Suicide is a prevalent cause of death worldwide and depression is a primary concern of many suicidal acts. It is possible that an individual during depression never has any suicidal thoughts at all. On the other hand, some individuals in stable condition with no apparent symptoms of depression feel urges to commit suicide (suicidal ideation). Many such individuals never let anyone know what they are feeling or planning. Suicidal ideation considered an important precursor to suicidal acts.
Detecting the suicide risk in individuals with mood disorders is a major challenge. The current clinical practice to assess suicide risk in these vulnerable individuals based on structured or semi-structured psychiatric interviews is inadequate as many of the suicidal behaviors often occur unpredictably especially during apparent clinical remission. Furthermore, some of these individuals are unable or unwilling to share their experiences with clinicians. An objective feature that can continuously monitor risk of suicidal thoughts would be advantageous in such situations.
Our research focused on finding objective features in activity data for detecting suicidal ideation in a sample of individuals diagnosed with Bipolar I, Bipolar II, or Unipolar who were currently in a euthymic state. Euthymic state is considered a non-depressed and reasonably positive mood state, but individuals in this state may still have suicidal ideation. Hence, our work explores detecting risk of suicidal thoughts in euthymic individuals in a group of mood disorder subjects using machine-learning approaches.
Statistically significant differences were observed between activity features of euthymic and depressed individuals. A strong negative correlation was observed between activity feature vulnerability index with self-rated suicidal ideation. This study demonstrates that we can use machine learning techniques to detect risk of suicide in euthymic individuals from activity data. The main advantage of using activity data is that it would be cost effective, since many people commonly use activity trackers.
Date of publication
Dr.Premananda Indic,Dr.Sarah Sass,Dr.David Beams
Master of Science in Electrical Engineering
Atluri, Pallavi, "Detecting Suicide Risk From Wristworn Activity Tracker Data Using Machine Learning Approaches" (2018). Electrical Engineering Theses. Paper 35.
Available for download on Wednesday, April 29, 2020
Bioelectrical and Neuroengineering Commons, Biological Psychology Commons, Biomedical Commons, Clinical Psychology Commons, Community Psychology Commons, Counseling Psychology Commons, Educational Psychology Commons, Health Psychology Commons, Signal Processing Commons, VLSI and Circuits, Embedded and Hardware Systems Commons