Can artificial intelligence help maintain your mental health? Can it predict and understand when we may need help?
‘Predictive maintenance’ is a concept gaining ground in manufacturing and business circles. Its essence is this: thanks to technologies like artificial intelligence and the internet of things, it is now possible to continuously monitor every aspect of a production line, collecting and processing data in near real-time to gain a complete picture of events. This means, in theory, that problems can be spotted – or perhaps more accurately put, sensed – before they arise and fixes can be made proactively. The result, so the thinking goes, is big savings of time and money because costly equipment failures and down time are eliminated.
Could the same logic be applied to humans? Inasmuch as our bodies can be considered biological machines whose every action can be measured, the answer appears to be a resounding ‘yes’. The plethora of apps, wearables and gadgets that have appeared in recent years that track everything from sleep patterns to blood glucose levels make it clear that, when it comes to physical health, the philosophy of proactive management through data-rich digital technologies is very much in vogue.
What about mental health? Could artificial intelligence monitor and predict potential mental health hiccups, eliminating the time it takes to acquire help and diminishing the downtime of recovery from serious mental problems, by offering mental health maintenance? When it comes to mental health, the picture is less straightforward than for physical health in general. While there may no longer be any debate over the efficacy of computers to monitor objectively observable aspects of our being, such as our heartbeat, or how many steps we take a day, it is far from universally accepted that digital technologies can, or should, be used to in any way track, let alone treat, disorders of emotion, cognition and mood. But that has certainly not stopped clinicians and entrepreneurs from trying.
Much more than telemedicine
Among those with the most success to date in commercializing the idea of digital mental health is the team behind Mindstrong. A startup founded in 2017, it is today riding high on a wave of support from venture capital. In its last funding round, Mindstrong raised $100 million from a star-studded list of investors that included Bezos Expeditions, the investment firm of none other than Amazon founder Jeff Bezos. Interest in the early-stage company has no doubt been bolstered by the presence of top figures from the world of medicine among its co-founders, including Thomas Insel, a former director at the US National Institute of Mental Health, and Rick Klausner, previously a top executive at the Bill and Melinda Gates Foundation.
Big names aside, Mindstrong’s chief selling point is its technology. The Mindstrong app, currently only available through certain US health insurance providers, offers two types of care. The first is, simply, telemedicine. This means that, through the Mindstrong app, users can schedule therapy sessions with a clinician which are then conducted through video, messaging or phone call. We might term this the ‘uberization’ of care, making it – like food, entertainment and taxi rides – convenient, smartphone-accessible and available on-demand. But useful though such a service is, it is not why investors are excited by Mindstrong; apps offering digital access to therapy have been around for nearly a decade (The digital therapy space is now a crowded market, with early entrants SilverCloud -founded 2011 – and Talkspace (2012) among the biggest names. Those US-based companies have since been joined by numerous competitors from around the world, including Spain’s Braingaze and the Berlin, Germany-based Selfapy).
What sets Mindstrong apart – or so its founders claim – is its use of data and artificial intelligence (AI). The app tracks how users interact with their smartphones, capturing data and aiming to provide insight into mental health. The Mindstrong website describes the concept thus:
“Once you download our app, our technology starts to get an idea for how you tap, scroll and type on your phone. How you passively use your smartphone…is a new way to measure things like your stress, mental health symptoms, and well-being. If you’re typing more slowly—even by a millisecond—it might mean there’s a change.”
The ‘technology’ Mindstrong is referring to is machine learning. The company’s algorithm analyzes the data collected through the app, reading changes in how a user interacts with their smartphone as a sign that something may be wrong with their mental health. That information is passed on to a user’s clinical team, who can then reach out to help. The promise of Mindstrong’s technology – and this, perhaps, is a big part of why it has won so much funding from venture capital – is that it may even be able to detect mental health issues before they arise in a kind of ‘predictive maintenance’ for the mind.
