Evaluating Music & Memory

“Early preferred music is thought to be retained in a part of the brain that’s preserved even into the later stages of dementia,” McCreedy said. “There’s been some exciting fMRI work that shows early preferred songs light up a part of the brain that’s affected later in the dementia course. The idea is that if you can find those early preferred songs, it can reduce agitation related to social isolation. People feel less lonely because they’re having autobiographical memories.”

McCreedy and her team worked with 54 nursing homes from four multi-facility nursing home corporations across the country. Of the 976 participants with ADRD, 66% had moderate to severe symptoms; the average age was 80.3 years.

We spoke with McCreedy about her work and its potential to improve the quality of life for older adults with ADRD.

Can you give us an overview of your study? 

McCreedy: We conducted two parallel trials—one before COVID-19 and one after vaccines became widely available in nursing homes. It wasn’t planned that way, but the pandemic disrupted our original timeline.

In the 2019 trial, we enrolled 54 nursing homes and conducted a cluster randomized trial. Half of the homes received a personalized music intervention we call “early preferred music,” meaning music that residents loved in their late teens and early twenties. This kind of music tends to stay with people, even in later stages of dementia, due to how our brains process memories tied to music. The intervention was delivered via headphones, not by researchers but by nursing home staff as part of usual care. Each nursing home involved 15 residents receiving the intervention, while the other half continued their standard care practices.

The idea was to address behavioral symptoms of dementia, like agitation, through a personalized music experience. While nursing homes already use music in various ways—such as group sessions or background music during meals—this intervention was different because it was tailored to the individual and delivered directly to them.

The idea of music unlocking autobiographical memories is interesting. How does that work?

It’s remarkable. Music engages neural pathways that are often preserved even in advanced dementia. Songs from someone’s youth can trigger vivid memories and emotional responses. For example, in our pilot study, a nonverbal man with advanced dementia lit up when he heard Hank Williams. He even started playing guitar and singing. Moments like that are rare but incredibly powerful.

Can you share how your background influenced this research?

Yes, I started as a music therapist, earning my undergrad at Michigan State and working in psychiatric and forensic settings. I saw firsthand how music could transform lives, especially in acute psychiatric care. Music brought organization and expression to people in ways nothing else could.

Later, I shifted to public health, earning a master’s in epidemiology and global health and a Ph.D. in health services research. This project brings my two worlds together. 

Professor Vincent Mor, the principal investigator, heard about my music therapy background and thought I’d be a perfect fit for this study. It’s been incredibly rewarding to use my public health training to scientifically validate something I’ve always been passionate about.

What were the main findings from the first trial?

We found no significant change in staff-reported behaviors over the preceding week using standard nursing home data collection tools. However, when we observed residents during structured time periods, we saw fewer verbally agitated behaviors among those receiving the music intervention. 

This suggests that while the music might not drastically change overall impressions of behavior, it does offer momentary relief from agitation—a meaningful finding. That temporary relief could potentially reduce the use of medications prescribed as-needed (PRN), which is an exciting avenue for future research. 

The study noted that verbal agitation decreased as a result of the intervention, but physical agitation didn’t. Why is that? 

That’s a key distinction. Verbal agitation often stems from causes like social isolation, loneliness, sensory deprivation or even pain. Personalized music helps address these by engaging and comforting the individual. In contrast, physically agitated behaviors—such as hitting or kicking—are often reactions to caregiving activities like bathing or dressing. These behaviors likely require different interventions tailored to those specific situations.

There are also repetitive, non-aggressive physical behaviors—like fidgeting or rocking—that remain less understood. Our intervention wasn’t designed to address physical aggression or these repetitive behaviors, so the lack of impact there aligns with our expectations.

How personalized does the music need to be for this intervention to work? 

That’s the million-dollar question. Personalization is key, but there are practical challenges. In institutional settings, family members may not know a resident’s early music preferences. When we can’t identify specific songs, we use billboard charts from the resident’s youth and test for positive reactions. If someone responds well to a genre, we tailor the playlist further. This approach works, but there’s still room for improvement. Future research could explore how much personalization is necessary to achieve meaningful results.

Your study focuses on momentary well-being. How should that shape future non-drug treatments?

It’s a critical area. Momentary effects, like temporary relief from agitation, are often missed in traditional evaluations that rely on routinely collected data. These brief improvements have inherent value, both for individuals and their caregivers. Even if personalized music does not decrease the need for PRN medication, providing people with a sense of connection to themselves and moments of relief from their symptoms is still profoundly valuable.  

But capturing these effects is a challenge. Structured time studies—where observers document behaviors at specific intervals—are effective but not scalable. That’s why we’re piloting technologies like emotion recognition apps and wearable devices. For example, smartwatches could detect agitation through changes in heart rate variability and even time the delivery of interventions. These tools might help us better measure and optimize momentary interventions. 

I would like to add a note of gratitude to the nursing home staff who made this study possible. These caregivers work in under-resourced and often underappreciated environments, yet they’re vital to implementing innovative interventions like this. As we advance this research, we must focus on making these solutions accessible to everyone, especially those in marginalized or low-resource settings. Music’s power should benefit all, not just those with the means to afford premium care.