on the EEG Rhythm in Meditation

Fig. 1

Locations of the 62 electrodes

2.2 Signal Processing

The raw EEG data were digitally referenced to the linked mastoids [9]. EEG with significant myoelectric artifacts was manually excluded. FFT, calculated in MATLAB software, was applied for time-frequency analysis on the EEG with 1,024 ms time window and sliding 512 ms. Power on each channel was computed by FFT for 1-min time window in each state. Mean energy ratio of five trials, energy of different frequency bands to total energy, for the meditator (the study subject) and non-meditators (control group) were compared. Topographic maps represented the power distribution of EEG over the scalp were plotted by using interpolation on a fine Cartesian grid. Mean coherence at different frequency for four pairs, F3-F4, P3-P4, F3-P3, and F4-P4 in 1 min EEG for five trials was estimated by a normalized cross spectral density function [8] implemented by function “mscohere” in MATLAB. Coherence coefficient, which varied between 0 and 1, increased with high coherence.

Meditative state is considered to be similar to a stage of sleep or drowsiness. For comparison, sleep EEG from ten male adults was recorded with the Neuroscan NuAmps Digital Amplifier on Fz channel, sampled at 250 Hz and referenced to linked mastoids followed by a band-pass filtering between 0.5 and 30 Hz. Meditation EEG on Fz channel was down-sampled to 250 Hz and band-pass filtered from 0.5 to 30 Hz before the power distribution of meditation and sleep in the low frequency band (<30 Hz) was compared.

3 Results

3.1 Power Distribution Analysis

The EEG of study subject (meditator) in normal resting state before meditation was different from control group (non-meditators). The power of delta band (<4 Hz) was lower while high frequency band (4–80 Hz) was higher for meditator than non-meditators as shown in Figs. 2 and 4.

Fig. 2

Average energy ratio in different frequency bands on channel CZ for 1 min in normal state of the meditator and non-meditators

Fig. 3

Average power of EEG recorded from all channels for 1 min in normal state of the meditator compared with non-meditator

Fig. 4

Topographic map of PSD of delta (1–3 Hz), theta (4–7 Hz), alpha (8–13 Hz), beta (14–30 Hz) and low gamma band (31–80 Hz) respectively in normal state of (a) control group and (b) the study subject

Alpha band (8–13 Hz) over the scalp was much more active for meditator than non-meditator in normal state and dominant alpha frequency decreased in meditator which was displayed in Fig. 3 and topographic maps Fig. 5. For normal resting state, EEG of meditator was dominated with parietal occipital alpha compared with midline frontal (7–9 Hz), parietal and occipital (9–13 Hz) alpha dominance in the control group. Active alpha rhythm was correlated with feeling of calm and positive affect in early biofeedback studies [10, 11]. Therefore, the meditator was much easier to get this positive state than non-meditator.

Fig. 5

Topographic comparison of mean alpha power over the scalp in 1 min normal state of (a) meditator and (b) control group for five trials

The active time and power of the alpha band (8–13 Hz) decreased, while the power of delta-theta rhythm (<8 Hz) increased in the first three phases of meditation, followed by significantly increased gamma activity in the fourth phase according to time-frequency analysis (Fig. 6). Phase4 was apparently different from normal state and other meditative states as the first three phases were searching process and the fourth phase was final subconscious state. The increased power in gamma band might be associated with thinking activities of the meditator. EEG rhythm of the meditator could quickly return to normal state after meditation. The dynamic quantitative power variation of EEG during the whole meditation experiment could see in Fig. 7. In contrast, the EEG power maintained stable during the whole resting experiment for control group.