Long-term effects of meditation and the differences between meditators and non-meditators have been studied extensively. In order to understand the corresponding mechanisms of action, brain areas activated while meditating need to be studied. Such studies are limited due to lack of a suitable experimental paradigm and analysis technique. We investigated experienced meditators who underwent a 20 minutes audio-guided meditation in the fMRI scanner. Using the Intersubject correlation (ISC) analysis, correlated activity across the subjects was found in frontal pole, middle frontal gyrus, precuneus, primary somatosensory cortex, visual cortex, lingual gyrus, and cerebellum.
In recent years, interest in meditation has increased dramatically. Many benefits such as increase in cortical thickness1, reduction in stress2, reduction in fatigue, anxiety and increased mindfulness3, increased positive mood states4,5 etc have been reported. Using fMRI, differences between long-term meditators and novices in terms of emotional regulation6 and emotional reaction7 to different stimuli have been established. Meditators have higher activation in prefrontal cortex (mPFC), ventromedial prefrontal cortex (vmPFC), hippocampus which signify a state of meta-awareness.
All these studies have observed the effect of long-term meditation on physiological and behavioral characteristics. But there have been a dearth of studies which observe the immediate effects of meditation on the brain. The dynamics of meditation are still unknown. An MRI scanner environment is not inherently conducive to a meditative practice. Also, supine position is not a default for meditation. We chose Yoga Nidra meditation, a unique practice meant to be done in the supine posture, to study the brain dynamics.
24 experienced yogic practitioners (average yogic experience of 2,989 hours, mean age 27.53 years, 20 males) volunteered from local meditation centers. The total yogic experience included time spent in yogic asanas (postures), pranayama (breathing exercises), meditation and chanting. Participants were subject to Mini International Neuropsychiatric Interview (MINI) screen and only those with no history of mental illness were recruited for the study. Ethical clearance was taken from the IIT Delhi Institute Ethical committee.
We acquired the MRI data on a Philips Ingenia 3.0T whole-body scanner at Mahajan Imaging, Gurugram. The anatomical data were acquired using a T1-weighted 3D MP-RAGE protocol (TR = 2250 ms, TE = 4.18 ms, FOV = 256 mm, flip angle = 9°, voxel size = 1×1×1 mm). Functional data were acquired using a multi-band echo-planar imaging pulse sequence (TR = 1000 ms, TE=25 ms, 65° flip angle) with a multi band factor of 8. Forty-eight 4 mm thick oblique-axial slices were acquired in head first order with no gap. The acquisition matrix was 73x73 with 3x3x4 mm voxels.
The protocol employed is given in Figure 1. We used FSL for preprocessing8. The brain was extracted from the high resolution structural file using the repeated version of the FSL-BET. Custom Python code was used for slice-time correction for multi band acquisition. This was followed by motion correction, temporal filtering (100s), spatial smoothing with 6mm and co-registration across the dataset. Normalization was done with an MNI 2mm template.
Auditory signal was downsampled, convolved with HRF, and then correlated with BOLD signal voxelwise across the whole brain to determine a subject’s auditory response to the guided instructions. Using a bootstrapping method to create a null distribution, correlation values were converted to p-values. A correlation greater than 0.15 corresponded to uncorrected p<0.05, 2 subjects having insignificant auditory correlation were excluded from further analysis Fig 2. shows the auditory correlations.
The Inter Subject Correlation9 analysis was performed on the selected subjects. A bootstrapping method was used to create a null distribution, the voxel wise correlation value was converted to a p-value which was corrected(p<0.05) for multiple comparisons. Figure 3 shows the corrected p-values for the correlated activity in the auditory cortex, primary somatosensory cortex, visual cortex, middle frontal gyrus, precuneus cortex, frontal pole, lingual gyrus, and cerebellum.