A listener hears a sequence of taps, then finds herself anticipating a tap that is never played. She can move her finger at the expected moment even though the sound itself has stopped. Beat perception is the inference of a recurring temporal pulse from an unfolding rhythm. It is more specific than hearing that sounds repeat, and it is not identical to producing a synchronized movement. A beat provides a reference for expecting when an event could occur; music may confirm, delay, or omit that event. The distinction matters because a response to a regular acoustic envelope does not by itself establish that a listener perceived an internally organized beat.
What makes a pulse perceptual?
In a simple metronome, each sound marks the pulse. In music, the pulse may be carried by only some notes, while other notes fall between them. Listeners may tap regularly through rests and syncopations because they infer a continuing temporal reference. This ability gives beat perception its explanatory role: it links incoming events to an expectation that can survive a missing event. Yet an even series of sounds can also drive a sensory response without the richer inference. A strong experiment separates the periodicity physically present in the stimulus from the pulse a listener reports or uses.
Grahn and Brett compared rhythms that induced a regular beat with control rhythms lacking an easy beat while people listened during functional imaging. Motor-related areas, including basal ganglia and supplementary motor regions, responded differently even without an overt tapping task. The result challenged a view that motor areas matter only after someone chooses to move. It does not prove that the basal ganglia alone create beat experience. Acoustic structure, expectation, and task demand may all influence activity, and imaging is correlational. The contribution is a carefully controlled association between beat-supporting rhythms and a network normally associated with movement.
Timing can be represented in more than one way. A person can estimate the duration of an isolated interval, or use a repeated pulse to anticipate the next moment. Beat-based timing is especially useful when intervals form a stable recurring pattern. Duration-based judgments may help when events are irregular. These approaches can interact; they should not be forced into a single internal clock. A listener who taps to a rhythm might use both a continuing beat and local adjustments to each sound. To infer the process, a study must manipulate regularity, missing events, tempo changes, and the timing of the response.
Beat perception differs from meter perception. A beat is a sequence of expected pulse points; meter groups those pulse points into patterns of relative strength, such as two or three. A listener can feel a beat without knowing which pulse begins a group. Conversely, an instruction to imagine groups can alter how identical events are organized. The companion meter article examines that hierarchy. Keeping the constructs separate avoids calling every response to a periodic stimulus “meter” and avoids assuming that a tapping frequency captures the listener’s interpretation of strong and weak beats.
Prediction, movement, and listening
Why should a motor network participate while the listener sits still? One proposal is that motor planning systems simulate possible movement times and thereby help predict auditory events. Grahn and Rowe compared musicians and nonmusicians during beat perception and reported interactions between premotor and striatal regions. The finding is consistent with a sensorimotor contribution. It is not a demonstration that an imagined limb movement is required for every beat percept. A neural network can have more than one function, and musical training changes experience as well as neural response.
Prediction makes the beat useful in real time. If an expected sound is delayed, the listener can detect its lateness relative to an internal reference. If a pulse is omitted, a person may continue to tap and then resynchronize when the pattern returns. This flexibility suggests that beat perception is not merely a passive imprint of sound energy. However, an expectation can be generated by learned statistical regularity without a single dedicated clock. Researchers should ask whether the participant’s behavior supports a persistent pulse across silence, rather than assuming persistence from a steady stimulus alone.
Movement can sharpen or alter beat experience. A small head movement or foot tap can provide proprioceptive feedback that stabilizes a pulse. Yet a listener can also perceive a beat without moving, and overt movement changes the task. Comparing still listening with movement conditions can help identify whether motor activity precedes or follows a perceptual organization. A stronger design records subtle movement and tests performance, because instructions to remain still do not guarantee absolute immobility. The purpose is not to remove the body from music, but to distinguish perception from a measured motor response.
Parkinson’s disease has been used to probe the basal ganglia contribution. Grahn’s review discusses imaging and patient evidence suggesting that beat processing can be affected by disruptions to basal-ganglia circuits. These studies matter because a patient contrast can add causal information to imaging associations. Still, diagnosis, medication, motor difficulty, and general cognitive changes can influence performance. A deficit on one rhythm task cannot be read as the loss of all musical timing. The strongest inference comes from matched tasks that separate perceiving a beat from executing a tap.
