A sequence of evenly spaced sounds can be heard as ONE-two, ONE-two or as ONE-two-three, ONE-two-three. The physical intervals have not changed, yet the listener has organized them into different repeating groups. If an experiment asks whether that listener perceives meter, a correct response to the timing of individual sounds will not suffice. Measuring meter perception means testing the organization of a pulse into stronger and weaker positions, while accounting for the acoustic cues, movement and prior experience that can shape that organization.
Meter is a hierarchy of recurring temporal expectations, commonly expressed as patterns such as duple or triple grouping. It is related to beat perception but not equivalent to finding a regular pulse. It is also distinct from tempo, the rate of events or beats, and from reproducing an entire rhythmic pattern. The Nature Reviews Psychology synthesis of musical rhythm, beat and metre provides an overview of these distinctions. This Methods & Measures entry asks how a study can make the metrical claim observable and where a task stops short of it.
State which level of temporal structure is tested
A listener can detect an isochronous beat without assigning strong positions to every second, third or fourth beat. Conversely, a listener may identify a duple feel from accents even when their tapping is imprecise. A meter task should specify the proposed grouping level, the event rate, the beat rate and any slower grouping rate. Otherwise an apparent difference between conditions may reflect tempo discrimination or memory for surface accents rather than metrical organization.
The stimulus needs an explicit acoustic description. List onset times, durations, pitches, loudness and timbres, not just a time signature label. A printed “three-four” notation can prompt an answer before any sound is heard. If the purpose is perception, do not give notation that reveals the intended category unless reading notation is itself the construct. A pattern can offer several plausible pulse levels; the report should say whether ambiguity was intentional.
Top-down attention also matters. A listener may focus on a bass tone, a high tone or a repeated phrase boundary. A PLOS ONE experiment using biphasic sounds found that attention to different acoustic features could shift reported meter. That result warns against treating the acoustic sequence as a single objective metrical answer when the listener can select different cues. In a trial, collect what the person heard, not only what the composer meant.
Choose a task according to the claim. A forced choice between duple and triple groupings asks for categorization. A continuation tap asks whether a participant can maintain a chosen temporal level after the sound stops. A probe-tone or deviant-detection task can ask whether an event at a strong position is processed differently from one at a weak position. These tasks may converge, but they include different demands and should not be silently merged into one universal meter score.
Separate perception from overt production
Tapping is useful because it gives a time-stamped behavior. Yet a tap can align to the beat while missing the metrical grouping. Recording only tap onsets may therefore underestimate or misdescribe meter. Tapping force offers another observable dimension: work on metrical aspects of rhythm found stronger or differentiated force profiles at positions within a hierarchy. This is evidence that movement can express meter under those task conditions, not proof that every listener’s fingers faithfully report every heard accent.
Ask participants what they are tapping to. One person may tap every pulse; another may tap every two pulses. Both could be stable, but the resulting rate means different things. If the task permits free tapping, estimate the chosen level and report its distribution. If participants are instructed to tap on strong beats, the instruction makes the target clear but may change the percept. A design should choose between observing spontaneous organization and testing compliance with a specified metrical level.
Movement is not merely output. Phillips-Silver and Trainor showed that bouncing infants on every second versus every third beat of an ambiguous rhythm influenced later rhythm perception. The specific infant experiment does not show that all adults require movement to hear meter. It does show that movement can alter the percept being measured. If one group taps during listening and another sits still, the groups differ in sensory and motor input as well as their ability to express an answer.
A tactile meter experiment likewise demonstrated that patterned touch and sound can be integrated in meter judgments. This widens the measurement question. A musician may use head movement, foot tapping or postural sway without an instruction to do so. Video or motion tracking can document such behavior when it matters, while privacy and participant comfort must be respected. The absence of recorded movement should not be mistaken for the absence of movement.
