Classifying meditative states: three meditative traditions under the scope of machine learning techniques

MARTÍNEZ VIVOT, ROCÍO; PALLAVICINI, CARLA; TAGLIAZUCCHI, ENZO

Garrison

Conferencia; Summer Research Institute: Exploring Mental Habits Contemplative Practices and Interventions for Individual and Social Flourishing; 2019

Mind and Life Institute

Literature that supports the benefits of meditative practices have been increasing for the past decades. There is abundance of studies that show enhances in cognition1, emotional processing2 and neuroplasticity3, to name a few, and most of the neuroimaging data are supported by fMRI-based studies that cannot capture real-time brain activity as the electroencephalography (EEG). In a recent work by Braboszcz et al.4 the authors stress that although EEG is one of the primary neuroimaging methods used to study meditation, to date no consensus has emerged on the basic effects of meditation on EEG activity. Hence they have analyzed EEG data obtained from 3 groups of expert meditators and a naïve population as control and we have performed a further analysis of their work searching for underlying information from the EEG data. Though they report a consistent difference between the three traditions studied in comparison to naïve meditators in the EEG activity of the gamma band, and assign a trait effect to the alpha band for the Vipassana tradition, their analysis fail to noticed other nuances that we could be described combining machine learning techniques with the individual channel signal entropy. The present work validates Braboszcz et al. findings and adds new emerging information lost in the average signal of each spectral band.