Music as Memory: Toward a Theory of Meta-Structure

Abstract

Music and memory are intrinsically intertwined, yet the nature of their relationship remains poorly understood. This article explores how auditory memory shapes the perception and experience of music, and how music may, in turn, influence memory more broadly. We propose a theoretical framework developed around four conceptual axes: music is virtual, music is future, music is embodied, and music is sensory. These dimensions are explored in terms of the acoustic, perceptual, and cognitive features of music, which require a particular kind of temporally extended perceptual and cognitive processing. We argue that music, as a complex and multifaceted experience, engages the activation and interaction of multiple brain functions as well as bodily arousal and motor activity, interconnecting them with the auditory system. From this distributed and embodied engagement emerges an interacting system, forming a functional meta-structure that extends beyond auditory processing. We suggest that this meta-structure may act as a “memory amplifier”: its temporal, embodied, sensory, and imagery characteristics may provide an alternative means of cognitive processing through music, granting access to multiple brain areas and functions even in circumstances where other cognitive faculties fail. We further propose that the repeated reactivation of this meta-structure may actively contribute to the consolidation and preservation of memory. The meta-structure might therefore constitute a strong case for the evolutionary value of music as a universal human capacity. In essence, music perception and cognition represent a form of memory: an ottofixation—a term coined to capture this capacity of musical experience to function as both a vehicle and an architecture for memory.

Music and Memory

Music and memory are inherently intertwined in ways that are still not clearly understood. Memory has an undeniable impact on our experience of music, influencing how we perceive boundaries and patterns within its structure over time. Musicality, understood as the capacity to make sense of and to be moved by music (Cross, 2012), also impacts neural processes and cognitive operations, granting access to memory and other cognitive functions even in the context of memory loss and other cognitive dysfunctions. But how is this relationship between music and memory established and structured?

Human memory is a complex phenomenon, difficult to decipher, that both encompasses the storage of information at different levels of abstraction and influences how we perceive, comprehend, and experience the world. From a physiological perspective, memory can be viewed as lasting changes in the nervous system, including permanent chemical modifications within nerve cells, the ability of neurons to alter the structure of their synaptic connections with each other, adjusting in both number and strength (stability), in ways that last over time (Snyder, 2000). From a cognitive perspective, memory is often described in terms of complementary systems, including declarative (explicit) memory for facts and events and nondeclarative (implicit) memory for skills and habits (e.g., Kandel et al., 2013; Squire, 2004). Together, these systems allow us to encode, retain, and retrieve information over time.

Within this broader architecture of human memory, sensory memory provides a very brief retention of incoming information in each modality (Snyder, 2000). For audition, this sensory store is referred to as echoic memory—a kind of echo—operating on an extremely short timescale, typically lasting only milliseconds (Snyder, 2000). Echoic memory, a sensory impression of sound, apprehends auditory information that is believed to persist in a continuous or “analogue” form, which has not yet undergone any kind of coding or processing: pre-categorical, pre-conceptual, raw (Snyder, 2000). Beyond this initial sensory stage, auditory memory also includes short-term or working memory, memory of the most recent past, that retains information for a few seconds, depending on the complexity and novelty of the information (Cowan, 2008). This type of memory does not cause permanent chemical changes in the brain, as long-term memories are thought to do. However, if repeated, short-term memories can be permanently inscribed in long-term memory. Finally, long-term memory retains all the knowledge acquired through auditory experience in a persistent way (Snyder, 2000). These types of permanent memories, formed through mere exposure, provide a usually unconscious contextual background for current musical experience, functioning as implicit knowledge. At the same time, they support explicit associations, whether involuntary or deliberate, with other musical memories and with memories in other modalities, particularly autobiographical memories (Kumar et al., 2016). Together, these interacting systems are embedded in wider networks for attention, perception, and decision-making (Kumar et al., 2016; Levitin, 2002) and underpin our ability to perceive, organize, and remember complex auditory events such as music.

Musical memory can be considered a particular case of auditory memory. It includes the ability to encode, store, and retrieve highly structured musical information such as melodies, harmonic progressions, rhythmic patterns, and timbres (Snyder, 2000; Tillmann et al., 2011). Musical memory allows us to recognize and reproduce a vast number of melodies and styles, often over many years, and it is closely linked to both implicit knowledge (e.g., expectations about tonal and rhythmic regularities) and explicit recollection of specific pieces or performances (Levitin, 2002; Snyder, 2000; Williamon et al., 2004).

