Phone app to detect mood changes in bipolar patients

Bringing together the expertise of mental health researchers, computer scientists, and sound engineers, the app could help people with bipolar disorder predict mood swings with enough time to receive help from their treating doctors. 

Participants in the pilot study all had a rapid-cycling form of type 1 bipolar disorder and a history of being prone to frequent depressive and manic episodes.

Running in the background on an ordinary smartphone, the program monitors the patients’ voice patterns during calls and analyses the characteristics of the sounds – and silences – of each conversation.

Patients in the study were also required to undertake standardised weekly mood assessments with a trained clinician, which provided a benchmark for the patient’s mood and was used to correlate the acoustic features of speech with their mood state.

The researchers showed that the app could accurately detect elevated and