Article Preview
TopIntroduction
Meditation is an ancient spiritual practice of achieving physical and mental wellness in the countries like India, China, Tibet, Japan etc. (McMahan & Braun, 2017). At present, it has growing recognition as a therapeutic intervention in western countries as well (King, 2019). The methods of meditation are classified vividly based on the countries of origin irrespective of their common outcome. These methods based on states of mindfulness are roughly categorized into two basic categories i.e., concentration meditation and insight meditation (Eddy, 2017). When a meditator focuses on his/her breathing, such type of meditation is called concentration meditation while when a meditator focuses on inner feeling, it is considered insight meditation. One of the very popular insight meditation methods is Sahaja Yoga meditation which is practised in more than 130 countries today (Dodich et al., 2019). This meditation method is based on the concept of a subtle system consisting of Kundalini and chakras. Through meditation, the awakening of Kundalini and the clearing of chakras is achieved (Hendriks et al., 2021). Along with insight meditation, music guided mediation using ragas and devotional music is also practised for swift and deep meditation (Dhir & Sharma, 2020).
Since the last decade, in place of rising health concerns, many research studies have been conducted for qualitative and quantitative affirmation of thoughtless awareness in the respect of medicine and psychology. In many such studies, it has been acclaimed that meditation affects the sympathetic and parasympathetic activities of autonomic nervous systems (ANS) which further improvises the regulation of the functions of human vital organs thereby controlling functions such as heartbeat, respiration, and digestion. For assessment of effects on ANS, heart rate variability (HRV) is considered as the best tool (Vigo et al., 2019) compared to activities of the brain, body temperature and other parts of the biological system using the latest technological and statistical tools to understand the effect of the various types of mediation on human psychology and physiology (Massaro & Pecchia, 2019). For HRV, direct heart rate (HR) measurements are done or derived from the ECG signals. In the second type, either R-peaks or QRS intervals are detected using automated methods. Unfortunately, there are great challenges for automated detection because of the morphologies and amplitudes in cases of abnormal QRS complexes. The superimposed noise in the ECG signal makes this even more tricky.
The existing methods applied on multiple samples or complete signals and hence also performed offline. Such methods are not suitable for real-time analysis, hence sample to sample-based semantic segmentation is being explored widely. In this scenario, sample-wise classification of the signal is led to extract P/QRS/T or neutral (n/a). Numerous research studies have conducted the ECG segmentation using the conventional segmentation approach in the past (Bayasi et al., 2014; Beraza & Romero, 2017; Campbell et al., 2017; Ghaffari et al., 2010; Gupta et. Al., 2011; Homaeinezhad et al., 2011; Kang et al., 2015; Karimipour & Homaeinezhad, 2014; Madeiro et al., 2012; Martinez et al., 2010; Mukhopadhyay et al., 2011; Sodmann et al., 2018; Vazquez et al., 2011; Vitek et al., 2010). However, these approaches are not applicable for continuous real-time signals as they only classify the samples into either P/QRS/T or n/a. Unlikely, semantic segmentation performs the multiclass classification to extract or locate multiple regions/objects from the single input data (image/signal). Such methods also allow assessing the cardiac effects using the local ECG components rather than analyzing the global parameter like heart rate (HR).