Attention
Prototype Distortion Task
Also called: dot-pattern prototype task, dot-distortion category task, prototype learning task, dot pattern classification task
Measure how people learn a category's central pattern: they learn three groups of dot patterns from feedback, then sort the groups' unseen prototypes and new distortions.
6-9 minUntimed
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Try it here first, with no account.
- Papers, 2021 to 2025
- 11
- Evidence grade
- D
- Registered clinical studies
- 1
What it measures
A measure of learning a category's central pattern from its distortions. Participants sort patterns of nine dots into three groups with feedback; each group's patterns are distortions of its own prototype, which is never shown. A test without feedback then shows the prototypes, the training patterns and new distortions. The score is the prototype advantage: accuracy on the unseen prototypes minus accuracy on new patterns as distorted as the training ones.
Constructs
- Category learning
- Prototype abstraction
- Learning from feedback
- Generalization
Research uses
- Research on implicit category learning and its brain systems
- Amnesia, Alzheimer's disease and autism research
- Within-study comparison of prototype and exemplar learning
How the paradigm works
The Prototype Distortion Task measures how people learn a category's central pattern from variations on it (Posner and Keele, 1968). Each group's patterns are distortions of one pattern of nine dots, its prototype, made by moving every dot at random by a set rule (Posner, Goldsmith and Welton, 1967). People who learn the groups from distortions alone later sort the never-shown prototype about as well as the patterns they practiced, and better than new distortions. Amnesic patients sort new patterns of a learned category about as well as controls while recognizing the training patterns poorly (Knowlton and Squire, 1993), so the paradigm is used to study category learning apart from memory for single items.
More about the paradigm
Three groups each have their own random prototype: nine dots in the central 30 by 30 cells of a 50 by 50 grid. Training shows four high-level distortions of each prototype, about 7.7 bits per dot in Posner's measure, so twelve patterns, each once per block in a seeded order for 6 blocks (standard; 4 to 10), 72 trials. Each pattern stays until the participant chooses group 1, 2 or 3, by a click, a tap or the 1, 2 and 3 keys, with no time limit; Correct or Wrong and the group then show for 1 second. The test, without feedback, has 42 trials in two halves: each half shows, for each group, two training patterns, the prototype, two new low-level distortions (about 5 bits per dot) and two new high-level ones. Every pattern is made from the session's seed by the published rule, which the task's own tests check. A practice of 6 trials teaches the buttons with numbers. About 6 to 9 minutes; it runs on computers, tablets and phones.
The main score is the prototype advantage: accuracy on the prototypes minus accuracy on the new high-level distortions, the training level, so neither was seen in training. Accuracy on each kind of test pattern, the prototypes against the training patterns, and training accuracy are also reported. Scores support research and do not diagnose.
Scores
- Prototype Advantage%
- Accuracy on the never-shown prototypes minus accuracy on new distortions as strong as the training ones.
- Above 0 means the prototypes were sorted better than equally new patterns; a difference between kinds of patterns
- Prototype Accuracy%
- Share of test answers that put a prototype in its group; each prototype is shown once in each half of the test.
- Chance is one in three
- Training Pattern Accuracy%
- Share of test answers that put a training pattern in its group, with no feedback.
- Chance is one in three
- New Low-Distortion Accuracy%
- Share right on new patterns distorted less than the training ones (about 5 bits per dot).
- Chance is one in three
- New High-Distortion Accuracy%
- Share right on new patterns distorted as much as the training ones (about 7.7 bits per dot).
- Chance is one in three
- Prototype Minus Training Patterns%
- Prototype accuracy minus training pattern accuracy in the test.
- Near 0 means the unseen prototypes were sorted as well as the trained patterns
- Training Pattern Advantage%
- Training pattern accuracy minus accuracy on new high-level distortions.
- Above 0 means what was trained was sorted better than new patterns of the same distortion
- Test Accuracy%
- Share right over all 42 test patterns.
- Higher is better; chance is one in three
- Training Accuracy%
- Share right over every training trial, with feedback.
- Higher is better
- Last Training Block Accuracy%
- Share right in the last training block.
- Under 0.5 is flagged for review
- Median Test Choice Timems
- Median time to choose a group in the test.
- Informational
- Anticipationscount
- Presses within 100 ms of a pattern appearing, in training and the test. They are not taken as choices.
- Quality check
- Test Choices Madecount
- Test choices in the scored run.
- Fewer than planned is flagged for review
- Planned Test Trialscount
- Test trials in the session: 42.
- Informational
- Training Choices Madecount
- Training choices in the scored run: 12 per training block.
