Depression is one of the leading causes of disability worldwide, affecting an estimated 300 million people. Evidence-based treatments are available and measurement-based care has been described as the gold standard. Monitoring of depressive symptoms is currently performed with self-administered and interview- based assessment methods conducted by clinicians in their offices. However, the shortage of mental health specialists and the limited resources available to primary care physicians who often manage patients with depression, prevent close monitoring of symptoms delaying optimal treatment potentially prolonging suffering. Passive recording of behavioral data (gathering information without individual's direct input) has been identified as a potentially feasible method for long-term monitoring of depression. To date, most studies have collected passive behavioral data in real time through mobile apps (i.e. accelerometer, phone clicks) with the goal of identifying potential markers of depression. However, this method lacks critical biological indicators of depression, including sleep, arousal, and motion. Recent development in wristband sensor technology developed by out lab has allowed to measure physiological parameters like gait, heart rate variability (HRV) and electrodermal activity (EDA) continuously in ?real time?, allowing a broader anatomical and neurophysiological understanding of emotion, behavior, and cognition in mood disorders as they occur during routine activity. During the past decade, along with the development of sensors, we have seen the progressive use of machine learning, a branch of artificial intelligence that enables the detection of complex patterns in multimodal data, allowing the development of complex models. The combination of sensor technology and machine learning allows detailed measurement in real time of a wealth of behaviors predicting mood variation. Over the past 2 years, our interdisciplinary team, including one of the leading lab on depression research, and one of the most innovative lab on affective computing, has conducted a study applying machine learning analytics to create a model combining wristband sensors data and phone- based passive measurements to assess severity of depressive symptoms. In our pilot study with depressed patients monitored over 8 weeks, we found that an algorithm based on biological and behavioral sensor data could estimate depression severity evaluated by a clinician with high accuracy. The proposed study will further refine our model in a sample of 100 adults with depression, assessed over 12 weeks. We anticipate that the proposed study will enable the development of an objective, passive, sensor-based algorithm able to measure depressive symptom severity. The identification of reliable, objective, passive assessment of depressive symptoms with biosensors will have significant ramifications for the monitoring of depression, early detection of response, remission and relapse and ultimately contribute to the advancement of precision medicine.

Public Health Relevance

The proposed project will create a novel method to assess depressive symptoms by harnessing modern machine learning analytics, phone sensors, and wrist-band sensors. Measurement-based treatment is considered optimal and the development of a valid passive, objective, behavioral and biological assessment of depressive symptoms that does not rely on clinician interviews will improve monitoring and ultimately improve treatment significantly. Ultimately, the ability to leverage a remarkable wealth of behavioral and biological data has the potential to transform the delivery of treatment for depression.

Agency
National Institute of Health (NIH)
Institute
National Institute of Mental Health (NIMH)
Type
Research Project (R01)
Project #
1R01MH118274-01
Application #
9643493
Study Section
Biomedical Computing and Health Informatics Study Section (BCHI)
Program Officer
Ferrante, Michele
Project Start
2019-01-01
Project End
2023-10-31
Budget Start
2019-01-01
Budget End
2019-10-31
Support Year
1
Fiscal Year
2019
Total Cost
Indirect Cost
Name
Massachusetts General Hospital
Department
Type
DUNS #
073130411
City
Boston
State
MA
Country
United States
Zip Code
02114