Executive Function
Action and Valence Go/No-Go Task
Also called: orthogonalized go/no-go task, motivational go/no-go learning task, valenced go/no-go task, Pavlovian bias task
Measure how winning and losing shape learning to act or hold back: four pictures teach, by probabilistic points, when to press to win or avoid losing and when not to, scored as the Pavlovian bias.
12-16 minTimed: desktop or laptop with a mouse or trackpad
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
- 12
- Evidence grade
- D
- Registered clinical studies
- 2
What it measures
Measures how reward and loss shape learning to act or to hold back. Four pictures teach, from points won and lost by chance, whether to press to win, press to avoid losing, not press to win or not press to avoid losing. The main score is the Pavlovian bias: how much more accurate choices are when pressing goes with winning and holding back with avoiding a loss than the other way round, with learning in each condition next.
Constructs
- Pavlovian-instrumental interaction
- Reward and punishment learning
- Action and inhibition
- Learning from feedback
Research uses
- Research on anxiety, depression and avoidance
- Dopamine and serotonin drug studies
- Developmental research on motivated learning
How the paradigm works
The action and valence go/no-go task measures how the prospect of winning or losing shapes learning to act or to hold back (Guitart-Masip, Huys, Fuentemilla and colleagues, 2012). Four pictures each stand for one condition: press to win, press to avoid losing, do not press to win, and do not press to avoid losing. Each round shows a picture, then a circle on the left or the right, and the participant either presses the key on the circle's side or presses nothing. The participant is not told what any picture means and learns it from the outcomes: a correct choice brings the better outcome, a win or no loss, on 80% of rounds, and an incorrect one on 20%. In the source, people learned to press for a win and to hold back to avoid a loss better than the reverse pairings, an interaction of action and valence that computational models attribute to a Pavlovian influence on instrumental choice. Accuracy was highest for press to win and lowest for do not press to win in Cavanagh and colleagues (2013).
More about the paradigm
Each round shows the picture for 1 second, a cross for 250 to 2,000 ms, then the circle for 1 second, in which the first press of F (left) or J (right) is the choice; a press during the picture or the cross does not count. After a 500 ms blank the outcome shows for 1 second: an upward green arrow with the points for a win, a downward red arrow for a loss, or a yellow bar for nothing. A press on the side away from the circle is not correct. The rounds come in sets of four, one of each picture in a seeded order, 36 rounds of each picture (20 to 60 in the settings; the source used 60), with the circle on each side in half of them and a pause halfway. The pictures are four abstract signs drawn by code for this task, and which picture stands for which condition comes from the session's seed. Twenty practice rounds of the circle alone, with feedback on the key, come first, as in the source. A win adds 10 points (a setting) and a loss takes them away; the points are not money unless the study says so. About 12 to 16 minutes, on a desktop or laptop with a keyboard.
The main score is the Pavlovian bias: half the sum of press-to-win accuracy minus press-to-avoid-losing accuracy and do-not-press-to-avoid-losing accuracy minus do-not-press-to-win accuracy. It equals the share of rounds with a press after win pictures minus that after loss pictures, from -1 to 1, and 0 means the two kinds of picture were learned alike. Accuracy in each condition, over the whole session and over each condition's second half, comes next, then the main effects of action and of valence, the Pavlovian performance bias of Cavanagh and colleagues (2013, 0.5 with perfect learning), the press rate, response times for correct presses, and quality checks. The scores describe choices in this task as research measures; they do not describe a person as biased, and they do not diagnose. In a large cohort, one person's modeled Pavlovian bias changed a great deal between sessions (Moutoussis and colleagues, 2018), so compare groups.
Scores
- Pavlovian Bias%
- Half of (press-to-win minus press-to-avoid-losing accuracy) plus (do-not-press-to-avoid-losing minus do-not-press-to-win accuracy): the press rate after win pictures minus that after loss pictures.
- Positive: pressing went with winning and holding back with avoiding loss. Research measure, no better direction
- Accuracy, Press to Win%
- Rounds with the press-to-win picture answered with the key on the circle's side.
- Higher is better
- Accuracy, Press to Avoid Losing%
- Rounds with the press-to-avoid-losing picture answered with the key on the circle's side.
- Higher is better
- Accuracy, Do Not Press to Win%
- Rounds with the do-not-press-to-win picture answered with no press.
- Higher is better
- Accuracy, Do Not Press to Avoid Losing%
- Rounds with the do-not-press-to-avoid-losing picture answered with no press.
- Higher is better
- Late Accuracy, Press to Win%
- Accuracy in the second half of the press-to-win rounds.
- Higher is better
- Late Accuracy, Press to Avoid Losing%
- Accuracy in the second half of the press-to-avoid-losing rounds.
- Higher is better
- Late Accuracy, Do Not Press to Win%
- Accuracy in the second half of the do-not-press-to-win rounds.
- Higher is better
- Late Accuracy, Do Not Press to Avoid Losing%
- Accuracy in the second half of the do-not-press-to-avoid-losing rounds.
- Higher is better
- Action Effect%
- Mean accuracy of the two press pictures minus that of the two do-not-press pictures.
