Meditation practice can be described as skill development—the more hours one practices, the more adept one becomes. to me, it seems plausible that time on the cushion with focused attention has resulted in development. Sometimes my motivation flags and I wonder whether I am wasting my time, but overall, I feel that my practice of zazen has strengthened my awareness of habits of the mind, my recognition of conscious choice and my ability to return to clarity. I presume that my practice is not that different from the practice of others.
So what evidence is there for skill development, beyond anecdotal reporting?
I believe that there must be physiological correlates of meditation that can be measured and that changes over time should be observable. If this is true, then we would expect to see observable differences between novices and expert practitioners of meditation.
Previously, I had found an increase in beta and gamma power in my own brainwaves during individual meditation sessions and a hint of possible increase in beta and gamma during the course of a 7-day meditation retreat. However, a brief look at the brainwaves of four long-time practitioners revealed no particular commonality.
To investigate this further, I extracted EEG data from recordings of 30 subjects. The recordings had been made in different contexts over the course of a two-year period; some were during or immediately after meditation retreats, others were made outside of formal retreats. Some recordings were made in studies where subjects were asked to meditate during part of the session and not meditate at other times. In all cases, data was extracted from segments during which subjects were asked to “count their breaths” or “meditate” or “do zazen.” The reason for merging data from several diverse investigations was to increase the number of subjects in the sample.
Prior to participation in any of these studies, subjects were asked to estimate the total number of hours of their meditation experience over their whole life. A paper form was given to each subject which included a table to aid in making the estimate. The table had columns for the average time of daily meditation and the number of years of this daily practice. It also included the number of meditation retreats the subject had participated in and the average number of retreat days and hours per day of meditation during the retreat. All the numbers were converted to hours and totaled. While this was a challenging exercise for people, it did help to ascertain a rough estimate for meditation experience.
Among these thirty subjects, experience reported ranged from 1 hour to 36,000 hours. It would be interesting if the EEG patterns for highly trained meditators were substantially different from those of novices. I initiated an effort to find out.
The next step was to quantify results from the EEG recordings. Using the Physiology Viewer application, I identified an interval of roughly 10-15 minutes with a stable EEG signal for each meditating subject. I chose the absolute mean value of EEG power for each of the standard frequency bands (delta, theta, alpha, beta and gamma). I decided to focus first on alpha power, with the goal of displaying a graph of power vs. hours of meditation experience.
Since the range of experience was so large, I chose the horizontal axis to be the logarithm (base 10) of the number of hours rather than a linear scale of hours. The values for absolute band power measured by the Muse are logarithms which can be negative (when the power values are less than 10), so I decided to plot the inverse log (base 10 exponential) of the absolute band power on the vertical axis. This way all ordinate values would be positive. Below are the results for each of the four sensors (lf, rf, lb, rb) for the alpha band.
The characteristic that most stands out is much stronger alpha power at the back sensors (lb and rb) compared to the front (lf and rf). This suggests that if we want to use the alpha signal as a correlate to practice, we should focus on the back electrodes (TP9 and TP10).
Examining the bottom two graphs, we see no appreciable difference between subjects with ~10,000 hours of experience vs those with ~1,000 hours. Comparing subjects having less than 100 hours of experience (three subjects) with subjects having more (twenty-seven subjects) may suggest that more experienced practitioners had stronger alpha power, but the correlation is weak. Unfortunately our sample included only three people with less than 100 hours experience. If we were to find half a dozen novices with strong alpha, then any correlation with experience would disappear. We just don’t know from the available data.
There are several highly experienced subjects in this sample who show no particularly strong alpha signal at the rear sensors. This particular measurement leaves our original question unanswered. Strong alpha power at locations TP9 and TP10 does not seem to be a predictable effect of long term meditation practice.
What about beta or gamma power? An earlier finding indicated that beta and gamma power increased in my own brainwaves over the course of a 7-day retreat. To find out, I did a similar analysis as above for the beta and gamma bands. Results for beta are shown below (gamma results were similar):
As in the case of alpha power, we see that subjects with the most experience in meditation exhibit a wide range of beta power: some show more than novices; others show comparable values. In any case, power in the beta band at either front or back sensors is not a good predictor of meditation experience.
There are several possible explanations for this lack of correlation, for example,
- The hours of meditation experience reported by subjects on our questionnaire is not a reliable independent variable–our estimates of how many minutes of meditation per day, days per year, and number of meditation retreats are extremely rough, especially when trying to recall many years in the past.
- Sitting on a cushion and “meditating” can mean different things to different people. Even within a single tradition (i.e., Zen), teachers have guided students with different instructions. Students have had varying degrees of success in applying those instructions.
- The limited number of four sensors in the Muse headband may be insufficient to register the electrical signature of meditation.
- Perhaps none of the EEG frequency bands used in this part of the investigation (alpha, beta, gamma) reflect meditation expertise.
- Perhaps the choice of using 10-15 minute intervals of meditation and averaging power over the whole interval was not the best strategy. Maybe it took several minutes for meditators to “get into the groove,” in which case, it might be more revealing to select a shorter interval (‘best’?) near the end of the meditation session.
The next part of the investigation examined power over these selected ‘best’ intervals.