After spending 24+ hours in a hot closet with strangers, I’ve learned a couple things.
Walk with me (like lovers do)
So far, I have recorded from a lot of people with a large range of experience levels in search for a biomarker for meditation- e.g. something more reliable than a person just “saying so” when they’re meditating.
When the meditators close their eyes, it’s almost like I’m leaving them in a closed room: there’s nothing telling me what they’re thinking and it’s hard to tell if they’re doing anything different from just sitting there with their eyes closed.
The known factors at that point in time are:
- They volunteered to participate in this meditation study.
- They live in Michigan.
- They have just heard me say something along the lines of, “For this meditation task you will adopt your normal meditation technique from when I say begin to when I say stop, close your eyes, begin.”
With EEG, we can’t see what you’re thinking, but hopefully we can differentiate between the kind of thinking you’re doing. I am going to see if alpha waves can be used as a lie detector for “real meditation”.
Quick and easy 40 minute Data Recipe
- Prepare the pan: low cost equipment for scalp recordings (a shameless BYB plug)
- Modify headband to record from locations that are thought to process different areas of interest.
- Capture ingredients
- Cast a wide net and record from as many people as possible!
- (scheduling is hard and people are flaky, tread strongly)
- Mix thoroughly and consistently
- Flip a coin to determine order of tasks – independently random so that any difference between meditation and rest is not just due to ordering
- A protocol script, to ensure that you are saying the same thing every time to induce the different mind states (*relies on participant obedience and self report)
- Take notes, never trust your memory
A page from my lab notebook where I write and draw observations during recordings.
- Put ‘em in the oven
- Record in a consistent environment
- A lil dark, a lil loud, a lot hot, but consistently so
- Acquire fresh, local, organic data
- eyes closed during all 2-4 tasks
- meditation = “adopt your normal meditation technique or focus on your breath”
- rest – “just sit with your eyes closed”
- rumination = “decide whether [the negative/neutral word] describes you and follow that train of thought”
- survey = “how old are you, describe your experience during this meditation, how many times a week do you meditate…”
Talk to me: Labeling Data Descriptively Is Mindfulness Based Stress Reduction
My big pile of data comes out to approximately 1000 minutes of recordings, with 2-4 tasks each, recording from 3-4 locations from meditators with 0-1000 hours of experience. With each added variable, the size of the dataset increases exponentially and gets hard to keep track of!
Let’s get wavy
Remember, frequency bands can be measured from any location and we are measuring from the scalp i.e. there is no physical contact with neurons, just big groups of them yelling at us!
- 0.5 – 3 Hz, Delta – large slow waves, less interesting things your brain is doing (like deep sleep)
- 3-8 Hz, Theta – drowsy, literally and theoretically in between sleep and awake
- 8-12 Hz, Alpha – idling mode of synchrony, ex: your visual cortex is not receiving complex information from your eyes
- 12-24 Hz, Beta – “many different conversations occurring between neurons” canceling each other out
Choose your fighter: L-gop vs Maria
Let’s look at the effect of condition (meditation) on the spectral power across the EEG bands. Spectral power means that for each range of frequencies, we’re looking at the amount of time that was spent at that frequency band. For example, according to my hypothesis, meditators will have more alpha frequency brain waves during meditation compared to rest. For this figure, I wasn’t interested in the differences between locations, so the EEG power from all brain regions was averaged.
When you un-omit their error bars, the data starts to look significantly less significant…
As you can see from our corresponding battle stats, Lagopoulos has me beat. There is a lot more variance in my data, in terms of error bars (how confident I am about my mean), equipment, and participant meditation technique and experience level. But even with this amount of control his error bars are still pretty big, and obviously overlapping (which is probably why he omitted the bottom half of his error bars) which is to say that his values (although more significant that mine) are still not quite significant.
It’s looking bleak for the alpha wave – if you try to use it as a lie detector for meditation it will probably be lying to you, but stay tuned for next week where I will attempt to “p-hack” the data and see if certain qualities about my meditators are responsible for the data being all over the place!
PS: Blink and you’ll miss it (again)
*In my figure, it looks the most difference between meditation and rest occurs in the delta band, but I am omitting all frontal delta data due to eye blink artifacts. In the Lagopoulos paper, “ten (10) minutes of eyes – closed EEG was recorded during each task”, meaning they recorded the eye artifacts and ended up cutting out approximately half of the recording when there was any eye activity.
I used a similar method to confirm my fake delta suspicions.
