Patient Monitoring Technologies for In-Situ Assessment of Emotional Processes In Naturalistic Settings

点击翻译按钮获取中文摘要

In-situ and ambulatory patient monitoring in naturalistic settings enables symptom observation outside controlled clinical settings, ultimately enhancing assessment and diagnostic processes. While a plethora of systems have been proposed for in-situ monitoring of physiological pathologies (e.g., cardiac abnormalities, seizures), technologies that monitor psychiatric- and mental health-related symptoms are scarce. These technologies can complement and enhance existing assessment methods which generally rely on highly subjective self- and care-giver reports. In this domain, monitoring of emotional processes is of specific interest as these processes play a key role in psychopathology (e.g., anxiety, depression, borderline personality disorder, bipolar disorders). The long-term objective of the proposed program is to develop intelligent, mobile, and wearable computing systems for continuous and in-situ monitoring of symptoms associated with psychiatric disorders as well as the related contextual factors. In this proposal, we will focus on emotion monitoring based on physiological signals as these variables can be non-invasively measured with wearable sensors in a variety of environments. Physiological change is also a key component of the emotional experience and atypical physiology has been associated with psychopathology. Existing systems for physiological monitoring of emotional processes have been developed and evaluated in highly controlled laboratory settings where emotional experiences are artificially induced. To allow the translation of these systems ambulatory and in-situ emotion monitoring in naturalistic settings, technical challenges related to modeling inter- and intra-personal variability as well as confounding affects of other cognitive and physical processes must be resolved. To address these gaps, the short-terms objectives of this program are to 1) develop algorithms for continuous estimation and adaptation of a time-varying physiological baseline, 2) develop classification algorithms for differentiating emotion-related physiological changes from those associated with cognitive and physical activity, and 3) develop a wearable and mobile system for continuous measurement and labeling of emotion-related physiological responses in real-life conditions. This research program will establish innovative new methods to advance emotion detection research from controlled laboratory settings to naturalistic environments. Ultimately, the outcomes of the proposed program can improve and enhance diagnosis and assessment of psychiatric and mental health disorders. The methods developed in this research program will be directly transferable to other application domains including affective computing and enriched human-computer interaction.