|Table I Group averages for median oscillation amplitude, Am (beats/min ± S.D.), and Fourier spectral power, S [(beat/min)2 ± S.D.].|
|Chi (n=8)||Kundalini(n=4)||Spontaneous nocturnal||Metronomic||Elite athletes|
|pre-meditation||during meditation||pre-meditation||during meditation||breathing (n=11)||breathing (n=14)||(nocturnal) (n=9)|
|Am=4.34 ± 1.03||6.37 ± 2.41*||3.41 ± 0.97||13.55 ± 3.41*||3.78 ± 1.39ª||3.20 ± 0.91©||2.58 ± 1.37§|
|S=22.27 ± 8.15||35.09 ± 19.59*||14.40 ± 5.23||98.30 ± 43.35*||18.71 ± 8.40ª||10.98 ± 4.99©||15.54 ± 11.41§|
* p<0.01 for comparison of meditation
groups before and during meditation (paired t-test).
ª p<10-4 for comparison of pooled meditation groups and spontaneous breathing control group (t-test).
© p<10-4 for comparison of pooled meditation groups and metronomic breathing control group (t-test).
§ p<10-4 for comparison of pooled meditation groups and elite athletes group (t-test).
Non-parametric comparison (Wilcoxon rank sums test) gives similar results. Note that the actual magnitude from minimum to maximum heart rate within one oscillatory cycle is equal to twice the amplitude computed with Hilbert transform technique used here.
Figure 2: Three hours of continuous heart rate spectra, before, during, and after Chi meditation. Spectra are calculated from the instantaneous heart rate for 15-minute intervals every 109 seconds using Lomb periodogram for unevenly sampled data , and are smoothed using a sliding window 50 points wide. Note compaction of spectral energy in the 0.05 Hz band during meditation.
Figure 3: Fifteen minutes of heart rate and respiration during Chi meditation. Both heart rate and ECG-derived respiration (EDR) [5,4] were resampled at 2 Hz, locally detrended with a sliding window 75 points wide, and smoothed with a sliding window 11 points wide. Spectra were calculated using a Fast Fourier Transform. The coherence plot confirms that the very prominent heart rate oscillations at Hz are correlated with respiration at same rate.