‘Measuring’ mental health
At the heart of Mindstrong’s thesis is the idea that the key to diagnosis and treatment of mental disorders is access to objective measures of mood and cognition. Such metrics, the startup’s founders believe, are sorely lacking in traditional psychiatry, where treatment proceeds on the basis of a patient’s own, self-reported thoughts and feelings.
Mindstrong’s turn to the smartphone is an attempt to gather that objective data in the same way that doctors gauge blood pressure to detect hypertension or look at blood sugar level to check for diabetes. Such metrics, described by the US Food and Drug Administration as “characteristics of the body you can measure”, are known as biomarkers. Mindstrong calls its version of these metrics ‘digital’ biomarkers, and they include many seemingly mundane activities such as how much time elapses between a press of the spacebar and the input of a letter when typing a text message, or the interval between scrolling through a web page and clicking a link.
Exactly how closely changes in that kind of digital metric correlate to the presence of mental illness is an area of active research and debate. One study found that when the severity of depressive symptoms increased in patients with bipolar disorder, the number of phone calls they would receive per day and the amount of time their smartphone screen was “on” increased too. At the same time, the number of outgoing calls decreased, as did the number of incoming calls that were answered. These results led the researchers to conclude that smartphones are “useful” when it comes to monitoring symptoms of patients suffering from bipolar disorder.
Mental health practitioners elsewhere, however, are more cautious. At the Beth Israel Deaconess Medical Center in Boston, Massachusetts, the digital psychiatry unit has been studying the links between mental health and digital biomarkers since 2018. Rashmi Patel, a Fulbright Scholar at the center, told outlet mobihealthnews that it is too early to say if the smartphone-centric approach to psychiatry is effective: “We’re still trying to understand what these biomarkers mean, how to interpret them,” he said. “So we don’t yet know what could indicate, for example, when an individual is becoming more unwell or when they are getting better.”
An emerging field of inquiry
The term ‘‘digital biomarker’ is still relatively new and lacks an authoritative definition. It is, arguably, a misnomer, having little to do with our biological makeup and instead seeking to tell us something about our emotional or psychological state. Nonetheless, it does appear to be well on its way to acceptance in research and academia, as does, more generally, the practice of measuring mental health by using digital data as a proxy.
Of course, the approach is not without its critics. One set of authors, writing in the journal “Digital Biomarkers”, point out that while “digital biomarkers have emerged as a scalable and practical tool”, very little is currently known about who uses mental health apps. Are they male or female? Old or young? As a rule, app developers do not share that kind of information and in many cases it is illegal for them to do so. This means that the field of digital biomarker research is on shaky ground, or so the paper’s authors suggest. “Without an understanding of which patients are using apps”, they write, “both the personal- and population-level benefits of mobile and connected mental health technology may be lost.”
Nonetheless, the big bets made on Mindstrong by venture capital reflect the belief that digital biomarkers are one of the possible futures of psychiatry. That view, naturally enough, has often been expressed by Mindstrong’s founders. They argue that by collecting patients’ digital data, clinicians stand to reap big benefits, such as gaining insight into hour-by-hour changes of a patient’s mood and mental state. Apart from the potential to predict mental health episodes, that level of granular information could also pave the way toward a new era of care that is highly customized and responsive to the needs of individual patients.
Populations at large stand to gain too. Imagine if everyone in the world carried a device in their pocket that could constantly monitor mental health and predict when they are at risk. Digital biomarker technology promises to bring exactly that service to the billions of people around the world who own a smartphone.
Despite the dream and the venture capital trust gained, many questions remain: Will digital biomarker be enough to signal when someone’s mental health has dipped? What biomarkers would be the most reliable? Are phone users likely to want to give access to their data in order to receive the benefit of apps accessing such biomarkers? Can the privacy of the users be protected, and can reliable scientific and user studies be conducted to improve the technology on privacy-enhanced data?
As it stands, many questions regarding the use of biomarker technology remain open. Both biomarker technology and the science behind it need to progress for a few more years before we have a clear answer to the real question: Will they help improve our wellbeing and bring us peace of mind?
For more articles subscribe here.
Author: Rahim Rahemtulla
Editor: Dr. Dr. Ana-Maria Olteteanu