Table 5 Examples of passive monitoring of patients with bipolar disorder related to smartphones, Internet activities, or wearables
Technology | Sensors | Aim | Primary measures | N | Findings | Study |
|---|---|---|---|---|---|---|
Ingestiblea | Ingestible sensor in tablets. Wearable sensor on torso | Measure medication adherence | Adherence metrics. Logs date and time of tablet ingestion | 28 | System is feasible in patients with BP and SCZ | Kane et al. 2013 |
Internet social media | Â | Differentiate depression subgroups by language use | Analyze topics and linguistic features in 24 online communities interested in depression | 5000 blog posts | Five distinct subgroups, one is BP. For those with BP, topics on medications and BP most important | Nguyen et al. 2015 |
Internet social media |  | Explore language differences among 10 mental health conditions | Using public Twitter posts 2008–2015, group by classifiers including self-reported diagnosis | >100 users/group; >100 posts/user | Language usage patterns differ by condition | Coppersmith et al. 2015 |
Smartphone | Accelerometer, GPS | Detect mood state | Daily mobility (physical motion), and travel patterns (number of locations visited, time outdoors) | 12 | Can detect a change in mood state. Less precise to detect mood state | Gruenerbl et al. 2014 |
Smartphone | Accelerometer; microphone | Detect mood state | Number of apps running; app usage patterns and selection. MONARCA software | 18 | Patterns of app usage vary with self-reported mood | Alvarez-Lozano et al. 2014 |
Smartphone | Accelerometer | Detect mood state | Overall activity levels | 9 | Substantial individual variation in activity levels, both daily and within intervals | Osmani et al. 2013 |
Smartphone | Â | Detect mood state | Number and duration of ingoing and outgoing calls; number of text messages. MONARCA software | 61 | Patterns of calls and texts vary in manic and depressive mood states | Faurholt-Jepsen et al. 2015 |
Smartphone | Microphone | Detect mood state | Phone call statistics; acoustic emotional analysis, and social signals from daily calls | 12 | Speaking length and call length among the most important predictors of mood | Muaremi et al. 2014 |
Smartphone | Recorder for outgoing speech | Detect mood state | Voice monitoring and acoustic analysis of speech patterns from continuously recorded outgoing calls | 6 | Can recognize manic and depressive mood states | Karam et al. 2014 |
Wearable (T-shirts) | Electrodes and sensors integrated into garment | Detect mood state | ECG and respiration. Long term heart rate variability analysis. PSYCHE monitoring system | 8 | Can differentiate mood states (depressed, manic, mixed, euthymic) | Valenza et al. 2014 |