From a functional perspective, the three processes of auditory memory described above can be applied directly to musical information. Echoic memory provides the immediate sensory trace of incoming sound. Short-term or working memory retains recent musical events over seconds, allowing the integration of tones and chords into local patterns. Long-term memory stores knowledge built up over repeated exposure to musical systems, such as tonal hierarchies, rhythmic schemas, and stylistic conventions (Edelman & Tononi, 2013; Vuust et al., 2022).

Memory is intrinsically linked to other cognitive functions such as attention, by selecting stimuli to be encoded and stored; perception, by providing a context for interpreting sensory stimuli; representation, by enabling the creation of mental models of the world; categorization, by organizing information into structured systems that facilitate the retrieval and efficient use of knowledge; and consciousness, by constructing and sustaining the continuity of identity (Kandel et al., 2013; Levitin, 2002). Given these interconnections and interdependencies, memory loss can have wide-ranging effects, compromising multiple cognitive capacities simultaneously.

Clinical and experimental studies indicate that musical abilities can show a relative preservation even when other cognitive functions are severely compromised. Dementia, a generic term that includes Alzheimer's disease, involves the progressive deterioration of memory, thinking and reasoning, orientation in time and space, and motor and language functions. Amnesia can arise from brain damage, illness, or emotional trauma, resulting in partial or total loss of the ability to recall past information or experiences (Kandel et al., 2013). Nevertheless, patients with Alzheimer's disease or other forms of dementia and certain amnesic syndromes frequently retain the ability to recognize and respond to familiar music, even when other cognitive skills are significantly compromised (Baird & Samson, 2015; Cuddy & Duffin, 2005; El Haj et al., 2012; Finke et al., 2012; Kaiser & Berntsen, 2023; Sánchez-Herrero et al., 2024). Neuroimaging studies suggest that familiar or personally meaningful music engages medial prefrontal, limbic, and motor regions that can remain relatively preserved or show compensatory connectivity in dementia, providing a possible neural basis for this relative resilience of musical abilities (e.g., Jacobsen et al., 2015; King et al., 2019). Listening to meaningful melodies can help these patients recover memories associated with personal experiences, autobiographical memories, which may otherwise remain inaccessible due to the progression of the disease (El Haj er al., 2012; Kaiser & Berntsen, 2023). In addition, musical experience can offer an alternative path to accessing other cognitive functions. For instance, music therapy has demonstrated effectiveness in promoting speech recovery for patients with aphasia. Music can also improve locomotion, motor coordination, and synchronization in individuals with neurodegenerative diseases (e.g., Parkinson's disease), facilitating movement through familiar rhythms and melodies (Altenmüller et al., 2015; Vuust et al., 2022).

This accumulated knowledge, supported by research and methodically articulated as presented here, calls for exploring why musical capacity remains, even when other capacities on which musicality relies, such as memory, seem to be lost. Could musicality have privileged access to memory? And, if so, why would this be? Why are music and memory indissociable, and how do they interact with and affect each other?

Therefore, in this article, we analyze, in a speculative exercise, why auditory memory may affect the perception and experience of music and, on the other hand, why musicality, as the set of cognitive and perceptual processes that underlie musical behavior, may affect memory in general. We present a hypothesis supported by four key proposals: music is virtual, music is future, music is embodied, and music is sensory.

Music is Virtual

Music is a complex sequence of acoustic events, often characterized by being extremely structured and hierarchical, which takes place on a timeline. The temporal nature of music suggests that, perceptually (and sensorially), we can only have access to a fraction of what constitutes the whole musical event, a fraction that is suspended in our echoic memory. As such, we never access music in its entirety, as a whole.

