Researchers use smartphone app to predict bipolar disorder symptoms with 71% accuracy

Researchers in Poland have developed a novel method to assess the severity of manic and depressive symptoms in individuals with bipolar disorder based on features of their speech. Using a smartphone app, the team collected voice data and analyzed its characteristics to predict mood states with an accuracy rate of around 71%. The study was published in Acta Psychiatrica Scandinavica.

Bipolar disorder is a serious mental health condition marked by extreme shifts in mood, ranging from emotional highs, known as mania or hypomania, to emotional lows, or depression. During manic phases, individuals may feel excessively energetic, euphoric, or irritable. In contrast, depressive episodes often involve feelings of deep sadness, hopelessness, fatigue, and a lack of motivation. These mood swings can severely impact a person’s daily life, affecting their ability to sleep, concentrate, and carry out daily activities.

It is estimated that between 0.3% and 3.5% of the global population will experience bipolar disorder at some point in their lives. Unfortunately, existing medications for the disorder are not always highly effective. Approximately 22% of individuals with bipolar disorder who regularly take prescribed medications experience a relapse in any given year. Each new episode tends to worsen the course of the illness, making early detection of symptom changes critical. Timely intervention could potentially prevent a full-blown episode, improving the overall management of the disorder and reducing the long-term impact on patients’ lives.

Recognizing this need, study lead author Katarzyna Kaczmarek-Majer and her colleagues sought to determine whether the severity of manic and depressive symptoms could be predicted using features of speech. Their aim was twofold: to identify speech characteristics most closely correlated with these mood states and to develop a statistical model that could predict the severity of symptoms based on those characteristics.

The study involved 51 patients diagnosed with bipolar disorder, with an average age of 36 years. Of the participants, 28 were female. The researchers recruited patients from two bipolar disorder treatment centers: the Institute of Psychiatry and Neurology (IPiN) in Warsaw and the Medical University of Poznan, both located in Poland.

Participants installed a smartphone application called the BDmon app, which continuously ran in the background. The app automatically activated whenever the participant made or received a phone call, recording the first five minutes of each conversation. During these calls, the app collected various speech parameters—such as pitch, loudness, and speech rate—and analyzed them in real time. To ensure privacy, the app deleted the recording after analyzing the speech characteristics, storing only the extracted features. This ensured that no actual conversation content was saved, preserving the participants’ confidentiality.

Participants used the app for an average of 208 days. At the same time, psychiatrists regularly assessed the severity of their manic and depressive symptoms using two well-established clinical tools: the 17-item Hamilton Depression Rating Scale and the Young Mania Rating Scale. These assessments took place every three months, providing a comprehensive understanding of each participant’s mental state over time.

The results revealed significant correlations between certain speech features and the severity of bipolar symptoms, with marked differences between males and females. In men with more severe depressive symptoms, several speech changes were observed: they tended to speak more quietly and with less energy, their speech was more slurred, their voice was smoother (less rough and irregular), and they often made longer phone calls, during which they spoke for longer periods. These features reflect the slowed, low-energy communication style often seen in severe depression, where speech tends to become more monotonous and effortful.

For women, however, the study found no significant correlations between speech characteristics and overall depression severity. The only exception was in cases of psychomotor retardation—a condition in which physical and mental processes slow down markedly. In women with this symptom, louder speech with more irregularities in voice intensity was observed. This suggests that speech features might reflect certain specific depressive symptoms in women, but the connection between speech and depression is less consistent compared to men.

In cases of mania, speech patterns were also distinct between genders. Men with severe manic symptoms spoke louder and more energetically, with rougher voices, more variability in voice intensity, and a sharper tone. Their speech was faster and clearer, a hallmark of the rapid, pressured speech often seen during manic episodes.

Women, in contrast, exhibited opposite speech patterns when experiencing severe mania: they tended to speak more quietly and with less energy, using a lower-pitched voice. Their speech was also slower, more slurred, and their voices were less rough and irregular. These findings highlight a pronounced gender difference in how mania affects speech patterns, with men and women showing almost opposite characteristics in their vocal expression during manic phases.

Using these speech patterns, the researchers developed a predictive model to estimate the severity of manic and depressive symptoms. The model was able to predict mood states with approximately 71% accuracy. This success suggests that voice analysis could be a powerful tool for monitoring patients with bipolar disorder, helping healthcare providers identify when a patient might be transitioning into a manic or depressive episode. Such predictions could prompt timely interventions, potentially preventing a full episode and improving the patient’s overall treatment outcome.

“Speech analysis provides physiological markers of affective symptoms in BD [bipolar disorder] and acoustic features extracted from speech are effective in predicting BD phases. This could personalize monitoring and care for BD patients, helping to decide whether a specialist should be consulted,” the study authors concluded.

The study makes a valuable contribution to developing novel and unobtrusive ways to assess the severity of mental health disorders. However, it remains unclear whether, and how, the use of medications would affect the results. The authors also note that patients with manic symptoms tended to turn off their phones or uninstall the app, leading to missing data that might have influenced the findings.

The paper, “Acoustic features from speech as markers of depressive and manic symptoms in bipolar disorder: A prospective study,” was authored by Katarzyna Kaczmarek-Majer, Monika Dominiak, Anna Z. Antosik, Olgierd Hryniewicz, Olga Kaminska, Karol Opara, Jan Owsinski, Weronika Radziszewska, Małgorzata Sochacka, and Łukasz Swięcicki.