Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Music Aesthetics: Rhythm, Expectation, and Emotional Reward

Abstract. Musical aesthetic experience arises through the interaction of organized sound, auditory prediction, emotional appraisal, reward valuation, embodied action, and personal and cultural learning. This fundamental research monograph examines rhythm and meter, the development of musical expectations, syncopation and groove, probabilistic accounts of harmonic surprise, and the neural mechanisms implicated in musical pleasure. A critical narrative synthesis integrates foundational theoretical work, original experiments, computational modeling, electroencephalography, functional neuroimaging, positron emission tomography, causal brain stimulation, and comparative and developmental research. The evidence demonstrates structured relations between temporal or harmonic expectations and pleasure, identifies auditory–striatal interactions and reward-related signals during music listening, and documents substantial variability across people, cultures, emotions, and social settings. Methodological analysis distinguishes statistical surprise from reward prediction error, perceived musical expression from induced emotion, and correlational brain activity from causal interventions. Particular attention is given to specific musical anhedonia, pleasurable sadness, the role of movement and synchrony, problems of measurement validity, and limitations of Western-centered stimulus sampling. The monograph concludes with a multilevel synthesis and independently testable research priorities relevant to human music experience and interactive artificial music systems. This work is a critical narrative review of published literature, not a systematic review, meta-analysis, or original empirical study.

Keywords: music aesthetics; music psychology; rhythm perception; beat induction; meter; syncopation; groove; music-evoked emotion; musical pleasure; musical reward; musical expectation; harmonic surprise; predictive coding; statistical learning; reward prediction error; dopamine; nucleus accumbens; musical anhedonia; music-induced chills; music and movement; music and social bonding; aesthetic judgment; cross-cultural music perception; neuroaesthetics; Artificial Era

Music Aesthetics as an Empirical Research Object

Music aesthetics studies how organized sonic events become objects of attention, evaluation, emotion, and meaning. Its central phenomena include musical beauty, interest, enjoyment, tension, being moved, the urge to move, expressive recognition, and reflective judgments of artistic value. These outcomes must be separated analytically: a listener can recognize sadness without becoming sad, admire a demanding work without immediate pleasure, enjoy a familiar rhythm without judging it artistically exceptional, or find a surprising chord interesting while disliking it. The research object of this dossier is the relationship among rhythmic organization, learned and immediate expectations, affective responses, and reward valuation. This relationship is investigated across perception, action, memory, culture, and neurobiology. Zentner et al. (2008) demonstrates why music-specific affective categories deserve measurement beyond a generic positive–negative scale, while Brattico et al. (2026) locates reward within a wider situated aesthetic encounter.

Rhythm denotes the temporal organization of musical events; beat denotes a perceived regular pulse; meter organizes pulses hierarchically; syncopation gives salience to events departing from an anticipated metric emphasis. Expectation is a graded prediction concerning forthcoming musical events, informed by acoustic regularities, exposure, stylistic learning, and current context. Emotional reward is the experienced and motivational value of musical engagement, investigated through pleasure ratings, willingness to listen or pay, autonomic responses, and activity within valuation networks. None of these measurements alone constitutes the entire phenomenon. Studies such as Witek et al. (2014), Salimpoor et al. (2013), and Cheung et al. (2019) operationalize distinct parts of this joint problem, allowing their findings to be compared without treating their dependent variables as interchangeable.

This monograph is a critical narrative synthesis of identified original experiments, foundational theoretical works, comparative evidence, psychometric research, and recent reviews. The guiding questions are how time is represented, how temporal and harmonic predictions become emotionally salient, what connects prediction and pleasure, how learning and culture alter aesthetic response, and which experimental designs identify causal mechanisms. Claims about neural correlation are distinguished from causal stimulation experiments; formal predictive models are evaluated by their actual behavioral and neural predictions. The evidence base is selective and explicitly identified, without a preregistered search protocol or an independent meta-analysis. The synthesis examines human musical experience and addresses computational and artificial systems where they provide testable models of musical prediction rather than equating algorithmic performance with felt reward.

Historical Foundations of Expectation and Musical Feeling

The psychological relation between musical expectation and emotion became a major explanatory program in Leonard B. Meyer's Meyer (1956) account of how learned stylistic probabilities organize listening. Musical patterns create tendencies toward continuation, completion, and closure; interruption and postponement can generate affective tension because the listener anticipates an outcome. Meyer's argument joined Gestalt principles, musical analysis, learning, and experience. Its historical significance lies in treating emotional responses as related to perceptual organization and acquired knowledge. It did not supply the later neuroscientific causal evidence or a complete computational measure of probability, so modern tests must translate its concepts into independently measured expectancies and emotional outcomes.

David Huron's Huron (2006) account expanded the expectation tradition into distinguishable response systems. Prediction responses concern the consequences of anticipated events, tension responses concern the interval before resolution, reaction responses can be rapid, and reflective appraisal revises the significance of an outcome. This theoretical separation is useful because a sudden musical event can produce an orienting response, a delayed evaluative reaction, and later pleasure for different reasons. Aesthetic suspense can also arise from the distribution of possible continuations rather than from an explicit verbal forecast. Huron's framework has generated testable hypotheses, although no single experiment establishes every proposed component or their interaction.

The experimental turn connected conceptual claims to manipulations of musical grammar. Steinbeis et al. (2006) altered individual chords in otherwise matched Bach chorales and measured subjective, autonomic, and event-related neural responses. Increasing harmonic unexpectedness was associated with greater tension, emotionality, and electrodermal activity. This is evidence that changes in local musical syntax can influence affective responses under laboratory conditions. It also establishes an important distinction: heightened tension or surprise is not equivalent to increased liking. The historical trajectory from Meyer to brain measurement therefore requires a multidimensional account in which expectancy, emotional activation, and aesthetic valuation remain separate variables.

Auditory Structure, Tonal Organization, and Perceptual Learning

A musical event is heard in a structured field rather than in isolation. Onset time, duration, timbre, pitch contour, spectral roughness, harmony, and repetition interact with the listener's grouping of phrases and accents. The auditory system encodes regularities over different time spans: a brief acoustic interval supports local grouping, a phrase carries melodic context, and extended listening establishes formal expectations. Zatorre & Salimpoor (2013) reviews auditory–frontal processing supporting the retention and recognition of tonal patterns and its proposed interactions with valuation systems. This pathway suggests a bridge from sensory prediction to reward while allowing reward valuation to depend on context and prior experience.

Tonal and harmonic familiarity is learned from musical environments that differ across communities. Western tonal harmony supplies a particularly tractable experimental syntax because corpus statistics and chord transitions can be quantified. Yet properties estimated from commercial Western popular music are neither universal psychological priors nor culturally neutral estimates of surprise. A convincing model estimates what a particular listener has learned and distinguishes acoustic availability from subjective probability. McDermott et al. (2016) found that listeners from a relatively isolated Amazonian population did not show the Western preference for consonance over dissonance, despite relevant acoustic discrimination. This finding bears directly on aesthetic valuation: sensitivity to a stimulus distinction does not determine a universal preference for one category.

Individual learning histories also change with musical expertise. Skilled musicians may anticipate elaborate progressions and parse rhythm into hierarchies that novices do not readily hear. In Steinbeis et al. (2006), both musicians and nonmusicians showed neural and emotional sensitivity to harmonic unexpectedness, while timing and magnitude of some event-related responses differed. Such results resist a simple expert–nonexpert binary. Training modifies representational efficiency and knowledge, whereas liking may depend on identification with a style, situational goals, and the balance between familiarity and novelty. Experimental sampling should measure these dimensions separately rather than use years of formal training as a proxy for every aspect of musical experience.

Beat Induction, Meter, and Neural Temporal Prediction

Beat induction is the inference of regular pulse positions from an acoustic stream, including positions at which no actual sound occurs. Meter embeds this pulse in stronger and weaker relational groupings. These processes are especially useful in experiments with omitted expected events: the brain's response can be compared with the response to acoustically present onsets. Winkler et al. (2009) reported neural evidence that newborn infants developed expectations about rhythmic downbeats; omission of a downbeat elicited a response associated with deviation from an anticipated pattern. The result makes early auditory expectation salient without settling the relative contributions of prenatal exposure, neural predispositions, and experience after birth.

Nozaradan et al. (2011) used electroencephalography and frequency tagging to examine how the brain represents beat and meter. Participants listened to rhythmic sequences and imagined alternative metrical groupings; activity at corresponding periodicities supported an account of neural alignment with perceived temporal structure. These spectral signatures demonstrate phase-locked responses to rhythmic regularities, yet their interpretation requires care. Frequency-domain peaks can arise from stimulus structure, nonlinear evoked responses, attention, or true sustained oscillatory coupling. A rigorous analysis tests whether brain activity tracks an internally inferred pulse after controlling the periodicities already present in the acoustic envelope.

Beat perception contributes to movement prediction because a listener can prepare an action at a future temporal position. This establishes a pathway from auditory grouping to motor readiness, making temporal attention available to dancing, tapping, and collective performance. The pathway is not limited to overt movement: imagined or inhibited actions can still organize perception. Vuust & Witek (2014) situates rhythmic complexity within predictive coding, explaining how a strongly expected beat and a displaced accent can coexist. The aesthetically relevant question becomes whether such tension is interpretable and manageable within the current meter, rather than whether a rhythm is simply regular or irregular.

