Differences in mobility patterns according to machine learning models in patients with bipolar disorder and patients with unipolar disorder

Patients with bipolar disorder showed statistically significantly lower mobility compared to those with unipolar disorder during depressive states.

• Patients with bipolar disorder had a total duration of moves per day that was lower than that of patients with unipolar disorder.
• In classification models, mobility data achieved a sensitivity of 0.70 and a specificity of 0.77 for distinguishing between bipolar and unipolar disorder.
• The area under the curve (AUC) for these classification models was 0.79, indicating good performance in differentiating the two disorders.
• The findings suggest that smartphone-based location data may assist in the diagnosis of bipolar disorder versus unipolar disorder.

AI simplified