The song you can still find when nothing else is left

A daughter sets a recording playing in her mother’s room. Her mother has not called her by name in more than a year. She does not always know that the person sitting with her is her daughter, or that she is in a care home, or what year it is. The song is a hymn from a country church, eighty years old, the kind of thing sung from memory by a girl who has long since stopped being a girl. The first line comes from the speaker. And the mother, without being asked, begins to sing. Not humming. The words. All of them, in order, in tune.

This happens often enough that the people who work in dementia care stop being surprised by it. It surprises everyone else.

What stays when the rest goes

Alzheimer’s disease does not take memory all at once, and it does not take it evenly. The losses tend to follow an order. Recent events go first. Then names. Then faces, including familiar ones. Then the long autobiography of a life, the chronological sense of who someone has been. Caregivers describe the disease as a slow retreat, a person moving back through their own history until very little of it is reachable.

And then there is music, which does not seem to follow the same retreat at all.

For a long time this was a clinical anecdote rather than a finding. Doctors and family members would notice that a patient who could no longer hold a conversation could still sing a song from childhood, or play a piece on the piano they had learned at six. The observation was common, and it was almost never studied directly. In 2005, the psychologist Lola Cuddy and the physician Jacalyn Duffin, both at Queen’s University, wrote a paper for the journal Medical Hypotheses that took the anecdote seriously. They described an eighty-four-year-old woman with severe Alzheimer’s whose musical recognition appeared, by her caregivers’ account, to be largely intact, and they argued that this sparing of musical memory was real, testable, and worth measuring properly rather than admiring from a distance.

Oliver Sacks had been circling the same observation for years. In his 2007 book Musicophilia, he gathered case after case of patients whose musical memory outlasted almost everything else, and he wrote about them the way he wrote about all his patients, with care and without overstatement. He noticed that music could reach people who seemed otherwise unreachable, and he was honest that he did not fully understand why. The cases accumulated faster than the explanations did.

What all of this was pointing at is the question underneath the daughter and the hymn. Why would a song survive when a child’s name does not.

Two different kinds of remembering

Part of the answer is that “musical memory” is not one thing, and the research has been careful about which part is preserved.

In 2009, Amee Baird and Séverine Samson reviewed what was then known and drew a useful line through it. The line falls between implicit memory, the kind that lets you do something without consciously recalling how, and explicit memory, the kind that lets you recognize that you have heard a particular tune before. Their reading of the evidence was that the implicit, procedural side can be remarkably spared. A trained musician with Alzheimer’s may still play. A person may still sing the words to a song that has been sung the same way for seventy years. The explicit side, knowing that a melody is familiar or telling two unfamiliar tunes apart, is more often impaired, much like other kinds of conscious recognition.

This is the honest version of the finding, and it matters. The popular story holds that people with dementia “still remember the music.” The more accurate story is that some musical abilities, especially the deeply practiced and the long familiar, ride on systems that the disease reaches late. The daughter’s mother is not recognizing the hymn the way she would recognize a photograph. She is running a pattern she has run thousands of times, and the running of it is still there.

The part of the brain the disease reaches last

In 2015, a team led by Jörn-Henrik Jacobsen at the Max Planck Institute, working with collaborators in France, published a study in the journal Brain that gave the anecdote an anatomy.

They started with healthy young volunteers in a high-resolution scanner, looking for the brain regions that light up when a person recognizes long-known music. They found that musical memory of this kind was encoded in part by the caudal anterior cingulate cortex and a nearby region called the ventral pre-supplementary motor area, sitting toward the front and middle of the brain. Then they did something the earlier work could not. They compared those regions, in patients with Alzheimer’s, against the pattern of damage the disease leaves behind.

The regions that carried musical memory were among the least damaged. They showed little of the shrinkage and little of the drop in glucose metabolism that mark the tissue Alzheimer’s has consumed. The areas that hold a lifetime of songs turn out to sit, more or less, in the part of the brain the disease reaches last.

That is the specific thing worth holding onto. It is not that music has some mystical immunity. It is that the neural real estate where deeply familiar music lives happens to be territory the disease is slow to take. The song you can still find is the song stored in the room that has not yet flooded.

It is worth being clear about how partial this picture still is. Jacobsen and his colleagues studied a small group, mapped musical memory in healthy brains, and then overlaid that map onto the known geography of the disease. That is a careful and elegant way to connect two things, and it stops short of proving that any individual patient sings because those exact regions are spared. The work shows a strong correspondence between where music is held and where the damage arrives late. It invites the explanation rather than settling it. The researchers say as much. What they have given is a plausible neural reason for a thing families have witnessed for decades, which is a real contribution and a modest one at the same time.

The honest size of it

It would be easy to make this finding carry more than it can, and the researchers are careful not to. What they have shown is where deeply familiar music is held, and that the disease is slow to reach there. Whether music does anything to the illness itself, earlier in a life or by some path no one has mapped, is a separate question, and an open one. The honest answer is that we do not know, and this letter does not pretend otherwise. It is sitting with a smaller and stranger fact.

What the music offers, in the moment a daughter presses play, is a meeting. For the length of a song, a person who has become hard to reach is reachable. A few minutes of presence. A mood that lifts. A voice that joins in. We told one version of this in an earlier letter, about a man named Henry whose face came back to him for a few minutes in a care home when the right music found him. The return was brief. It was also unmistakably real, and the people in the room with him would not let you call it nothing.

There is a separate strand of research worth keeping beside this one. Sometimes a familiar song pulls a whole scene back with it, a summer, a kitchen, a person. That vivid, place-and-time memory has been traced by Petr Janata, in 2009, to a different region, the medial prefrontal cortex. That is the experience of a song carrying you somewhere. The finding at the heart of this letter is about something more basic and more durable, the bare ability to hold the music itself, which can stay even after the scenes it once summoned are gone.

The song you can still find

So the daughter sits with her mother, and the hymn plays, and her mother sings every word.

What the science can now say is modest and precise. The words are stored in a part of the brain the disease is slow to reach. The singing is a pattern run so many times that it became something closer to motion than to memory. None of it means her mother is coming back, and the daughter, who has been living this for years, already knows that.

It means something smaller and harder to dismiss. There is a room in her mother that is still lit. For the length of a song, the door to it opens, and a girl from a country church eighty years ago sings in a voice that everyone in the present can hear. The reach is brief. The reach is real. And for as long as the music lasts, there is somewhere left to meet.

If you’d like a few of these letters delivered each month, you can subscribe here.