From the perceptual perspective, this chain of temporal events is primarily characterized by the analysis of its spectra and how it changes over time. Therefore, apart from timbre, which requires the simultaneous extraction of the spectral composition of sound (a frequency-to-place mapping into the basilar membrane (Plack, 2014)), the perceptual processing of acoustic information, such as music, relies on a sequential parsing of events over time (Joris et al., 2004). At a fundamental structural level, this involves the detection of the beginning of events (onsets), their termination (offsets), the perception of brief silent intervals between sounds (gap detection) and modulations in sound features such as frequency, intensity or spectral content, which all together track sound durations and sequences (Haumann et al., 2021; Irsik et al., 2021; Meehan et al., 2024).

However, it is not possible to extract the musical structure solely by analyzing the immediate information that is being acquired. Understanding the musical structure requires articulating newly acquired information with past events and projecting future events (Huron, 2006; Snyder, 2000; Vuust et al., 2022). The perceptual and cognitive processing of this sequence of acoustic events thus involves processing the local (present) and the global, implying comparisons and associations with the past and conjectures for the future (Snyder, 2000; Tillmann et al., 2011). It involves the coding of individual elements acquired through echoic memory and combining this information with data stored in short- and long-term memory (Cowan, 2008; Snyder, 2000). This integration is crucial in assessing and predicting the immediately following event and understanding the overall structure of the stimulus (Huron, 2006; Vuust et al., 2022).

Music experience is therefore directly linked to and dependent on memory. Music experience does not exist without memory. Music exists virtually in the present, as a suspension: the acquisition of the auditory world is only an impression. Everything else is an interference, a construction of memory, and a hypothesis for the future.

Music is Future

Music unveils in time. The temporal structure of the acoustic world has implications for music perception and experience, as mentioned previously. The basic structural level of detection of durations and sequences, such as the detection of onsets, offsets, gaps, and the modulation of sound features are not enough to process patterns and sequences (rhythmic and melodic) in the more complex context of music. Auditory memory is crucial for this process.

Auditory memory, particularly long-term memory, results from the accumulation of knowledge built up through mere exposure to the acoustic environment that surrounds us. Thus, auditory memory is a constructive process, an editing territory that generates a subliminal unconscious context in constant interference with current musical perceptions (Bianco et al., 2020; Snyder, 2000).

In this process of accumulating information, implicitly and through exposure, rules underlying the structure of sound information are extracted: regularities, hierarchical structures, schematic knowledge of the material, for example. It is with this information that a system of perceptual representation of the musical system is built. These long-term memory representations incorporate invariant properties of auditory stimuli, which allow the creation of categories and concepts (Edelman & Tononi, 2013; Hawkins & Blakeslee, 2004). Accessing long-term memory implies accessing this representational system. The musical experience shapes these representations and these representations shape the musical experience.

These representations do not just have an impact on the immediate. They also shape the way future events are predicted. Organisms are equipped with the ability of anticipation with the purpose of preparing them for the future (Friston, 2010; Huron, 2006). Anticipating future events means preparing actions and perceptions that are better suited to the context. Predicting the future is an evolutionary biological adaptation. Musical experience, due to its temporal structure, can only happen with the anticipation of the future auditory event (Vuust et al., 2022). At the same time, musical experience provides the environment and the structure for these anticipations to be acquired and applied.

Anticipating a certain event involves creating an expectation about when and how that event will occur. The outcome of such expectations depends on the degree to which the initial anticipation is fulfilled or not. From a phenomenological perspective, expectations are a subjective experience that generates emotions. Emotions, triggered by expectations, reward correct anticipations, promote preparation for future events, and increase the likelihood of favorable future outcomes (Huron, 2006). Music perception can only happen by producing expectations regarding the future auditory event. It is through these expectations that we experience pleasure, surprise, and strangeness while listening. Therefore, there can be no music experience without expectation and, consequently, no music experience without emotion.

The consolidation of auditory memory depends on the structure of the information to be memorized—structure in terms of regularities, schemes, and hierarchical organization, but also the structure or emotional charge of the stimulus. Environmental information that induces a change in the internal affective state of the organism—an ‘emotional stimulus’—affects the quantity and quality of information that is retained. The cognitive mechanisms of memory favor the encoding and retrieval of emotionally charged information, whether positive or negative, to the detriment of nonemotional or “neutral” information, which is regarded as less relevant (Kensinger & Schacter, 2008). In this framework, neutrality is not a category of emotional response but the absence of one: either the stimulus elicits an internal affective change, in which case it is emotional, or it does not, in which case it is nonemotional.

