Harvard Medical Student: THIS is the future of brain disease [Part 1]
Researcher Meg Majumder shares her groundbreaking work on neurodegenerative disease with us
Meg Majumder is a Harvard Medical student who researches neurodegenerative disorders like Alzheimer's and Parkinson's. Her research started as an undergraduate at UC Berkeley and includes the use of artificial intelligence to diagnose and treat these diseases, plus the ethical implications that follow. You can follow Meg on Twitter here.
We were fortunate to grab Meg for an exclusive interview. The original transcript has been edited for clarity and ease of reading. Stay tuned for Part 2 later this week.
This is part of our ongoing series with doctors and researchers at the cutting edge of medical science and theory. You can find the last installment, an interview with orthopedic surgeon and diet advocate Dr. Shawn Baker, right here.
CHN: Thanks so much for taking the time to speak with us. Can you give our audience a quick background?
Meg Majumder: Prior to Harvard Med, I attended UC Berkeley where my research was in neurology, and I’m continuing, even though I'm also a student, to do research at Beth Israel Hospital in Boston.
CHN: Can you speak specifically about what your research is in?
Meg Majumder: Sure. I study the complications associated with early diagnosis of multiple sclerosis, Alzheimer's disease, and Parkinson's disease, especially in relation to perceptions and portrayals of neurodegenerative conditions.
These are all conditions that affect elderly populations.
So this is something that I want people to be concerned about and ready for, because AI (artificial intelligence) is coming to these populations in the form of diagnostic tools, and it can seem intimidating.
So what I wanted to do is not just develop the tools, but also study the ethics, the underlying concerns that these populations might experience. I'm doing this at Harvard, where ethics is all the rage. Whereas in the Bay Area (where I studied at UC Berkeley), tech is all the rage.
So I feel like I have a split personality here.
CHN: Can you explain more about the ethics of artificial intelligence in medical diagnostics?
Meg Majumder: I’m exploring who to blame when the artificial intelligence-driven diagnostics tool comes up with a false prediction. Because these AIs and these tools, they are not the same as other diagnostic systems.
So, there's a sort of software black box here that gives you a prediction but doesn't tell you how it comes to a conclusion.
And black box medicine is when you use these models to make decisions that are related to health care. We use high-quality data sets and we combine those with predictive algorithm.
But when all of the data sets get taken into account, we're not really asking, whose data is it?
Who is privileged enough to participate in these clinical trials so the data can be used by these systems to create conclusions?
Many programmers are not asking those questions. They kind of just blindly create algorithms, if that makes sense.
CHN: Can we talk a little bit about the upside, why this is worthwhile, and then get into the dangers of these AI diagnostics?
Meg Majumder: Yeah, the upside is that early diagnosis and diagnostic tools are extremely beneficial to people who are concerned about their long-term finances and whether or not they believe that certain cures might come out, given that medical technology is developing exponentially, basically more evolved health medicine is following Moore's law.
This is controversial, but in my opinion, early diagnosis is the way to go.
It is extremely beneficial to all ages and all populations. I've studied a vast array of perceptions and portrayals of Alzheimer's diagnosis and patients who have MS and have been diagnosed with Parkinson's.
And I've reflected on the possible impact of early diagnoses in different ways, like in terms of how it might affect people psychologically, emotionally.
How it might affect inter-family relationships and roles, like responsibilities in groups.
And it's pretty clear to me, often the consequences of a certain way of proceeding or portraying a neurodegenerative disorder like Alzheimer's or MS, might be beneficial in some ways and harmful in others.
So, it's the cultural climate that often defines whether or not we're able to accept early diagnosis, but the cultural climate is leaning towards general acceptance.
CHN: It sounds like you're saying early diagnosis has significant medical benefits but there are concerns that come with it. For example: Could it affect your health insurance? Could it affect your financial and legal status?
Meg Majumder: Yeah, definitely. So, when it comes to long-term care insurance, it varies depending on what exactly you purchase and the coverage selected. Like for example, someone might choose to purchase comprehensive care with a lifetime MS benefit.