Musical expertise adds another layer. A musician may have practiced identifying and coordinating with complex pulses, while a nonmusician may rely on familiar simple patterns. Grahn and Rowe found differences in network interactions associated with training, but selection also matters: people who enjoy or perceive rhythms well may be more likely to become musicians. A cross-sectional difference is not proof that training caused the neural pattern. For education, it invites an experiment in which comparable groups practice specified skills and are tested on untrained rhythms.
What neural frequency measures reveal
Nozaradan and colleagues used electroencephalography to measure periodic responses while people listened to rhythms and were asked to imagine different metrical groupings. Peaks in the recorded signal at beat- and meter-related frequencies suggested that brain activity reflected more than a simple sequence of sound onsets. Such frequency tagging offers a way to compare physical and perceived periodicity. Yet a peak is not a transparent reading of subjective experience. Repeated sensory responses can themselves create spectral peaks, and analysis choices affect their size.
A subsequent multi-laboratory registered-replication preprint of a related 2011 frequency-tagging effect found much smaller and less reliable effects than the original report across its individual samples, while some pooled exploratory analyses found effects. That result does not erase all EEG evidence for beat processing. It warns against treating one large frequency peak as a settled neural signature of an imagined pulse. The interpretation needs converging behavior, stimulus controls, and attention to replication. A good article can preserve the original insight while reporting that the effect size and generality remain under examination.
Another strategy tests a rhythm whose acoustic energy does not contain a strong pulse-frequency component, then asks whether neural activity nonetheless tracks the perceived pulse. Such “missing-pulse” designs target a specific alternative explanation: that the brain signal merely follows a physical modulation already present. When a pulse-related response appears under those conditions, it strengthens an inference about endogenous organization. It still leaves open which neural operation produced the signal and whether every listener experienced the same pulse. Asking participants to report or reproduce the beat remains important.
Brain recordings and tapping serve different questions. An EEG peak can reveal periodic neural activity, while tapping demonstrates a chosen movement aligned to a temporal interpretation. A person may perceive a beat but avoid tapping because of motor difficulty or instruction; another may imitate repeated sound timing without a stable beat. Combining measures is useful only when their differences are respected. The measurement companion article treats task design and confounds in detail. Here, the mechanism claim is that an expected pulse can organize perception and action, with multiple candidate neural routes.
Where the beat comes from
The sound provides constraints. Onset timing, accent, pitch, timbre, and harmonic change can all suggest a pulse. A repeated drum pattern may invite one rate, while a slow melody invites a subdivision or a broader cycle. A listener brings learned conventions and current attention to that evidence. This does not imply that any interpretation is equally likely. Some rhythms strongly favor one pulse; others permit several plausible levels. Experiments should state which cues are present and whether they are controlled across conditions.
Tempo influences what is comfortable to maintain. Extremely slow events can be hard to connect without subdivision; extremely fast events can be grouped into a slower pulse. The preferred level may change with context, movement, and training. A laboratory task using one convenient tempo may not generalize to all music. Researchers can vary tempo continuously and ask where a listener switches levels. Such switches reveal the interaction between acoustic timing and perceptual organization more clearly than a single accuracy score.
Accents complicate interpretation. A sound can be louder because the instrument actually emphasizes it, or it can seem stronger because it falls on an expected pulse. The latter is sometimes called a subjective accent. To distinguish them, researchers can hold physical sound levels constant while varying the induced pattern or the listener’s instruction. That contrast is central to meter studies, but it also shows how a beat can shape hearing. A participant’s impression of emphasis is not always a direct measurement of acoustic amplitude.
Culture and experience matter, though not as a simple division between people who have and lack rhythm. Exposure to a musical tradition changes which groupings feel familiar and which syncopations are easy to follow. A single Western drum pattern is not a neutral test of a universal beat faculty. Comparative studies can use multiple musical systems, measure familiarity, and separate detection of periodicity from knowledge of a style. The general capacity to coordinate with temporal regularity may coexist with culturally learned preferences about where the pulse is placed.
Testing an explanation rather than a metaphor
“The brain entrains to the beat” can mean several things. It may describe a neural oscillation adjusting its phase to sound, repeated evoked responses that form a rhythmic waveform, a motor prediction that aligns action, or a behavioral tendency to tap at regular intervals. These possibilities make different predictions when sounds are omitted, shifted, or acoustically matched. A useful mechanism study states which one it tests. Frequency matching alone is insufficient to show that an endogenous oscillator caused the percept.
A causal account should also explain error correction. When a pattern accelerates gradually, a listener may shift the predicted pulse; when one note arrives early, the listener may treat it as an expressive deviation and retain the old pulse. The difference depends on evidence across several events. A model that updates after every deviation could lose the beat too easily, while one that never updates would fail to follow tempo change. Trial-by-trial tapping and perception reports can test how people balance stability and adjustment.
Individual variability deserves more than a label of “good” or “bad” rhythm. Hearing status, attention, familiarity, motor ability, and the task’s chosen tempo can affect a score. Someone who cannot tap steadily might still identify a missing beat; someone who taps well might be relying on the visible metronome. A study of beat perception should include tasks that do not require movement, then compare them with synchronization tasks. This separates a perceptual question from the additional demands of action.
The practical appeal of rhythm interventions should not outrun evidence. A beat can support coordinated group action and make a timed task easier, but that does not prove that any rhythmic training improves language, cognition, or health. An Institute synthesis is to design activities with clear timing cues and to test the intended outcome in the intended population. Transfer beyond practiced rhythms needs its own comparison and follow-up. The laboratory literature establishes mechanisms worth testing; it does not certify a universal educational product.
What the evidence does not show
Current evidence does not show that a single basal-ganglia structure acts as a universal metronome, that every EEG frequency peak is an internal beat, or that a listener must move to perceive pulse. Imaging, patient studies, frequency tagging, and behavior each constrain an explanation while leaving alternatives. The 2025 multi-laboratory replication preprint particularly cautions against treating the original EEG effect as a fixed quantitative signature. The strongest claims name the rhythm, task, participant group, and control condition.
Beat perception is useful because it explains anticipation across sound and silence. A person can hear a pattern as organized in time and prepare for an event that may not occur. That simple experience connects auditory evidence, prediction, and possible movement. Its limits are just as instructive: a beat is not a proof of global timing ability, a moral measure of musicality, or a guarantee of training transfer. Research advances when it asks exactly what pulse the listener inferred and which observation would distinguish that inference from a response to the sound alone.
Research references
- Grahn, J. A., & Brett, M. (2007). Rhythm and beat perception in motor areas of the brain. Journal of Cognitive Neuroscience, 19(5), 893–906. View source
- Grahn, J. A., & Rowe, J. B. (2009). Feeling the beat: Premotor and striatal interactions in musicians and nonmusicians during beat perception. Journal of Neuroscience, 29(23), 7540–7548. View source
- Grahn, J. A. (2009). The role of the basal ganglia in beat perception: Neuroimaging and neuropsychological investigations. Annals of the New York Academy of Sciences, 1169, 35–45. View source
- Nozaradan, S., Peretz, I., Missal, M., & Mouraux, A. (2011). Tagging the neuronal entrainment to beat and meter. Journal of Neuroscience, 31(28), 10234–10240. View source
- Nave, K. M., Hannon, E. E., & Snyder, J. S., with the Replication of Auditory Frequency Tagging Consortium. (2025). Registered report: Replication and extension of Nozaradan, Peretz, Missal and Mouraux (2011) [bioRxiv preprint; not peer reviewed]. View source
- Tal, I., et al. (2017). Neural entrainment to the beat: The “missing-pulse” phenomenon. Journal of Neuroscience, 37(26), 6331–6341. View source