Social Cognition
Gaze and Expression Theory of Mind Task
Also called: cognitive and affective theory of mind task, eye gaze theory of mind task, mentalistic gaze task
Measure cognitive and affective theory of mind: participants choose what a drawn face is thinking of or likes from its gaze and expression, at first and second order, with physical control screens.
7-10 minUntimed
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Preview any test free after you sign up.
- Papers, 2021 to 2025
- 17
- Evidence grade
- C
- Registered clinical studies
- 0
What it measures
Measures cognitive and affective theory of mind, in the style of the Yoni task (Shamay-Tsoory and Aharon-Peretz, 2007). A drawn face sits in the middle with four pictures or four smaller faces around it, and a sentence such as "The face in the middle is thinking of ___". Participants choose what fits from where the faces look and whether they smile or frown. The main score is percent correct on the thinking and liking screens.
Constructs
- Theory of mind
- Cognitive theory of mind
- Affective theory of mind
- Gaze understanding
- Social cognition
Research uses
- Cognitive and affective theory of mind research
- Aging and social cognition research
- Research on frontal lobe damage and social cognition
- Within-study group comparison
How the paradigm works
The Gaze and Expression Theory of Mind Task measures how well people infer what someone is thinking of, and how they feel about it, from where a face looks and its expression. It follows the design of the Yoni task of Shamay-Tsoory and Aharon-Peretz (2007), which grew from the "Charlie" task of Baron-Cohen and colleagues (1995), with our own name, drawings and sentences. A drawn face sits in the middle of the screen. On first-order screens a picture sits in each corner, and a sentence with a blank asks what the face is thinking of (cognitive), what it likes or does not like (affective), or what it is next to (a physical control). On second-order screens a smaller face sits in each corner with a picture between each two, and the sentence asks about the corner faces: which one wants what the middle face is thinking of, likes what it likes, or has the same thing. Cognitive screens rest on gaze, with neutral faces and sentences; affective screens add smiles and frowns.
More about the paradigm
Every 64 screens follow the original's design: 24 cognitive, 24 affective and 16 physical, half first order and half second order. On 20 of them the middle face looks straight ahead, as in the original, and on 8 physical screens it looks at a wrong picture or face, so following its gaze alone does not answer. The session's seed draws every screen from rules that let exactly one answer fit, with the answer in each corner equally often. The faces are line drawings our code makes, the pictures are Noto emoji of five everyday groups, and the sentences are built from eight templates written in each of the app's eight languages; no drawing, slide or sentence of the original task is used. Part 1 holds the first-order screens and part 2 the second-order ones, each after three practice screens with feedback. Each screen allows 60 seconds by default, as in Isernia and colleagues (2023).
The main score is percent correct on the cognitive and affective screens, with a missed answer counted as not correct. The test also reports accuracy for each kind and order of screen, the affective minus cognitive difference, physical control accuracy, how often a misleading gaze was followed, and median times to a right answer, the measures Shamay-Tsoory and Aharon-Peretz (2007) report. In published studies the second-order screens separate groups more than the first-order ones (Isernia and colleagues, 2023), and the affective second-order screens are the ones on which people with orbitofrontal damage and offenders with psychopathic traits fell behind (Shamay-Tsoory and colleagues, 2010). A session needs answers on at least 75% of the screens and at least half of the physical screens right, the original's rule for excluding a participant. Scores support research and do not diagnose.
Scores
- Theory of Mind Accuracy%
- Percent correct on the cognitive and affective screens. A missed answer is not correct.
- Higher is better
- Cognitive Accuracy%
- Percent correct on the screens about what a face is thinking of or wants.
- Higher is better
- Affective Accuracy%
- Percent correct on the screens about what a face likes or does not like.
- Higher is better
- Affective Minus Cognitive%
- Affective accuracy minus cognitive accuracy.
- Below zero means the affective screens were harder
- First-Order Accuracy%
- Percent correct on the first-order cognitive and affective screens, about the middle face alone.
- Higher is better
- Second-Order Accuracy%
- Percent correct on the second-order cognitive and affective screens, about the middle face and a corner face together.
- Higher is better
- Cognitive First-Order Accuracy%
- Percent correct on first-order cognitive screens.
