Music-evoked autobiographical memories (MEAMs) refer to the recollection of personal experiences or past events that are triggered when hearing music or some musical stimulus. While there is a degree of inter-individual variation in music listening patterns and evoked responses, MEAMs are generally triggered in response to a wide variety of music, often popular or classical genres, and are estimated to occur in the range from one to a few times per day, regardless of formal instrumental practice or music lessons. Consistent with the hallmarks of general autobiographical memories, everyday MEAMs similarly exhibit a recency effect, a reminiscence bump (later discussed in the section "The reminiscence bump"), and childhood amnesia, encoding autobiographical knowledge at several levels of specificity and across several common social and situational contexts. The phenomenon of MEAMs has been widely studied in the fields of psychology, neuroscience, and musicology. In recent years, the subject has garnered significant interest from researchers and the general public alike due to music's capacity to evoke vivid, emotional, and episodically rich autobiographical memories.
History The relationship between music and memory has long been recognized, with music's ability to elicit emotional responses and trigger memories dating back to ancient times. In ancient Graeco-Roman society, for instance, musical memory formed a fundamental part of social, cultural, and political life.
The scientific study of music's deep correlation with autobiographical memories gained prominence in the early 2000s, when the term "music-evoked autobiographical memory" was coined by cognitive neuroscientist Petr Janata and colleagues, who first described the phenomenon in healthy undergraduate students. At this time, Janata conducted a pioneering experiment in which his team developed a novel approach to studying autobiographical memory using a range of popular music. Analyses indicated that 30% of the presented songs evoked autobiographical memories, demonstrating the paradigm's efficacy and justifying its use in subsequent neuroimaging work.
Since Janata's work, numerous studies have contributed to the understanding of MEAMs. Researchers have investigated factors that influence the formation and retrieval of MEAMs, from the characteristics of the music stimulus to its personal relevance, with many contemporary studies citing his work.
Recording techniques
fMRI Functional magnetic resonance imaging (fMRI) serves as a neuroimaging technique which is used by researchers aim to identify the neural mechanisms underlying this experience and clarify the interplay between music, memory, and emotions.
In studies exploring MEAMs, participants are typically placed inside an MRI scanner while they listen to music that holds personal significance to them. The fMRI scanner measures the blood oxygen level-dependent (BOLD) signals, which provide an indirect measure of neural activity. Researchers analyze the fMRI data to identify brain regions that exhibit increased activity during MEAM retrieval compared to control conditions. These regions are believed to play crucial roles in memory processing, emotional responses, and self-referential processing.
Beyond fMRI measurements, studies on MEAMs often incorporate additional behavioral measures such as questionnaires or interviews to capture the subjective experience and vividness of the retrieved memories. These measures help validate the connection between music and autobiographical memories and provide a richer understanding of the phenomenon.
Since fMRI measures blood oxygenation level-dependent signals, this showcases an indirect line to neural activity, which can be seen as a limitation to this recording system.
EEG Electroencephalography (EEG) has emerged as a valuable tool in the investigation of MEAMs. EEG allows researchers to examine the temporal dynamics of neural activity associated with MEAMs, providing insights into the underlying cognitive and emotional processes.
In studies focusing on MEAMs, participants wear a cap embedded with multiple electrodes that record electrical brain activity. EEG captures the millisecond-level fluctuations in neural oscillations, revealing the brain's real-time response to MEAMs. Researchers analyze the EEG data to identify specific patterns of brain activity, such as event-related potentials (ERPs) and oscillatory activity.
ERPs, derived from averaging the EEG signal time-locked to specific events, offer insights into the brain's response to musical stimuli. In MEAM research, researchers examine ERPs related to memory retrieval processes, such as the P300 component, which reflects attention allocation and memory updating. By comparing ERPs between music-evoked and control conditions, researchers gain insights into the neural mechanisms underlying MEAMs.
Oscillatory activity, as measured by EEG, provides further understanding of the neural dynamics during MEAMs. Researchers investigate changes in frequency bands, such as alpha, beta, and theta, associated with memory processes and emotional responses. For example, enhanced theta activity has been observed during successful memory retrieval, suggesting its involvement in encoding and recollection of music-evoked autobiographical memories.
EEGs have a hard time localizing the signals they receive from the brain which brings forth a limitation of being unable to capture information in a precise manner.
Deep learning architecture Deep learning has also been a tool that is used to study MEAMs. Deep learning, a subset of machine learning, involves training neural networks with multiple layers to automatically learn and extract complex patterns from data. In the context of MEAMs, deep learning architectures are used to analyze and model the relationships between music and autobiographical memories, providing valuable insights into the underlying mechanisms.