In this context, music cannot be a neutral stimulus. Music perception, grounded in the temporal structure of sound, continuously points to the future and evokes expectations. These expectations create a demand for an ongoing emotional response. Musical experiences become, therefore, inevitably prioritized in the encoding and consolidation of memory: long-lasting impressions recorded in both the mind and the body.

Music is Embodied

Musical experience comprises a complex interplay of cognitive and perceptual processes that simultaneously activate multiple brain areas and networks, which are related to other vital functions (Koelsch, 2014; Vuust et al., 2022). In this sense, by embodied we mean that musical experience is not confined to auditory and “mental” processing, but engages motor systems, autonomic responses, and bodily reactions as integral parts of how music is perceived and understood. Consequently, music is a profoundly subjective human experience which is grounded in a range of cognitive and perceptual functions, including (but not limited to) memory, attention, planning, as well as motor control and coordination, involving brain structures implicated in action, emotion, and learning.

Throughout the nineteenth century, the predominant paradigm in neuroscience was influenced by studies of brain lesions, particularly the work of Broca and Wernicke. This approach, characterized by the principle of cerebral localization, which explored the concept of a modular and compartmentalized brain, posited that the brain is functionally segmented into distinct regions, each responsible for specific cognitive processes (Rutten, 2022). However, with the development of imaging technologies and the introduction of sophisticated tools such as electroencephalography (EEG), functional magnetic resonance imaging (fMRI), and magnetoencephalography (MEG), cognitive neuroscience has brought a considerably transformed perspective on how the brain functions. These new tools have fundamentally challenged the traditional notion of cerebral localization of human cognitive faculties, contradicting the previous idea of a single specific area of the brain dedicated to processing music. We are now faced with a more refined understanding of the brain, which reveals the distributed and integrated nature of brain functions and brings with it a conceptual change based on function versus structure. In this way, it is possible to observe that many cognitive and perceptual functions, including musicality, involve complex networks of interconnected brain areas, challenging the idea of a single area dedicated to each function (Vuust et al., 2022).

This distributed character of musical processing is not merely anatomical, but also the very condition that makes embodiment possible, since the recruitment of motor, autonomic, and limbic systems alongside auditory cortex is what allows musical experience to extend beyond hearing into the body. This conceptual model is particularly evident when examining how the brain processes the various dimensions of sound and musical experience. The processing of sound characteristics, including pitch, rhythm, and timbre, primarily occurs within the auditory cortex. However, other areas and circuits are also recruited, engaging many other cortical and subcortical regions. Examples are the motor cortex and cerebellum, which play pivotal roles in the execution and coordination of movements; the parietal cortex, which contributes to spatial perception and attentional mechanisms; the frontal and prefrontal cortex, implicated in higher-order functions such as planning and decision-making; areas associated with speech, such as the Broca and Wernicke areas; and regions involved in visual processing such as the occipital lobe (Loui & Przysinda, 2017; Peretz & Zatorre, 2005; Warren, 2008; Zatorre et al., 2007). Work on entrainment and sensorimotor synchronization further shows that listeners’ movements and internal timing tend to align with musical rhythms, highlighting the tight coupling between auditory perception and bodily action in music (Demos et al., 2014; Large & Jones, 1999; Repp & Su, 2013).

Music experience also activates complex brain structures. The limbic system, related to emotion, motivation, and basic vital functions for survival, interconnects auditory experiences with areas related to emotion processing and regulation (the amygdala and the cingulate gyrus) and memory formation (the hippocampus). The reward system, a complex network of brain structures related to basic survival functions also activated by musical experience, includes a network of brain structures (e.g., the prefrontal cortex, the nucleus accumbens, and dopaminergic pathways) with the main function of regulation through pleasure and reward (Brown et al., 2004; Koelsch, 2010; Menon & Levitin, 2005; Zatorre, 2024). The activation of these brain structures is accompanied by bodily manifestations. Music-evoked bodily reactions comprise physiological changes in heart rate, blood pressure, respiration, electrodermal activity, and posture or movement, which further illustrates that musical experience is instantiated in the body as well as in the brain (Juslin & Västfjäll, 2008; Koelsch, 2010; Tschacher et al., 2023).

