Mindfulness Meditation Decreases EEG Functional Connectivity

Abstract

Introduction The brain is composed of enormous neurons which can only function by their connections. These connections form a complex networks both at microcircuits and macrocircuit levels. The neural circuits constitute the underlying biological machinery for the genesis of functional rhythmicity which may mediate various brain functions such as working and resting. The brain states can be adjusted by external and internal information. Meditation and mindfulness have long exist and been assumed to bring the mind to unique state of peace and lucidity. One important feature of mindfulness is to be non-judgmental, non-reactive. We hypothesize that the brain may merely overview the ongoing information flow without being fully affiliated in its processing, thus the interactions/coherence between brain regions may be less intense during this mental practice. Method To test this assumption, we used a unique coherence analysis to investigate the brain connectivity during mindfulness meditation. Specifically, minimum variance distortionless response (MVDR) based coherence was applied to analyze different spectrum of six EEG band, including delta (0.5~4.0Hz), theta (4.0~8.0Hz), alpha1 (8.0~10.0Hz), alpha2 (10.0~12.0Hz), beta (12.0~30.0Hz), gamma (30.0~45.0Hz); see figure 1A. MVDR method can minimize the variance distortionless of the output signal through a bank of filters, with a constraint g_k^H u_k=u_k^H g_k=1 (Benesty,Chen et al.2005). According to this constraint, the process signal passed through the filter g_k have no distortion along u_k. In the generalized MVDR method, the magnitude-squared coherence can be expressed as: C_xy^2 (f)=|S_xy (f)|^2/(S_xx (f) S_yy (f) )=C_xy^2 (u_k )=|u_k^H R_xx^(-1) R_xy R_yy^(-1) u_k |^2/([u_k^H R_xx^(-1) u_k ][〖u_k^H R〗_yy^(-1) u_k]) where R_xx is the covariance of x, R_xy is the cross-correlation matrix of x and y. R_xx=E{x(n)x^H (n)} , R_xy=E{x(n)y^H (n)}. A 128-channel NeuroSCAN EEG system was used to collect the data. Closed-eye EEG data were collected from nine participants during mindfulness breathing versus during normal rest, both at the beginning (pre-) and after 8-week standard mindfulness training (post-). For convenient of data analysis, the channels were averaged to 22 areas. Results Significant difference was always found between mindful-state breathing and rest-state; while no difference was found between the pre-and post- data, thus they were combined together and there were 18 data sets. Please refer to our previous work (Gao, Fan et al. 2016). We also separately analyze the first-, second- and last- 2 minutes of mindful breathing. As shown in figure 1B, there is a decreasing tendency of coherence between FLR and TAR (the locations can be found at figure 2) areas during time. The overall results (Figure 2) show significant decrease in brain connectivities in most of the spectrum band, including delta 1-4Hz (hubs mainly in the right frontal lobe, right parietal and bilateral occipital lobe), theta 4-8Hz (mainly in the right frontal lobe, right parietal lobe and bilateral occipital lobe), alpha 8-12Hz (mainly in the right parietal lobe and slightly in the right frontal lobe) and beta 12-30 Hz (mainly in the left temporal lobe). Only sporadic brain regions have increased connectivities mainly in the alpha1 and gamma band. Conclusion The current coherence analysis shows clearly a trend of reduced brain connectivity during mindfulness meditation. It indicates that the brain regions are less engaged in the spontaneous processing on the ongoing information, especially the right hemisphere, thus the individuals may become less reactive to the external and internal information. Further study with better signal processing method, more psychological, together with different stages of meditation can better demonstrate this issue (Zhang, Cai et al. 2009).