- Positive: pressing was learned better than holding back
- Valence Effect%
- Mean accuracy of the two win pictures minus that of the two loss pictures.
- Positive: winning was learned better than avoiding loss
- Pavlovian Performance Bias%
- The mean of the share of presses that came after win pictures and the share of non-presses that came after loss pictures (Cavanagh and colleagues, 2013).
- 0.5 with perfect learning; higher means choices followed the pictures' valence more. Research measure
- Overall Accuracy%
- Correct rounds over all four pictures.
- Higher is better
- Press Rate%
- Rounds with a press while the circle showed, on either side.
- 0.5 when every picture is learned
- Press Rate, Second Half%
- Rounds with a press in the second half of the session.
- 0 is flagged for review
- Median Press Time, Press to Winms
- Median time from the circle to a correct press for the press-to-win picture.
- Informational
- Median Press Time, Press to Avoid Losingms
- Median time from the circle to a correct press for the press-to-avoid-losing picture.
- Informational
- Wrong-Side Presses%
- Presses on the side away from the circle, among all presses.
- Above 0.1 is flagged for review
- Anticipationscount
- Presses within 100 ms of the circle appearing. They still count as the round's press.
- Quality check
- Early Pressescount
- Presses while the picture or the cross showed, before the circle. They do not count as choices.
- Quality check
- Pointspoints
- Points won minus points lost over the session.
- Informational
- Rounds Completedcount
- Rounds in the scored run.
- Fewer than planned is flagged for review
- Planned Roundscount
- Rounds in the session: four times the rounds per picture.
- Informational
Use and evidence
12 papers from 2021 to 2025 name this task in their title or abstract. That ranks it 441 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 6 |
| 2024 | 1 |
| 2023 | 4 |
| 2022 | 1 |
| 2021 | 0 |
| 2020 | 1 |
| 2019 | 0 |
| 2018 | 0 |
| 2017 | 0 |
| 2016 | 1 |
| 2015 | 1 |
| 2014 | 1 |
| 2013 | 0 |
| 2012 | 0 |
| 2011 | 0 |
| 2010 | 0 |
| 2009 | 1 |
| 2008 | 0 |
| 2007 | 0 |
| 2006 | 0 |
| 2005 | 0 |
| 2004 | 0 |
| 2003 | 0 |
| 2002 | 0 |
| 2001 | 0 |
| 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
- 1
- Registered clinical studies using it as an outcome
- 2
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 31 October 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Conditions
- Four abstract pictures drawn by code for this task stand for press to win, press to avoid losing, do not press to win and do not press to avoid losing. Which picture is which comes from the session's seed, and the participant learns it from the outcomes.
- Round
- The picture for 1,000 ms, a cross for 250 to 2,000 ms, then a circle on the left or the right for 1,000 ms, in which the first press of F (left) or J (right) is the choice. A 500 ms blank, the outcome for 1,000 ms, then a 500 ms blank. A press during the picture or the cross does not count.
- Outcomes
- A correct choice brings the better outcome on 80% of rounds and an incorrect one on 20%, from a seeded draw for each round. Win pictures pay a win or nothing; loss pictures give nothing or a loss. An upward green arrow, a downward red arrow or a yellow bar, each with the points. A press on the side away from the circle is not correct.
- Rounds
- 36 of each picture (standard; 20 to 60), in sets of four with one of each picture in a seeded order, never the same picture twice in a row, the circle on each side in half of each picture's rounds, and a pause halfway.
- Points
- A win adds 10 points and a loss takes them away (standard; 1 to 100). The standard payment line says the points are not exchanged for money.
- Practice
- 20 rounds of the circle alone, with Right key, Wrong key or Too slow after each, as in the source.
The source used fractal pictures and showed the circle for 1,500 ms with 1,000 ms to respond, and 1,000 ms passed between the circle and the outcome. This test uses abstract signs of its own, shows the circle for the 1,000 ms response window and shortens that pause to 500 ms, so its accuracies are not directly comparable with published ones.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: action_valence_gonogo_trial from the scored run, a press of the key on the circle's side, or no press, within 1 second after each cue. Which cues need a press is learned from probabilistic points, so neither accuracy nor the split of presses between cues is judged against chance: pressing for the wrong cues is what the test measures. A session with no press at all is invalid. Responses under 100 ms follow the timed-test rule. No press in the second half of the session, more than 10% of presses on the side away from the circle, and a run short of its planned rounds are flagged for review and withhold the percentile. A hidden tab withholds the percentile.