Syncopation, Groove, and the Pleasure of Movement

Groove is commonly studied as a felt desire to move with music and a related sense of rhythmic engagement. It is not identical to beat accuracy or generic positive affect. Syncopation introduces metric deviations against a pulse that remains perceptually available. The experimental logic therefore involves two simultaneous demands: preserve enough temporal organization to support prediction and introduce departures that elicit interest or anticipatory engagement. Witek et al. (2014) manipulated the syncopation of drum breaks in an online listening experiment and found that moderate degrees of syncopation produced higher ratings of pleasure and desire to move than very low or high degrees in their stimulus set. This inverted-U association supports a contextual rather than a monotonic effect of complexity.

An inverted-U curve is an empirical description conditional on a measurement scale, stimulus family, and listener distribution. It is not a universal law specifying a preferred percentage of off-beat attacks. Witek and colleagues found moderation by dancing enjoyment; the association of moderate syncopation with pleasure was particularly pronounced for listeners who enjoyed dancing Witek et al. (2014). Related personal variables, including familiarity with a style and abilities to generate a stable pulse, can alter the experienced balance of challenge and coherence. Researchers should fit alternative nonlinear functions and assess whether individual preference peaks survive independent samples, rather than infer an optimal rhythm from an aggregate curve.

Groove also clarifies why the experience of reward cannot be located entirely in passive auditory prediction. A listener imagines movement, times bodily preparation, and sometimes joins other people in synchrony. This sensorimotor opportunity may add value beyond the detection of acoustic patterns. Laboratory headphone studies can isolate timing manipulations, while naturalistic performance studies can study reciprocal bodily coordination; the two approaches answer different causal questions. Brattico et al. (2026) describes how contemporary music neuroaesthetics increasingly treats the aesthetic encounter as situated, embodied, and shared. The direct evidence for groove-related pleasure remains strongest when the measured stimulus manipulations and listener reports are specified.

Predictive Coding and Computational Surprise

Predictive-processing accounts describe perception as inference over hidden causes of incoming signals. A listening system assigns probabilities to future notes, chords, accents, and phrase boundaries; unexpected events generate information that can revise these probabilities. In information-theoretic terms, the information content of an event can be represented as the negative logarithm of its predicted probability, while uncertainty before the event can be described by the entropy of the predictive distribution. These quantities concern different moments of listening: uncertainty characterizes a forecast before an event, and information content characterizes the realized event. Treating them as interchangeable can conceal the interaction responsible for aesthetic response.

Cheung et al. (2019) estimated expectations over a large corpus of Western commercial chord progressions and experimentally probed pleasure using harmonic events. Musical pleasure depended nonlinearly on the interaction of uncertainty and surprise: a strongly unexpected event in a relatively certain context and an expected event within a highly uncertain context could both be pleasurable. The neuroimaging results associated the interaction with auditory and limbic regions but found a different relationship for the nucleus accumbens, which primarily reflected uncertainty in that experiment. The findings require a refinement of simplified accounts in which every pleasurable harmonic surprise produces the same neural reward prediction-error signature.

Computational surprise depends on a model's training distribution, representation, temporal horizon, and handling of repetition. A corpus-derived estimate is not automatically a listener's experienced surprise; a richly trained model may assign low probability to a familiar passage, or high probability to an unfamiliar tradition's pattern. Model comparison should use listener-specific ratings of expectedness, independent validation corpora, held-out predictions, and appropriate competing baselines. The strongest test asks whether estimated information quantities explain a new behavioral or neural response beyond acoustic covariates and familiarity. Gold, Mas-Herrero et al. (2019) provides a different experimental approach, fitting reinforcement-learning reward prediction errors rather than only informational deviation. The two programs overlap but are not mathematically identical.

Pleasure, Reward Prediction Errors, and Learning

Reward prediction error denotes a difference between expected and obtained reward value; it is distinct from the surprise of a chord considered solely as a statistical event. Learning systems can update future valuation based on that difference. Gold, Mas-Herrero et al. (2019) used a decision-learning paradigm in which musical cues and outcomes could be modeled as reward prediction errors, reporting nucleus accumbens responses and associations with learning. This establishes that music can function as a valued outcome in a reward-learning task. The causal interpretation remains conditional on the chosen reward model and on whether value-based errors can be distinguished from novelty, arousal, or information surprise in the experimental design.

Prediction and value can form a feedback loop. Familiarity with patterns supports anticipation; outcomes update perceived probability and desirability; repeated exposure can alter both liking and expectations. A listener may enjoy a delayed resolution because it is learnably intelligible, while a violation devoid of interpretable context may remain aversive or confusing. Gold, Pearce et al. (2019) examined the relation between informational predictability and musical preference across two behavioral studies and reported nonlinear preference for intermediate complexity. The results support a balance of uncertainty and learning opportunity, but the experiments do not justify equating subjective pleasure with an objectively calculated amount of information gain in every style.

Competing accounts make different predictions. A familiarity account expects liking to rise with exposure until satiation or other factors intervene; processing fluency emphasizes ease of recognition; curiosity and learning-progress accounts predict special value in reducible uncertainty; tension-resolution accounts emphasize temporal dynamics of anticipation. A design that only compares two melodies differing in multiple acoustic properties cannot separate these explanations. Factorial manipulations, repeated-measures trajectories, individual expectation estimates, and computational model recovery can identify which account adds explanatory power. A unified model is useful only insofar as it produces distinct falsifiable predictions across structural and behavioral dimensions.

Affective Mechanisms Beyond Musical Expectation

Expectancy is one channel among several routes to musical feeling. Juslin & Västfjäll (2008) distinguished mechanisms involving brain-stem reflexes, evaluative conditioning, emotional contagion, visual imagery, episodic memory, and musical expectancy. These mechanisms operate at different representational and temporal levels: a loud transient can provoke a rapid orienting response, a remembered song can retrieve an autobiographical scene, and expressive performance can invite perception of an emotion in the music. Their simultaneous operation can yield complex episodes in which nostalgia, tenderness, sadness, and reward co-occur. A theory restricted to syntactic violation cannot explain these cases simply by adding further surprise terms.

Physiological arousal also requires interpretation. Changes in heart rate, skin conductance, breathing, or pupil diameter can indicate autonomic engagement but do not uniquely identify pleasure, awe, sadness, or fear. Steinbeis et al. (2006) combined electrodermal measures with tension and emotionality reports precisely because an unexpected harmonic event can alter more than one outcome. Trost et al. (2012) examined brain activity during a wider set of aesthetic musical emotions and found distinguishable associations with arousal-related and affective categories. Such studies support distributed emotion-related processing while cautioning against one-to-one mappings between a named emotion and a localized brain region.

Affective mechanisms are also sensitive to task instructions. Asking whether music sounds sad invokes perceived expression; asking how sad the listener feels measures experienced affect; asking whether the listener wants the song repeated probes reward and motivation. These questions can yield different results for the same excerpt. Zentner et al. (2008) developed a framework for music-evoked emotions that includes experiences such as wonder, transcendence, nostalgia, and peacefulness. It helps research capture aesthetic responses that conventional basic-emotion inventories may undersample. Construct validity therefore begins with matching each measure to a precisely defined claim.

The Neurobiology of Musical Pleasure and Chills

Powerful musical enjoyment can be accompanied by chills, goosebumps, changes in heart rate, altered respiration, and self-reported intense pleasure. These bodily events furnish observable time markers for studying otherwise continuous musical experiences. In a foundational positron emission tomography study, Blood & Zatorre (2001) related intense pleasurable responses to activity in neural regions previously implicated in reward and emotion. Participants listened to music that reliably elicited their own chills, with control pieces reducing some generic auditory confounds. The resulting pattern established a biological association between musical pleasure and reward-related networks, while the correlational imaging design left open which network components generated the feeling and which reflected associated autonomic, memory, or motor processes.

Salimpoor et al. (2011) combined positron emission tomography with functional magnetic resonance imaging to study anticipation and peak emotional moments in the same listeners and musical excerpts. The PET measure indicated endogenous striatal dopamine release during intensely pleasurable music listening, while functional imaging associated anticipation more strongly with caudate activity and peak response more strongly with the nucleus accumbens. The findings identify temporal differentiation within a distributed valuation system. They do not establish that every musical pleasure episode requires measurable chills or that dopamine alone carries the entirety of subjective enjoyment. The inference rests on converging neurochemical, hemodynamic, and physiological evidence, each with different temporal and mechanistic resolution.

Reward systems acquire value from sensory and cognitive information that is represented elsewhere. Auditory cortices parse pitch, timbre, timing, and learned regularities; prefrontal regions contribute context and evaluation; striatal circuitry participates in motivation, valuation, and learning. Zatorre & Salimpoor (2013) proposes interactions among these regions as a framework for explaining how an aesthetic stimulus can become rewarding. This network-level formulation is more informative than localizing music appreciation to one brain center. It also makes clear predictions about connectivity, individual differences, and experimentally induced changes in valuation that can be investigated independently of a listener's verbal interpretation.

