The Ayahuasca Experiment
Ayahuasca is a hallucinogenic tea that has been consumed by indigenous tribes in South America for many centuries. More recently, it has gained popularity and mainstream appeal evidenced by increased ayahuasca tourism in countries such as Peru and Brazil. Sometimes fondly referred to as “Mother Ayahuasca”, the active hallucinogenic compound is DMT (N, N-dimethyltryptamine) which causes vivid visualisations.
Intense visuals aside, it is often claimed that ayahuasca has the power to heal psychological ailments such as depression, anxiety, and post-traumatic stress disorder (PTSD). Furthermore, almost everyone who consumes ayahuasca claims it was a life-changing experience (myself included).
Finally and most relevant in our case, it is believed by many that taking ayahuasca can catalyse an enlightenment experience or “kundalini awakening”. As far as I know, no one has tested this notion and this case study attempts to do exactly that.
Can ayahuasca catalyse higher levels of consciousness?
I do not know. Fortunately, we do not need to know what it is to measure its change. For this, I propose the concept of a “Consciousness Fingerprint”.
I define the consciousness fingerprint as a set of pre-defined metrics recorded at a regular frequency that represents the brain state for that period of measurement. In other words, it is a snapshot of the brain. This set of measurements does not address the nature of consciousness itself, but rather provides us a way to track its change over time.
Theoretically, the metrics that compose the consciousness fingerprint are completely arbitrary. For example, you could decide that your emotional state ranging from 1 (very sad) to 10 (very happy) is the consciousness fingerprint for the day. Indeed for some purposes and with sufficient data, this could even be an effective consciousness fingerprint (if the subjectivity of rating one’s own emotional state can be ignored). However, for this case study I have the hardware and software to create a much more rigorous consciousness fingerprint and have chosen the following set of metrics to be recorded daily:
| a) Average power spectral density (read more about this here) of the brain waves produced during Anuloma Viloma Pranayama |
b) Average power spectral density (PSD) of the brain waves produced during Ajapa Japa meditation |
c) Left/Right Brain hemisphere correlation during Anuloma Vilomna Pranayama |
| d) Left/Right Brain hemisphere correlation during Ajapa Japa meditation |
| e) Average breath retention time over 5 rounds of Wim Hof Breathing |
These five measurements were recorded daily over the 30-day experiment and enable us to track changes in the brain. The more dissimilar our metrics are, the better resolution lens we have into the state of consciousness on that day.
Measurements (a)-(d) were all taken with the Muse2 Brain Sensing Headband, however, they are capturing different types of meditations and at different times of the day.
Furthermore, measurements (c) and (d) were not concerned with the PSD, rather the correlation of the raw electrical data coming from each brain hemisphere. Wim Hof breath retention (e) was an intelligent addition to the mix as it does not even use the Muse2 and thereby further reduces the correlation between the metrics themselves and the risk of systemic hardware/software error.
BTW, WTF is Power Spectral Density?
For our purposes, it suffices to think of it as a heuristic for the power or amplitude of a particular type of brain wave. Another way to think of power spectral density is through the following analogy from James Clutterbuck, founder of MindMonitor:
“To understand what this does, think about colour. Any specific colour can be made by mixing red, green and blue. So if you see purple paint, you know it was made by mixing Red and Blue. How much red and how much blue is the relative power density. Now imagine you have thousands of colours and you’re getting closer to what the FFT math does with the RAW EEG signal”
Again, I do not know. However, there is a popular theory that gamma waves are associated with states of enlightenment, expanded intelligence, and interconnectedness. Since the Muse2 Brain Sensing Headband can detect gamma waves, we can choose to define “higher states of consciousness” as an increased power spectral density (PSD) in the gamma spectrum of the brain waves.
Another way to measure higher consciousness would be left/right brain hemisphere coherence. Thanks to the placement of the nodes on the Muse2, we can analyse the raw electrical data coming from each sensor on each side of the head and calculate hemispheric correlation.
Now that we have a way to track changes in consciousness over time and we have defined what is meant by “higher consciousness” we can proceed.