Then the monthly premium would be determined by the applicant's age at purchase and remain constant over time.
So this person would receive a certain amount of money every month during their illness or disability until reaching the maximum allotment. But there are a few limitations to this, so if you're living a long-term care facility and you have a dementia or Alzheimer's diagnosis or you need assistance with activities with daily living, you would most likely be denied coverage.
In many of these cases, early diagnosis of the risk of developing Alzheimer's would be crucial because the individual has to plan for the future and purchase insurance before some things would appear.
Once the plan is approved, coverage can't be canceled due to health issues.
So it's like being educated about somebody's personal risk for developing a condition like Alzheimer's and then deciding to purchase early. That can make all the difference in the ability to pay for long-term care.
I have to think about insurance a lot.
CHN: So it sounds like early diagnosis is GOOD for financial and legal status, if I'm hearing you correctly. And additionally preventative measures taken earlier could really benefit people and keep treatment costs down in the long run.
Meg Majumder: Exactly. Yeah, that's pretty much the gist of my spiel.
It's like if you can identify that you have it earlier on or you run the risk of having it earlier on, and you can purchase insurance coverage then, then the insurance companies can't jack up the prices once you're diagnosed with the condition.
The over-use of antibiotics is leading to an epidemic of drug-resistant superbugs. Before you even think about using antibiotics, please read our new report about the 10 safe and natural antibiotics you can find right in your own kitchen. These “astringent” foods naturally cleanse your blood without harmful side effects
Become a Clear Health Now reader today and claim your free copy of the “Top Ten Natural Antibiotics Hiding in Your Kitchen.”
CHN: I guess so, a lot of your work then, is specific to the United States.
Meg Majumder: Yeah, definitely. I mean, I'm of Indian descent, but this technology has not disseminated into India.
And I mean, I hope to bring it there after I graduate, but right now, there is a severe lack of any kind of AI and precision medicine in India itself. And Brazil, as well.
CHN: So it sounds like you're saying the technology is already here, it's just unevenly distributed. We already have good AI diagnostic tools?
Meg Majumder: Definitely. We have plenty of tools. There's plenty of clinical trials underway, especially with regard to neurodegenerative diagnostic tools that integrate AI or are fueled by AI.
Yeah, it's not evenly distributed in the States, let alone developing countries.
CHN: Do you want to talk more about your personal background and how you came to go down this path?
Meg Majumder: So I'm a first-generation student. But it also kind of means that I've seen two sides of the picture, in which the communities in which I grew up have such a need for equitable access to health care.
[In India] you can see from across the street a brilliant shining hospital and know, just know that you'll never, ever be able to step foot in it because you can't afford it.
And health insurance is a whole other monster.
But that's where here it's a lot better.
So, yeah. And then, how I got started as a student in this particular research, I had a professor at Berkeley, her name's Jennifer Doudna, she's amazing. She literally came up with CRISPR.
So when I was working under her and I was her student, she cautioned me about dealing with technology.
She emphasized how we ought to consider the ethics and set some kind of understanding before we actually begin to deploy the technology into clinical testing. So that was an epiphany moment.
You realize you're sitting in a lecture hall with a professor talking to you and a few friends, but like really you're in the midst of one of the greatest minds of the century and it's a historic moment.
Whenever I was in her presence, it was brimming with potential.
CHN: It seems like we hear these stories a good bit in science — the right mentor at the right time.
Meg Majumder: Absolutely true. That's absolute true. If I hadn't gone to Berkeley that year, I would never have gotten interested in any of this, let alone medicine.
I was definitely not on the track for medicine. I was thinking about photography and film.
CHN: Have these issues touched your own family?
Meg Majumder: Yeah, definitely. Alzheimer's, MS, those run pretty deep in my ancestry. I've had a grandparent pass away from Alzheimer's and two cousins from MS, so I'm hoping that there's going to be future genomic studies that we can identify stuff, like the genes for MS.
Because I'm pretty sure that a genetic correlate is there; we just haven't been able to identify it.
I mean, I've worked on that before, but now that there are tools in which I can actually help prevent MS. That's part of what I'm working on now.