Auditory–Striatal Connectivity and Assigned Musical Value

Music can elicit considerable variation in reward even when listeners hear similar acoustic material. Salimpoor et al. (2013) investigated how auditory cortex interactions with the nucleus accumbens predicted the amount listeners were willing to pay for unfamiliar musical excerpts. As the listeners encountered new pieces and evaluated them, functional connectivity between reward-sensitive and auditory regions covaried with value. This operationalization is distinctive: willingness to pay captures a decision about prospective access to music, potentially linking immediate pleasure with expected future enjoyment. It does not measure a universal monetary price of musical beauty, because demand depends on budgets, task incentives, preferences, and experimental context.

The auditory–striatal account supplies a plausible route for predictions to influence valuation. If an acoustic structure carries recognizable probabilities, auditory systems may generate information about anticipated continuations that valuation systems use when estimating the significance of upcoming events. Yet functional connectivity inferred from fMRI is a statistical relationship between blood-oxygen-level-dependent signals, not a direct anatomical proof of information transfer or a causal direction. Common inputs, attention, and arousal can produce correlated fluctuations. Salimpoor et al. (2013) therefore supports a mechanistic framework while leaving important questions for perturbation, computational modeling, and temporal measurements.

An important analytic distinction concerns observed value and latent utility. A participant may spend money for music because it seems intriguing, socially meaningful, or useful for future listening, even if the excerpt did not produce chills. Conversely, a highly emotional event may resist monetary valuation. Studies benefit from measuring separate dimensions of current pleasure, desire for repetition, and choice behavior, then testing whether a common factor or partially independent processes explain the covariance. Zatorre & Salimpoor (2013) describes the progression from auditory perception to emotional reward as a network process rather than a single perceptual feature with a fixed hedonic sign.

Causal Perturbation of Fronto–Striatal Reward Circuits

Neuroimaging links between musical pleasure and valuation circuitry are observational, and an experimental perturbation offers a stronger test of causal involvement. Mas-Herrero et al. (2018) used transcranial magnetic stimulation over the left dorsolateral prefrontal cortex, aiming to influence fronto-striatal function, during music listening. Relative to sham conditions, excitatory and inhibitory stimulation produced bidirectional changes in subjective pleasure, psychophysiological response, and assigned monetary value. The convergence of affective and motivational outcomes strengthens the claim that a fronto-striatal circuit participates causally in music reward, while keeping the scope limited to the stimulation parameters and tested listeners.

Transcranial stimulation changes the activity of interacting networks and is not a selective switch for a single psychological mechanism. Changes in attention, anticipation, decisional policy, or general arousal could contribute to the observed effects. The study also cannot establish a precise neurotransmitter mechanism directly, because the intervention altered cortical excitability rather than measuring dopamine release under every condition. The strongest interpretation is causal participation of the manipulated circuit in rewarding musical experience. Salimpoor et al. (2011) provides complementary neurochemical and anticipatory evidence, but triangulation across studies must preserve their distinct tasks and populations.

Future causal tests can combine within-person crossover designs, blinded sham controls, physiologically verified intervention effects, and simultaneous estimates of musical prediction. The decisive question is whether perturbation changes the relationship between expected harmonic or rhythmic information and experienced reward, rather than changing all pleasantness ratings indiscriminately. An individual difference design could test whether intervention effects are strongest when auditory–striatal communication is initially weak or when musical prediction is particularly reliable. These are research proposals derived from the findings of Mas-Herrero et al. (2018) and Salimpoor et al. (2013), not established treatment recommendations.

Specific Musical Anhedonia and Reward Dissociations

Music reward varies widely across people and may be selectively diminished even when general reward processing and basic musical perception are preserved. Mas-Herrero et al. (2014) studied individuals reporting specific musical anhedonia and described dissociations between musical and monetary reward responses. Their evidence shows why low enjoyment of music should not automatically be interpreted as hearing impairment, general depressed mood, inadequate expertise, or indifference to every form of pleasure. The construct depends on converging measurements of musical perception, reward sensitivity, and relevant alternative explanations.

Specific musical anhedonia is theoretically revealing because it tests whether perceptual analysis alone produces pleasure. Individuals may identify tonal changes and recognize emotional expression while showing attenuated motivational or physiological reward responses to music. Such findings support the view that aesthetic valuation depends on interactions between sensory representations and reward networks. Salimpoor et al. (2013) connects this hypothesis with functional connectivity during valuation, while Mas-Herrero et al. (2014) demonstrates that association between perception and liking can be unusually weak in particular listeners. Neither result provides a single diagnostic threshold for clinical use.

Individual reward sensitivity also introduces sampling bias. Experiments recruiting people who regularly experience musical chills may yield much stronger correlations with dopamine-related signals than a representative population study. Study designs should identify musical anhedonia, general anhedonia, hearing status, musical training, cultural familiarity, and subjective listening motives rather than assume uniform responses. Separate outcomes for pleasure, motivation, and perceived emotion help distinguish multiple pathways to engagement. A theory robust across ordinary enjoyment, exceptionally intense pleasure, and low musical reward is scientifically stronger than one fitted only to self-selected enthusiasts.

Aesthetic Emotion, Perceived Expression, and Felt Emotion

A musical passage can be recognized as mournful, agitated, tender, or triumphant without inducing the corresponding emotion in every listener. Perceived emotion concerns the expressive character attributed to the musical object; induced emotion concerns the listener's own experienced response. Juslin & Västfjäll (2008) distinguished several mechanisms capable of generating emotional responses, including conditioning, episodic memory, imagery, contagion, and expectancy. Aesthetic judgment adds another evaluative level: a listener may consider a tragic performance artistically successful while feeling sadness and pleasure together. Experimental prompts need to specify which of these levels participants are rating.

Music-specific emotions are often described with terms that cross ordinary categorical boundaries. Zentner et al. (2008) developed and investigated a taxonomy of emotions evoked by music, including wonder, transcendence, tenderness, nostalgia, peacefulness, power, joyful activation, tension, and sadness. Such profiles help researchers capture the coexistence of activation, positive valuation, and complex affect. A scale is a theoretically organized instrument whose factors are supported to the extent that sampling, model fit, reliability, and cross-group validation justify them. Its categories must not be treated as natural kinds independent of language, listening context, or culture.

Trost et al. (2012) used functional imaging to examine neural activity associated with differentiated musical emotions. The reported patterns implicate distributed neural systems rather than isolated modules for each feeling label. The findings contribute a bridge from psychometric differentiation to neural response, but multivariate or region-specific activation cannot conclusively decode the full subjective meaning of a passage. A scientific account of musical feeling therefore combines what participants report, what the music affords, how the body responds, and which neural processes are engaged under explicit tasks.

The Pleasure of Sad Music and Being Moved

Sad music poses a central aesthetic problem: listeners sometimes seek out a stimulus they recognize as sad because the encounter is rewarding. A simple account equating negatively valenced expression with avoidance cannot explain the choice. Vuoskoski & Eerola (2017) examined pleasure in response to sad music and reported that feelings of being moved mediated its relationship with enjoyment. The evidence redirects attention toward tenderness, social significance, and meaningful engagement, while requiring separation between the statistical role of a measured mediator and a demonstrably causal pathway. Feeling moved may combine emotional intensity and positive evaluation without removing the recognized sadness.

Several mechanisms can plausibly contribute to pleasurable sadness. Music can evoke autobiographical memory, provide a safe imaginative space for emotion, elicit empathic involvement with expressive gestures, and create a sense of meaningful connection. Schubert (2016) considers parallel processing as an account of why sad expression and enjoyable experience may coexist. These mechanisms are distinguishable through interventions: manipulate familiarity, lyrical content, narrative framing, social context, or opportunities for autobiographical recall, then examine changes in reported sadness, being moved, and pleasure. A single rating of whether the music is 'sad' leaves the explanatory problem unresolved.

Study design must also consider selection into listening. People who voluntarily choose melancholic pieces may differ from people assigned unfamiliar sad music in a laboratory, and a preferred recording may carry relationships and memories that a standardized stimulus cannot reproduce. Zentner et al. (2008) suggests why music-evoked emotions deserve distinctive language, while Vuoskoski & Eerola (2017) shows how a specific affective construct can organize a testable account. The paradox becomes more tractable when the representation of sadness, experienced affective mixture, and evaluation of the episode are modeled separately.

Physiological Arousal, Frisson, and Measurement Specificity

Musical frisson refers to brief episodes of intense affective and bodily response, often accompanied by chills or piloerection. It offers a useful event-related marker for research because participants can identify peaks during a continuous recording. Blood & Zatorre (2001) studied music selected for the capacity to elicit chills, enabling correlations between reported intensity and regional brain activity. Salimpoor et al. (2011) used pleasurable peaks to study temporal differentiation in reward-related circuits. Frisson nevertheless represents a subset of aesthetic encounters; quiet absorption, soothing pleasure, rhythmic enjoyment, and reflective admiration can occur without a conspicuous autonomic peak.