For a perception-focused result, pair production with a judgment that does not require precise motor timing. For a motor-focused result, retain the tap and describe its force and phase. When tasks disagree, resist choosing the result that best fits the story. A participant may hear a grouping but be unable to produce stable taps, or tap a pattern by imitation without identifying the grouping when asked. That difference is substantive.
Design acoustic controls for the intended inference
Accents can be created through loudness, duration, pitch, harmony or silence. These are not perceptually interchangeable. If a metrical preference appears only when every strong position is louder, a listener may be tracking intensity rather than inferring a hierarchy beyond the signal. A more probing test uses a rhythm whose physical envelope does not fully specify the metrical level, or asks whether learned grouping changes the response to an identical acoustic event.
Nozaradan and colleagues used EEG frequency tagging to examine responses at beat- and meter-related frequencies. Their work provides one route to studying internal periodic organization, particularly where a metrical frequency is not prominent in the stimulus spectrum. Still, a neural peak alone does not read out a complete musical interpretation. Acoustic regularities, attention and analysis choices can produce periodic brain responses. Relate the EEG frequencies to both the sound envelope and the participant’s behavior.
Varying the acoustic cue is a useful test of generality. If a participant classifies duple grouping from loudness accents, can they do so when pitch or duration carries the pattern? Performance across cue types would support a broader skill, but cue changes also alter difficulty and familiarity. Pretest the stimuli and avoid interpreting one hard condition as a pure deficit in meter perception.
Another control compares metrically congruent and incongruent events while matching local physical properties. A deviant at a strong position may be more surprising, but only if the listener established the hierarchy and if local transition probabilities are controlled. If strong positions also always contain more salient sounds, a larger response to a deviant may reflect ordinary sensory adaptation. The design should explain why its contrast isolates position within an inferred pattern.
Meter can operate over several cycles. One brief sound pair may be insufficient for a stable grouping, whereas a long repeating pattern allows adaptation and strategy. Report exposure length, number of cycles and how early and late responses are analyzed. A listener may switch interpretation halfway through. Averaging across the whole trial can hide that change and make an ambiguous experience look like a weak, stable preference.
Consider expertise and cultural familiarity
Musical training changes the repertoire of rhythms and cues a participant can use. A trained drummer may detect a syncopated grouping that a novice experiences as irregular. This does not mean either response is invalid. It means the stimulus and the sample jointly determine what the test measures. Record relevant training and experience rather than collapsing everyone into a single normative ranking.
Familiarity is broader than formal lessons. A participant’s music environment can privilege different metrical patterns and movement practices. A task using only Western instrumental samples may test familiarity with that material as much as a general human capacity. Translate instructions carefully, allow practice trials and assess whether participants understand the response options. If the study claims cross-cultural generality, recruit and analyze across those contexts rather than importing a laboratory norm.
Age and development affect the feasible response. An infant cannot explain duple versus triple grouping; the Phillips-Silver and Trainor design used later preference as a bounded indicator of encoding. A school-age child may understand a simple tap game but not technical meter labels. Adults may adopt verbal counting strategies that change performance. The test should fit the population while acknowledging that different age-appropriate tasks are not numerically interchangeable.
Hearing ability and motor ability can also determine test performance. A participant who cannot reliably detect a pitch accent is not necessarily unable to group time. A participant with motor difficulty may give an inaccurate tap despite a clear percept. Match sensory cues to accessibility, offer a non-motor response where appropriate, and avoid clinical inferences from a single musical task.
Analyze the answer rather than the task label
For forced-choice judgments, report accuracy relative to a defensible target, response bias and ambiguous-trial choices. When the stimulus legitimately supports two interpretations, the target may be a distribution of judgments rather than one correct category. For taps, report phase relation to the candidate beat and grouping level, tempo drift and variability. A participant who taps exactly every third pulse expresses a different pattern from one who taps every pulse with stronger force every third tap.
For force data, normalize cautiously. Differences in hand strength, fatigue and device sensitivity can create apparent group effects. Inspect whether force modulation follows the intended metrical hierarchy and whether it predicts independent judgment under the same stimulus. The 2021 tapping-force study links force patterns with perceptual precision in its design; it does not validate force as a stand-alone meter diagnosis for every population.
