Probabilistic Reward Task

Background

The Probabilistic Reward Task (PRT) is a behavioral tool to measure people's responsiveness to rewards ('hedonic capacity') using differential reward contingencies to classifications of ambiguous stimuli. The task was developed by Diego Pizzagalli and colleagues in 2005 to quantify anhedonia, the inability of feeling pleasure, a key deficit in people suffering from Major Depressive Disorder.

The PRT consists of a simple categorization task: decide whether a 'mouth' (aka a horizontal black bar) is short or long. The difficulty of the task lies in the ambiguity and short presentation duration of the two stimuli: 'short' and 'long' mouths are close in size, with the short mouth being about 88% of the length of the long mouth. Additionally, the presentation duration of each mouth is very short. Reward learning is manipulated using a different reward schedule for correct responses to one type of stimuli (e.g. the short mouths) compared to the other. For example, correct responses to short mouths are rewarded with a 60% probability ('frequent'); whereas correct responses to long mouths get rewarded with only a 20% probability ('infrequent').

Pizzagalli and colleagues could show that healthy participants typically develop a quantifiable response bias to err on the side of the more frequently rewarded category. This, however, was not the case for patients suffering from depression.

Task Procedure

Each trial starts out with a fixation cross in the center of the screen (500ms), followed by a mouthless 'face' (also 500ms). The mouth is then added for a brief duration and is erased after 100ms. At this point, participants are asked to press the designated 'short mouth' and 'long mouth' response keys (e.g. press 'E' for short mouth; press I for long mouth). The next trial starts once a response is registered.

After a short demonstration of the two mouths and two practice trials, participants work on three test block, each consisting of 100 trials (50 long and 50 short mouths). Correct responses to the frequently rewarded stimulus (e.g. short mouth) get rewarded with 5cents in 60% of the trials. Correct responses to the infrequently rewarded stimulus (e.g. long mouth) get rewarded in only 20% of the trials. Reward trials are randomly determined (Note: if a randomly selected reward trial is answered incorrectly, the reward designation carries over to the next trial). The designated response keys as well as the designated frequently and infrequently rewarded stimuli can be counterbalanced across participants.

What it Measures

The Probabilistic Reward Task (PRT) measures people's responsiveness to rewards

Psychological domains

• Reward Sensitivity: Degree to which rewards influence subsequent choices
• Decision Making under Uncertainty: Choice behavior when faced with ambiguous information
• Anhedonia: Inability to feel pleasure from being rewarded

Main Performance Metrics

• Measure of Discriminability: Non-parametric measure of people's ability to distinguish between short and long mouths
• Measure of Response Bias: Non-parametric measure of a response bias; primary PRT measure

Psychiatric Conditions

The following patient groups show a blunted response bias on the PRT

• Major Depressive Disorder (MDD)
• Bipolar Disorder
• Schizophrenia
• Post-Traumatic Stress Disorder (PTSD)

References

Search Google Scholar for peer-reviewed, published research using the Inquisit Probabilistic Reward Task.

---

Pizzagalli, D.A., Jahn, A.L, & O’Shea, J.P. (2005). Toward an Objective Characterization of an Anhedonic Phenotype: A Signal-Detection Approach. Biol Psychiatry, 57(4), 319–327.