Researchers employ deep learning architectures, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), to process and analyze musical features and associated autobiographical memory data. CNNs excel at capturing spatial patterns and are commonly used to analyze spectrograms or other representations of music. By training CNNs on large datasets of music and associated memory information, researchers can uncover patterns and correlations that may be indicative of MEAMs.
RNNs, on the other hand, are specifically designed to model sequential data and are well-suited for analyzing the temporal aspects of music and memory. Researchers use RNNs to capture the dependencies and dynamics between musical elements and autobiographical memories. By training RNNs on sequences of music and associated memory descriptions, researchers can generate predictions and gain insights into the temporal dynamics of MEAMs.
Deep learning architectures in MEAM research are not limited to CNNs and RNNs alone. Variants such as long short-term memory (LSTM) networks and generative models, such as generative adversarial networks (GANs), are also employed. LSTM networks are capable of capturing long-range dependencies and are particularly useful when modeling the temporal aspects of music-evoked memories. GANs, on the other hand, can generate new music or memory samples based on learned patterns, allowing researchers to explore the generative aspects of MEAMs.
Memory type In general, musical memory can be split into both explicit and implicit memory systems. "Implicit" memory refers to the unconscious, automatic storage of information, encompassing procedural memories and motor skill learning. In the context of music, implicit memory accounts for the ability to play an instrument or sing a song without consciously bringing into memory how to play it. This differs from explicit memory, a declarative memory consisting of conscious recall of a specific piece of information. Explicit memory can be further split into episodic and semantic memory. Semantic memories are stored general knowledge and facts, accounting for a general sense of familiarity of music. In contrast, episodic memories are memory for the associated context of an event, accounting for the time, place, people, emotions, and personal experiences surrounding a musical experience. By these definitions, MEAMs are typically characterized as an explicit and episodic memory type. The distinct neural activation patterns associated with semantic vs episodic memories, as evidenced by a fMRI study looking into the neural correlates of music memory, helps explain some of the observed characteristics of MEAMs. This is further discussed in the next section "Brain regions involved".
Brain regions involved The act of listening to music elicits a wide range of neural responses and engages various cognitive processes, encompassing motor, auditory, emotional, memory, and reward-related functions within the brain. Processing of musical qualities in general such as timbre, tonality, and rhythm are associated with activation in temporal (superior and middle temporal gyri, insula), frontal (superior and middle frontal gyri, cingulate gyrus, precentral gyrus), parietal (inferior parietal gyrus, precuneus, postcentral gyrus), and cerebellar regions. In terms of forming and retrieving MEAMs specifically, the medial prefrontal cortex, encompassing both dorsal and ventral regions, emerges as a central hub. The dorsal MPFC responds to the autobiographical salience of music excerpts and tracks their movements through tonal space. This region exhibited a familiarity effect and showed properties of associating structural aspects of retrieval cues with episodic memories. Additionally, rostral and ventral aspects of the MPFC also exhibited responses related to familiarity, autobiographical salience, and positive valence.
The involvement of prefrontal regions in MEAMs was not limited to the MPFC. Both lateral and medial prefrontal areas showed responses, suggesting the usage of multiple memory retrieval processes to assemble semantic and episodic content into a recollective experience with an affective component. Also, the ventrolateral prefrontal cortex (VLPFC) has been implicated in processing musical structure, integrating musical events into syntactically coherent sequences.
Retrieval mechanisms
Involuntary vs voluntary retrieval MEAMs can occur in a variety of settings, both through deliberate and spontaneous recall. There has been some debate in the classification of MEAMs as voluntary or involuntary autobiographical memories. Voluntary autobiographical memories are "personal memories that follow a controlled, strategic retrieval process". Involuntary autobiographical memories are "memories of personal events that come to mind spontaneously-that is, with no conscious initiation of the retrieval process". They are considered to be more specific, retrieved significantly faster, and accompanied by a stronger emotional response than voluntary autobiographical memories.
Current research supports the idea that MEAMs aligns more often with the latter memory retrieval mechanism, as MEAMs are similarly more specific, retrieved faster, and are accompanied by greater emotional impact than memories retrieved in silence. A recent diary study also provides support, with a higher proportion of MEAMs being reported by participants to be involuntarily rather than deliberately retrieved. However, as 20% of participants in this study used deliberate recall, involuntary retrieval is not an exclusive feature of MEAMs.
Situational context