How many of you have taken care of a family member or a friend suffering from cancer? We might see quite a few hands in the crowd. How many of you have taken care of someone suffering from heart disease? Again, there might be many of you out there. Now, let me ask you this: How many of you have taken care of a friend or a family member suffering from a mental illness? or a member looking for mental health treatment or mental health facilities? The number here is not that big as you would expect it to be. About 1 in 23 people suffer from cancer, 1 in 10 people suffer from heart diseases and about 1 in 6 people suffer from mental illnesses and seek mental health treatment and mental health facilities. While most of you understand and most of us know what it means to suffer from various types of physical illness, the truth is only a few of us understand what it means to suffer from a mental illness and acknowledge it. And the greatest challenge here is that there is no blood test or a medical exam that will give you a “Yes” or “No” answer for mental illness. But there are some significant cues in how we behave in the world that can be in a way predictive of a person’s mental state. The following are a few examples and these are observations made from actual patients who suffered from mental illness:
- “My wife can tell by my walk.” (bipolar disorder)
- “I didn’t want to wake up. I was having a much better time asleep.” (clinical depression)
- “My legs bounce, speech goes fast….even eat too fast.” (mania)
- “I barely have any social contact.” (schizophrenia)
The point being, our physical activities, our sleep patterns, our speech, and our social contact provide meaningful cues about our mental state. And these are just a few examples from the wide range of mental illnesses that are out there, it covers the broad spectrum of where we see changes in their pattern in day-to-day activities that we perform and their susceptibility to risk for decline in mental health.
But there’s good news, we can measure each one of these behaviors using something that’s each in our pockets: Smartphones. We have our smartphones all the time, they have sensors that can monitor our speech, our movement that can be used to measure our sleep and this information can be used to correlate and relate to a person’s mental well-being.
Let’s look at an example:
So, let’s talk about Sleep. The above image displays the sleep pattern of two individuals. The top graph represents an individual who doesn’t suffer from any mental illness and the bottom graph displays the sleep pattern of a person suffering from depression and needs mental health treatment. The graphs show the brain activity over time and the differences are quite evident, the duration of sleep in a depressed person is much shorter and the activity is dampened which often relates to the quality of sleep. So in depressed individuals as well as more broadly in mental health, we see either insomnia, disturbed sleep, or hypersomnia and in some cases excessive sleep than usual. We can use sensors that are in our phones, and just by the way you use your phone it can measure your sleep duration, sleep quality and sleep interruptions very robustly. In addition to that it can also predict the kind of body clock that you have: are you an early riser or a late riser?
The image below shows 8-days of data of a person- the blue represents solid bars of a person’s sleep and the red bars represent phone usage.
As you can see from the image, some patterns are quite clear. For example, on day-7 the person had an interruption at 3:00 am, in this case, they had to wake up to drop a friend at the airport in the middle of the night which leads to him catching up by sleeping more on day-8.
Learning from the data with the help of Machine Learning, can help reliably pick up interruptions and the sensors can predict what is interrupting the person’s sleep- is it something in the environment around them or is it their behavior that’s actually causing them to sleep less.
How we speak matters. We can use cell phones which have sensors that can reliably capture speech. Tanzeem Choudhury, an associate professor in Computer and Information Science at Cornell University mentions that they have developed some privacy sensitive techniques that are able to detect when someone speaks, how someone speaks- whether someone is speaking fast or slow, in a monotone or has an infection. Depressed people have been measured to speak slower and walk slower, these behaviors often known as “Behavioral Indicator” to see if someone is declining.
From this information we can figure out how many conversations a person has, the people they are talking to and the amount of duration they’re talking for. In the image below, the lines indicate different conversations with the length being the duration they’re talking for.
And after a certain duration if this is how the lines look, we know something has happened. The image tells us that the person is not engaging with the community or the rest of the world. Such information can really provide important cues about a person’s mental health state.
In fact, this kind of information has been evaluated and validated with the help of clinicians and health doctors. What was observed is that just looking at simple measures of how someone speaks can be revealing of a person’s mental health state, correlations were observed with depression, general mental health and also how sociable a person is.
This is very simple information, collectable from devices that are with us 24/7 and can be used to predict a person’s risk for decline.
For the longest time it was impossible to even measure these behavioral changes unless you have someone following you around and taking notes 24/7 and a person’s memory is often compromised when they’re suffering from a mental illness and needs mental health treatment, so now we can actually get reliable measurements of behavior but that alone is not enough, the right intervention has to be provided. So you might ask: “How do we go from data to intervention?” Access to mental health in countries like the U.S.A and India is still poor because of the stigma that revolves around mental health but research has shown the right behavioral intervention of bringing the person’s behavior back to normal can sometimes help prevent mental health decline or stabilize the person’s mental state and look for mental health facilities. So we really need to look at intervention as well as measurement.
There’s some very simple technology in terms of therapist in a pocket. If your phone knows you’re having some irregularity with your sleep that is associated with mental health decline, it can give you suggestions that a therapist might give, at the time the person needs it the most.
A person’s behavior can also be mined and based on their behavior.
Mental Health Treatment
Mental Health Treatment can be quite complex, and we need to have social support which is something that’s lacking in most of the societies. Despite acknowledging the problem, one of the most important factors of mental health treatment is our behavior, and behavior tracking and behavior sensitive intervention which has been really challenging so far, and we hope technology can actually tackle that and deliver that measurement and intervention in a scalable and accessible manner. Finally, there’s medication, just like you would treat heart disease or diabetes – mental health is no different. We need to make changes in our lifestyle and also have access to the right kind of medication.
There are many apps designed for this purpose- Check out these mental health apps.
It’s high time we stop treating mental health problems as an isolated issue and look for mental health treatment as soon as possible. Thinking of mental illness as a sign of weakness or lack of willpower is something that needs to change as well, just like any other physical illness it is an illness.
I would like to end this article with a quote from Mike Wallace, who was an American journalist, game show host, actor and media personality:
