Permuted Rules Operation Task (PRO)
AKA: Rapid Instructed Task Learning - RITL
Background
The Permuted Rule Operations (PRO) task is a cognitive test of how the brain flexibly combines simple rules into complex, novel tasks. It helps scientists understand how the brain transitions from processing individual rule components to executing a fully integrated, practiced task Michael W. Cole and colleagues (2010) introduced the task in 2010 to understand the foundations of Rapid Instructed Task Learning (RITL), the uniquely human ability to quickly learn new behaviors (e.g. playing board games with others) based only on instructions, completely bypassing trial-and-error learning.
The PRO presents participants with two words from different semantic categories (e.g. 'microfiber'-'alligator') and mixes and matches response rules from three different discrete domains: logical decisions, sensory features, and motor responses. For example, one response rule could be "If the answer 'Is it SWEET (=sensory feature)?' is the SAME (=logic decision) to both items, press the 'LEFT MIDDLE FINGER' key (= motor response); if it is not the same, press the LEFT INDEX finger instead'. Participants practice a fixed set of rule combinations during a training session before they are tested on all possible combinations during the test.
The task was updated to run 'concrete' audio-visual stimuli such as horizontal or vertical blue and red rectangles, each combined with sounds varying in pitch and continuity in 2017 and is known as the Concrete-PRO (C-PRO). By swapping out abstract, linguistic prompts (like reading words) for physical, multi-sensory features, the C-PRO minimizes complex language-processing of the stimuli and provides a more direct mapping of the stimuli to their sensory features. Takuya Ito and colleagues, using the C-PRO in 2017 in an fMRI imaging study, demonstrated how brain networks — especially the frontoparietal network (FPN)— act as a global coordinator for the knowledge transfer of task-rule information from previously learned rules to novel ones.
The Millisecond PRO task is based on the original language-based PRO paradigm.
Task Procedure
The PRO task is separated into a training and a test session. The training session familiarizes participants with the task by repeatedly using a subset of all possible rule combinations. The test session, on the other hand, runs all possible rule combinations.
Each rule combination consists of three components: a sensory one, a logical and a motor one, that can use one of four rules each:
- sensory: sweet, green, soft, loud
- logical: same, different, true for second word, not true for second word
- motor: use left middle finger (D-key), left index finger (F-key), right index finger (J-key), right middle finger (K-key) Combining all possible 4x4x4 rules results into a total of 64 rule combinations.
The training session is run with a subset of 4 of these rule combinations, introducing each of the 12 individual rules to the participants (each rule is used in a different combination). The training runs 6 blocks with 36 trials each. Each trial presents a rule combination screen, followed by two word probes presented one above the other. Participants get as much time as needed to process each rule combination (= self-paced), but are time-restricted to respond to the two probe words. For the first 35 trials, the probe words are presented for 3 seconds and participants get 4 seconds to decide which of the four possible response keys to use, for the remaining trials the presentation duration and response window are cut down to 1.5 seconds and 2 seconds, respectively, which are the timings also used during the test. Feedback is provided for the first 35 trials for 1000ms and then replaced by a fixation cross. Note that the rules mention only the 'affirmative' response keys. If the affirmative response key does not apply, participants are always instructed to use the opposite response key of the same hand. For example, if participants see the rule combination "SOFT, LEFT-INDEX FINGER, SAME", followed by the words "computer-bunny", the affirmative response key "left-index finger" does not apply because at least one of the words represents something that is clearly not soft. In this case the other response key from the same hand, the left-middle finger (resting on the D-key), is the correct response finger.
The stimuli set uses 180 nouns for the four different sensory rules (Color,Sweetness,Softness,Loudness). The stimuli were selected such that each word can be used by only 2 of these sensory rules. For example, the word 'sugar' is used by 'color' (it's white, not green) and by sweetness (it is in fact sweet) but not by Softness nor Loudness. This was done to ensure that each word is not 100% associated with a particular semantic category while at the same time avoiding ambiguous semantic situations, such as whether an apple is green (sometimes it is, sometimes it is not).
The test session starts out with a short practice block to remind participants of the practiced training rule combinations. Afterwards, the test runs 10 blocks of 36 trials each. Each block presents 18 training rule combination and 18 novel ones. Across all 10 blocks, each of the 4 training rule combinations is presented 45 times and each of the 60 novel rule combinations is presented 3 times.
Note that the Millisecond PRO can be run with one of 16 different group numbers. Each of the 16 groups uses a different set of training rule combinations. Thus, it's important that test sessions are always run with the same group number as used during the training session.
What it Measures
The Permuted Rule Operations (PRO) tests how practiced rule components are flexibly combined into novel task sets on the fly.
Psychological Domains
- Rapid Instructed Task Learning (RITL): Ability to quickly process and execute a novel set of instructions
- Working Memory: Ability to keep the rules active during performance [1.303, 1.158].
- Executive Control: Ability to continuously update working memory while suppressing irrelevant information (e.g. the previous sum)
- Cognitive Flexibility & Rule Switching: Inhibiting old rule combinations to seamlessly shift into an entirely new permutation
- Processing Speed: Ability to quickly process information held in working memory
- Sustained Attention: Ability to maintain focus during a long task
Main Performance Metrics
- Proportion Correct Responses: Accuracy measure for training and novel sets, overall and across the different training/test blocks
- Response Time: Mean correct response times; main indicator of processing speed and working memory function
Psychiatric Conditions
The PRO/C-PRO have been used to study the following patient groups
- Schizophrenia
- Psychosis
A measure of learning and retention of goal-oriented rules designed by Cole et al (2010).
References
Search Google Scholar for peer-reviewed, published research using the Inquisit Permuted Rules Operation Task (PRO).
Cole, M. W., Bagic, A., Kass, R., & Schneider, W. (2010). Prefrontal dynamics underlying rapid instructed task learning reverse with practice. Journal of Neuroscience, 30(42), 14245–14254.
Ito, T., Kulkarni, K. R., Schultz, D. H., Mill, R. D., Chen, R. H., Solomyak, L. I., & Cole, M. W. (2017). Cognitive task information is transferred between brain regions via resting-state network topology. Nature Communications, 8(1), Article 1027. https://doi.org/10.1038/s41467-017-01000-w
Cole, M.W., Patrick, L.M., Meiran, N., & Braver, T.S. (2018). A role for proactive control in rapid instructed task learning, Acta Psychologica, 184, 20-30.
Cocuzza, C. V., Ito, T., Schultz, D., Bassett, D. S., & Cole, M. W. (2020). Flexible Coordinator and Switcher Hubs for Adaptive Task Control. The Journal of Neuroscience, 40(36), 6949–6968. https://doi.org/10.1523/JNEUROSCI.2559-19.2020