Detecting cognitive decline before its symptoms start
Top illustration: iStock
In his research on the brain, Daniel Gustavson looks for clues about when cognitive decline begins
According to Daniel Gustavson, assistant research professor in the Institute for Behavioral Genetics, much of the research on cognitive decline starts late.
“A lot of studies of older adults—too many, in my opinion—focus on when some cognitive decline has already happened,” he says. “It's clear that a lot of the disease, or even just normal aging, has already taken place by the time somebody comes into a clinic and says, ‘I'm worried about my brain.’”
Gustavson wants to dig deeper into the timeline and see if cognitive decline can be spotted before its telltale signs arise. he coauthored and recently published in Neurobiology of Aging makes headway toward accomplishing that goal.
鶹ѰBoulder researcher Daniel Gustavson notes that a lot of cognitive decline, or even just normal aging, has already taken place by the time "somebody comes into a clinic and says, ‘I'm worried about my brain.’”
The cognitive gas tank
Gustavson’s study—which used twin research, genetic analysis and magnetic resonance imaging (MRIs), among other methodologies—examines the relationship between brain reserve in middle age and executive function later in life.
“Brain reserve,” says Gustavson, “is a bit like a gas tank. You have a certain amount of gas built up when you’re a young adult, when your brain is at its healthiest, and as you age, you start to lose some of that fuel.”
Executive function, he adds, refers to complex goal management or attentional control. “It captures higher-level cognitive processes, where you have to be controlling other sub-processes.”
An example of executive function in action is asking someone to memorize and reorder a string of letters and numbers.
“You might have people listen to a list like X, six, B, Y, seven, J, and then they’d have to remember that list in their head and repeat the numbers back in numerical order and the letters in alphabetical order,” Gustavson says. “It’s a little more complicated than just repeating what someone said.”
Using data from the Vietnam Era Twin Study of Aging (), which includes more than 1,600 subjects who have undergone various cognitive assessments at regular intervals over the past 20 years, Gustavson and his coauthors concluded that higher brain reserve at the age of 56 was associated with better executive function at the age of 68.
Looks can be revealing
Brain reserve, says Gustavson, is a proxy for brain thickness, and brain thickness is determined through MRIs.
To analyze the hundreds of MRIs of VETSA subjects, Gustavson and his coauthors used a developed by , professor of neuroimage computing at the , which was trained in much the same way Google trains its search algorithms.
“You can train it over and over again,” Gustavson says. “The more data you have”—that is, MRIs—“and the more times you tell it, ‘You were wrong this time. You were right this time,’ the better it gets at classifying this brain as one age versus that brain as another age.”
The algorithm assesses plump, padded brains as younger and atrophied, motheaten brains as older, regardless of the chronological age of the people in whose heads those brains reside. That means, for example, that a 56-year-old can have a brain that appears 60 and a 60-year-old can have a brain that appears 56.
And this matters, Gustavson says, because how a brain looks in an MRI predicts its executive function years later.
“Controlling for their actual age, people with younger-looking brains had much shallower decline in executive function over the subsequent 12 years, and people whose brains appeared older than average had steeper drops in executive function.”
Yet the cause of this discrepancy—genetics? environment? trauma?—is something the algorithm alone can’t explain. That’s where twin research comes in.
Same genes, different story
One of the benefits of twin studies like VETSA, Gustavson says, is their ability to separate environmental influences on a person’s health—things like diet, exercise and place of residence—from genetic influences.
“Brain reserve is a bit like a gas tank. You have a certain amount of gas built up when you’re a young adult, when your brain is at its healthiest, and as you age, you start to lose some of that fuel,” says Daniel Gustavson. (Illustration: iStock)
“Those two things aren't fully separable, but basic twin studies give us some idea of how inherited different constructs are—not only cognitive abilities, like memory or speed, but also changes in those abilities. Twin studies help us quantify how much those changes are due to genetics and how much are due to environment.”
If one twin experiences cognitive decline faster than the other, in other words, researchers can confidently point to environment as the reason, since twins share the same genes.
But twin studies can go only so far, Gustavson says, as they tend to paint with a broad brush. “You often can't pinpoint specific genes or specific environments that matter, because it's all statistical.”
That’s why Gustavson and his team incorporated genetic analyses in their study. They wanted a higher-resolution snapshot of the genetic influences on cognitive decline, specifically by seeing if the APOE genotype, which is strongly associated with Alzheimer’s, predicted a drop in executive function.
What they found is that, although APOE alone did not fully explain changes in subjects’ executive function over time, those subjects’ genes taken as a whole did.
“Most of the association between people's brain health and their future cognitive decline, about two-thirds, was explained by genetics,” Gustavson says.
But that’s not to dismiss the other third as inconsequential.
“Things like healthy lifestyle, diet, smoking and alcohol use, social engagement—those things don't seem like they relate to cognitive changes, but they might impact your brain health in the first half of your life, and then your brain health in midlife will impact your cognition later,” says Gustavson.
The fourth wave
Gustavson and his fellow researchers just completed the fourth wave of data collection, when the VETSA subjects were 74 years old, and are therefore currently working to build upon their findings.
“We would like to expand our models to capture the cognitive changes even further out,” he says.
Gustavson would also like to deepen his understanding of what exactly the brain-age algorithm is detecting. “Is it capturing something new to midlife, or is it capturing something from young adulthood, the consequences of which are only becoming apparent in midlife?”
He suspects it’s the latter, but he’s not yet sure. “I really want to look at that in more detail.”
Jeremy A. Elman, Chandra A. Reynolds, Lisa T. Eyler, Christine Fennema-Notestine, Olivia K. Puckett, Matthew S. Panizzon, Nathan A. Gillespie, Michael C. Neale, Michael J. Lyons, Carol E. Franz and William S. Kremen contributed to this research.
Did you enjoy this article? Passionate about behavioral genetics? Show your support.