If you have watched TV since the 1990s, the sitcom theme song, “I’ll Be There for You,” has likely been stuck in your head at one point or another.
“Scientists have known for some time that music evokes autobiographical memories, and that those are among the emotional experiences with music that people cherish most,” said Petr Janata, UC Davis professor of psychology and co-author on a new study.
“What hasn’t been understood to date is how those memories form in the first place and how they become so durable, such that just hearing a bit of a song can trigger vivid remembering,” said Janata.
The paper, “Spontaneous Mental Replay of Music Improves Memory for Incidentally Associated Event Knowledge,” was published online in the Journal of Experimental Psychology: General. Co-authors are Janata and Benjamin Kubit, a postdoctoral researcher in cognitive neuroscience, both of the UC Davis Department of Psychology, and Center for Mind and Brain.
This new research offers an initial glimpse into these mechanisms and, somewhat surprisingly, finds that the songs that get stuck in your head help that process of strengthening memories as they first form, the authors said. Thus, this is the first research to link two of the most common phenomena people experience with music – earworms (having a song stuck in your head) and music-evoked remembering.
For their latest study, the researchers worked with 25 to 31 different people in each of three experiments, over three different days, spaced weeks apart. Subjects first listened to unfamiliar music, and then, a week later, listened to the music again, this time paired with likewise unfamiliar movie clips. In one instance, movies were played without music.
The research subjects, all UC Davis undergraduate and graduate students, were subsequently asked to remember as many details as they could from each movie as the music played. They were also quizzed about their recollection of the associated tunes and how often they experienced each of the tunes as an earworm. None of them had formal music training.
Repetition and accuracy
The results: the more often a tune played in a person’s head, the more accurate the memory for the tune became and, critically, the more details the person remembered from the specific section of the movie with which the tune was paired.
With only one week between when they saw the movie, and when they were asked to remember as many details from the movie as they could while listening to the movie soundtrack, the effect of repeatedly experiencing a tune from the soundtrack as an earworm resulted in near-perfect retention of the movie details.
These people’s memories, in fact, were as good as when they had first seen the movie. Additionally, most subjects were able to report what they were typically doing when their earworms occurred, and none of them mentioned the associated movies coming to mind at those times.
“Our paper shows that even if you are playing that song in your mind and not pulling up details of memories explicitly, that is still going to help solidify those memories,” Janata said.
Future help for memory loss?
The authors said they hope the research, which is ongoing, could eventually lead to the development of nonpharmaceutical, music-based interventions to help people suffering from dementia and other neurological disorders to better remember events, people and daily tasks.
Emotional events from our life are more likely to be later recollected than similar, non-emotional events. This emotional enhancement of memory has been extensively studied using words and pictures (for review, see Yonelinas and Ritchey, 2015). Fewer studies have however examined the influence of emotion on musical memory. ù
And yet, the emotional power of music is well-established (Koelsch, 2014), and musical memory may in fact benefit from such emotional enhancements, explaining why certain pieces of music frequently become unforgettable. Is this related to the particularly efficient post-encoding processes involved in memorizing musical information?
In order to investigate this question, we examined musical memory abilities in adult listeners by manipulating the emotional characteristics of musical excerpts and the lengths of time separating the study phase and a recognition test.
Emotion theorists often take the position that affective experience can be described according to two orthogonal dimensions, namely arousal and valence (Russell, 1980). Arousal refers to a continuum ranging from calm to excitement, whereas valence is measured along a continuum ranging from positive to negative.
The impact of these emotional dimensions on different forms of memory, including declarative (explicit) memory, has been explored at great length. In particular, evidence has shown that high-arousal information is better remembered than low-arousal information (for review, see Hamann, 2001) in immediate (Bradley et al., 1992) or delayed memory tests administered after 1 day (Hu et al., 2006), a year (Bradley et al., 1992) or several years (Wagner et al., 2006). By increasing attention and elaboration at time of encoding, high-arousal stimuli are thought to be more deeply processed than low-arousal stimuli.
Their memory trace would be subsequently enhanced by post-encoding processes, including stress hormone release, leading to better consolidation over a long period of time (Hamann, 2001; Sharot and Phelps, 2004; Payne et al., 2008; Sharot and Yonelinas, 2008). Emotional valence has also been reported to influence memory, with stimuli with a negative or positive valence being better memorized than neutral stimuli (Ely et al., 1999; Cabeza and Dolcos, 2002; Mickley and Kensinger, 2008).