For EEG frequency tagging, compare neural spectral peaks with the stimulus spectrum. A periodic sound can drive periodic neural activity without the listener inferring a metrical hierarchy. A response at a metrical subdivision or grouping frequency is more informative when the acoustic component is controlled and when attention or reported grouping changes the neural pattern. Specify frequency resolution, epoch length, artifact handling and electrode aggregation before interpreting a peak.
Assess reliability where the intended use requires it. A group effect may be replicable while an individual’s score fluctuates across days, materials or task instructions. Repeatability across stimulus sets matters if the claim concerns a stable capacity. A training study should compare changes with practice effects from repeating the same rhythms. If a participant improves at one set of sounds, that is not yet proof of general transfer to music understanding.
Keep the unit of analysis clear. Multiple trials from one person do not create multiple independent people. Multiple pieces of music may share a composer, groove or tempo. Statistical models should represent such nesting when possible. A large collection of repeated taps can yield tiny uncertainty around a participant’s timing under one condition while saying little about their response to unfamiliar meters.
What the evidence does not show
The cited studies do not establish one gold-standard meter test. A correct categorization, an accented tap and an EEG frequency peak are evidence at different levels. None by itself proves that a listener experiences a particular time signature in the way a musician describes it. Meter can be ambiguous, and tasks can make one interpretation more likely by directing attention or movement.
Neither do these studies show that weak performance means a broad inability to enjoy music, learn rhythm or coordinate socially. Task comprehension, acoustic salience, cultural familiarity, hearing, movement and practice all affect scores. A child who does not provide an adult-like verbal label may still organize sound in a way the chosen response cannot capture.
The neural frequency-tagging results support experimental contrasts, not a simple mind-reading instrument. A spectral response at a grouping rate can have sensory and analytical explanations unless the design rules them out. Movement-induced changes in infant listening do not establish that every metrical percept is motor-generated. The defensible conclusion is tied to the tested stimulus and comparison.
Describe the pattern a reader could reconstruct
A strong report gives the complete sound sequence, intended candidate meters, acoustic cue structure, presentation duration, participant instructions, response method and preprocessing. It shows which level the person tapped or selected, how ambiguous answers were handled and whether independent measures agreed. Where interpretation depends on movement or neural data, it presents those traces alongside the acoustic structure instead of replacing them with a single score.
The opening sequence can be heard in more than one grouping. A good meter measure does not conceal that possibility. It uses the ambiguity to ask which cues and actions shape the listener’s organization, and it stops at the level its evidence can support.
Research references
- Snyder, J. S., Gordon, R. L., & Hannon, E. E. (2024). Theoretical and empirical advances in understanding musical rhythm, beat and metre. Nature Reviews Psychology, 3, 449–462. View source
- Phillips-Silver, J., & Trainor, L. J. (2005). Feeling the beat: Movement influences infant rhythm perception. Science, 308, 1430. View source
- Nozaradan, S., Peretz, I., Missal, M., & Mouraux, A. (2011). Tagging the neuronal entrainment to beat and meter. Journal of Neuroscience, 31, 10234–10240. View source
- Nozaradan, S., Peretz, I., & Mouraux, A. (2012). Selective neuronal entrainment to the beat and meter embedded in a musical rhythm. Journal of Neuroscience, 32, 17572–17581. View source
- Matsuo, H., & Sakaguchi, Y. (2024). Effects of rhythm and accent patterns on tempo-keeping property of finger tapping. Music & Science. View source
- Kondoh, S., Okanoya, K., & Tachibana, R. O. (2021). Switching perception of musical meters by listening to different acoustic cues of biphasic sound stimulus. PLOS ONE, 16, e0256712. View source
- Benedetto, A., & Baud-Bovy, G. (2021). Tapping force encodes metrical aspects of rhythm. Frontiers in Human Neuroscience, 15, 633956. View source