- Informational
Use and evidence
11 papers from 2021 to 2025 name this task in their title or abstract. That ranks it 450 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 4 |
| 2024 | 1 |
| 2023 | 1 |
| 2022 | 1 |
| 2021 | 4 |
| 2020 | 2 |
| 2019 | 2 |
| 2018 | 2 |
| 2017 | 1 |
| 2016 | 2 |
| 2015 | 2 |
| 2014 | 1 |
| 2013 | 1 |
| 2012 | 4 |
| 2011 | 5 |
| 2010 | 3 |
| 2009 | 0 |
| 2008 | 4 |
| 2007 | 0 |
| 2006 | 4 |
| 2005 | 0 |
| 2004 | 2 |
| 2003 | 0 |
| 2002 | 0 |
| 2001 | 2 |
| 2000 | 0 |
Use grade E
Fewer than 30 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 0
- Papers on reliability or validity
- 0
- Registered clinical studies using it as an outcome
- 1
How the grades work
Use: A is 1,000 or more papers from 2021 to 2025, B 300 to 999, C 100 to 299, D 30 to 99 and E fewer than 30. Evidence: A means at least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome; B, at least 8 test-retest papers, and either 5 registered clinical studies or 150 reliability or validity papers; C, at least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies; D, little published measurement evidence found. The counts measure use. They include every paper that names the task in its title or abstract. Grades describe the published research on this task, not ConductCognition's version of it.
Source: ConductCognition test census 0.22.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Patterns
- Each group's prototype is 9 dots in distinct cells of the central 30 by 30 cells of a 50 by 50 grid, drawn at random from the session's seed. Dots are 8 px across in a 300 px box, one cell every 6 px.
- Distortions
- Each dot moves on its own over a 400-cell template numbered in a spiral, into areas of 1, 8, 16, 75 and 300 cells: with probabilities 0, .24, .16, .30 and .30 for high-level distortions (about 7.7 bits per dot) and .20, .30, .40, .05 and .05 for low-level ones (about 5 bits per dot).
- Training
- 4 high-level distortions of each of the 3 prototypes, 12 patterns, each once per block in a seeded order: 6 blocks (standard; 4 to 10), 72 trials. Correct or Wrong and the group show for 1,000 ms, then a 500 ms blank.
- Answer
- Click or tap group 1, 2 or 3, or press the 1, 2 or 3 key, with no time limit. Presses under 100 ms are not taken.
- Test
- 42 trials with no feedback, in two halves; each half shows, for each group, 2 training patterns, the prototype, 2 new low-level and 2 new high-level distortions.
- Practice
- 6 trials in which the box shows a group's number, with feedback.
- Score
- The prototype advantage: accuracy on the never-shown prototypes minus accuracy on the new high-level distortions, the training level.
The patterns are made by our code with the published distortion rule, from random prototypes, so no published dot pattern is shown. Percentiles are not computed: the prototype advantage is a difference between two kinds of new patterns. The scores support research and do not diagnose.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: prototype_distortion_test from the scored run, a choice of group 1, 2 or 3 with no time limit and no feedback (chance one in three; each group has 14 of the 42 test patterns). Accuracy below chance invalidates and accuracy not above chance is flagged. One group on 90% of answers invalidates when accuracy is not above chance. Answers under 200 ms on at least half the patterns invalidate. A run short of its planned test patterns, or a last training block under 50% right, is flagged for review and withholds the percentile.
Why these rules
Accuracy significantly below chance (one-sided exact test, p < .05) cannot come from guessing; it means the response mapping was reversed or the task was not followed (Slick, Sherman and Iverson 1999 treat below-chance forced-choice performance as definite response bias). Chance-level accuracy can be a real result in impaired participants, so it is only flagged. One option on at least 90% of answered trials, when the correct options are balanced and accuracy is not above chance, means the answers did not depend on the items (the task counterpart of long-string responding in surveys; Curran 2016). An answer under 200 ms is faster than a simple visual reaction, so it cannot reflect the item. Following response-time effort screening for untimed tests (Wise and Kong 2005), the session is invalid only when at least half the answers are that fast. The test asks about groups learned in training, so a last training block near chance means the groups were not learned and the prototype advantage cannot be read.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Training blocks | 6 | 4 to 10 |
Yearly plans can change settings within these limits. Free and pay-as-you-go studies use the standard settings.
Need a setting this test does not have?
Devices and languages
- Untimed test: runs on desktops, laptops, tablets and phones. Participant devices
- Languages: English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese.
- Available on: Every plan, including the free plan.
- No percentile. Scores and quality flags only. How norms work
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Prototype Distortion Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Posner and Keele, 1968; Posner, Goldsmith and Welton, 1967), with the standard settings (training blocks: 6). The main score was prototype advantage (accuracy on the never-shown prototypes minus accuracy on new distortions as strong as the training ones).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Posner, M. I., & Keele, S. W. (1968). On the genesis of abstract ideas. Journal of Experimental Psychology, 77(3, Pt.1), 353-363. https://doi.org/10.1037/h0025953The task: categories learned from distortions of their prototypes, then the prototypes, the training patterns and new distortions sorted without feedback.
- Posner, M. I., Goldsmith, R., & Welton, K. E. (1967). Perceived distance and the classification of distorted patterns. Journal of Experimental Psychology, 73(1), 28-38. https://doi.org/10.1037/h0024135The statistical distortion rule and its levels in bits per dot.
- Knowlton, B. J., & Squire, L. R. (1993). The learning of categories: Parallel brain systems for item memory and category knowledge. Science, 262(5140), 1747-1749. https://doi.org/10.1126/science.8259522Amnesic patients classified new patterns of a learned category like controls but were impaired at recognizing the training patterns.
- Smith, J. D., & Minda, J. P. (2002). Distinguishing prototype-based and exemplar-based processes in dot-pattern category learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28(4), 800-811. https://doi.org/10.1037/0278-7393.28.4.800Steep typicality gradients and large prototype effects in dot-pattern category learning.
- Smith, J. D., Chapman, W. P., & Redford, J. S. (2010). Stages of category learning in monkeys (Macaca mulatta) and humans (Homo sapiens). Journal of Experimental Psychology: Animal Behavior Processes, 36(1), 39-53. https://doi.org/10.1037/a0016573The grid, the five areas and the probabilities of level 5 distortions.
- Coutinho, M. V. C., Couchman, J. J., Redford, J. S., & Smith, J. D. (2010). Refining the visual-cortical hypothesis in category learning. Brain and Cognition, 74(2), 88-96. https://doi.org/10.1016/j.bandc.2010.07.001The probabilities of high-level distortions.
- Gastgeb, H. Z., Dundas, E. M., Minshew, N. J., & Strauss, M. S. (2012). Category formation in autism: Can individuals with autism form categories and prototypes of dot patterns? Journal of Autism and Developmental Disorders, 42(8), 1694-1704. https://doi.org/10.1007/s10803-011-1411-xThe 400-cell spiral template and its five areas.
- Zaki, S. R., & Nosofsky, R. M. (2004). False prototype enhancement effects in dot pattern categorization. Memory & Cognition, 32(3), 390-398. https://doi.org/10.3758/BF03195833High-level distortions at 7.7 bits per dot, and the prototypes' compactness, which can favor them without learning.
Source papers
- Posner MI, Keele SW (1968). On the genesis of abstract ideas. Journal of Experimental Psychology, 77(3):353-363.Paradigm source, Adults
- No normative data for this task.Paradigm source, N/A
Questions
How long does the Prototype Distortion Task take?
About 6 to 9 minutes, as listed in the test library.
What device does a participant need?
Untimed test: runs on desktops, laptops, tablets and phones. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Prototype Advantage, Prototype Accuracy, Training Pattern Accuracy, New Low-Distortion Accuracy, New High-Distortion Accuracy, Prototype Minus Training Patterns, Training Pattern Advantage, Test Accuracy, Training Accuracy, Last Training Block Accuracy, Median Test Choice Time, Anticipations, Test Choices Made, Planned Test Trials and Training Choices Made. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Prototype Distortion Task used in research?
11 papers from 2021 to 2025 name it in their title or abstract. That ranks it 450 of 495 test families in the ConductCognition test census (use grade E). Source: ConductCognition test census 0.22.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Prototype Distortion Task have?
In the published research, 0 papers report its test-retest reliability and 1 registered clinical study uses it as an outcome (evidence grade D in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Training blocks (4 to 10). Settings can be changed on yearly plans; free and pay-as-you-go studies use the standard settings.
Which languages can participants use?
Participants can take this test in English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. English is on every plan, and the other languages come with yearly plans. Have a native speaker read the machine-translated text before launch.
Is this a clinical test?
No. ConductCognition is for research use only. It runs, scores and exports the test; clinical interpretation is outside the platform.
Every test: the full test list.
Run this test in your next study
Start free with 10 free participants on any test, then use participant credits or a yearly plan.