Electrodermal activity reflects sympathetic sweat gland activation and can rise during emotionally intense, surprising, attention-demanding, or unpleasant moments. Heart-rate and respiratory responses likewise depend on multiple regulatory processes. The studies of harmonic violation in Steinbeis et al. (2006) and of intensely pleasurable music in Salimpoor et al. (2011) illustrate why physiological records require simultaneous subjective reports and stimulus-level analysis. Without those links, elevated arousal might be mistakenly interpreted as positive reward. Temporal alignment is indispensable: an anticipation-related response preceding an event should not be assigned the same function as an evaluation following it.

Robust physiological experiments register acoustic intensity, movement artifacts, time from stimulus onset, baseline state, and responsiveness of each participant. Where feasible, continuous affect ratings, recorded muscle activity, and objective markers of piloerection can complement retrospective chill reports. Data analysis should account for temporally autocorrelated measurements rather than treating every time point as independent evidence. When samples are small or selected for high sensitivity, confidence intervals and transparent preprocessing decisions matter as much as statistical significance. These principles follow from the multiple explanatory routes documented in Juslin & Västfjäll (2008) and the convergent-measure logic of Blood & Zatorre (2001).

Music, Memory, Autobiographical Meaning, and Mood Regulation

Emotional responses to music often arise from the history carried by a recording rather than its formal syntax alone. A melody associated with a particular person, place, ritual, or period of life can evoke memories with strong affective content. Juslin & Västfjäll (2008) identifies episodic memory among the routes through which music evokes emotion. This mechanism predicts that the same excerpt may provoke divergent responses in people who share knowledge of its harmonic structure but differ in autobiographical associations. Aesthetic research therefore gains explanatory power by recording familiarity, personal significance, and current goals alongside objective musical properties.

Music is also used deliberately to manage affect, energy, and social participation. Saarikallio & Erkkilä (2007) investigated music's role in adolescents' mood regulation, identifying varied ways listeners use music in everyday life. Such findings situate rewarding listening within goal-directed behavior: a person may choose an intense song to energize action, a sad piece to reflect, or familiar music to feel continuity. Motivational function and immediate pleasantness need not coincide. The evidence from self-report and qualitative accounts identifies meaningful patterns, while experimental work is necessary to determine how particular selection strategies causally alter subsequent mood.

An adequate account of emotional reward links within-piece musical events with long-term memory and situated use. Anticipation and resolution operate on seconds; autobiographical associations may extend across decades; social narratives can be repeatedly reconstructed. Designs that ignore these timescales risk interpreting personally significant music as evidence for universal acoustic reward parameters. Zatorre & Salimpoor (2013) and Saarikallio & Erkkilä (2007) support a multilevel explanation in which neural value assignment interacts with learned goals and memory. Quantitative experience sampling offers a way to study these relationships outside the laboratory.

The Development of Musical Temporal Expectations

Musical timing develops through interaction between early perceptual capacities and exposure to patterned sound. Infants detect temporal regularities before acquiring explicit theoretical knowledge of rhythm; however, their later representations increasingly reflect the timing practices encountered at home and in the wider community. Winkler et al. (2009) reported responses to expected rhythmic positions in newborns, supporting early sensitivity to pulse structure. This evidence does not specify a fully formed adult metrical system at birth. Developmental research must distinguish detection of acoustic regularity, representation of an inferred beat, and culturally learned metric weighting.

Hannon & Trehub (2005) studied sensitivity to musical meters from different cultural traditions in infants and adults. Six-month-old infants displayed relatively broad discrimination across the tested rhythmic contexts, whereas twelve-month-old infants showed more culture-specific responding; brief exposure enabled younger listeners to adapt to less familiar metrical patterns more readily than adults in the tested paradigm. The results make perceptual tuning measurable and show why an adult's difficulty with unfamiliar timing should not be interpreted as an absence of rhythmic structure in the music. They also illustrate the need for carefully matched stimuli and participants' listening histories.

Longitudinal models should track exposure to caregivers' songs, participation in dance, language rhythm, formal instruction, and media environments across development. Theories of prediction learning would expect repeated regularities to guide perceived meter and subjective fluency, while allowing novelty to remain valuable when listeners can assimilate it. The relation between learned meter and musical pleasure remains less directly established in infancy than the relation between exposure and discrimination. Hannon & Trehub (2005) supports the latter claim; translating it into later aesthetic preference requires additional studies that measure both expectation and reward over time.

Cultural Systems, Consonance, and Musical Preference

Musical cultures provide distinctive distributions of scales, timbres, tuning, metric structures, performance practices, and social meanings. The concept of an unexpected event is therefore defined relative to a learned repertoire and local expectations. In McDermott et al. (2016), listeners from a relatively isolated Amazonian population showed indifference to consonance and dissonance as an aesthetic contrast, unlike comparison populations exposed to Western harmony. This observation challenges the assumption that Western consonance preference is an automatic universal output of human hearing. It does not deny acoustic properties of beating or harmonicity; rather, it separates detection of such properties from learned hedonic ranking.

Cross-cultural work must avoid taking Western notation as the only precise description of musical organization. Irregular meter can be highly regular relative to another tradition's timing, while a tuning system described as 'dissonant' under Western conventions may fulfill the structural role expected in a particular practice. Savage et al. (2015) examined cross-cultural musical features across a global sample and found statistical regularities without absolute universals in the sampled recordings. Such findings motivate comparative modeling of recurrent human constraints and locally learned conventions rather than a single global likelihood model trained on one commercial repertoire.

A more defensible universality claim concerns capacities for learning patterns, coordinating action, recognizing expressive cues, or valuing socially meaningful sound, not necessarily the same preferred chords everywhere. Testing this claim requires independently collected corpora, sensitive translations of emotional scales, local involvement in selecting stimuli, and explicit evaluation of measurement equivalence. The contrast between McDermott et al. (2016) and Cheung et al. (2019) is productive: one tests culturally specific preferences and the other explains pleasure in quantified Western harmonic contexts. They address different populations and stimulus distributions, so the findings can be integrated without forcing a universal rule.

Statistical Regularities and the Diversity of Human Song

Comparative musical research asks which properties are frequent across societies, what functions songs serve, and how listeners learn the relevant expectations. Savage et al. (2015) identified cross-regional statistical regularities in a global recording corpus, while explicitly reporting that the observed features were not absolute universals. The sample and coding scheme define what can be inferred: recurrent forms of group performance and rhythmic organization are empirical tendencies, not proof that every musical tradition contains identical structure or rewards. These distinctions are important when researchers explain the popularity of rhythmic repetition through species-wide psychological mechanisms.

Mehr et al. (2019) combined ethnographic and musical analysis to investigate universality and diversity in human song, connecting behavioral regularities with the social contexts of singing. The findings reinforce a dual problem for aesthetics: listeners can respond to musical structures with substantial shared capacities while learning what a given pattern means through particular practices. Song use for infant care, healing, dance, or other functions involves an expectation about an activity as well as an expectation about a note. A pleasing melodic continuation in a ritual setting may depend on fulfilling a social pattern that cannot be inferred from audio acoustics alone.

Comparative evidence can improve computational models when cultural sampling is genuinely independent of the training data. It also helps distinguish a musical feature's prevalence from an individual listener's preference. A pattern may be widespread because it supports coordination or is easy to transmit, even if it is not invariably rated most beautiful. Savage et al. (2015) thus provides population-level context, whereas Witek et al. (2014) estimates preferences under controlled rhythmic manipulations. The difference in levels of analysis should be maintained when claiming that groove, meter, or tonal organization has evolved or acquired its aesthetic value.

Musical Expertise, Improvisation, and Listener Expectations

Expertise changes how musical events are grouped, predicted, and interpreted. A trained listener can hear a phrase in relation to harmonic conventions or rhythmic substitutions that a novice may treat as isolated surprises. Expertise is heterogeneous: reading notation, performing an instrument, improvising, dancing, and listening extensively to a genre cultivate overlapping but distinct competencies. Steinbeis et al. (2006) investigated harmonic expectancy violations in musicians and nonmusicians, making expertise relevant to both cortical signatures and affective response. An expert's expectation of a dissonance can therefore coexist with enjoyment of its creative placement.

Improvisational listening adds uncertainty about the performer's next action and about the relationship between phrase, ensemble, and style. An unexpected event may be aesthetically effective if it reveals a new organizational possibility or coherently transforms an anticipated motif. Predictive-coding accounts such as Vuust & Witek (2014) offer concepts for quantifying this balance, but practical application requires models of richer event sequences than isolated note transitions. A jazz solo and a classical cadence may produce formally similar estimates of low event probability while involving distinct social norms, goals, and expressive interpretations.

Study designs should measure knowledge and preference separately and train computational models on the musical materials familiar to each participant. An advantage in predicting a style might increase processing fluency, but it could also reduce novelty and change the desired level of surprise. Repeated exposure experiments can identify evolving curves rather than assuming that years of training linearly increase liking. The relevance of Gold, Pearce et al. (2019) lies in specifying testable relations among predictability, uncertainty, and pleasure; expertise remains a proposed moderator that should be evaluated rather than treated as an automatic explanation.

