Experimental App Listens to Your Voice to Predict Bipolar Episodes
Photo by Thinkstock
Your smartphone can tell you where you are, how many steps you’ve walked, and whether your flight is on time. But how well can it understand the state of your mental health?
As Slate’s Future Tense reports, a new app named PRIORI is aiming to ameliorate the emotional ups and downs that come with being bipolar in a novel way: by monitoring the quality of patients’ voices as they talk on the phones.
The program, which was designed by University of Michigan psychiatrist Melvin McInnis, aims to predict changes in mood. McInnis, with the help of computer scientists from the school’s College of Engineering, designed the app based on a common symptom used to diagnose patients with mental disorders: speech characteristics. Bipolar patients experiencing a manic phase tend to speak in rapid bursts and struggle to focus on just one idea, while those in depressive states are usually more irritable and less energetic in their diction. Several recent studies, Slate points out, have shown that pauses, pitches, and jitters in speech all tend to be indicative of a bipolar patient’s current mental state.
Once you have downloaded and set up PRIORI on your smartphone, it’ll begin recording the audio of your calls. Obviously, to avoid violating others’ privacy — and several state laws — it’s designed to monitor only the patient’s side of the conversation. As that audio is collected, it’s encrypted and then sent to a University of Michigan lab, where a program analyzes it.
At no time does the program take the actual substance of a conversation into account. It’s all about the volume, stops, starts, and intonations of the person’s voice.
Though the app is still making its way through preliminary tests, a recent group study showed that it could predict an upcoming change in mental state about 65 percent of the time.
This isn’t the first app that aims to monitor and improve a person’s mental state. As my colleague Rob Walker noted earlier this year, “happiness apps” are proliferating at an alarming rate. Last year researchers at Dartmouth developed an app called FOCUS to help schizophrenia patients manage their illness. Slate also mentions asimilar project by MIT fellow Max Little, which was able to diagnose Parkinson’s disease through 10-second voice recordings with 98.6 percent accuracy.
It may take a leap of faith to send all your personal conversations to a computer database far, far away, but the fact that PRIORI is a passive tool, rather than one that requires a patient’s constant attention, is key. As anyone who has struggled with depression knows, updating an app with your current mood would be the first responsibility to go as your life started falling apart.
That being said, a new generation of bipolar millennials are coming, and they’re much less likely to have meaningful conversations via phone calls, as compared to texting, instant messenger apps, emails, and social media posts. Just the other day I — a 20-something reporter who speaks with people on the phone regularly for my job — realized that I hadn’t spoken on the phone with a good friend even once in our three years of friendship. It’s admirable that researchers have taken it upon themselves to adapt to the technology-centric lives of their patients. But without a program that can analyze text-based conversations, apps like PRIORI might not be of much help for future generations.
[Via Slate]
Follow Alyssa Bereznak on Twitter or email her here.
