Predicting depression from daily gross motor activity
Actigraphy is a non-invasive(no bodily intervention) method of monitoring human rest/activity cycles. A small device (looks like smart watch), also called an actimetry sensor, is worn for a week or more to measure gross motor activity. This kind of feature (measuring actigraph) is easily found in latest smartwatches available in the market.
We aim to predict the depression state of a person by analysing the sensor data recorded by the actigraph watch worn by each subject.
Depresjon: a motor activity database of depression episodes in unipolar and bipolar patients. The dataset consists of sensor recorded motor activity data collected from actigraph watches worn by 23 unipolar and bipolar depressed patients and 32 healthy controls. The gross motor activity is recorded continously for several days.
On performing exploratory data analysis, we observed that the gross motor activity for depressed patients is significantly lesser than any healthy person.
- Random Forest Classifier
- XG Boost Classifier
- LSTM
For more information please refer to the codes folder.
| Title | Link |
|---|---|
| Depresjon: a motor activity database of depression episodes in unipolar and bipolar patients | https://dl.acm.org/doi/pdf/10.1145/3204949.3208125 |
| One-Dimensional Convolutional Neural Networks on Motor Activity Measurements in Detection of Depression | https://dl.acm.org/doi/pdf/10.1145/3347444.3356238 |
| Applying machine learning in motor activity time series of depressed bipolar and unipolar patients compared to healthy controls | https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0231995 |
| Monitoring Motor Activity Data for Detecting Patients’ Depression Using Data Augmentation and Privacy-Preserving Distributed Learning | https://ieeexplore.ieee.org/abstract/document/9630592 |