But early diagnosis, it would definitely help, or at least affect the health of an early-stage Alzheimer's or MS patient, like besides treatment effects, in so many amazing ways.
It could differ between cultures and like different roles of the elderly in families and societies.
Hypothetically, the way that I perceive all of this, early diagnosis would be beneficial in individualistic societies where the cognitive-deficient elderly are less supported by families and they have more to gain from early intervention from the health care system.
But ultimately, I think every patient deserves a personal analysis before disclosing any kind of early diagnosis. It's definitely crucial at every stage to involve patients in this process.
CHN: And obviously, we'll be very clear that this is not medical advice, it's informational.
Meg Majumder: Yes!
CHN: So informationally, for someone who's reading this and has either a loved one or themselves dealing with neurodegenerative issues, what's the very first thing they can do? A test to request?
Meg Majumder: Yeah, there are tests. So I would just have a general checkup.
I would encourage a patient or a prospective patient to consult with their doctor about the possibility of a neurodegenerative disease.
Doctors are generally able to determine risk based on a simple series of questions, like that's as good as we can do right now. That's the first approach.
But hopefully with the AI software tools we're developing, they will soon be available in general clinics.
So you don't have to go to a neurologist to get access to these tools. Like they'll just be wham, bam: This is what your risk is of Alzheimer's; it's literally a three-step process.
For now, it's absolutely critical that everyone asks. At a certain age, there's such a high percentage of being affected and afflicted with these disorders and diseases that it's an imperative. Not just for the patient, but for the provider.
CHN: And just asking doesn't mean that you'll get a formal diagnosis?
Meg Majumder: Yep, just starting that conversation, there shouldn't be any financial or legal implications or risk.
It's just something that somebody can do, not just to ease their own psyches but the psyches of their loved ones and even maybe be a burden on a provider.
Because how can you use early diagnosis that would enable your provider to initiate early therapies, and that's been shown to delay the onset of plenty of disabilities and it can even possibly slow the onset of disabilities, at least in the context of multiple sclerosis, so. Yeah.
With regard to preventing these things earlier on, it's difficult, because it's complex.
These disorders and conditions, they're embedded in your genetics. So the way in which your genetics interact with your environment, which is a whole new area called epigenetics, which is something we're just beginning to get in to and understand.
What my research is on, we're addressing issue of diagnosis.
So it's not necessarily what you think of; it's more of like a... It's like eye tracking, gait technology, integrating virtual reality, voice processing, and classifying, in order to create an experimental paradigm, through which we can compare and contrast whether or not a certain patient's data is akin to the normal and how much variance would be involved, and whether or not that indicates they're at risk for developing a certain neurodegenerative disorder.
CHN: Broadly, are you optimistic for treating and preventing neurodegenerative issues?
Meg Majumder: Definitely. Especially amongst elderly populations, and that affects a whole range of people starting from like age 30, and there's even earlier onset.
It's harder to do with MS patients, because the majority of our trial applicants and trial members are older. So, there is a bias towards elderly patients, simply because the bulk of people who are associated with neurodegenerative conditions are elderly. But that's something that I definitely want to acknowledge.
Still, I'm hopeful with regard to the technology, not because I'm on the team who's making it, but because we're also considering ethical implications every step of the way. Like, every single step.
I'm trying to cover everything I can.
CHN: So you think that we'll have some exciting new options, even for people who may currently be suffering?
Meg Majumder: People in the earlier stages, for sure, I think we're going to have some interesting options. Just because, oftentimes... This is just a general trend I've observed, but oftentimes with the proliferation of early diagnostic technology, we also have earlier treatment methods because they come hand in hand with one another.
And there are. Like UCSS is coming out with a whole new slate of drugs under Dr. Houser, who used to run the department of neurology there. And his last contributions before he retired were like basically a panacea for mid-stage MS patients, which is incredible, like this whole new treatment regimen.
It's not just UCSS, either.
Harvard is spearheading this as well in conjunction with MIT.
This is Part 1 of a longform interview. Stay tuned for Part 2!