PEBL Hub — IRB Fact Sheet — 2026-08-23
Probabilistic Reversal Learning
Executive Function
Learning
Category:Executive Function
Duration:~15 minutes
Parameters:6 configurable parameters available
Version:1.0
Description
A reversal learning task where participants learn stimulus-response rules that change unpredictably, with probabilistic feedback.
Two abstract visual stimuli are presented on each trial. One is designated as 'correct' according to a hidden rule. Participants choose one using the left or right shift keys and receive smiley (correct) or frowny (incorrect) face feedback. Feedback is probabilistic: occasionally correct responses receive negative feedback. After a criterion number of correct responses, the rule reverses. Measures cognitive flexibility and sensitivity to feedback.
Measures: Cognitive flexibility, reversal learning, feedback sensitivity
Key Parameters
| numblocks | Each block is around 150 trials. (default: 4) |
| numstages | Number of learning 'stages' per block; after which a reversal happens (default: 10) |
| stagemin | Smallest stage size (default: 10) |
| stagemax | Largest stage size (default: 15) |
| errorMin | smallest number of invalid per block (default: 0) |
| errorMax | largest number of invalid per block (default: 5) |
Scoring & Interpretation
The primary measure is the switch cost: how many extra trials are needed in the alternating block compared to pure blocks. Mean errors per reversal indicates how quickly participants adapt to rule changes. Lower values indicate better cognitive flexibility. By-block performance can reveal learning trends across the session.
Data Output Columns
| sub | Participant ID |
| block | Block number |
| stage | Reversal stage within the block |
| totaltrial | Cumulative trial number |
| trial | Trial number within stage |
| corrstim | Currently correct stimulus (A/B/C/D) |
| order | Correct stimulus side (1=left, 2=right) |
| resp | Participant response (left/right) |
| valid | Feedback validity (1=valid, 0=probabilistic error) |
| corrfeedback | Feedback shown (1=positive, 0=negative) |
| acc | Cumulative correct in stage |
| corr | Response correctness (1/0) |
| rt | Response time in milliseconds |
References
Cools, R., Clark, L., Owen, A. M., & Robbins, T. W. (2002). Defining the neural mechanisms of probabilistic reversal learning using event-related functional magnetic resonance imaging. Journal of Neuroscience, 22(11), 4563-4567.
Mueller, S. T. (2011). PEBL's Probabilistic Reversal Learning Task. Computer software retrieved from http://pebl.sf.net/battery.html