Social media data could help researchers understand and predict Alzheimer’s disease

Senior research fellow Dr. Jude Mikal finds social media to be a kaleidoscope of data for studies, tracing back to his dissertation in 2008—a few years after the creation of Facebook—which assessed social media’s impact on cultural adjustment for students studying abroad. The project led him to question more deeply what can be learned about an individual from the way they use social media. Could researchers examine someone’s online activity as a form of clinical observation? 

The issue with using social media for data is precisely its kaleidoscopic nature— “the landscape of internet research changes so quickly that the picture behind it is always changing,” Mikal explained. By the time a researcher has thoroughly studied a particular platform or user, the internet has already changed and moved on. So, instead of trying to understand how the platform affects social relationships, Mikal focused on what could be learned by the way people engage with each other online. 

Mikal, who’s a researcher in the Pharmaceutical Care and Health Systems department, has conducted other studies through this lens, examining the digital footprints of breast cancer survivors before and after their diagnoses, for example. But the study he has been actively seeking funding for takes notes from a study famous amongst dementia researchers known as The Nun Study. Researchers conducted a longitudinal study of several hundred Catholic sisters in their early twenties entering the same convent who agreed to donate their brains to science after death. The circumstances created a controlled experiment in which all subjects in the study had a similar diet, sleeping arrangement, and lifestyle— factors that, when varied, typically affect research findings. Upon entering the convent, all the sisters wrote autobiographical essays, which researchers examined for linguistic measures like idea density and vocabulary variability.   

“The short version is, these writing samples [at an average age of 22] had significant predictive power in terms of whether these women…would end up with Alzheimer's disease or Alzheimer's symptoms by the time they were in their 70s and 80s,” Mikal said. “And that’s especially interesting in the context of the world that we live in now, where we have writing samples like that, not just for women entering a convent, but everyone…[with social media] we have a trove of longitudinal written data.”

Researchers still have much to learn about the progression of dementia. The nature of Alzheimer’s makes it difficult to track— researchers can study following the date of diagnosis, but that could be months after the disease onsets. In Alzheimer’s research, screening determines who needs to be tested for the disease while the test itself determines whether or not an individual has a particular variety of dementia. Knowing who to screen is difficult, especially due to compensatory behaviors people develop that mask developing symptoms, Mikal said. His example: a baker with declining memory whom Mikal interviewed for a book chapter started pre-preparing and measuring out ingredients in individual containers after finding himself digging through mixed ingredients to see whether he remembered to include sugar. 

“What we have in this trove of social media data is an ability to look at progression over time,” Mikal said. “Social media reaches into pockets of the population that often are not getting their annual exams or seeing doctors as frequently as they should, but it also allows you to compare one person to themselves over time to look for evidence of any kind of changes over time.”

Together with collaborators Dr. Laura Hemmy and Dr. Kelvin Lim at the University of Minnesota and Dr. Mike Conway at the University of Melbourne, Mikal has pursued federal funding for this project for upwards of five years, getting close and then falling back to square one. Mikal attributes these challenges to the stereotype that older adults aren’t online enough (“It’s actually not true— boomers tend to be quite plugged in to social media accounts.”) In the meantime, his team is looking at publicly available data on blogging sites and on X, previously Twitter, though they have to sift through jokes and exaggerations for authentic disclosures. 

“By finding those people, can we look at their speech patterns to determine who hasn't disclosed or may not know, or who might benefit from additional testing? It expands the clinical window of observation,” Mikal explained. “There's no way to line them up like we did with breast cancer research, but we can look at it longitudinally. Often people began the blogs when they found out they had dementia. Once you disclose that you have dementia, how do your social relationships change? How does your engagement online change?”

The point of his studies is to intervene earlier on in the progression of dementia and to develop tools to help people identify whether they have dementia, though this poses its own ethical dilemma, Mikal said. Without a cure to Alzheimer’s, does earlier intervention needlessly decrease an individual’s quality of life by subjecting them to tests, monitoring, and stress before the onset of symptoms? His team posed a hypothetical scenario to a sample of individuals without Alzheimer’s: If you had a button to press right now that would tell you whether you will develop Alzheimer’s disease, would you press it? Everyone said they would press the button, citing opportunities to participate in clinical trials or to get affairs in order. 

“I do worry sometimes about the prospect of developing something that in the wrong hands could raise people's insurance premiums or deny them coverage… Am I feeding the beast? At the same time, I don't think it's right to stymie science because of bad actors,” Mikal said. “I think we have a right to know things about ourselves and my hope is that, armed with more knowledge about yourself, about what's to come, you can prepare better and ideally, as a society…we can also prepare by developing better therapies, better interventions.”