Working Memory
Pitch Memory Task
Also called: tonal memory task, tone memory task, pitch short-term memory, Deutsch tone interference paradigm, delayed pitch comparison, auditory memory interference
Measure memory for the pitch of a tone over 5 seconds, in silence or with six other tones in between, and the cost of those tones (Deutsch, 1970).
8-10 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
- 25
- Evidence grade
- C
- Registered clinical studies
- 0
What it measures
Measures memory for the pitch of a tone over 5 seconds, in silence or while other tones play (Deutsch, 1970). Each trial plays a first tone and, 5 seconds later, a last tone through headphones; participants say whether the last tone had the same pitch or a different one. On half the trials six other tones play in between. The score is the share right with tones in between, with the cost of those tones against silence. Needs headphones and normal hearing; results are research measures.
Constructs
- Auditory short-term memory
- Pitch memory
- Interference in memory
Research uses
- Auditory working memory research
- Music cognition and congenital amusia research
- Within-study group comparison
How the paradigm works
The Pitch Memory Task measures memory for the pitch of a single tone, and how much other tones heard in between get in the way. Deutsch (1970) found that six tones played between two test tones raised errors in a same-or-different judgment from 2.4 percent to 32.3 percent, while six spoken numbers that had to be remembered barely did. Interference from tones has been central to research on pitch memory since, including in congenital amusia (Williamson and colleagues, 2010; Jiang and colleagues, 2013).
More about the paradigm
Each trial plays a first tone of 200 ms and, 5 seconds later, a last tone of 200 ms through headphones: the same pitch, or a semitone higher or lower (2 semitones as a setting). On half the trials the 5 seconds are silent. On the other half, six other tones of 200 ms play 300 ms apart, starting 300 ms after the first tone, then a 2 second pause comes before the last tone, as in Deutsch (1970). The participant presses Play, hears the trial, then answers Same or Different, with no time limit. The tones lie on a grid of semitones from 300 Hz: the first tone is one of 12 pitches from 300 to 566 Hz, and the six intervening tones come from the same 12, never the first or last tone's pitch and never repeated within a trial. There are 24 trials of each kind (standard; 12 to 40), half ending on the same pitch, mixed in one seeded order after 4 practice trials with feedback, for 48 trials. The tones play at the level the participant set as comfortable in a loudness step, after a headphone check (Woods, Siegel, Traer and McDermott, 2017). Desktop, laptop, tablet or phone, with headphones; about 8 to 10 minutes.
The main score is accuracy with tones in between: the share of same-or-different answers that were right when other tones played. Accuracy with a silent interval is the control, and the interference cost is the difference between the two. d prime and the response criterion are given for each condition. The test needs normal hearing and headphones; scores support research and do not diagnose.
Scores
- Accuracy with Tones in Between%
- Share of usable answers that were right when six other tones played between the first and the last tone.
- Higher is better; guessing gives 0.5
- Accuracy in Silence%
- Share of usable answers that were right when the 5 seconds between the tones were silent.
- Higher is better; under 0.75 is flagged for review
- Interference Cost%
- Accuracy in silence minus accuracy with tones in between.
- Higher means the intervening tones cost more
- Accuracy%
- Share of usable answers that were right, over both conditions.
- Higher is better; guessing gives 0.5
- d-Prime with Tones in Betweend'
- Sensitivity to a change of pitch when tones played in between, from the hit and false alarm rates, log-linear corrected.
- Higher is better
- d-Prime in Silenced'
- Sensitivity to a change of pitch with a silent interval, from the hit and false alarm rates, log-linear corrected.
- Higher is better
- Response Criterion with Tones in Betweenz
- Minus half the sum of the two z values behind d prime. Positive values mean a leaning to answer Same.
- Informational
- Response Criterion in Silencez
- Minus half the sum of the two z values behind d prime. Positive values mean a leaning to answer Same.
- Informational
- Hit Rate with Tones in Between%
- Share of changed-pitch trials answered Different when tones played in between.
