Fitbits help detect mood episodes in people with bipolar disorder

Using machine learning models, researchers have found that fitness trackers may help detect episodes of mania and depression in people with bipolar disorder.

Researchers from a hospital in Boston have found that using passive sensor data from digital devices like smartwatches can accurately detect changes in mood in those with bipolar disorder (BD).

Bipolar disorder is characterized by extreme mood swings that fluctuate between periods of depression and mania.

Depression is often described as the feeling of extreme and persistent sadness, hopelessness, or indifference. Meanwhile, mania is sometimes characterized by feelings of intense happiness and/or euphoria. However, manic episodes can sometimes be accompanied by psychotic symptoms.

The study collected data from 54 adults diagnosed with BD who wore Fitbits for nine months.

Researchers then applied machine learning models to the data, which was collected over two-week observation periods, to detect changes in mood.

What they found was that Fitbits could accurately detect changes in mood, including episodes of (hypo)mania and depression.

By applying a machine learning algorithm to the data, they were able to identify symptoms of depression with over 80% accuracy and (hypo)mania with over 89% accuracy.

This is particularly useful as researchers could apply these algorithms to predict when patients with BD may experience a depressive or manic episode in between appointments with their psychologist, psychotherapist, or healthcare physician.

How can smartwatches track mood changes?

The researchers focused on passive sensor data gained from fitness trackers to identify when a person was going through a depressive or manic episode.

Passive sensor data includes data collected without input from the user. For example, wearable devices like Fitbits or Apple Watches collect data such as your activity, heart rate, sleep patterns, blood oxygen levels, and more.

Depending on the type of episode, whether that be depressive or manic, certain physical changes can occur.

For example, those experiencing a depressive episode might move less than usual or sleep more than usual. This is the type of data collected passively from smartwatches and fitness trackers.

In contrast, those who are experiencing a manic episode could have an increased heart rate, move more, and sleep less. However, symptoms will vary from person to person.