Social Synchrony, Dance, and Shared Musical Reward

Music is encountered in concerts, rituals, festivals, rehearsal rooms, and informal gatherings where other people's movements become part of what a listener anticipates. A common beat permits temporal coordination, and successful coordination may itself be rewarding. Tarr et al. (2014) proposed that self–other merging and neurohormonal mechanisms contribute to music-related social bonding, outlining a framework that spans synchronous movement and interpersonal connection. The theory makes social context a candidate cause of reward rather than a decorative backdrop to private acoustic experience.

Tarr et al. (2015) experimentally varied synchrony and exertion during dance and observed independent effects on pain threshold and reported social bonding. Pain threshold served as an indirect proxy in the authors' neurohormonal account; the experiment did not directly establish endorphin concentrations or isolate a single biochemical mechanism. The factorial structure nevertheless provides stronger leverage than a simple association between liking music and feeling close to others. It supports a social pathway through coordinated bodily action, with exertion and synchrony as distinguishable variables.

Social musical reward requires a multilevel design. One must measure beat tracking and interpersonal movement synchrony, distinguish shared timing from mutual attention and emotional contagion, and assess whether group identification precedes or follows the activity. Witek et al. (2014) links rhythm to movement desire at the individual level, while Tarr et al. (2015) examines social outcomes during coordinated dance. Integrating their findings suggests that an attractive groove can facilitate a context for group reward, although direct mediation from syncopation manipulation through interpersonal synchrony to bonding requires its own experiment.

Live Performance, Attention, and Ecological Aesthetic Encounters

Live performance transforms musical expectation because events emerge from visible bodily action and real-time coordination. A performer's gesture can forecast an impending accent, a dynamic change can prepare a climactic entrance, and audience reactions can reinforce the value attributed to a passage. A laboratory recording allows acoustic control but omits many of these contingencies. Brattico et al. (2026) identifies situated and multidimensional musical aesthetic experience as a continuing challenge for neuroaesthetics. The review supports broadening research beyond passive presentation of brief isolated excerpts.

Ecological designs can combine recorded audio, timing markers, continuous ratings, physiological sensors, and unobtrusive estimates of movement while retaining a detailed record of performance features. They can also compare participation with listening to prerecorded performances. Nonetheless, live settings introduce interpretive hazards: louder sound, social pressure, visual attention, acoustic reverberation, familiarity, and interpersonal attraction may change together. Statistical adjustment alone may be insufficient when factors are tightly coupled; randomization and carefully staged contrasts are needed where feasible.

An experience has an extended temporal form. Anticipation can precede a performance, memories can shape attention during it, and retrospective judgments may depend on the ending. The scientific target therefore includes changes in valuation across time rather than only a postconcert score. Saarikallio & Erkkilä (2007) demonstrates how musical experiences operate within everyday emotion regulation, while Trost et al. (2012) shows the diversity of emotion-related brain patterns. Combining such perspectives encourages studies that ask when, why, and for whom a performance becomes meaningful.

Aesthetic Appreciation, Value Judgments, and Artistic Context

Aesthetic appreciation includes pleasure but also interpretation, recognition of form, evaluation of performance, and judgments of artistic merit. A listener can prefer an easily predictable song in daily life while valuing a technically intricate work for its expressive or structural achievement. Nadal & Skov (2025) develops a sensory valuation account of aesthetic experience, placing perceptual processing within a broader framework of assigned value. Such approaches encourage differentiation among liking, aesthetic interest, perceived beauty, and decisions to engage again. The domain requires evidence about each construct rather than assuming that any pleasant neural response proves appreciation of artistic merit.

Artistic context can alter the same acoustic event's meaning. Attribution to a performer, knowledge of a composition's historical role, attention to virtuosity, and a listener's current purpose can change the reference set used for expectation and evaluation. A 'surprise' in an avant-garde piece may be stylistically predictable to a specialist, while a perfectly conventional cadence can be emotionally powerful in a remembered song. Meyer (1956) located musical meaning partly in culturally acquired patterns of expectation; contemporary research can operationalize context by changing information given before listening and measuring ensuing probability judgments and affect.

Experimental aesthetics benefits from designs that cross stimulus properties with task framing. Present the same excerpt as part of a stylistic learning task, a performance competition, or an emotionally meaningful story; separately measure recognized structure and inferred artistic value. Comparing conditions would help determine whether context changes prediction itself, evaluation of the predicted outcome, or both. This is a proposed approach based on Nadal & Skov (2025) and Cheung et al. (2019), not evidence that a single laboratory task already captures the social institution of music.

Music, Artificial Prediction, and the Limits of Computational Analogy

Machine-learning models can quantify patterns in music corpora and generate conditional forecasts about future events. Their value for psychology is methodological: formal probabilities can be compared with human expectancy judgments and with the effects of controlled violations. Cheung et al. (2019) illustrates a corpus-based approach in which musical uncertainty and surprise become experimentally testable predictors. More advanced models can represent several timescales, timbre, and rhythmic organization, but architectural sophistication does not by itself validate psychological relevance. A model must outperform appropriate acoustic and familiarity baselines on held-out human observations.

An artificial system that predicts a chord accurately demonstrates a competence on a forecasting task. Musical emotional reward is a distinct construct operationalized through behavior, valuation, physiological response, and for humans subjective report. A probability distribution produced by software cannot be directly substituted for a listener's reported pleasure. The question of artificial aesthetic evaluation is therefore approached through explicit observable criteria, computational constraints, and reproducible experiments. Gold, Mas-Herrero et al. (2019) shows that learning depends on value-specific prediction errors, reinforcing the need to distinguish statistical surprise from motivated reward even within human experiments.

In the Artificial Era, human–artificial musical interaction offers a further test bed. Adaptive accompaniment could systematically adjust rhythmic displacement or harmonic uncertainty in response to a musician's predictions, allowing counterbalanced tests of preferred complexity and shared timing. Such designs would examine coadaptation and distributed performance rather than presume that artificial generation and human emotion are equivalent. The evidence in Witek et al. (2014), Tarr et al. (2015), and Salimpoor et al. (2013) motivates hypotheses concerning embodied coordination and value, but causal claims about artificial partners require studies that actually manipulate the interactive system and compare outcomes with appropriate human and noninteractive baselines.

Experimental Manipulation of Rhythm, Harmony, and Expectation

A decisive test of musical expectancy requires control of the event that was expected and the event actually delivered. Harmonic studies can embed alternative chords in a shared tonal progression; rhythmic studies can alter syncopation while maintaining a comparable tempo, instrumentation, and beat representation. Steinbeis et al. (2006) manipulated local harmonic expectancy violations in chorales, and Witek et al. (2014) parametrically varied syncopation in drum breaks. The two experiments show how carefully defined deviations can be related to listener responses. However, adding an unexpected chord may also alter acoustic roughness and affective coloration, while modifying syncopation can change accent energy and auditory density. Independent acoustic matching or statistical modeling is essential.

Experiments should use more than one stimulus exemplar and model the structure of repeated responses. A design with many listeners but one tune cannot establish that a result generalizes across compositions; a design with many excerpts from one genre may fail to generalize across cultures. Crossed random effects for participants and musical items, preregistered primary outcomes, and holdout sets for computational predictions reduce stimulus-specific and analytic overfitting. Information-theoretic conditions should be manipulated or measured with adequate coverage of low and high uncertainty rather than testing only 'surprising' and 'unsurprising' events.

External validity depends on preserving recognizable musical form. Extreme laboratory simplification can help identify a sensory mechanism while removing the melodic development and contextual meaning that produce aesthetic experience. A promising sequence uses controlled stimuli to establish a local effect, then tests it with longer unfamiliar compositions, familiar repertoire, and live contexts. Cheung et al. (2019) and Gold, Pearce et al. (2019) illustrate distinct approaches to probability estimation; their convergence is informative only when both operations are specified. A formal model should predict the manipulation's effects before the response data are used to tune it.

Behavioral and Psychometric Assessment of Musical Reward

Musical pleasure can be assessed through immediate ratings, continuous hedonic responses, selection of a preferred excerpt, willingness to pay, or decisions to replay and share a recording. These measures sample partially different functions. In Salimpoor et al. (2013), willingness to pay captured assigned monetary value for unfamiliar music; in Witek et al. (2014), ratings of pleasure and desire to move captured separable aspects of rhythmic response. Agreement between indicators supports a common motivational interpretation, while disagreement can reveal a distinction between wanting and liking, expressive appraisal and bodily readiness, or enjoyment and artistic admiration.

Emotion measures also require construct specificity. The music-evoked emotion framework investigated by Zentner et al. (2008) offers items for nuanced states beyond generic valence and arousal. It requires empirical checks for reliability, factor structure, translation equivalence, and temporal sensitivity in any new population. State ratings during listening can differ from reports about a person's typical relationship with music; retrospective measures may privilege memorable peaks and endings. Researchers should make their unit of measurement explicit—moment, excerpt, session, person, or social group—and align their claims with that unit.

Subjective reports remain central because reward includes experienced value, but demand characteristics must be addressed. Music expertise, familiarity with experimental language, perceived investigator preferences, and the ordering of questions can change ratings. Blinding hypotheses where practical, randomizing stimulus order, collecting open-ended descriptions, and triangulating with choices and physiology help reduce interpretive ambiguity. Mas-Herrero et al. (2014) makes this especially clear: a low pleasure score requires checking whether the listener understands the stimulus, perceives the intended structural contrast, and has adequate opportunities to express a different kind of engagement.