The multiple activation and interaction of different brain areas and networks reveal music perception as a complex and multifaceted experience that interconnects the primary auditory cortex with different vital functions and consequently arouses the body. This embodied engagement, recruiting emotional, autonomic, and motor systems alongside auditory and cognitive processing, contributes to the richness and durability of musical memories. Musical experience, thus, acts like a parasite in the cognitive system, which builds on existing cognitive functions and, in doing so, transforms the way we experience the world. In this sense, music operates as what Patel (2008) describes as a “transformative technology of the mind.” Therefore, there is no single, specific area of the brain dedicated to processing music. Instead, we engage with music with our whole body, our whole brain—not just our ears. Cognitively, music is not just hearing, it is an embodied experience. Music experience is multimodal and multisensory.

Music is Sensory

Auditory memory has a peculiarity: remembering an auditory event involves fundamentally different mental processes than remembering the capital of a country or a mathematical formula. The latter examples, which are propositional, do not seem to engage any specific sensory modality. Conversely, auditory memory is intrinsically sensory. The process of recalling an auditory event requires imagining the original experience, simulating it as if it were being lived. It is as if this process of auditory imagery gave us access to that experience, independently of it being induced by external sensory information. Remembering an auditory event therefore engages the sensory system involved in the original experience, which is responsible for coding the mental representation of that information (Zatorre et al., 1996) and this reactivation can also give rise to the corresponding bodily reactions, reflecting the embodied experience previously discussed.

Although music can often be associated with linguistic information, such as lyrics, titles, semantic facts or autobiographical narratives that are verbally expressible, the core musical content itself is primarily nonlinguistic. Similarly to muscle memory, musical memories are pre-linguistic memories that do not possess a linguistic component and thus do not match any abstract linguistic category. Therefore, musical imagery processes seem to require mechanisms for representing sensory information in a continuous, “analog” format, which is impossible to discretize or whose linguistic description, in its entirety, inherently lacks a precise definition (Pylyshyn, 1981).

Thus, access to auditory memory requires imagery processes of auditory events that may be mediated by neuronal mechanisms similar to those used in the perception of the same events, activated endogenously, in the absence of external sensory stimulation. Such imagery processes not only enable us to recall past experiences but also allow us to manipulate these representations, enabling us to imagine experiences we have not encountered before, including potential future scenarios.

Because musical memories are encoded in a continuous, sensory format that preserves the contextual richness of the original experience, they are especially effective at recalling the multimodal and emotional texture of past events. This is precisely what characterizes autobiographical recall. Musical experience is also a catalyst of memories associated with personal experiences, where certain events, people, places, and emotions are evoked. We thus associate music with episodic and autobiographical memories (Svoboda et al., 2006). These memories are characterized by being involuntary and from distinct areas of experience. Notably, such memory activations by music are observed not only in the general population but also in individuals with dementia and other memory disorders. In many cases, musical engagement allows access to memories that were previously thought to be irretrievably lost, which would otherwise be impossible to recover (El Haj et al., 2012).

Like musical experience itself, autobiographical and episodic memories also co-activate an interconnected complex network of brain regions that integrate sensory, emotional, and contextual information (see, e.g., Svoboda et al., 2006). When triggered by music, these involuntary autobiographical memories tend to be especially vivid, rich in sensory detail, and emotionally intense, reflecting the multimodal and distributed nature of such recollections (Belfi & Jakubowski, 2021). Moreover, it is plausible that the repeated co-activation of such complex networks may, over time, facilitate their functional expansion, recruiting additional cortical and subcortical regions into the memory trace. This expansion could, in turn, reinforce the robustness and durability of the encoded memories themselves. Such a mechanism would be consistent with evidence that emotionally salient and personally meaningful musical experiences produce stronger and more enduring memory traces than neutral stimuli, as discussed earlier, and with the broader principle that richly interconnected neural representations tend to be more resistant to forgetting (Nadel & Moscovitch, 1997). If this is the case, music may not merely serve as a retrieval cue for pre-existing memories, but also actively contribute to the consolidation and preservation of autobiographical knowledge by expanding the very neural architecture that sustains it. In this way, the catalyzing force of music reveals an intrinsic relationship between musical experience and memory.

