair within each type of trial.
Thus, there are a total of four types of trials:
- Win/Not Win at 90/10 probability distribution;
- Win/Not Win at 80/20 probability distribution;
- Not Lose/Lose at 90/10 probability distribution; and
- Not Lose/Lose at 80/10 probability distribution.
To generate multiple parallel versions, we developed four different sets of stimuli.
The task starts with a 20 trial session (10 trials each "Win or Not Win" and "Lose or Not Lose" at the 90/10 probability distribution) to ensure task comprehension, using different stimuli than the actual task.
The first trial is always "Win or Not Win" and participants are guaranteed to experience a win on the first trial by mapping that stimulus to the optimal stimulus category.
Participants have to achieve 60% accuracy for both types of trials in order to proceed to the real task.
If they do not achieve this accuracy, there are asked to repeat the practice (with the same stimuli and mappings) for up to a total of six practice sets.
If they still do not achieve the target accuracy, the task is terminated.
We have versions with differing lengths of training
(2 block versus 4 blocks) to determine if effects could be achieved in a shorter time than the standard version.
Following training a transfer
test phase is presented.
In these 72 trials, the original 4 training pairs are each presented 4 times, and novel pairings re presented on 58 trials.
For novel pairings, each trained item is presented with every other trained item.
Of most interest are pairings that pit stimuli that have experienced different types of reinforcement histories against each other (referred to as pairings).
Participants are instructed to pick the item in the pair that they think is "best" based on their earlier learning. No feedback is administered during this phase.