Neuroimaging, Electrophysiology, and Neurochemical Evidence

Functional MRI measures hemodynamic signals related indirectly to local neural activity. It provides spatial coverage of auditory, cortical, and striatal networks and supports hypotheses about regional responses and connectivity, but its temporal resolution does not match the millisecond scale of note anticipation. Salimpoor et al. (2013) related functional connectivity to music valuation and Cheung et al. (2019) mapped responses associated with statistical properties of harmonic progressions. These are powerful correlational observations whose interpretation depends on task design, model specification, and the validity of hemodynamic preprocessing.

Electroencephalography is well suited to responses tied to rhythm, prediction violations, and temporally structured attention. Nozaradan et al. (2011) reported beat- and meter-related steady-state activity, while Winkler et al. (2009) investigated responses to omitted rhythmically expected sounds in infants. Event-related potentials and frequency-tagged signals provide high temporal precision but limited spatial specificity. When the auditory envelope itself is periodic, spectral peaks in EEG do not automatically prove spontaneous oscillatory mechanisms; appropriate acoustic modeling and control sequences are required. Theories of entrainment should specify whether they predict phase alignment, enhanced sensory response, selective attention, or motor preparation.

PET imaging with radioligands offers an indirect way to investigate neurochemical processes under particular assumptions. The endogenous dopamine signal in Salimpoor et al. (2011) provides evidence unavailable from fMRI alone, yet the method integrates activity over a longer period and cannot identify the subjective content of every musical event. Causal techniques such as the stimulation used in Mas-Herrero et al. (2018) provide additional leverage but also affect distributed circuits. Triangulation across PET, fMRI, EEG, psychophysiology, and behavior is therefore essential; each method contributes a different constraint to the mechanism rather than serving as a universal measure of musical pleasure.

Information-Theoretic Models and the Problem of Model Validation

A probabilistic music model requires an alphabet of events, a representation of context, and a method for estimating transition probabilities. For tonal harmony, the events may be chord classes; for rhythm, they may be onset times relative to a meter; for melody, pitch intervals, scale degrees, or contour changes. Information content is computed from the model's assigned probability to the realized event, while uncertainty describes the breadth of predictions before that event. Cheung et al. (2019) explicitly separated these quantities and showed that their interaction mattered for pleasure under experimental harmonic conditions. A model's choice of representation can change both variables, so explanatory claims must include that choice.

Validation must be independent of the data used for parameter selection. A model can fit enjoyment ratings by adding enough nonlinear terms without recovering a psychologically plausible predictive mechanism. Held-out participant data, new musical corpora, benchmark baselines, and cross-cultural transfer tests help identify genuine generalization. In Gold, Pearce et al. (2019), relations between predictability and pleasure were investigated at a behavioral level, but an estimated preference curve remains conditional on task, stimuli, and musical environment. Mechanistic claims need evidence that the model predicts when listeners actually expect events, ideally before receiving the enjoyment rating.

Information surprise and reward prediction error require separate formal treatment. Surprise is large when an event is improbable regardless of desirability; a reward prediction error depends on how good the obtained event is relative to anticipated value. Gold, Mas-Herrero et al. (2019) modeled learning from preferred and dispreferred musical endings, whereas Cheung et al. (2019) assessed statistical harmonic uncertainty and surprise. These constructs can covary when surprising outcomes are rewarding, but they are not the same mathematical object. Their integration requires experiments in which value and probability are varied orthogonally and learning trajectories are observable.

Evidence Quality, Replication, and Generalizability

Music neuroscience includes small neuroimaging samples, selected listeners with intense responses, repeated-measures acoustic manipulations, and populations heavily exposed to Western popular or classical traditions. Each choice answers particular questions while narrowing the population to which findings can be generalized. Blood & Zatorre (2001) and Salimpoor et al. (2011) studied intense music-induced responses; their results have high relevance to chills and reward but cannot establish the incidence of chills in all listeners. McDermott et al. (2016) demonstrates why cross-cultural comparison is needed when preference for harmonic relations is assumed to be species-wide.

Replication requires more than obtaining a statistically significant association in another dataset. Direct replication tests a specified protocol and outcome; conceptual replication tests the hypothesized mechanism with different operations; transportability examines whether an effect persists across stimuli, demographics, cultures, and settings. A study of harmonic surprise in short unfamiliar chorales might replicate behaviorally yet show different response functions in rhythmic improvisation or a familiar song. Good synthesis therefore reports effect directions and boundary conditions instead of treating heterogeneous measurement as one pooled variable without justification.

Selective publication and analytical flexibility can inflate reported effects, especially when many brain regions, time windows, emotional scales, and stimulus transformations are examined. Preregistered contrasts, transparent data exclusions, correction for multiplicity, publicly available stimulus annotations, and interval estimates strengthen reliability. Independent laboratories and multisite samples are valuable for distinguishing stable signal from local conventions. The absence of large preregistered replication evidence for a specific claim should be reported as a limitation rather than filled with a consensus statement. The generalizable conclusion supported by Zatorre & Salimpoor (2013) is a network-level relation between music perception and reward, with many details open to refinement.

Competing Explanations and Critical Theoretical Disputes

One explanatory dispute concerns whether musical pleasure is driven chiefly by prediction fulfillment, prediction violation, or opportunities for learning. Meyer's Meyer (1956) expectation account foregrounded delayed or altered continuation; Huron (2006) differentiated several temporally organized responses; Cheung et al. (2019) found an interaction between uncertainty and surprise rather than a simple main effect of either. These positions suggest distinct causal processes, not merely alternative words for the same phenomenon. A design must separate immediate orienting or tension from later positive valuation to discriminate them.

Another dispute concerns whether reward prediction errors in a laboratory task establish the mechanism of pleasure during ordinary listening. Gold, Mas-Herrero et al. (2019) reported nucleus accumbens signals consistent with formally modeled value-based learning during choices involving consonant and dissonant endings. de Fleurian et al. (2019) challenged the inference from that paradigm to musical pleasure, emphasizing limits in what a reward-learning task establishes. The debate illustrates a crucial distinction: demonstrating that musical feedback can support reinforcement learning does not alone prove that all instances of pleasurable listening are generated by the same trial-by-trial value-learning process.

A third dispute concerns adaptation and biological universality. Statistical regularities of music across societies Savage et al. (2015) can reflect broad perceptual or social constraints, whereas the differing consonance preferences documented by McDermott et al. (2016) show considerable learned variation. These findings are compatible if universality is stated at the appropriate level, but they warn against treating local tonal rules as innate aesthetic categories. The strongest scientific program compares multiple models on common data while allowing certain mechanisms to explain only well-defined portions of an experience. No single theory currently accounts equally well for chills, dancing, nostalgia, pleasurable sadness, and reflective artistic judgment.

Methodological Integration Across Timescales and Levels

Music unfolds in nested timescales from individual onsets to phrases, movements, and long-term personal relationships with repertoire. Predictive mechanisms may be rapid enough to anticipate a drum hit, whereas remembered emotional significance can persist across years. A unified measurement strategy aligns an acoustic timeline with behavioral responses and physiology while collecting independent measures of familiarity, social setting, and motivation. Salimpoor et al. (2011) differentiates anticipation from peak experience within a listening episode; Saarikallio & Erkkilä (2007) describes music as a resource for regulating emotion across everyday contexts. Each finding occupies a different explanatory level.

A multilevel statistical model can include trial-level harmonic and metric variables, excerpt-level genre and performance features, person-level training and preference, and group-level cultural and social context. Temporal autocorrelation and repeated encounters should be modeled explicitly. In such analyses, a relationship between complexity and pleasure within individuals may differ from the association between expertise and typical musical preference across individuals. The inversion of these levels is a common route to misleading causal inference. Witek et al. (2014) offers a starting point for investigating within-stimulus syncopation while recognizing individual differences in dancing enjoyment.

Meaningful scientific integration also requires distinctions among evidence types. A review may organize theoretical possibilities; a computational simulation demonstrates consequences of a particular model; a behavioral experiment tests effects of a manipulation; and an intervention can supply causal leverage under its design assumptions. Brattico et al. (2026) highlights contemporary conceptual breadth in music neuroaesthetics, while Mas-Herrero et al. (2018) provides a focused perturbational result. Their combination produces a more complete research map without implying that a recent synthesis is itself an independent experiment or that correlational data inherit causal status from another methodology.

Research Ethics, Transparency, and Open Questions

Ethical music research treats listeners' cultural background, identity, accessibility, and expectations with care. Emotional induction can evoke distressing autobiographical material; experiments involving loud sound or intense physiological measurement require informed consent and appropriate safety limits. Studies involving minors require developmental safeguards, and interpretations of musical anhedonia must avoid stigmatizing ordinary preference diversity. The primary ethical obligation in a scientific monograph is accurate representation of evidence, with clinical claims confined to validated indications and procedures. Current reward-neuroscience findings do not authorize prescribing music as a replacement for established medical treatment.

Research transparency should include stimulus files or reproducible descriptions where licensing permits, exact preprocessing protocols, code for probability models, preregistered comparisons, and descriptions of participant recruitment. Cross-cultural studies should involve local expertise in stimulus selection and translation. A claim of universality requires genuine cultural coverage and culturally appropriate response tasks, while a claim of causal circuitry requires manipulation or other defensible causal identification. McDermott et al. (2016), Savage et al. (2015), and Mas-Herrero et al. (2018) exemplify how different designs supply complementary rather than interchangeable evidence.

The unresolved questions include how music-specific expectations are updated over the lifespan, how reward signals incorporate social meaning, how aesthetic interest relates to pleasure in complex repertoires, how diverse musical cultures shape neural valuation, and whether interpretable computational models transfer across listeners. Particularly valuable experiments would cross uncertainty with musical desirability, manipulate access to movement and social coordination, and follow developing listening habits longitudinally. Cheung et al. (2019) and Gold, Mas-Herrero et al. (2019) provide operational starting points for these questions; Nadal & Skov (2025) suggests a wider valuation perspective. These are directions for future evidence, not results already obtained.

An Integrated Account of Rhythmic, Predictive, and Reward Processes

The evidence supports a layered account of musical aesthetic response. Acoustic events supply onset, pitch, timbre, and intensity information; perceptual systems form structured rhythmic and tonal representations; learning adjusts the probabilities assigned to possible continuations; affective and valuation systems register the consequences in relation to current goals and prior experience. Nozaradan et al. (2011) and Hannon & Trehub (2005) illuminate temporal organization and learning, Cheung et al. (2019) quantifies an interaction of harmonic uncertainty and surprise, and Salimpoor et al. (2013) links music valuation with auditory–striatal communication. This account is a scientific synthesis of separate operations supported by distinct studies. It does not assert that every layer is active to the same degree in every listener or piece.

The strongest support for causal reward circuitry comes from the transcranial stimulation study Mas-Herrero et al. (2018), while Salimpoor et al. (2011) contributes neurochemical evidence about intense musical responses. Gold, Mas-Herrero et al. (2019) demonstrates that musical outcomes can support reward prediction-error signaling during choice learning. These findings constrain theories of pleasure, but none makes music appreciation reducible to one neurotransmitter, one brain region, or one prediction error. Emotion induction, aesthetic judgment, and social participation must be measured on their own terms. Variation across individuals and cultures is a necessary component of the model rather than statistical noise.

The account generates concrete tests. First, manipulate rhythmic or harmonic uncertainty while measuring listener-specific forecasts before reward judgments. Second, independently alter task goals and narrative contexts to distinguish changes in event probability from changes in assigned value. Third, investigate whether the presence of dance, shared performance, or collaborative improvisation alters the reward function even when the sound sequence is held comparable. Fourth, examine whether identified perturbations of reward networks shift the effect of information content or produce a general hedonic change. These proposed experiments connect Witek et al. (2014), Tarr et al. (2015), and Mas-Herrero et al. (2018) in a testable research program.

Theoretical Boundaries and Alternative Pathways

Expectation-based explanations have greatest direct empirical purchase where measurable musical events generate predictions that can be compared with actual outcomes. This domain includes harmonic closure, metric accents, omitted beats, and the information-theoretic properties of local progressions. It is broader than simple anticipation of a note because uncertainty extends over several possible continuations and time spans. Meyer (1956) and Huron (2006) provide historical and theoretical foundations, while Cheung et al. (2019) supplies experimentally estimated quantities. Their shared interest in expectation does not require identical assumptions about emotion or neural implementation.

Other affective pathways operate through culturally shared social meanings, personal memory, expressive identification, and regulation goals. The sad-music findings in Vuoskoski & Eerola (2017) and the mood-regulation evidence in Saarikallio & Erkkilä (2007) demonstrate why an exclusively local acoustic explanation will be incomplete. Intense arousal elicited by surprise is also neither necessary nor sufficient for reward; a predictable childhood song may be comforting, while a dazzling unfamiliar rhythm may be difficult to enjoy. Differentiating these pathways supplies useful boundary conditions for predictive theories and protects their distinctive explanatory claims.

An aesthetic response may continue beyond listening as reflective judgment about beauty, historical originality, or performance significance. Nadal & Skov (2025) situates aesthetic experience within broader valuation, and Brattico et al. (2026) reviews the transition toward treating music as art in situated contexts. These conceptual accounts complement experimental reward studies without independently proving that all judgments arise from the same mechanism. The resulting scientific position is pluralistic in a precise sense: multiple processes can contribute, interactions can be tested, and the explanatory scope of each mechanism depends on a defined outcome and manipulation.

Priority Research Questions for the Next Generation of Music Aesthetics

A first priority is transportable computational prediction. Researchers should develop models calibrated to individuals and musical traditions rather than importing Western chord probabilities as default measures of human expectation. These models should predict independent expectancy ratings and emotional outcomes in unfamiliar recordings, new listener groups, and contrasting cultural settings. McDermott et al. (2016) establishes that aesthetic preferences can vary across populations, while Savage et al. (2015) demonstrates that cross-cultural regularities can nevertheless be quantified. The methodological objective is to explain both patterns with explicitly stated uncertainty and scope.

A second priority is causal differentiation of anticipation, hedonic evaluation, and learning. The combination of Salimpoor et al. (2011), Gold, Mas-Herrero et al. (2019), and Mas-Herrero et al. (2018) establishes relevant neural evidence but leaves open how different task and physiological operations map onto episodes of musical enjoyment outside the laboratory. Designs combining carefully varied musical probabilities, behavioral learning, repeated reward ratings, and network perturbations could identify more precise pathways. Longitudinal within-person designs can further separate immediate surprise from gradual changes in familiarity, preference, and expected value.

A third priority is a science of the whole musical encounter. Live performance, social coordination, movement, visual attention, memory, and artistic narratives can alter aesthetic response, yet these factors are often simplified away. Tarr et al. (2015) demonstrates measurable consequences of synchronized movement, and Brattico et al. (2026) identifies emerging research on naturalistic and shared encounters. Future studies should retain experimental contrasts while using multimodal measurement and culturally meaningful repertoire. Such work would clarify which forms of musical value generalize beyond short recordings, and what changes when human performance is mediated by artificial interactive systems.

Evidence Synthesis and Scope of the Findings

At the perceptual level, the presence of learned musical expectations is supported by rhythm discrimination, beat-related neural responses, and responses to harmonic violations. Hannon & Trehub (2005), Nozaradan et al. (2011), and Steinbeis et al. (2006) employ complementary methods and populations. At the aesthetic level, controlled rhythmic and harmonic work demonstrates that pleasure depends on relational properties such as degree of syncopation and the interaction of uncertainty with realized surprise Witek et al. (2014), Cheung et al. (2019). These effects are conditional on stimulus construction and participants' repertoires, so they do not provide a single species-wide formula for beautiful music.

At the neural valuation level, Blood & Zatorre (2001) and Salimpoor et al. (2011) connect intense pleasurable episodes with reward circuitry and striatal dopamine, Salimpoor et al. (2013) relates auditory–striatal interactions to willingness to pay, and Mas-Herrero et al. (2018) provides causal evidence for a fronto-striatal contribution under its stimulation protocol. Gold, Mas-Herrero et al. (2019) identifies value-learning signals in a specific experimental choice task, while de Fleurian et al. (2019) illustrates the need to interpret these findings without extending them indiscriminately to all listening. Together, the studies establish a research program with distinct observational, neurochemical, computational, and perturbational constraints.

At the level of emotion, culture, and social use, evidence becomes more heterogeneous but indispensable to the object of music aesthetics. The differentiation of music-evoked emotions Zentner et al. (2008), being moved by sad music Vuoskoski & Eerola (2017), intercultural differences in consonance preference McDermott et al. (2016), and synchronized dance effects Tarr et al. (2015) show that aesthetic reward depends on what is valued, remembered, understood, and shared. The major methodological challenge is to integrate these dimensions in models that preserve their distinct measurements. A complete account must explain the relationship among layers as rigorously as it explains the success of each specialized experiment.

Conclusion

Music aesthetics brings auditory organization, temporal anticipation, emotional experience, and reward valuation into a single empirical field. Musical rhythm makes future positions perceptually available; learned melodic and harmonic regularities establish probabilistic expectations; deviations, delays, fulfillment, and uncertainty can shape tension, interest, and pleasure. Experimental findings show that these relationships are structured and context dependent. The pleasure associated with syncopation can vary nonlinearly with rhythmic complexity, and harmonic enjoyment can depend jointly on uncertainty before an event and surprise after it. Such results explain a substantial part of musical engagement while leaving room for memory, culture, expression, and artistic judgment.

Neuroscientific research connects intense musical pleasure with interactions among auditory, cortical, and reward systems. Evidence includes striatal dopamine release, functional connectivity during valuation, reward prediction-error signals in decision learning, and bidirectional effects of fronto-striatal stimulation. The convergence is meaningful because the methods constrain different parts of the explanation. Their differences are equally informative: a neurochemical signal accompanying chills, a listener's monetary valuation, and a computational model of reward learning are distinct measurements of related processes. Scientific synthesis preserves those distinctions and tests their relationships.