Experimental Design
- Average power spectral density (PSD) of the brain waves produced during Anuloma Viloma Pranayama
- Average power spectral density (PSD) of the brain waves produced during Ajapa Japa meditation
- Left/Right Brain hemisphere correlation during Anuloma Vilomna Pranayama
- Left/Right Brain hemisphere correlation during Ajapa Japa meditation
- *Average breath retention time over 5 rounds of Wim Hof
*Unfortunately, after a software update, all the data I recorded in the Wim Hof app during the 30-day case study has been deleted. I have contacted support and will update this case study with the Wim Hof breath retention data as soon as the issue is resolved.
The case study was divided into three phases:
| Day 1-10 | Control phase | Set Baseline (no ayahuasca) |
| Day 11-20 | Test phase | Take Ayahuasca |
| Day 21-30 | Integration phase | Continue daily measurements (no ayahuasca) |
As the only bodily organ that studies itself, the brain is notoriously biased and often paradoxical. Here is one of my favourite books on the subject.
One great example of a potential bias and how I circumvented it: during course of the case study I only collected data, I never analysed or processed it in any way. I never knew if the experiment was “going my way” or not until after. That is because even the act of knowing how know well the experiment was going while taking measurements would have influenced my brain waves! The brain is indeed a wily beast.
To collect clean and reliable brain data was quite challenging. The brain is a very noisy place and is affected by absolutely everything. Identifying a repeatable brain signal-response is often akin to finding a needle in an electromagnetic haystack.
Furthermore, I collected this data in the jungle which made it even more challenging and prone to experimental error. Empirically however, ingesting ayahuasca has an extremely profound effect on the brain-state. Therefore, I hoped that the significance of this factor would outweigh the noise and produce some detectable signal.
Nonetheless, I followed very austere rules for the duration of the 30-day case study in order to de-noise the retrieved data and produce more statistically valid results:
- Vegetarian/Vegan Diet
- No smoking
- No sex
- No alcohol
- No Coffee / Stimulants
- 16/8 Intermittent fasting
- No Media (Netflix, Youtube, etc.)
- Binduasana (I will expound on this in future content)
- No artificial sugar
- No fluoride
- No vitamins/supplements
In a further effort to regulate the data, I followed a very strict daily schedule; standardising both the time of day I took the measurements and (as much as possible) the physical state of my body. The daily schedule was as follows:
- Wake up no later than 8 am
- First thing after waking → 40 minutes of Hatha Yoga (same sequence every day). Pushup Pranayama would be an excellent (and more time-efficient) option for future experiments.
- Immediately following Hatha yoga, 1st measurement→ 5 rounds of Wim Hof, record average breath retention
- Immediately preceding breaking my fast (between 1 pm – 2 pm), 2nd measurement session → Anuloma Viloma pranayama measurement with the Muse2 Brain Sensing Headband paired with MindMonitor software
- Immediately preceding sleep (at latest 11 pm), 3rd measurement session → Ajapa Japa meditation again with the Muse2 Brain Sensing Headband combined with MindMonitor software
Hypothesis
The power spectral density (PSD) of the gamma wave segment of the brain wave spectrum and the brain hemispheric coherence will increase during the Test phase (Day 11-20) compared to the Control (Day 1-10). It will subsequently decrease in the Integration phase (Day 21-30) but remain elevated compared to the Control.
Hardware
The hardware used for this experiment was the Muse2 Brain Sensing Headband. The Muse2 is a 4-node EEG device with nodes located at AF7, AF8, TP9, and TP10 according to the 10-20 international node placement system. While the hardware is fairly robust, the associated software is not as it is tailored to a retail meditation market and not scientific inquiry. For that reason, I paired the Muse2 hardware with more advanced software, described next.
Sensor placement on the Muse2 Brain Sensing Headband
Software
The software used in this experiment is from MindMonitor. MindMonitor allows the user to pair the Muse2 and collect raw brain electrical data from each of the nodes. Furthermore, via the Fast Fourier transform, it decomposes the overlapped frequencies entering each node into its constituent parts thereby capturing the power spectral density (PSD) of each brain wave type.
Screenshot of MindMonitor’s real-time EEG software in action #allwavesmatter
Video Diary: A bit of unhinged psychedelic fun
In addition to all the quantitative measurements, I thought it would be fun and informative to make a short video on each day of the experiment. All videos were created in one attempt with no re-takes to, as authentically as possible, demonstrate my psychological state and insights for that day (which certainly do noticeably change over the 30-day period). Check it out here.