- Higher is better
- False Alarm Rate with Tones in Between%
- Share of same-pitch trials answered Different when tones played in between.
- Lower is better
- Hit Rate in Silence%
- Share of changed-pitch trials answered Different with a silent interval.
- Higher is better
- False Alarm Rate in Silence%
- Share of same-pitch trials answered Different with a silent interval.
- Lower is better
- Median Answer Timems
- Median time from the answer buttons turning on, once the last tone has played, to the answer.
- Informational; the answer has no time limit
- Anticipationscount
- Answers within 100 ms of the answer buttons turning on, which leave the score.
- At least half the answers that fast is invalid
- Completed Trialstrials
- Trials answered in the scored run.
- Fewer than planned is flagged for review
- Planned Trialstrials
- Trials the run set out to play.
- Informational
- Headphone Check Passed0/1
- 1 when the headphone check before the test passed: at least 5 of its 6 answers right.
- 0 is flagged; a failed check that stopped the test leaves no scores
- Sound Output Delayms
- The output delay the device reported for its sound, saved once per session; empty when the browser does not report it.
- Informational; answers are untimed
Use and evidence
25 papers from 2021 to 2025 name this task in their title or abstract, down 19% compared with 2016 to 2020. That ranks it 316 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 6 |
| 2024 | 3 |
| 2023 | 5 |
| 2022 | 3 |
| 2021 | 8 |
| 2020 | 3 |
| 2019 | 6 |
| 2018 | 5 |
| 2017 | 11 |
| 2016 | 6 |
| 2015 | 10 |
| 2014 | 10 |
| 2013 | 11 |
| 2012 | 13 |
| 2011 | 9 |
| 2010 | 5 |
| 2009 | 5 |
| 2008 | 9 |
| 2007 | 6 |
| 2006 | 8 |
| 2005 | 2 |
| 2004 | 5 |
| 2003 | 13 |
| 2002 | 3 |
| 2001 | 3 |
| 2000 | 5 |
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
- 23
- 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.
- Trial
- A first tone of 200 ms, then 5,000 ms, then a last tone of 200 ms: the same pitch, or 1 semitone higher or lower (standard; 2 semitones as a setting).
- In between
- On half the trials the 5,000 ms are silent. On the other half, six tones of 200 ms play 300 ms apart, starting 300 ms after the first tone, and a 2,000 ms pause comes before the last tone, as in Deutsch (1970).
- Tones
- Pure tones with 10 ms raised-cosine ramps on a grid of semitones from 300 Hz. The first tone is one of 12 pitches from 300 to 566 Hz, each used equally often. The six intervening tones come from the same 12, never the first or the last tone's pitch, and none repeats within a trial. All play at the level the participant set as comfortable in the loudness step.
- Answer
- Press Play, hear the trial, then choose Same or Different, with no time limit. The answer buttons turn on once the last tone has played.
- Trials
- 24 with each kind of interval (standard; 12 to 40): 48 trials mixed in one seeded order, half of each kind ending on the same pitch, a quarter higher and a quarter lower.
- Practice
- 4 trials with feedback for 1,500 ms: two silent, then two with tones in between, one same and one different in each pair.
- Headphones
- A loudness step and a six-trial headphone check come first. A failed check ends the test, or, as a setting, lets it run. The test needs normal hearing and headphones.
- Score
- The share right with tones in between, the share right in silence and the cost of the tones (silence minus tones), with d prime and the response criterion for each. Answers under 100 ms from the buttons turning on leave the score.
Pitch memory studies differ in the interval, the number and pitch of the intervening tones and the size of the change. This version keeps the timing of Deutsch (1970) with a grid of tones of its own, so its accuracies are compared within this version. 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: pitch_memory_test from the scored run, Same or Different after the last tone, with no time limit (chance 50%; half the trials in each condition end on the same pitch), graded from the button pressed and the saved change. Below chance invalidates and not above chance is flagged; one answer on 90% of trials invalidates when accuracy is not above chance. Answers under 100 ms from the buttons turning on leave the score, and at least half the answers that fast invalidate. A failed headphone check that stopped the test invalidates; one the study ran past, a run short of its planned trials and under 75% right with a silent interval are flagged for review.