Furthermore, remembering negative information yielded even higher memory performance than remembering positive information, whether using pictures (Comblain et al., 2004; Waring and Kensinger, 2009) or words (Inaba et al., 2005; for review, see Kensinger, 2007). This effect would persist over days (Denburg et al., 2003; Kensinger, 2007; Waring and Kensinger, 2009) or weeks (Ely et al., 1999; Ochsner, 2000; Pierce and Kensinger, 2011), suggesting a negativity bias in emotional memory, at least in young adults.
Few studies have conjointly manipulated arousal and valence within a single experiment. In some cases, valence and arousal had their own respective effects on memory. Arousing pictures were better remembered than non-arousing ones and negative pictures were better memorized than positive ones (Comblain et al., 2004). In other investigations, the effect of arousal on memory performance was found to be modulated by valence (Kensinger, 2008; Li et al., 2018).
However, although studies show interaction effects, the nature of the interaction differs. This interaction and its impact on memory performance are still debated in the behavioral literature (Ochsner, 2000; Dolcos et al., 2004). However, the joint effect of valence and arousal on recognition has been demonstrated in numerous neuroimaging studies, suggesting that these two dimensions influence memory through distinct neural mechanisms (Dolcos et al., 2004; Kensinger and Corkin, 2004). These results therefore justify the importance of considering these two dimensions in studies exploring the impact of emotions on memory.
Interestingly, prior investigations suggest that the emotional characteristics of stimuli may shape the retention of information differently over long delays, compared with short delays. In some studies, emotional enhancement was observed after a long delay, but not a short delay (Yonelinas and Ritchey, 2015).
In other cases, the effects of arousal on memory were enhanced after a long delay, relative to a shorter one (Sharot and Phelps, 2004; Payne et al., 2008; Sharot and Yonelinas, 2008; Kensinger, 2009). Finally, Waring and Kensinger (2009) reported a complex interaction between all these factors in a visual scene recognition task. After a short delay, they showed that the level of arousal modulated the effect of valence on memory.
Thus, an emotional enhancement of memory was observed in all conditions, except for positive low-arousal scenes, at least in young adults. After a long delay, the pattern of results was different. The enhancement of memory was greater for negative than for positive scenes, and greater for arousing than for non-arousing scenes underlying once more the negativity bias and the effect of arousal in emotional memory. Furthermore, the emotional enhancement effect was greater after a long delay than a short delay, emphasizing the role of post-encoding processes in memory consolidation of emotional information.
Taken as a whole, extant literature underlines the complex interplay between emotional features of stimuli and study-test intervals, and how memory performance can be shaped by post-encoding actions.
An absence of agreement persists among musical domain scholars about how these factors contribute to the emotional memory for music. Indeed, as an enjoyable human activity that is present in all cultures, capable of generating strong and varied emotions, it seems to be a privileged medium for studying this effect. One series of studies used the exact same prototypal clips intended to express happy, fearful, peaceful, or sad emotions, emotional characteristics which had been validated previously (Gosselin et al., 2005; Vieillard et al., 2008).
Musical recognition of these four emotional categories of computer-generated MIDI musical excerpts was generally assessed after a short delay (Aubé et al., 2013; Vieillard and Gilet, 2013; Narme et al., 2016), though sometimes after a longer delay of 24 h (Samson et al., 2009).
Except for one study that did not report any difference in explicit recognition, at least in young adults (Narme et al., 2016), the other investigations showed better recognition of musical clips expressing fear and, to some extent, happiness (Samson et al., 2009; Aubé et al., 2013; Vieillard and Gilet, 2013), suggesting a benefit for high-arousal stimuli. However, this finding should be interpreted cautiously, as it could be explained by different levels of difficulty between recognition of high and low-arousing musical excerpts, or of each prototypical emotional category.
As demonstrated by a control study (Samson et al., 2009), the perceptual distinctiveness (or dissimilarity) between the scary music was larger than the distinctiveness between the three other categories, explaining, at least in part, the superior recognition scores obtained with scary musical excerpts. However, the high recognition score obtained with happy musical excerpts, in particular after a 24-h delay, might nonetheless reflect the effect of arousal and/or valence on consolidation of musical memory, though it is too early to draw any firm conclusions.
Other studies reported in the literature have used instrumental classical music. Eschrich et al. (2005) examined the relationship between emotion and music recognition. In this study, piano pieces by Bach were rated in terms of arousal (from very pacifying to very arousing) and valence (from negative to positive valence) during encoding.
Recognition tests conducted 2 weeks later revealed that well-recognized pieces were associated with higher arousal ratings and received a higher positive valence rating. In a subsequent study, symphonic film music eliciting positive feelings (i.e., little positive to very positive) were repeated twice over different days (Eschrich et al., 2008).
The authors confirmed an effect of valence on recognition with better memory for very positive music excerpts than for less positive ones, although no effect of arousal on recognition was obtained in this case. These findings suggest that emotional valence, as rated by the participants, appears to influence long-term memory for music, underlying once more the impact of positive valence music on memory consolidation.
However, the lack of negative valence music in this study limits the interpretation of the results. By addressing again this question in a neuroimaging study, the authors failed to replicate their previous finding (Altenmüller et al., 2014). Participants’ single exposure to the musical pieces during encoding may have been insufficient to induce a memory enhancement, thereby explaining the lack of behavioral results. Based on all these data, it remains unclear whether only positive valence or both positive and negative valence provides a memory advantage.
Prior investigations into emotional memory in music have assessed memory at only one delay. To our knowledge, in the music domain, only one study has examined the effects of delay interval upon emotional memory (Alonso et al., 2015). Making use of the parsimonious model of emotion, which defines emotional spectrum into two dimensions, its authors manipulated valence and arousal levels in symphonic musical excerpts.
Participants were requested to rate these two emotional feelings induced by listening to music before taking two recognition memory tests administered immediately after the study and 24 h after the study session. The results indicated that arousal and valence interacted differently with memory performance at each study-test delay. In immediate recognition, the effect of valence varied as a function of arousal. Whereas valence did not interfere with the remembering of high-arousal excerpts, it did modulate recognition of low-arousal stimuli such that positive excerpts were better recognized than negative excerpts.
In contrast, in the delayed condition, the results revealed no interaction between the two emotional dimensions. Only independent effects of arousal and valence were reported, such that high-arousal excerpts and negative excerpts were better memorized than low-arousal and positive ones, respectively, confirming the memory advantage for high arousing and negative stimuli already reported in non-musical domains (Ochsner, 2000; Waring and Kensinger, 2009).
Unlike findings obtained with words and pictures, there was no loss of memory in delayed, as compared to immediate recognition, indicating no deleterious impact of delay interval in music. Yet this recognition test, frequently used in psychology, presents a methodological bias. By presenting the target stimuli once again in the delayed recognition test, this condition benefits from an additional exposure compared to the immediate condition. It is therefore difficult to disentangle the effect of delay from the effect had by number of presentations in post-encoding processes.
To overcome this methodological limit and to clarify the ways emotional dimensions (i.e., valence and arousal) affect musical memory as study-test delay intervals increase, we designed a new study using the symphonic musical excerpts selected by Alonso et al. (2015). However, we manipulated the study-test interval while keeping the number of exposures constant. For this purpose, the encoding phase was distributed over two distinct sessions, one session on day 1 allowing the first half of the target stimuli to be encoded and the other session on day 2 allowing the other half to be encoded.
Immediately after this second session on day 2, a recognition test including all target stimuli mixed with foils was presented. The recognition of the target stimuli presented just before (on day 2) or after 24 h (on day 1) provided respective memory performances after a short delay and a long-delay retention without changing the number of exposures to the targets. In line with the literature, we hypothesized an emotional enhancement of music memory induced by post-encoding processes leading to better recognition of musical excerpts in delayed condition as compared to immediate one. The effects of arousal and valence should become exaggerated after a long delay. Finally, we also predicted that the two emotional dimensions would be differently affected by the study-test intervals.
reference link : https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7488292/
More information: Benjamin M. Kubit et al, Spontaneous mental replay of music improves memory for incidentally associated event knowledge., Journal of Experimental Psychology: General (2021). DOI: 10.1037/xge0001050