Results
The data collected over the 30-day experiment was rich, providing potential insights on the full spectrum on brain waves (#allwavesmatter) and more. Check out the full raw data set here. For this case study, I will only focus on the data that concerns the hypothesis, specifically the gamma wave PSD and the left/right brain hemisphere coherence.
Below is the average PSD of the gamma waves produced by my brain during each phase of the case study. Recall that my hypothesis was that the PSD of the gamma waves in the Test phase would exceed the Control baseline. However, here we observe a different story:
In fact the PSD of the gamma waves throughout the Control and Test periods were almost exactly the same. Furthermore, the standard deviations of these averages, 0.09 and 0.08 respectively, were very close. Therefore, the PSD of the gamma waves throughout the Control and Test phases had essentially the same statistical distribution – the act of taking ayahuasca did not catalyse any change.
Interestingly however, the PSD of the gamma waves did seem increase in the Integration phase in a statistically significant way (see Conclusion).
A quick note about negative PSD values: don’t worry about. It is simply how the folks at MindMonitor decided the scale the raw data and is totally arbitrary.
Next, let’s take a look at how brain coherence behaved during the course of the case study. Brain coherence was measured by calculating the correlation between the raw electrical data from the left side of the brain (nodes TP9 and AF7) and the right side of the brain (nodes TP10 and AF8).
My hypothesis seems at least partially correct. Brain coherence increased in the Test phase compared to the Control. However, to my surprise, it apparently continued increasing in the Integration phase as well! A seemingly repeating theme across both the PSD and the coherence results is not to underestimate the importance of the Integration phase. However, we will re-visit this in the next section where we test for statistical significance and determine how well the hypothesis held up.
Conclusion
In order to draw conclusions from the collected data, averages are not sufficient – we must test for statistical significance.
I will now revisit the hypothesis using the data collected and use the one-tailed two-sample t-test to investigate the statistical significance of the difference in means of the gamma wave PSD and the brain coherence during the Control, Test and Integration phases of the case study.
Although the 0.05 significance level is standard in most experiments, I have chosen to use the 0.1 significance level to account for the fact that I am collecting the data in the jungle (this ain’t yo momma’s lab). Despite my best efforts, there is much that could have added variance to the data (mosquitoes, power outages, snakes, giant beetles flying into my face, etc.), so I think p-value = 0.1 is reasonable.
Starting with the gamma brain wave power spectral density, we can ask:
| Null Hypothesis | There is no difference between the means of the gamma wave PSD in the Control and Test Phases |
| Alternative Hypothesis | The gamma wave PSD in the Test Phase is higher than the gamma wave PSD in the Control Phase |
| Significance Level | 0.1 |
| Control Phase Sample Size | 10 days |
| Test Phase Sample Size | 10 days |
| Control Phase Sample Mean | -0.05563540256 |
| Test Phase Sample Mean | -0.05583854617 |
| Control Phase Sample Standard Deviation | 0.09107687979 |
| Test Phase Sample Standard Deviation | 0.08410346376 |
| Standard Error | 0.03920253901 |
| Degrees of Freedom | 17 |
| t-stat | -0.00518189931 |
| p-value | 0.502 |
| Conclusion | We can not reject the null hypothesis |
| Null Hypothesis | There is no difference between the means of the gamma wave PSD in the Control and Integration Phases |
| Alternative Hypothesis | The gamma wave PSD in the Integration Phase is higher than the gamma wave PSD in the Control Phase |
| Significance Level | 0.1 |
| Control Phase Sample Size | 10 Days |
| Integration Phase Sample Size | 10 Days |
| Control Phase Sample Mean | -0.05563540256 |
| Integration Phase Sample Mean | 0.04764017146 |
| Control Phase Sample Standard Deviation | 0.09107687979 |
| Integration Sample Standard Deviation | 0.1184136483 |
| Standard Error | 0.04724065002 |
| Degrees of Freedom | 16 |
| t-stat | 2.186159039 |
| p-value | 0.022 |
| Conclusion | We can not accept the null hypothesis |
| Null Hypothesis | There is no difference between the mean brain hemispheric coherence in the Test and Control phases |
| Alternative Hypothesis | The mean brain hemispheric coherence is higher during the Test phase compared to the Control |
| Significance Level | 0.1 |
| Control Phase Sample Size | 10 days |
| Test Phase Sample Size | 10 days |
| Control Phase Sample Mean | 3.32% |
| Test Phase Sample Mean | 8.13% |
| Control Phase Sample Standard Deviation | 6.32% |
| Test Phase Sample Standard Deviation | 5.40% |
| Standard Error | 2.63% |
| Degrees of Freedom | 17 |
| t-stat | 1.8303 |
| p-value | 0.0424 |
| Conclusion | We can not accept the null hypothesis |
The brain coherence during the Integration phase is quite surprising. I had hypothesised that brain coherence should increase while taking ayahuasca (which it did), but would decrease during the integration phase (although remaining elevated compared to the Control). Instead, left-right brain hemisphere coherence continued to increase materially during the Integration phase. Astoundingly, my brain coherence increased by about 462% over the course of the 30-day case study.