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). The answer buttons turn on once the last tone has been heard, so an answer under 100 ms came before the buttons could be seen; only a session where at least half the answers are that fast is invalid. Accuracy with a silent interval is the control: when it is low, the tones were hard to tell apart and the cost of the intervening tones cannot be read. A hidden tab lets the 5 second interval run without the participant, so a percentile is withheld after one. The test needs headphones, so a failed headphone check stops it (Woods, Siegel, Traer and McDermott, 2017).
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Trials per condition | 24 | 12 to 40 |
| Change on different trials | 1 semitone (standard) | 1 semitone (standard), 2 semitones |
| When the headphone check fails | Stop the test | Stop the test, Continue the test |
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. Needs headphones. 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 Pitch Memory Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Deutsch, 1970), with the standard settings (trials per condition: 24; change on different trials: 1 semitone; when the headphone check fails: Stop the test). The main score was accuracy with tones in between (share of usable answers that were right when six other tones played between the first and the last tone).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Deutsch, D. (1970). Tones and numbers: Specificity of interference in immediate memory. Science, 168(3939), 1604-1605. https://doi.org/10.1126/science.168.3939.1604The timing, six intervening tones and the semitone change; errors rose from 2.4 to 32.3 percent with tones in between.
- Deutsch, D. (1972). Mapping of interactions in the pitch memory store. Science, 175(4025), 1020-1022. https://doi.org/10.1126/science.175.4025.1020Interference depends on how close an intervening tone is to the first tone.
- Williamson, V. J., McDonald, C., Deutsch, D., Griffiths, T. D., & Stewart, L. (2010). Faster decline of pitch memory over time in congenital amusia. Advances in Cognitive Psychology, 6, 15-22. https://doi.org/10.2478/v10053-008-0073-5The same timing with 0 to 8 tones in a fixed 5 second interval, no tone repeated, and a whole-tone change.
- Jiang, C., Lim, V. K., Wang, H., & Hamm, J. P. (2013). Difficulties with pitch discrimination influences pitch memory performance: Evidence from congenital amusia. PLoS ONE, 8(10), e79216. https://doi.org/10.1371/journal.pone.0079216Silent and filled intervals mixed in one order, scored with d prime in each.
- Woods, K. J. P., Siegel, M. H., Traer, J., & McDermott, J. H. (2017). Headphone screening to facilitate web-based auditory experiments. Attention, Perception, & Psychophysics, 79(7), 2064-2072. https://doi.org/10.3758/s13414-017-1361-2The headphone check before the test.
Source papers
- Deutsch D (1970). Tones and numbers: specificity of interference in immediate memory. Science, 168(3939):1604-1605.N 24, Adults
Questions
How long does the Pitch Memory Task take?
About 8 to 10 minutes, as listed in the test library.
What device does a participant need?
Untimed test: runs on desktops, laptops, tablets and phones. Needs headphones. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Accuracy with Tones in Between, Accuracy in Silence, Interference Cost, Accuracy, d-Prime with Tones in Between, d-Prime in Silence, Response Criterion with Tones in Between, Response Criterion in Silence, Hit Rate with Tones in Between, False Alarm Rate with Tones in Between, Hit Rate in Silence, False Alarm Rate in Silence, Median Answer Time, Anticipations, Completed Trials, Planned Trials, Headphone Check Passed and Sound Output Delay. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Pitch Memory Task used in research?
25 papers from 2021 to 2025 name it in their title or abstract, down 19% compared with 2016 to 2020. That ranks it 316 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 Pitch Memory 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?
Trials per condition (12 to 40), change on different trials (1 semitone (standard), 2 semitones) and when the headphone check fails (Stop the test, Continue the test). 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.