ITHACA, N.Y. — Although mental health awareness is a major factor in society today, spotting the signs that someone is suffering is not always easy. Could a device most people use every day also alert someone when they are about to have a problem? Researchers at Cornell University say smartphone data can actually predict when patients with schizophrenia will have a relapse, even a month before an episode occurs.
Schizophrenia is a chronic and serious mental health condition that affects around 20 million people globally. According to the World Health Organization, schizophrenics suffer distortions in their thinking, perception, emotions, language, and behavior. They commonly experience hallucinations and delusions, such as hearing voices and developing false beliefs. WHO adds schizophrenics are two to three times more likely to die prematurely than the rest of the population.
The study looked at how data on movement, ambient sound, and sleep patterns could help record drastic changes in a schizophrenic patient’s behavior. Sudden shifts could then throw up a red flag that a relapse is about to take place.
“The goal of this work was to predict digital indicators that are early warning signs of relapse, but these symptoms or changes can be very, very different from one individual to another,” says Dan Adler, a doctoral student at Cornell Tech in a university release.
“We tried to create an approach where we could tell a clinician: Not only is this participant experiencing unusual behavior, these are the specific things that are different in this particular patient,” the study’s first author adds. “If we can predict when someone’s symptoms are going to change before relapse, we can get them early treatment and possibly prevent an inpatient visit.”
What signals a schizophrenic episode?
Researchers reviewed data from the phones of 60 participants over the span of one year. Eighteen of the patients had a schizophrenic relapse during that time.
Using encoder-decoder neural networks, study authors could build behavioral patterns from a person’s sleep habits, the numbers of calls they miss, and even the length of their conversations. This neural network is a type of machine learning system that takes all this irregular data that may seem unconnected and detects trends amid the jumble of information.
The results show there was a 108-percent average increase in behavioral abnormalities up to 30 days before a schizophrenic relapse.
This data also helped another study focusing on how body rhythms predict certain symptoms of schizophrenic behavior. These rhythms can also be monitored by a smartphone. That study, published in Scientific Reports, reveals circadian rhythms (which regulate the sleep-wake cycle) influenced a patient’s ability to sleep, feel calm, and socialize.
Ultradian rhythms, which encompass all the biological clock functions that occur during the day like your heart beat, have more influence over seeing things and hearing voices.
“Taken together, these different types of rhythms provide a more intuitive way to interpret the relationship between a patient’s behaviors and their symptoms,” study authors write. “This can determine when and the type of intervention to be delivered to avoid certain symptoms or prevent them from worsening.”
Can this prevent other mental health incidents?
Researchers add this technology isn’t just for spotting possible emergencies, smartphones can also play an active role in stopping them. The study suggests that if a smartphone is monitoring for schizophrenia and detects spikes in the level of ambient noise — a problem that can affect hallucinations — the system can warn that person to move to a quieter location.
“We wanted to provide some actionable steps or clinically interpretable features so we can either tell the patient to take some actions or tell the clinician to suggest some early interventions,” Cornell Tech doctoral student Vincent Tseng explains.
The study finds that such learning tools could also be modified to work with other disorders like depression. Since extreme behavioral changes tend to precede various mental health episodes, researchers believe their predictions can spot these changes too.
The study appears in the Journal of Medical Internet Research mHealth and uHealth.