Episodic Memory
Repeat Detection Task
Also called: continuous recognition task, continuous recognition paradigm, continuous recognition memory, repeat detection paradigm, old-new picture recognition
A continuous recognition test: pictures appear one at a time, some come back after 2, 8 or 32 others, and participants say whether each one is new or seen before. Scored as d-prime.
5-7 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
- 54
- Evidence grade
- D
- Registered clinical studies
- 2
What it measures
A continuous recognition test. Pictures appear one at a time and some come back after 2, 8 or 32 others; participants say whether each picture is new or seen before. The main score is d′ for telling repeats from new pictures.
Constructs
- Recognition memory
- Episodic memory
- Visual memory
Research uses
- Research on recognition memory in aging, mild cognitive impairment and dementia
- Studies of forgetting over short and longer lags
- Experimental research on picture memory and interference
How the paradigm works
The Repeat Detection Task is a continuous recognition test: memory is tested while new pictures keep arriving, rather than in a separate test after a study list. Pictures appear one at a time in one long stream. Some come back once, and for each picture the participant says whether it has already appeared in the task (Seen before) or is new. Shepard and Teghtsoonian (1961) introduced the procedure, Hockley (1982) used it to show how answers slow as more items come between the two showings, and Friedman (1990) ran it with pictures that came back after 2, 8 or 32 other pictures.
More about the paradigm
In this version, the standard stream has 118 pictures of everyday objects, animals and places: 48 pictures come back once, 16 after 2 other pictures, 16 after 8 and 16 after 32, and 22 are shown only once. The pictures shown once keep the end of the stream from turning into repeats alone, and the places of the repeats are drawn so that their share levels off after the first quarter of the stream. No more than 4 repeats or 8 new pictures come in a row. Each picture stays on screen until it is answered, with no time limit and no feedback, and 300 ms of blank screen follow. A setting gives 12 to 16 repeats at each lag (94 to 118 pictures). A practice stream of 8 pictures, 3 of which come back, gives feedback after each answer.
The pictures are 75 Noto Emoji pictures that the Associative Recognition Test or the Remember and Know Test also shows, so a study runs this test and those two in different sessions. Each session gives every picture its role at random: shown twice at one of the lags, or shown once.
The main score is d′ for telling repeats from new pictures, with the log-linear correction (Hautus 1995). The task also reports the hit and false alarm rates, the criterion c, accuracy, the hit rate at each lag and its drop from the shortest lag to the longest, the false alarm rate in each half of the stream, and answer times. Pictures are remembered well: in Brady and colleagues' (2008) stream of object pictures, repeats were detected near ceiling with up to 63 pictures between the two showings, so young adults' scores can come near the top of the scale, and lag effects show more in older adults. Scores support research and do not diagnose.
Scores
- d-Primed'
- z(hit rate) minus z(false alarm rate) for Seen before answers, with the log-linear correction.
- Higher means repeats were told from new pictures better
- Hit Rate%
- Share of repeats answered Seen before.
- Higher is better
- False Alarm Rate%
- Share of new pictures answered Seen before.
- Lower is better
- Criterionz
- Minus the mean of z(hit rate) and z(false alarm rate), with the log-linear correction.
- Positive means a stricter Seen before; context for the rates
- Accuracy%
- Share of pictures answered correctly: Seen before for repeats and New for new pictures.
- Higher is better; about 6 in 10 pictures are new, so read it with d-prime
- Hit Rate, Lag 2%
- Share of repeats answered Seen before when 2 pictures came between the two showings.
- Higher is better
- Hit Rate, Lag 8%
- Share of repeats answered Seen before when 8 pictures came between the two showings.
- Higher is better
- Hit Rate, Lag 32%
- Share of repeats answered Seen before when 32 pictures came between the two showings.
- Higher is better
- Lag Effect%
- Hit rate at lag 2 minus hit rate at lag 32.
- Positive when repeats were missed more after more pictures
- False Alarm Rate, First Half%
- Share of new pictures in the first half of the stream answered Seen before.
- Context for the false alarm rate
- False Alarm Rate, Second Half%
- Share of new pictures in the second half of the stream answered Seen before.
- Context for the false alarm rate
- Answer Time, Hitsms
- Median time to answer Seen before to repeats.
- Context for the hit rate
- Answer Time, Correct Rejectionsms
- Median time to answer New to new pictures.
- Context for the false alarm rate
- Anticipationscount
- Answers faster than 100 ms, left out of every score.
- Lower is better
- Pictures Showncount
- Pictures shown in the scored run (118 at the standard settings).
- Informational
Use and evidence
54 papers from 2021 to 2025 name this task in their title or abstract, about the same as from 2016 to 2020. That ranks it 246 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 12 |
| 2024 | 7 |
| 2023 | 11 |
| 2022 | 13 |
| 2021 | 11 |
| 2020 | 11 |
| 2019 | 12 |
| 2018 | 11 |
| 2017 | 10 |
| 2016 | 10 |
| 2015 | 13 |
| 2014 | 9 |
| 2013 | 10 |
| 2012 | 13 |
| 2011 | 10 |
| 2010 | 13 |
| 2009 | 9 |
| 2008 | 13 |
| 2007 | 6 |
| 2006 | 4 |
| 2005 | 14 |
| 2004 | 6 |
| 2003 | 5 |
| 2002 | 5 |
| 2001 | 5 |
| 2000 | 7 |
Use grade D
30 to 99 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 1
- Papers on reliability or validity
- 15
- 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 0.14.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Pictures
- 75 emoji pictures of everyday objects, animals and places from Noto Emoji, all from the Associative Recognition and Remember and Know tests: 70 for the test and 5 for practice.
