Removal of Subject-Dependent and Activity-Dependent Variation in Physiological Measures of Stress

Abstract

The ability to monitor stress levels in daily life can provide valuable information to patients and their caretakers, help identify potential stressors, determine appropriate interventions, and monitor their effectiveness. Wearable sensor technology makes it now possible to measure non-invasively a number of physiological correlates of stress, from skin conductance to heart rate variability. These measures, however, show large individual differences and are also correlated with the physical activity of the subject. In this paper, we propose two multivariate signal processing techniques to reduce the effect of both forms of interference. The first method is an unsupervised technique that removes any systematic variation that is orthogonal to the dependent variable, in this case physiological stress. In contrast, the second method is a supervised technique that first projects the data into a subspace that emphasizes these systematic variations, and then removes them from the data. The two methods were validated on an experimental dataset containing physiological recordings from multiple subjects performing physical and/or mental activities. When compared to z-score normalization, the standard method for removing individual differences, our methods can reduce stress prediction errors by as much as 50%.

Detecting cocaine use with wearable electrocardiogram sensors

“…Higher frequencies are displayed as darker colors. We can see that some features like RR, TH and QT show distinct 1 We note that the electrocardiology literature recommends slightly different procedures for computing the QT, PR and QRS interval lengths based on where tangent lines to the individual waves intersect the baseline. We do not use this approach because numerical estimates of the tangent lines remain noisy even for our smoothed samples.…”

“…While most of the previous studies, involving mobile and on-body physiological sensing, used heart rate and heart rate variability as features [1,7,11] only few studies used limited forms of morphological features [2,3]. To the best of our knowledge our ECG feature set is the most extensive ever used.…”

“…Wearable ECG for Health and Activity Sensing: A substantial body of work has explored the use of ECG using wearable chestbands, primarily in the context of understanding physiological stress [1,11], assessing cognitive load [8] and detection of arrhythmias caused specifically by atrial fibrillation [2,3,12]. The work on physiological stress focus primarily on the use of heart rate-based features to determine an individual's stress level.…”

“…We define the QT, PR and QRS interval to be the distance between the identified locations of the Q and T peaks, P and R peaks and Q and S peaks, respectively. 1 These distances are all measured in the standardized waveform space where the RR interval length has been completely normalized away. To calculate QTc in its standard units, we then need to multiply our QT interval (where RR has be normalized away) by the square root of the original RR interval length.…”

Modeling the user state for context-aware spoken interaction in ambient assisted living

“…These new technologies create a network of communication channels and heterogeneous systems that help individuals in a variety of situations, such as detecting the needs of health professionals and patients [115,141], healthcare interdisciplinary design [12,93], gamificatio and virtual environments [13,17,99], physical activity and behavior evaluation [107,126,149], therapy and rehabilitation [143,166], supporting the elderly [82,91], monitoring sleep and affect [38,147], gait and exercise evaluation [39,126], in-home monitoring and mobile healthcare [12,71,76,146,162], supporting medical decision [101], or psych-physiological sensing [4,92].…”