Predicting Calmness, Anxiety, and Depression using Wearable Sensors

Abstract

One of the main reasons for the high prevalence of mental disorders is that there is no technology to aid diagnosis or to report on recovery factors like effect of therapeutic interventions and medicines. To enable faster access to screening and to measure recovery, we propose a wearables-based framework for the automatic prediction of the states of anxiety, depression and calmness in individuals. The framework called H2SEC is based on the integrated measurements of Habituation, Hypoactivity, Synchronization, Experience and Calmness (H2SEC) through a unified framework. The framework also enables the tracking of these states in realtime, showing transition from one state of consciousness to another. To build and validate the H2SEC framework, we collected Electrodermal Activity (EDA) and Blood Volume Pulse (BVP) data, sampled at 100 Hz using a wearable device from 61 Neutral, 60 Depressed and 110 subjects with Anxiety, while they performed the experimental task. By tracking the timeseries data in centiseconds, we identified segments of time in the subjects’ data where their per-minute state of Sympathetic Nervous System (SNS) was lower compared to the previous minutes. Within these segments called Sympathetic Transition Points (STPs), we calculated scores related to the H2SEC parameters of Habituation, Hypoactivity, Synchronization, Experience and Calmness and compared them against similar scores from non-STP segments. We thus arrived at the physiological coordinates of the Neutral, Anxiety, and Depression groups, apart from identifying the intrinsic nature of calmness in each group. We implemented multi-output, multi-label machine learning (ML) algorithms to predict mental states(Neutral, Anxiety, and Depression) along with the nature of calmness (Calm, Approaching Calmness, and Not Calm) in each subject. We report 100% F1, along with 100% Precision and Recall in identifying both the states. Our methodology is the state-of-the-art in terms of mental health monitoring, and to the best of our knowledge we are the first to report on mental health disorders (anxiety and depression) and recovery mechanism (calmness) using an integrated methodology.

Real-Time Physiological Signal Feedback for Anxiety Intervention Music System for Postpartum Women

“…The emotion recognition system designed by Udovi i et al through the Shimmer3 device to collect GSR and PPG is used to identify surprise, disgust, joy, fear, sadness and anger [30]. An H2SEC framework was constructed to predict calmness, anxiety, and depression [31]. Mai et al designed an ear-electroencephalogram to identify negative emotions and alert users on their phones [32].…”

Heartfulness Meditation: A Yogic and Neuroscientific Perspective

“…Moreover, recent studies on Heartfulness have found a rebalancing of the Autonomic Nervous System (ANS), even in novice practitioners. Indeed, Rajlakshmi Borthakur ( Patel, 2021 ; Borthakur et al, 2022 ), who has been studying the effects of Heartfulness practices using wearable devices that measure Heart Rate Variability (HRV) and Electrodermal Activity (EDA), found that in novice meditators, high frequency power was higher than controls when meditating with Transmission, while Low Frequency power was lower. These findings point at a sympatho-vagal balance that shifts to the parasympathic side ( Patel, 2021 ; Borthakur et al, 2022 ).…”

“…Indeed, Rajlakshmi Borthakur ( Patel, 2021 ; Borthakur et al, 2022 ), who has been studying the effects of Heartfulness practices using wearable devices that measure Heart Rate Variability (HRV) and Electrodermal Activity (EDA), found that in novice meditators, high frequency power was higher than controls when meditating with Transmission, while Low Frequency power was lower. These findings point at a sympatho-vagal balance that shifts to the parasympathic side ( Patel, 2021 ; Borthakur et al, 2022 ). Their study also found that experienced practitioners were longer able to hold a state of calmness, meditators and trainers had increased heart rate variability, but decreased skin conductance response (SCR), and during deep meditation experienced practitioners had a decrease in their sudomotor nerve activity (SMNA).…”

“…Heart resonance between trainer and practitioner may also be investigated. Finally, wearable devices for practitioners to measure heart rate variability and skin conductance response, as used by Borthakur et al (2022) , will make it possible to assess the effects of Meditation and Cleaning on day-to-day well-being ( Balconi et al, 2019 ). The extent to which practitioners either experience or actively practice constant remembrance is also of interest to investigate, as we expect that this will greatly affect the effectiveness of the meditation practice, both in terms of well-being, as well as in speed of development of consciousness.…”