Why these rules
Learning which cues to press for is the measure, and studies of this task find participants who never learn some conditions (Guitart-Masip and colleagues, 2012; Cavanagh and colleagues, 2013), so wrong choices cannot mean the instructions were reversed. A session with no press, or none after halfway, was not engaged in the rounds; presses on the side away from the circle were under 0.1% in the source. Responses under 100 ms follow the timed-test rule: a press that soon cannot be a reaction to the circle. A hidden tab lets the rounds run on without the participant, and every round without a press then counts as a choice.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Rounds per picture | 36 | 20 to 60 |
| Points per win or loss | 10 | 1 to 100 |
| Payment line | Hypothetical points (standard) | Hypothetical points (standard), The study pays for points |
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
- Timed test: participants need a desktop or laptop with a mouse or trackpad. 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 Action and Valence Go/No-Go Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Guitart-Masip, Huys, Fuentemilla et al., 2012; Cavanagh, Eisenberg, Guitart-Masip et al., 2013), with the standard settings (rounds per picture: 36; points per win or loss: 10; payment line: Hypothetical points). The main score was pavlovian bias (half of (press-to-win minus press-to-avoid-losing accuracy) plus (do-not-press-to-avoid-losing minus do-not-press-to-win accuracy): the press rate after win pictures minus that after loss pictures).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Guitart-Masip, M., Huys, Q. J. M., Fuentemilla, L., Dayan, P., Duzel, E., & Dolan, R. J. (2012). Go and no-go learning in reward and punishment: Interactions between affect and effect. NeuroImage, 62(1), 154-166. https://doi.org/10.1016/j.neuroimage.2012.04.024The task: four cues, the circle's side as the press, the timings, 80% probabilistic outcomes, 60 rounds per condition, 20 practice rounds, and the action by valence interaction in learning.
- Guitart-Masip, M., Fuentemilla, L., Bach, D. R., Huys, Q. J. M., Dayan, P., Dolan, R. J., & Duzel, E. (2011). Action dominates valence in anticipatory representations in the human striatum and dopaminergic midbrain. Journal of Neuroscience, 31(21), 7867-7875. https://doi.org/10.1523/JNEUROSCI.6376-10.2011The earlier form with instructed cues.
- Cavanagh, J. F., Eisenberg, I., Guitart-Masip, M., Huys, Q., & Frank, M. J. (2013). Frontal theta overrides Pavlovian learning biases. Journal of Neuroscience, 33(19), 8541-8548. https://doi.org/10.1523/JNEUROSCI.5754-12.2013The Pavlovian performance bias, and accuracy by condition with 40 rounds each.
- Guitart-Masip, M., Economides, M., Huys, Q. J. M., Frank, M. J., Chowdhury, R., Duzel, E., Dayan, P., & Dolan, R. J. (2014). Differential, but not opponent, effects of L-DOPA and citalopram on action learning with reward and punishment. Psychopharmacology, 231(5), 955-966. https://doi.org/10.1007/s00213-013-3313-4The task in a drug study, with 60 rounds per condition.
- Moutoussis, M., Bullmore, E. T., Goodyer, I. M., Fonagy, P., Jones, P. B., Dolan, R. J., Dayan, P., & the NSPN Consortium (2018). Change, stability, and instability in the Pavlovian guidance of behaviour from adolescence to young adulthood. PLoS Computational Biology, 14(12), e1006679. https://doi.org/10.1371/journal.pcbi.100667936 rounds of each stimulus in a large community cohort, and the low stability of one person's modeled bias between sessions.
- Raab, H. A., & Hartley, C. A. (2020). Adolescents exhibit reduced Pavlovian biases on instrumental learning. Scientific Reports, 10, 15770. https://doi.org/10.1038/s41598-020-72628-wThe Pavlovian performance bias across age, with 45 rounds per condition.
Source papers
- Guitart-Masip M, Huys QJM, Fuentemilla L, Dayan P, Duzel E, Dolan RJ (2012). Go and no-go learning in reward and punishment: Interactions between affect and effect. NeuroImage, 62(1):154-166.N 47, 18 to 35
- Cavanagh JF, Eisenberg I, Guitart-Masip M, Huys Q, Frank MJ (2013). Frontal theta overrides Pavlovian learning biases. Journal of Neuroscience, 33(19):8541-8548.N 34, 18 to 34
- No normative data for this task.Paradigm source, N/A
Questions
How long does the Action and Valence Go/No-Go Task take?
About 12 to 16 minutes, as listed in the test library.
What device does a participant need?
Timed test: participants need a desktop or laptop with a mouse or trackpad. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Pavlovian Bias, Accuracy, Press to Win, Accuracy, Press to Avoid Losing, Accuracy, Do Not Press to Win, Accuracy, Do Not Press to Avoid Losing, Late Accuracy, Press to Win, Late Accuracy, Press to Avoid Losing, Late Accuracy, Do Not Press to Win, Late Accuracy, Do Not Press to Avoid Losing, Action Effect, Valence Effect, Pavlovian Performance Bias, Overall Accuracy, Press Rate, Press Rate, Second Half, Median Press Time, Press to Win, Median Press Time, Press to Avoid Losing, Wrong-Side Presses, Anticipations, Early Presses, Points, Rounds Completed and Planned Rounds. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Action and Valence Go/No-Go Task used in research?
12 papers from 2021 to 2025 name it in their title or abstract. That ranks it 441 of 495 test families in the ConductCognition test census (use grade E). Source: ConductCognition test census 0.22.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Action and Valence Go/No-Go Task have?
In the published research, 0 papers report its test-retest reliability and 2 registered clinical studies use 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?
Rounds per picture (20 to 60), points per win or loss (1 to 100) and payment line (Hypothetical points (standard), The study pays for points). 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.