Did left-right brain hemisphere coherence increase during the Integration phase compared to the Test phase (while taking ayahuasca)?
Short answer: Yes
Long answer:
| Null Hypothesis | There was no difference between mean brain hemispheric coherences in the Test and Integration phases |
| Alternative Hypothesis | Mean brain coherence increased during the Integration phase compared to the Test phase |
| Significance Level | 0.1 |
| Test Phase Sample Size | 10 days |
| Integration Phase Sample Size | 10 days |
| Test Phase Sample Mean | 8.13% |
| Integration Phase Sample Mean | 15.34% |
| Test Phase Sample Standard Deviation | 5.40% |
| Integration Phase Sample Standard Deviation | 13.05% |
| Standard Error | 4.47% |
| Degrees of Freedom | 11 |
| t-stat | 1.6141 |
| p-value | 0.0674 |
| Conclusion | We can not accept the null hypothesis |
Sources of Error
As mentioned earlier, this is jungle science baby (this ain’t yo momma’s lab, think this is game? This the jungle). Although I took extensive steps to de-noise the data (see Experimental Design), there are many factors that could have introduced error into the dataset. Here are some of them that immediately come to mind:
- Due to internet or power outages, on some days I was only able to capture data from one meditation session. On these days the average of PSD or brain coherence was actually just from one reading (Anuloma or Ajapa Japa, but not both)
- Moon cycles. From past experience, I strongly suspect that the phase of the moon has a statistically significant effect on brain wave behaviour. It’s common knowledge that the gravitational pull of the moon regulates tides in the ocean. Since our brain is mostly water, it would make sense that the moon’s gravity affects it – I will investigate this further in a future case study.
- Mosquitoes. So many mosquitoes. In one session, a mosquito bit me on my penis. Although I finished the meditation, I am sure this caused anomalous brain wave patterns (i.e., it really hurt). In general, bugs were quite vicious and their effect must be considered. Check out this video on a mosquito that climbed the kundalini corporate ladder:
Further Research
Thus far, I focussed only on the data subset relevant to the hypothesis, namely, gamma wave PSD and brain hemispheric correlation. However, the full dataset is quite rich. Much further research can be done.
The Muse2 Brain Sensing Headband together with MindMonitor can capture the full spectrum of brain wave activity ranging from low-frequency delta waves to high-frequency gamma waves. I am very confident that there were statistically notable changes in the other parts of the brain wave spectrum. For example, I observed that that PSD of the delta waves decreased significantly over the course of the case study. I am not sure what this means, but it’s certainly interesting.
Furthermore, lest we forget – this case study covered 3 different measurements: Wim Hof, Anuloma Viloma and Ajapa Japa meditation. Unfortunately, due to a buggy Wim Hof app software update, I was not able to include the breath retention results. However, I am in contact with the Wim Hof technical team to retrieve that data and will update this case study when it becomes available.
Anuloma Viloma and Ajapa Japa are traditional tantric meditations focussed on the ajna (3rd eye) and anahata (heart) chakras, respectively. It would be interesting to investigate if the brain behaved differently during each type of meditation.
Raw Data
Subscribe
Keep up-to-date with all the cool research, meditation, tantra, life-hacks, psychedelics, kundalini and beautiful mathematics by subscribing to my channel here