- Higher is better
- Affective First-Order Accuracy%
- Percent correct on first-order affective screens.
- Higher is better
- Cognitive Second-Order Accuracy%
- Percent correct on second-order cognitive screens.
- Higher is better
- Affective Second-Order Accuracy%
- Percent correct on second-order affective screens.
- Higher is better
- Physical Control Accuracy%
- Percent correct on the screens about what is next to the face or what it has, which need no mental state.
- Below 50% invalidates
- Gaze-Following Errorscount
- Physical screens where the middle face looked at a wrong picture or face and that one was chosen.
- Lower supports interpretation
- Median Time, Cognitivems
- Median time to a right answer on cognitive screens.
- Informational
- Median Time, Affectivems
- Median time to a right answer on affective screens.
- Informational
- Median Time, Physicalms
- Median time to a right answer on physical control screens.
- Informational
- Median Time, First Orderms
- Median time to a right answer on first-order cognitive and affective screens.
- Informational
- Median Time, Second Orderms
- Median time to a right answer on second-order cognitive and affective screens.
- Informational
- Answer Rate%
- Answers as a percent of the screens set.
- Below 75% invalidates
- Missed Answerscount
- Screens with no answer in the time allowed.
- Lower supports interpretation
- Screens Shownitems
- Screens the session reached.
- Informational
- Screens Setitems
- Screens the session set out to show.
- Informational
- Anticipationscount
- Answers under 100 ms.
- Lower supports interpretation
Use and evidence
17 papers from 2021 to 2025 name this task in their title or abstract, about the same as from 2016 to 2020. That ranks it 343 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 4 |
| 2024 | 4 |
| 2023 | 5 |
| 2022 | 2 |
| 2021 | 2 |
| 2020 | 2 |
| 2019 | 3 |
| 2018 | 3 |
| 2017 | 6 |
| 2016 | 3 |
| 2015 | 3 |
| 2014 | 2 |
| 2013 | 1 |
| 2012 | 1 |
| 2011 | 1 |
| 2010 | 1 |
| 2009 | 0 |
| 2008 | 0 |
| 2007 | 0 |
| 2006 | 0 |
| 2005 | 0 |
| 2004 | 1 |
| 2003 | 0 |
| 2002 | 0 |
| 2001 | 0 |
| 2000 | 0 |
Use grade E
Fewer than 30 papers from 2021 to 2025.
Evidence grade C
At least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies.
- Papers on test-retest reliability
- 3
- Papers on reliability or validity
- 7
- Registered clinical studies using it as an outcome
- 0
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 0.14.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Screens
- 64 by default, or 32 or 96 as a setting. Every 64 hold 24 cognitive, 24 affective and 16 physical control screens, half first order and half second order, as in Shamay-Tsoory and Aharon-Peretz (2007).
- First order
- A drawn face in the middle and a picture in each corner. The sentence asks what the face is thinking of, likes or does not like, or is next to.
- Second order
- A smaller face in each corner and a picture between each two. The sentence asks which corner face wants what the middle face is thinking of, likes what it likes or does not, or has the same thing.
- Gaze control
- On 20 of every 64 screens the middle face looks straight ahead, as in the original, and on 8 physical screens it looks at a wrong picture or face, so following its gaze alone does not answer.
- Pictures and sentences
- 4 Noto emoji a screen from 5 groups of 6: animals, fruit, vehicles, clothes and musical instruments. The sentences come from 8 templates written in each of the 8 app languages.
- Order
- The answer is in each corner equally often, with no more than 3 screens of one kind or 3 answers in one corner in a row. The session's seed draws every screen, and all of it is saved.
- Answer time
- Up to 60 seconds a screen by default, 20 to 60 as a setting. Isernia and colleagues (2023) allow 60 seconds.
- Practice
- 3 screens before each part, with feedback. The test has no feedback.
