A Voice Detecting Depression? Lindsey Venesky, PhD, Discusses New Data

A recent study showed the promise of Sonde Health’s mental fitness vocal biomarker tool in mental health tracking, making it possible to identify depression from the sound of a voice.^{1, 2} The tool can work in any voice-enabled device.³

Earlier this year, another study was published regarding the app, MoodCapture: the first smartphone application using AI to detect the onset of depression based on facial cues alone.⁴ MoodCapture accurately diagnosed 75% of participants with early symptoms in the clinical trial.

The arrival of Sonde Health’s mental fitness vocal biomarker tool offers another way to detect depression with one of the senses, but this time, the voice. Led by Erik Larsen, PhD, from Sonde Health in Boston, investigators examined 104 outpatient psychiatric patients at the Cognitive Behavioral Institute in Pittsburgh who recorded 30-second voice journal entries on Sonde’s Mental Fitness smartphone app over 4 weeks.¹ On the app participants would talk about their thoughts and feelings.

The biomarker tool analyzed the recordings for 8 acoustic features—jitter and shimmer, pitch variability, energy variability, vowel space, phonation duration, speech rate, and pause duration—and gave people the following scores: 80 – 100 (excellent), 70 – 79 (Good), 0 – 69 (pay attention). The analysis identified increased or decreased mental health risks, consistent with the M3 Checklist.

Participants were twice as likely to report increased mental health symptoms if they received a “pay attention” score versus an “excellent” score over 2 weeks, ranging from an improved risk from 1.53 (95% confidence interval [CI], 1.09 – 2.14; P = .0138) to 2.00 (95% CI, 1.21 – 3.30; P = .0068). Furthermore, using the tool 5 – 6 times per week had a greater likelihood of increased symptoms for those who had received an 8.5 score (95% CI, 2.31 – 31.25; P = .0013).

Participants reported positive experiences with the tool. Approximately 40% changed their behavior or lifestyle and 30% noticed benefits to their well-being.²

In an interview with HCPLive, Lindsey Venesky, PhD, NCSP, from the Cognitive Behavioral Institute, was fascinated to learn about MoodCapture and wondered if someone were to adjust their facial expression, the app would provide a different response.

“What I really liked about the vocal biomarkers is there’s not much that you can really change about your voice to change the results,” Venesky said. “We did [have] some people that would try to change the pitch or how quickly they were speaking, to see if they would get different results, and it didn't always make too much of a difference. They were able to really get an accurate assessment of what works.”

Venesky said the app was “insightful” for her as a therapist to see some clients did not believe their score, thinking it was inaccurate until they talked about their week and realized it was more difficult than they believed.

“And their score was more accurate than they thought it would be,” Venesky said. “Even sometimes, there were some interesting situations where the app really provided some insight for the client that they may not have had otherwise. I think that might be a really good way of seeing how [useful it could be]."

There are no reported disclosures for Lindsey Venesky.

^References

1. ^{Larsen, E, Murton, O, Song, X. Validating the Efficacy and Value Proposition of Mental Fitness Vocal Biomarkers in A Psychiatric Population: Prospective Cohort Study. 2024; 15. https://doi.org/10.3389/fpsyt.2024.1342835.}
2. ^{Study Confirms Sonde Health's Voice-Based Mental Fitness Solution Accurately Identifies Individuals with Elevated Mental Health Symptoms. Sonde Health. March 11, 2024. https://www.sondehealth.com/news-blog/sonde-healths-voice-based-mental-fitness-solution-accurately-identifies-individuals-with-elevated-mental-health-symptoms. Accessed March 13, 2024.}
3. ^{Unleash the Power of Voice as a Vital Sign. Sonde Health. https://www.sondehealth.com/. Accessed March 13, 2024.}
4. ^{Derman, C. Smartphone App with AI Detects Depression Onset from Facial Expressions. HCPLive. February 28, 2024. https://www.hcplive.com/view/smartphone-app-ai-detects-depression-onset-facial-expressions. Accessed March 13, 2024.}