Bipolar Disorder Predictive Model: A Study to Analyze and Predict Emotional Change Using Physiological Signals
2019-07
Loading...
View/Download File
Persistent link to this item
Statistics
View StatisticsJournal Title
Journal ISSN
Volume Title
Title
Bipolar Disorder Predictive Model: A Study to Analyze and Predict Emotional Change Using Physiological Signals
Authors
Published Date
2019-07
Publisher
Type
Thesis or Dissertation
Abstract
Bipolar disorder, a chronic mental illness, most common among people of ages 18 years and older, affects over 2.6 percent of the United States population alone. Although this illness cannot be cured, it can be managed by continuous tracking and monitoring. Hence, if a manic or depressive episode can be identified or predicted in its early stages, severe damage can be minimized, if not prevented. This thesis proposes the design of an objective sensor based system that is based on physiological predictors for the continuous and autonomous monitoring of bipolar patients. This system consists of a pulse rate sensor to record heart rate, and an electrodermal activity sensor to trace the emotional and cognitive state changes and does not rely completely on self-assessment or reporting. Furthermore, it investigates how psychological changes affect physiological responses, such as Heart rate variability (HRV) and Electrodermal activity (EDA). We conducted a study with 50 healthy participants, where each participant was subjected to a certain degree of image and audio induced emotion. Baseline and the emotional stimuli data was collected. Time- domain and non-linear analysis of HRV was then performed on the collected HRV data. In addition, EDA data analysis was performed by decomposing it into tonic and phasic components. We investigated the ability of HRV and EDA to identify the activity of the au- tonomic nervous system in response to emotional stimuli. Moreover, the extracted features from the data were then used to build machine learning models to predict the given psychological change in response to the emotional stimuli. Our results showed that these combination of HRV and EDA features from the study we conducted yielded an average accuracy of up to 73% with Support vector machine, and 68.3% with Discriminant analysis for predicting emotional change in healthy individuals.
Keywords
Description
University of Minnesota M.S. thesis. July 2019. Major: Computer Science. Advisor: Arshia Khan. 1 computer file (PDF); viii, 95 pages.
Related to
Replaces
License
Series/Report Number
Funding information
Isbn identifier
Doi identifier
Previously Published Citation
Suggested citation
Anwar, Yumna. (2019). Bipolar Disorder Predictive Model: A Study to Analyze and Predict Emotional Change Using Physiological Signals. Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/206714.
Content distributed via the University Digital Conservancy may be subject to additional license and use restrictions applied by the depositor. By using these files, users agree to the Terms of Use. Materials in the UDC may contain content that is disturbing and/or harmful. For more information, please see our statement on harmful content in digital repositories.