The faces are drawn by our code and the sentences were written for this test; no drawing, slide or sentence of the original task is used. 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: gaze_expression_tom_item test rows from the scored run, one choice of four corners per screen, the answer in each corner equally often, so chance is 25%. Accuracy significantly below chance invalidates. One corner on at least 90% of answers invalidates unless accuracy is above chance. Answers under 200 ms on at least half of the answers invalidate. Answers on fewer than 75% of the screens set invalidate. Fewer than half of the physical control screens right invalidates. A hidden tab 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. A missed screen gives no answer to score, and with fewer than three quarters of the screens answered the accuracies rest on too few screens to compare with a full session. The physical screens need only the sentence and where things are, not a mental state; Shamay-Tsoory and Aharon-Peretz (2007) excluded participants below 50% on them, because the sentences were not being read. A hidden tab lets the study, delay or response timers run without the participant, so a percentile is withheld after one.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Screens | 64 screens (standard) | 32 screens, 64 screens (standard), 96 screens |
| Answer time (seconds) | 60 | 20 to 60 |
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.
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Gaze and Expression Theory of Mind Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Baron-Cohen S, Campbell R, Karmiloff-Smith A, Grant J, Walker J, 1995; Shamay-Tsoory SG, Aharon-Peretz J, 2007; Shamay-Tsoory SG, Harari H, Aharon-Peretz J, Levkovitz Y, 2010; Isernia S and colleagues, 2023), with the standard settings (screens: 64 screens; answer time (seconds): 60). The main score was theory of mind accuracy (percent correct on the cognitive and affective screens).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Shamay-Tsoory, S. G., & Aharon-Peretz, J. (2007). Dissociable prefrontal networks for cognitive and affective theory of mind: A lesion study. Neuropsychologia, 45(13), 3054-3067. https://doi.org/10.1016/j.neuropsychologia.2007.05.021The task's design: cognitive, affective and physical screens at first and second order, with gaze controls.
- Shamay-Tsoory, S. G., Harari, H., Aharon-Peretz, J., & Levkovitz, Y. (2010). The role of the orbitofrontal cortex in affective theory of mind deficits in criminal offenders with psychopathic tendencies. Cortex, 46(5), 668-677. https://doi.org/10.1016/j.cortex.2009.04.008Affective and cognitive scores that separate groups.
- Baron-Cohen, S., Campbell, R., Karmiloff-Smith, A., Grant, J., & Walker, J. (1995). Are children with autism blind to the mentalistic significance of the eyes? British Journal of Developmental Psychology, 13(4), 379-398. https://doi.org/10.1111/j.2044-835X.1995.tb00687.xReading what a face wants from where it looks.
- Isernia, S., Rossetto, F., Shamay-Tsoory, S., Marchetti, A., & Baglio, F. (2023). Standardization and normative data of the 48-item Yoni short version for the assessment of theory of mind in typical and atypical conditions. Frontiers in Aging Neuroscience, 14, 1048599. https://doi.org/10.3389/fnagi.2022.1048599Answer time, reliability and the larger group differences on second-order items.
Source papers
- Isernia S, Rossetto F, Shamay-Tsoory S, Marchetti A, Baglio F (2023). Standardization and normative data of the 48-item Yoni short version for the assessment of theory of mind in typical and atypical conditions. Frontiers in Aging Neuroscience, 14:1048599.N 235, 18 and older
Questions
How long does the Gaze and Expression Theory of Mind Task take?
About 7 to 10 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?
Theory of Mind Accuracy, Cognitive Accuracy, Affective Accuracy, Affective Minus Cognitive, First-Order Accuracy, Second-Order Accuracy, Cognitive First-Order Accuracy, Affective First-Order Accuracy, Cognitive Second-Order Accuracy, Affective Second-Order Accuracy, Physical Control Accuracy, Gaze-Following Errors, Median Time, Cognitive, Median Time, Affective, Median Time, Physical, Median Time, First Order, Median Time, Second Order, Answer Rate, Missed Answers, Screens Shown, Screens Set and Anticipations. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Gaze and Expression Theory of Mind Task used in research?
17 papers from 2021 to 2025 name it in their title or abstract, about the same as from 2016 to 2020. That ranks it 343 of 395 test families in the ConductCognition test census (use grade E). Source: ConductCognition test census 0.14.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Gaze and Expression Theory of Mind Task have?
In the published research, 3 papers report its test-retest reliability and 0 registered clinical studies use it as an outcome (evidence grade C in the ConductCognition test census).
Which settings can a researcher change?
Screens (32 screens, 64 screens (standard), 96 screens) and answer time (seconds) (20 to 60). 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, with its words in the same language. 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.