Study of Heart Rate Variability to Comprehend the Significance of Singing Bowl Meditation on the Functioning of the Autonomic Nervous System

Ritika  Upadhyay; Biswajeet  Champaty; Suraj Kumar  Nayak

doi:10.18502/fbt.v12i2.18276

Ritika Upadhyay School of Engineering, Ajeenkya DY Patil University, Pune, Maharashtra 412105, India
Biswajeet Champaty School of Engineering, Ajeenkya DY Patil University, Pune, Maharashtra 412105, India
Suraj Kumar Nayak Department of Electrical and Electronics Engineering, School of Engineering and Technology, ADAMAS University, Kolkata, West Bengal, 700126, India

DOI: https://doi.org/10.18502/fbt.v12i2.18276

Keywords: Meditation; Heart Rate Variability; Autonomic Nervous System; Machine Learning; Radial Basis Function.

Abstract

Purpose: This study aims to determine whether Himalayan singing bowl vibrations could lead to deeper and faster relaxation than supine silence. Numerous civilizations have used singing bowls, gongs, bells, didgeridoos, and voice sounds and chants as instruments for sound healing for ages in religious rites, festivals, social celebrations, and meditation activities.

Materials and Methods: The effect of sound vibrations on physical and mental wellness is supported by scientific research. Although various pieces of research have demonstrated the effect of meditation on humans, very few studies have been done on the beneficial effects of singing bowls on the body and the mind (decrease in unease and temperament, Electroencephalogram, etc.). This study suggests two Machine Learning (ML) models for the automatic classification of the meditative state from the normal state using the Heart Rate Variability (HRV) data.

Results: To pick suitable inputs for the ML models a statistics-based t-test and Principal Component Analysis (PCA) was applied. In the statistics-based t-test method, the HRV parameters were subjected to choose appropriate input for the ML model.

Conclusion: In this case study there are two models that were considered the most effective models based on their accuracy, that are MLP 31-13-2 and RBF 31-17-2 model having a training accuracy of 83.75% and 68.75% respectively. In the second case study, the PCA approach was applied to the HRV parameters, and as a result MLP 4-6-2 and MLP 4-10-2 were the most effective models, with an accuracy of 69.6% and 71.4% respectively.