Conclusion

Supported by four main arguments, we have explored acoustic, perceptual, and cognitive features of music that may contribute to understanding of the relationship between memory and music. Specifically, these arguments may contribute to addressing, on the one hand, why auditory memory may interfere with the perception and experience of music and, on the other hand, why music perception and experience may affect memory in general.

First, we have highlighted the temporal acoustic nature of music. Because of this temporal nature, music is never available to the perceptual apparatus in its totality. As such, interpreting and understanding music (and its structure) demands a specific sort of perceptual and cognitive processing which is intricately tied to memory: attending to the local (present) and the global, implying comparisons and associations with the past and forming conjectures for the future. Therefore, past acoustic events must be stored in memory—music cannot be interpreted or exist cognitively without memory.

The need to anticipate future acoustic events, rooted in previously acquired knowledge, stored in memory, reveals another cognitive processing form intricately tied to memory. These anticipations lead to the formation of expectations, which inevitably result in an emotional charge of musical experience. This nonneutrality of music makes it a priority stimulus for memory encoding and retrieval.

Moreover, musical experience co-activates multiple brain areas and networks, interconnecting the auditory system with a vast diversity of functions, some of them quite primitive (e.g., the limbic and reward systems), related to basic biological functions and survival. This distributed character of musicality expands to embodied experiences and has direct consequences for memory. Because musical experiences are encoded not only as acoustic patterns but also as patterns of bodily arousal, motor patterns, emotions, sensory information and contextual states, they generate rich, multicomponent memory traces with multiple routes for retrieval.

In this sense, musicality's distributed engagement can be seen as an interacting system, an intricate set of communicating vessels forming a functional meta-structure that extends music perception and cognitive processing to brain regions, vital functions, and embodied experience, beyond the auditory system. This interconnected quality might be determinant on three levels. First, this meta-structure may be spreading to other cognitive brain functions the particular auditory processing that the temporal acoustic feature of music requires, extending its memory amplifier power for its temporal, sensory, and imagery characteristics and assigning an alternative means of cognitive processing, through music. Second, this meta-structure, which connects the cognitive with the embodied experience of music may have direct consequences for memory, modulating arousal and reward, processes known to facilitate the encoding and consolidation of salient events. Third, and also as a consequence of what was said, this meta-structure might be granting an alternative access, through this “auditory cognitive processing,” to multiple brain areas and functions, in circumstances where other cognitive faculties fail.

In addition, recalling a musical experience requires imagining it in its entirety, simulating it and engaging the sensory system, the cognitive apparatus and bodily arousal involved in the original experience. Evoking musical experiences requires reactivating the same brain regions, network circuits, and embodied responses involved in the original experience—the meta-structure referred to previously. Consequently, the recall of musical experience engages the sensory system, demands the reactivation of the meta-structure that includes other cognitive functions and may induce the same physiological (bodily) reaction patterns and trigger the same emotional reactions as the original experience. The reactivation of these interconnected complex networks, driven by the sensory features of musical experience, may actively contribute to the consolidation and preservation of memory.

Finally, the meta-structure that musical experience (whether intrinsic or extrinsically evoked) requires, in which a particular kind of cognitive processing runs, may provide a strong hypothesis for justifying its evolutionary value in order to become a universal human capacity. Its processing triggers a mental (cognitive) structure that bridges different brain (and vital) functions and fosters memory across those functions. The meta-structure, thus, acts as a memory amplifier and enhancer, providing an alternative access, through music, to multiple brain areas and functions, in circumstances where other cognitive faculties fail.

This may also be why musicality so often endures where other faculties falter. In dementia, amnesia, aphasia, or Parkinson's disease, the meta-structure still holds, continuing to bind the sensory to the bodily, the remembered to the anticipated, the cognitive to the emotional, and offering a route back to memory, recognition, movement, and self when other routes have closed. Music as virtual, as future, as embodied, as sensory: the four axes converge precisely here, each a thread of the meta-structure, in the preservation of access when other faculties fail.

In essence, music perception and cognition represent a form of memory: an ottofixation—from the Greek otos (ear) and the Latin fixatio (fixing, anchoring)—denoting the capacity of musical experience, through its temporal, embodied, sensory, and predictive nature, to function as both a vehicle and an architecture for memory.