Abstract
Comparisons between involuntarily and voluntarily retrieved autobiographical memories have revealed similarities in encoding and maintenance, with differences in terms of specificity and emotional responses. Our study extended this research area into the domain of musical memory, which afforded a unique opportunity to compare the same memory as accessed both involuntarily and voluntarily. Specifically, we compared instances of involuntary musical imagery (INMI, or "earworms")—the spontaneous mental recall and repetition of a tune—to deliberate recall of the same tune as voluntary musical imagery (VMI) in terms of recall accuracy and emotional responses. Twenty participants completed two 3-day tasks. In an INMI task, participants recorded information about INMI episodes as they occurred; in a VMI task, participants were prompted via text message to deliberately imagine each tune they had previously experienced as INMI. In both tasks, tempi of the imagined tunes were recorded by tapping to the musical beat while wearing an accelerometer and additional information (e.g., tune name, emotion ratings) was logged in a diary. Overall, INMI and VMI tempo measurements for the same tune were strongly correlated. Tempo recall for tunes that have definitive, recorded versions was relatively accurate, and tunes that were retrieved deliberately (VMI) were not recalled more accurately in terms of tempo than spontaneous and involuntary instances of imagined music (INMI). Some evidence that INMI elicited stronger emotional responses than VMI was also revealed. These results demonstrate several parallels to previous literature on involuntary memories and add new insights on the phenomenology of INMI.
Introduction
The distinction between involuntary and voluntary autobiographical memories has long been a subject of debate in cognitive psychology. While some researchers argue for separate memory systems, others propose a unitary system where differences arise from retrieval mechanisms rather than storage. Musical imagery—specifically "earworms" or involuntary musical imagery (INMI)—offers a unique methodological advantage. Unlike general autobiographical memories, music has a "canonical" ground truth (the original recording), allowing for precise measurement of recall accuracy. This study compares INMI episodes with deliberate voluntary musical imagery (VMI) of the same tunes to investigate temporal accuracy and emotional phenomenology.
Methodology
The study employed a within-subjects design with 20 participants over two three-day periods.
- INMI Task: Participants wore a GENEActiv accelerometer on their dominant wrist. When an earworm occurred, they tapped the beat on their leg and logged details (tune name, mood, arousal) in a digital diary.
- VMI Task: Participants were prompted via SMS at the same times they previously reported INMI to deliberately imagine the same tunes, repeating the tapping and logging procedure.
Tempo (BPM) was extracted from accelerometer data using a MATLAB tap-detection algorithm. For tunes with definitive recordings, "canonical" tempos were established using streaming services and audio analysis software to serve as a ground truth for accuracy comparisons.
Results and Discussion
Temporal Consistency and the Unitary System
A primary finding was the high degree of consistency between involuntary and voluntary recall. The raw correlation between INMI and VMI tempos was $r = .625$. However, when accounting for participants tapping at different metrical levels (e.g., tapping at half-time or double-time), the correlation rose significantly to $r = .917$.
This strong relationship supports the Unitary System Hypothesis, suggesting that both spontaneous earworms and deliberate recall access the same stable long-term memory representations.
Precision and the "Slow-Down" Effect
Participants demonstrated remarkable temporal precision. The median deviation from the original recording's tempo was approximately 7.3% for INMI and 7.7% for VMI.
Interestingly, both imagery types tended to be slightly slower than the original recordings. This "slow-down" effect likely reflects motor constraints; tapping a fast beat physically is more demanding than simply hearing it internally. Crucially, there was no significant difference in accuracy between the two modes, indicating that "push" (involuntary) and "pull" (voluntary) retrieval are equally faithful to the source material.
Emotional Responses
While temporal accuracy was similar, emotional responses differed. INMI episodes elicited marginally more positive valence and higher arousal than VMI. This suggests that the spontaneous nature of earworms may enhance their perceived vividness or emotional impact, a phenomenon often noted in involuntary autobiographical memory research. Furthermore, tempo was a significant predictor of arousal only in the INMI condition, suggesting a tighter coupling between musical features and emotional state during spontaneous recall.
Conclusion
This research demonstrates that involuntary and voluntary musical memories are remarkably similar in their temporal characteristics, supporting the idea of a shared cognitive architecture. While the retrieval mode does not significantly impact accuracy, it does influence the emotional experience of the memory. Future research should expand these findings by examining other musical dimensions like pitch and timbre, and by exploring these phenomena in more diverse populations.
Summary updated