Executive Function
AX Continuous Performance Test (AX-CPT)
Also called: AX version of the continuous performance test
Measure how people use a cue to prepare a response, with letter pairs scored by d'-context and the proactive behavioral index.
16-19 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. Preview any test free after you sign up.
- Papers, 2021 to 2025
- 180
- Evidence grade
- C
- Registered clinical studies
- 191
What it measures
Measures how people keep a cue in mind and use it to prepare a response. Letters appear in pairs; participants press one key for an X that follows an A and another key for every other letter. Errors on B-X and A-Y pairs show how much they rely on the cue.
Constructs
- Context processing and goal maintenance
- Proactive and reactive control
- Cognitive control
Research uses
- Cognitive control research
- Within-study group comparison
- Individual-differences and intervention studies
How the paradigm works
The AX continuous performance test (AX-CPT) measures how people keep a cue in mind and use it to prepare a response, the proactive control described by Braver et al. (2009). Letters appear in pairs: a cue letter, a blank delay and a probe letter. The target is an X that follows an A. As in the version of Braver et al. (2009), every letter gets a response: J for the X of an A-X pair and F for every other letter, cues included.
More about the paradigm
Each pair shows the cue for 300 ms, a 4,900 ms blank delay and the probe for 300 ms. A response counts within 1,300 ms of each letter's onset, and pairs are 1,000 ms apart. 70% of pairs are A-X, and A-Y, B-X and B-Y pairs are 10% each: 84, 12, 12 and 12 of the standard 120 pairs, as in Braver et al. (2009). Cues and probes other than A and X are capital letters except A, X, K and Y (Gonthier et al., 2016) and except the response keys F and J. The session starts with 10 practice pairs with feedback, and every session runs a recorded random order.
d'-context, the primary score, compares hits on A-X probes with false alarms on B-X probes: Z(hit rate) minus Z(false-alarm rate), with the log-linear correction (count + 0.5) / (trials + 1) used by Gonthier et al. (2016) after Braver et al. (2009). Higher values mean the cue separated the X of an A-X pair from the X of a B-X pair. The proactive behavioral index, (A-Y minus B-X) / (A-Y plus B-X), is computed on corrected error rates and on mean correct response times (Braver et al., 2009); positive values mean more reliance on the cue. Error rates and response times are also reported for each pair type.
Scores
- d'-Contextd'
- Z(A-X hit rate) minus Z(B-X false-alarm rate), log-linear corrected.
- Higher values mean better use of the cue
- Proactive Index (Errors)score
- (A-Y minus B-X) / (A-Y plus B-X) corrected error rates.
- Positive values mean more reliance on the cue
- Proactive Index (RT)score
- (A-Y minus B-X) / (A-Y plus B-X) mean correct response times.
- Positive values mean more reliance on the cue
- A-X Error Rate%
- Share of A-X probes answered with F or not answered.
- Lower is better
- A-Y Error Rate%
- Share of A-Y probes answered with J or not answered.
- Higher values mean the A cue led to more target responses
- B-X Error Rate%
- Share of B-X probes answered with J or not answered.
- Higher values mean the cue was used less
- B-Y Error Rate%
- Share of B-Y probes answered with J or not answered.
- Lower is better
- A-X RTms
- Mean correct response time on A-X probes, from the probe onset.
- Lower is faster
- A-Y RTms
- Mean correct response time on A-Y probes, from the probe onset.
- Lower is faster
- B-X RTms
- Mean correct response time on B-X probes, from the probe onset.
- Lower is faster
- B-Y RTms
- Mean correct response time on B-Y probes, from the probe onset.
- Lower is faster
- Cue Accuracy%
- Share of cue letters answered with F.
- Higher is better
Use and evidence
180 papers from 2021 to 2025 name this task in their title or abstract, up 16% compared with 2016 to 2020. That ranks it 132 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 43 |
| 2024 | 35 |
| 2023 | 31 |
| 2022 | 36 |
| 2021 | 35 |
| 2020 | 40 |
| 2019 | 37 |
| 2018 | 39 |
| 2017 | 18 |
| 2016 | 21 |
| 2015 | 25 |
| 2014 | 20 |
| 2013 | 16 |
| 2012 | 9 |
| 2011 | 13 |
| 2010 | 10 |
| 2009 | 13 |
| 2008 | 18 |
| 2007 | 7 |
| 2006 | 10 |
| 2005 | 8 |
| 2004 | 5 |
| 2003 | 6 |
| 2002 | 2 |
| 2001 | 1 |
| 2000 | 6 |
Use grade C
100 to 299 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
- 43
- Registered clinical studies using it as an outcome
- 191
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 31 October 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Display
- One capital letter at a time in the middle of the screen.