- Stream
- 118 pictures one at a time (standard; 94 to 118): 48 pictures shown twice, 16 at each lag (standard; 12 to 16), and 22 shown once. No more than 4 repeats or 8 new pictures in a row.
- Lags
- A picture comes back after 2, 8 or 32 other pictures.
- Answer
- Seen before or New for each picture, with no time limit and no feedback, then 300 ms of blank screen.
- Practice
- 8 pictures, 3 of which come back after 2 or 3 others, with 1,500 ms of feedback after each answer.
- Roles
- Each session gives every picture its role at random: shown twice at one of the lags, or shown once.
The pictures are also shown by the Associative Recognition Test and the Remember and Know Test, so run those in another session. Pictures are remembered well, so young adults' hit rates can come near the top of the scale (Brady et al. 2008). 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: continuous_recognition_test from the scored run, one Seen before or New answer per picture. Repeats are fewer than new pictures, so accuracy is not judged against 50%. Instead, new pictures answered Seen before significantly more often than repeats (one-sided Fisher exact test, p < .05, with at least 8 of each answered) invalidate, and repeats not answered Seen before more often than new pictures beyond chance are flagged. The same answer for every picture invalidates. Answers under 200 ms on at least half of 4 or more answered pictures invalidate. A hidden tab withholds the percentile.
Why these rules
About 4 in 10 pictures are repeats, so always answering New is right on 6 in 10 and accuracy against 50% would pass it. Comparing Seen before answers on repeats with those on new pictures separates memory from a habit of one answer, whatever the share of repeats: Responses concentrated on trials that needed none (one-sided Fisher exact test, p < .05) mean the instructions were reversed and invalidate; responses that do not separate the two trial types are flagged. One answer for every picture means the pictures were not looked at; a participant who rarely recognizes a repeat but still answers Seen before mostly to repeats keeps a valid session. 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 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 |
|---|---|---|
| Repeats at each lag | 16 | 12 to 16 |
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 Repeat Detection Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Shepard & Teghtsoonian, 1961; Hockley, 1982), with the standard settings (repeats at each lag: 16). The main score was d-prime (z(hit rate) minus z(false alarm rate) for Seen before answers, with the log-linear correction).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Shepard, R. N., & Teghtsoonian, M. (1961). Retention of information under conditions approaching a steady state. Journal of Experimental Psychology, 62(3), 302-309. https://doi.org/10.1037/h0048606The continuous recognition procedure: one long series in which each item is new or a repeat.
- Hockley, W. E. (1982). Retrieval processes in continuous recognition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 8(6), 497-512. https://doi.org/10.1037/0278-7393.8.6.497A yes or no answer to every item, with lag as the number of items between the two showings.
- Friedman, D. (1990). Cognitive event-related potential components during continuous recognition memory for pictures. Psychophysiology, 27(2), 136-148. https://doi.org/10.1111/j.1469-8986.1990.tb00365.xPictures answered new or old, repeated after 2, 8 or 32 other pictures.
- Brady, T. F., Konkle, T., Alvarez, G. A., & Oliva, A. (2008). Visual long-term memory has a massive storage capacity for object details. Proceedings of the National Academy of Sciences, 105(38), 14325-14329. https://doi.org/10.1073/pnas.0803390105Repeats of object pictures detected near ceiling up to 63 intervening pictures.
- Poon, L. W., & Fozard, J. L. (1980). Age and word frequency effects in continuous recognition memory. Journal of Gerontology, 35(1), 77-86. https://doi.org/10.1093/geronj/35.1.77Age differences in continuous recognition across lags.
- Hautus, M. J. (1995). Corrections for extreme proportions and their biasing effects on estimated values of d'. Behavior Research Methods, Instruments, & Computers, 27(1), 46-51. https://doi.org/10.3758/BF03203619The log-linear correction for d-prime.
Source papers
- Shepard RN, Teghtsoonian M (1961). Retention of information under conditions approaching a steady state. Journal of Experimental Psychology, 62(3):302-309.Paradigm source, Adults
- Friedman D (1990). Cognitive event-related potential components during continuous recognition memory for pictures. Psychophysiology, 27(2):136-148.Paradigm source, Young adults
- Poon LW, Fozard JL (1980). Age and word frequency effects in continuous recognition memory. Journal of Gerontology, 35(1):77-86.Paradigm source, Men with median ages of 20, 52 and 63
- ConductCognition platform normsParadigm source, N/A
Questions
How long does the Repeat Detection Task take?
About 5 to 7 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?
d-Prime, Hit Rate, False Alarm Rate, Criterion, Accuracy, Hit Rate, Lag 2, Hit Rate, Lag 8, Hit Rate, Lag 32, Lag Effect, False Alarm Rate, First Half, False Alarm Rate, Second Half, Answer Time, Hits, Answer Time, Correct Rejections, Anticipations and Pictures Shown. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Repeat Detection Task used in research?
54 papers from 2021 to 2025 name it in their title or abstract, about the same as from 2016 to 2020. That ranks it 246 of 395 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.14.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Repeat Detection Task have?
In the published research, 1 paper reports its test-retest reliability and 2 registered clinical studies use it as an outcome (evidence grade D in the ConductCognition test census).
Which settings can a researcher change?
Repeats at each lag (12 to 16). 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.