The central conclusion is that emotional reward in music arises from the organization of a listening encounter across perceptual, predictive, affective, bodily, cultural, and social processes. The mechanisms by which music becomes pleasing or moving are measurable through explicit constructs and experimentally differentiated evidence. Progress requires accurate models of musical expectation, culturally diverse data, transparent replications, causal investigation of valuation networks, and methods capable of studying complete performances and collaborative listening. In the Artificial Era, computational musical prediction and interactive performance expand the experimental possibilities, while research on experienced aesthetic reward continues to demand independent observable evidence. The present dossier provides a critical narrative synthesis and a traceable foundation for those investigations.

References

  • Blood, A. J., & Zatorre, R. J. (2001). Intensely pleasurable responses to music correlate with activity in brain regions implicated in reward and emotion. Proceedings of the National Academy of Sciences, 98(20), 11818–11823. https://doi.org/10.1073/pnas.191355898
  • Brattico, E., Marzano, D., & Carraturo, G. (2026). Music as a form of art: Current findings and challenges of music neuroaesthetics research. Current Opinion in Psychology, 73, Article 102407 (online ahead of issue). https://doi.org/10.1016/j.copsyc.2026.102407
  • Cheung, V. K. M., Harrison, P. M. C., Meyer, L., Pearce, M. T., Haynes, J.-D., & Koelsch, S. (2019). Uncertainty and surprise jointly predict musical pleasure and amygdala, hippocampus, and auditory cortex activity. Current Biology, 29(23), 4084–4092.e4. https://doi.org/10.1016/j.cub.2019.09.067
  • de Fleurian, R., Harrison, P. M. C., Pearce, M. T., & Quiroga-Martinez, D. R. (2019). Reward prediction tells us less than expected about musical pleasure. Proceedings of the National Academy of Sciences, 116(42), 20813–20814. https://doi.org/10.1073/pnas.1913244116
  • Gold, B. P., Mas-Herrero, E., Zeighami, Y., Benovoy, M., Dagher, A., & Zatorre, R. J. (2019). Musical reward prediction errors engage the nucleus accumbens and motivate learning. Proceedings of the National Academy of Sciences, 116(8), 3310–3315. https://doi.org/10.1073/pnas.1809855116
  • Gold, B. P., Pearce, M. T., Mas-Herrero, E., Dagher, A., & Zatorre, R. J. (2019). Predictability and uncertainty in the pleasure of music: A reward for learning? Journal of Neuroscience, 39(47), 9397–9409. https://doi.org/10.1523/JNEUROSCI.0428-19.2019
  • Hannon, E. E., & Trehub, S. E. (2005). Tuning in to musical rhythms: Infants learn more readily than adults. Proceedings of the National Academy of Sciences, 102(35), 12639–12643. https://doi.org/10.1073/pnas.0504254102
  • Huron, D. (2006). Sweet anticipation: Music and the psychology of expectation. The MIT Press. https://mitpress.mit.edu/9780262083454/sweet-anticipation/
  • Juslin, P. N., & Västfjäll, D. (2008). Emotional responses to music: The need to consider underlying mechanisms. Behavioral and Brain Sciences, 31(5), 559–575. https://doi.org/10.1017/S0140525X08005293
  • Mas-Herrero, E., Zatorre, R. J., Rodriguez-Fornells, A., & Marco-Pallarés, J. (2014). Dissociation between musical and monetary reward responses in specific musical anhedonia. Current Biology, 24(6), 699–704. https://doi.org/10.1016/j.cub.2014.01.068
  • Mas-Herrero, E., Dagher, A., & Zatorre, R. J. (2018). Modulating musical reward sensitivity up and down with transcranial magnetic stimulation. Nature Human Behaviour, 2, 27–32. https://doi.org/10.1038/s41562-017-0241-z
  • McDermott, J. H., Schultz, A. F., Undurraga, E. A., & Godoy, R. A. (2016). Indifference to dissonance in native Amazonians reveals cultural variation in music perception. Nature, 535, 547–550. https://doi.org/10.1038/nature18635
  • Mehr, S. A., Singh, M., Knox, D., Ketter, D. M., Pickens-Jones, D., Atwood, S., Lucas, C., Jacoby, N., Egner, A. A., Hopkins, E. J., Howard, R. M., Hartshorne, J. K., Jennings, M. V., Simson, J., Bainbridge, C. M., Pinker, S., O’Donnell, T. J., Krasnow, M. M., & Glowacki, L. (2019). Universality and diversity in human song. Science, 366(6468), Article eaax0868. https://doi.org/10.1126/science.aax0868
  • Meyer, L. B. (1956). Emotion and meaning in music. University of Chicago Press. https://press.uchicago.edu/ucp/books/book/chicago/E/bo28551887.html
  • Nadal, M., & Skov, M. (2025). The sensory valuation account of aesthetic experience. Nature Reviews Psychology, 4, 49–63. https://doi.org/10.1038/s44159-024-00385-y
  • Nozaradan, S., Peretz, I., Missal, M., & Mouraux, A. (2011). Tagging the neuronal entrainment to beat and meter. Journal of Neuroscience, 31(28), 10234–10240. https://doi.org/10.1523/JNEUROSCI.0411-11.2011
  • Saarikallio, S., & Erkkilä, J. (2007). The role of music in adolescents’ mood regulation. Psychology of Music, 35(1), 88–109. https://doi.org/10.1177/0305735607068889
  • Salimpoor, V. N., Benovoy, M., Larcher, K., Dagher, A., & Zatorre, R. J. (2011). Anatomically distinct dopamine release during anticipation and experience of peak emotion to music. Nature Neuroscience, 14, 257–262. https://doi.org/10.1038/nn.2726
  • Salimpoor, V. N., van den Bosch, I., Kovacevic, N., McIntosh, A. R., Dagher, A., & Zatorre, R. J. (2013). Interactions between the nucleus accumbens and auditory cortices predict music reward value. Science, 340(6129), 216–219. https://doi.org/10.1126/science.1231059
  • Savage, P. E., Brown, S., Sakai, E., & Currie, T. E. (2015). Statistical universals reveal the structures and functions of human music. Proceedings of the National Academy of Sciences, 112(29), 8987–8992. https://doi.org/10.1073/pnas.1414495112
  • Schubert, E. (2016). Enjoying sad music: Paradox or parallel processes? Frontiers in Human Neuroscience, 10, Article 312. https://doi.org/10.3389/fnhum.2016.00312
  • Steinbeis, N., Koelsch, S., & Sloboda, J. A. (2006). The role of harmonic expectancy violations in musical emotions: Evidence from subjective, physiological, and neural responses. Journal of Cognitive Neuroscience, 18(8), 1380–1393. https://doi.org/10.1162/jocn.2006.18.8.1380
  • Tarr, B., Launay, J., & Dunbar, R. I. M. (2014). Music and social bonding: ‘Self-other’ merging and neurohormonal mechanisms. Frontiers in Psychology, 5, Article 1096. https://doi.org/10.3389/fpsyg.2014.01096
  • Tarr, B., Launay, J., Cohen, E., & Dunbar, R. (2015). Synchrony and exertion during dance independently raise pain threshold and encourage social bonding. Biology Letters, 11(10), Article 20150767. https://doi.org/10.1098/rsbl.2015.0767
  • Trost, W., Ethofer, T., Zentner, M., & Vuilleumier, P. (2012). Mapping aesthetic musical emotions in the brain. Cerebral Cortex, 22(12), 2769–2783. https://doi.org/10.1093/cercor/bhr353
  • Vuoskoski, J. K., & Eerola, T. (2017). The pleasure evoked by sad music is mediated by feelings of being moved. Frontiers in Psychology, 8, Article 439. https://doi.org/10.3389/fpsyg.2017.00439
  • Vuust, P., & Witek, M. A. G. (2014). Rhythmic complexity and predictive coding: A novel approach to modeling rhythm and meter perception in music. Frontiers in Psychology, 5, Article 1111. https://doi.org/10.3389/fpsyg.2014.01111
  • Winkler, I., Háden, G. P., Ladinig, O., Sziller, I., & Honing, H. (2009). Newborn infants detect the beat in music. Proceedings of the National Academy of Sciences, 106(7), 2468–2471. https://doi.org/10.1073/pnas.0809035106
  • Witek, M. A. G., Clarke, E. F., Wallentin, M., Kringelbach, M. L., & Vuust, P. (2014). Syncopation, body-movement and pleasure in groove music. PLOS ONE, 9(4), Article e94446. https://doi.org/10.1371/journal.pone.0094446
  • Zatorre, R. J., & Salimpoor, V. N. (2013). From perception to pleasure: Music and its neural substrates. Proceedings of the National Academy of Sciences, 110(Suppl. 2), 10430–10437. https://doi.org/10.1073/pnas.1301228110
  • Zentner, M., Grandjean, D., & Scherer, K. R. (2008). Emotions evoked by the sound of music: Characterization, classification, and measurement. Emotion, 8(4), 494–521. https://doi.org/10.1037/1528-3542.8.4.494