- Pairs
- A cue letter for 300 ms, a 4,900 ms blank delay, then a probe letter for 300 ms; pairs are 1,000 ms apart.
- Responses
- J for an X after an A and F for every other letter, cues included, within 1,300 ms of each letter's onset.
- Pair types
- 70% A-X, and 10% each of A-Y, B-X and B-Y.
- Letters
- Cues and probes other than A and X are capital letters except A, X, K, Y, F and J.
- Length
- 120 pairs (60 to 200), with a break after every 40.
- Practice
- 10 pairs with feedback: 4 A-X and 2 of each other type.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: ax_cpt_test probes from the scored run, J for an X right after an A and F for any other probe (two keys, chance 50%) within 1,300 ms of the probe onset. Below chance invalidates and not above chance is flagged. A-X probes are 70% of the pairs, so the one-key check is replaced: a session that pressed J on A-X probes no more often than on B-Y probes is invalid. Missed responses and responses under 100 ms follow the timed-test rule.
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. Missed responses and responses under 100 ms keep the platform's timed-test rule: more than max(3, 25%) missed responses or more than max(2, 20%) responses under 100 ms invalidate, and any is flagged. A response under 100 ms cannot be a reaction to the stimulus. With 70% targets, pressing J on every probe scores above chance, so the J rate on A-X probes must exceed the J rate on B-Y probes, the pairs where neither letter is part of the target.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Letter pairs | 120 | 60 to 200 |
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.
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the AX Continuous Performance Test in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Servan-Schreiber, Cohen and Steingard, 1996; Braver et al., 2009), with the standard settings (letter pairs: 120). The main score was d'-context (Z(A-X hit rate) minus Z(B-X false-alarm rate), log-linear corrected).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Servan-Schreiber, D., Cohen, J. D., & Steingard, S. (1996). Schizophrenic deficits in the processing of context: A test of a theoretical model. Archives of General Psychiatry, 53(12), 1105-1112. https://doi.org/10.1001/archpsyc.1996.01830120037008
- Braver, T. S., Paxton, J. L., Locke, H. S., & Barch, D. M. (2009). Flexible neural mechanisms of cognitive control within human prefrontal cortex. Proceedings of the National Academy of Sciences, 106(18), 7351-7356. https://doi.org/10.1073/pnas.0808187106Timing, pair frequencies, a response to every letter and the proactive behavioral index.
- Gonthier, C., Macnamara, B. N., Chow, M., Conway, A. R. A., & Braver, T. S. (2016). Inducing proactive control shifts in the AX-CPT. Frontiers in Psychology, 7, 1822. https://doi.org/10.3389/fpsyg.2016.01822Letter set and the log-linear correction for d'-context.
Source papers
- Braver TS, Paxton JL, Locke HS, Barch DM (2009). Flexible neural mechanisms of cognitive control within human prefrontal cortex. Proceedings of the National Academy of Sciences, 106(18):7351-7356.Paradigm source, Adults
Questions
How long does the AX Continuous Performance Test (AX-CPT) take?
About 16 to 19 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?
d'-Context, Proactive Index (Errors), Proactive Index (RT), A-X Error Rate, A-Y Error Rate, B-X Error Rate, B-Y Error Rate, A-X RT, A-Y RT, B-X RT, B-Y RT and Cue Accuracy. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the AX Continuous Performance Test (AX-CPT) used in research?
180 papers from 2021 to 2025 name it in their title or abstract, up 16% compared with 2016 to 2020. That ranks it 132 of 395 test families in the ConductCognition test census (use grade C). Source: ConductCognition test census 0.14.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the AX Continuous Performance Test (AX-CPT) have?
In the published research, 3 papers report its test-retest reliability and 191 registered clinical studies use it as an outcome (evidence grade C in the ConductCognition test census).
Which settings can a researcher change?
Letter pairs (60 to 200). 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.