Hello dear reader! It’s Maria again and my meditation project is picking up speed. I’ve now recorded ~brain waves~ from 20 different people! If I’ve learned anything so far, it’s that in science it’s data or bust. I’ve been soliciting people over email and clipboarding more than a Jehovah’s Witness. If this science thing doesn’t work out, at least I’ll have canvassing skills… For the past couple weeks I’ve gone to so many open meditation sessions that I’ve started to make friends! There are weekly “sits” with the Ann Arbor mindfulness community, University of Michigan, and a local Zen Buddhist temple.
The Zen Buddhist Temple
From these communities and around town, I’ve been gathering up people with a large range of meditation experience, from self proclaimed yogis to novices who have never meditated before.
There’s a mind reader in your closet
And so! I decorate the volunteers with sweatbands and wires and let bake for approximately 45 minutes in Greg’s closet. With these headband scalp recordings, the most reliable signal that we can pick up is the alpha wave, a frequency of oscillation that was discovered by Hans Berger in the 1920’s. As a student around my age, he fell off of a horse and in this moment of mortal danger believed that he had spontaneously transmitted his thoughts to his sister. He later became obsessed with proving the physiological basis for “psychic energy” and developed a tool to record the electrical current of the brain from the surface of the scalp using rubber bands and silver foil (not so different from the sweatbands and buttons I’m using this summer).
Electrode setup used during meditation task
I’m recording from the frontal, temporal, occipital, and left parietal areas – locations related to vision, complex object recognition, planning, and tactile attention that I have taken from a corresponding EEG and FMRI studies. The goal is to compare the alpha waves from these areas during different tasks. When the eyes are closed and the visual cortex is not receiving information, and the neurons oscillate slowly at the same rate to produce a summed wave called the alpha frequency. In contrast, when they eyes are open and the visual cortex is active and receiving various types of information, the neurons will fire out of phase and no distinct waveform will be observed (Backyard Brains has a stellar visual explanation of this here). Because alpha waves are thought to correspond to the idling state of neurons, I am curious to see if the waves be affected by more than just visual stimuli. Perhaps meditation is more or less “idle” than other states of rest.
A problem that I’ve run into is what to compare meditation to. What is a consistent control that I can measure meditation against? What the heck is your resting state, the state your brain defaults to when you are not consciously concentrating or directing your thoughts like with meditation or math? So far, I’m testing out what the prompt of “eyes closed rest” will bring about in both the EEG data and what they report to have been thinking about during the open-ended task. So far, over half of all participants have explicitly mentioned “what I need to do today” as the content of their thoughts during the rest task.
I am also testing a more specific contrasting task that is a theoretical opposite to meditation: rumination – a psychological state that is associated with depression. In terms of thinking about yourself, meditation is a kind of experiential self focus in the present moment, whereas rumination is often references past experience in a narrative way. I also leave the electrodes on during the survey questions and ask them questions verbally as another comparative task.
Blinking is not sleeping
I wrote a little bit of code to take calculate the amount of time during each task that the brain waves were in a certain frequency i.e. the amount of time alpha waves are present during meditation.
Positive self reported meditation, mantra technique, 3120 hours of meditation experience
Looking at the data from individual recording sessions so far, you can see that in the frontal region, the delta frequency band has more power on average, but this is actually due to eye blinks and not real delta waves!
A recording of real delta waves and eye blinks that occur at a similar frequency
The FFT algorithm generates a spectrogram (blue boxes) by converting the electrical signal into frequency components of Hz or cycles per second. Spontaneous blinks, which happen at around 0.3 Hz or 15-20 times per minute are often mistaken for delta frequency waves that occur during deep sleep (0.5 to 3 Hz). Although the difference in signal is from eye muscles and not the brain, it suggests that for some reason people blink differently during different tasks. Eye blinks may not be the biomarker that I’m looking for, but it does indicate that there is some difference between the mind states.
PS: here’s a sneak peek of a little side project of portable EEG recordings.
Since I have been struggling with what to use as a control, I figured that I should see if I can measure all mind states at all times!
Ramiz Kara is a freshmen at Emory University as a Pre-Med track, and he had some questions. He wanted to know the effects of meditation on brain waves and abstract problem solving, so he decided to make his own experiment to measure this, using our EEG device, the Heart and Brain SpikerShield. Although he didn’t find conclusive results regarding the brain activity, starting the research and taking different tools to experiment towards solving this questions is the most important step and the essence of science.
Below are pictures and the presentation of the whole experiment process and methodology Ramiz used. Congratulations Ramiz!
If you are interested in seeing the brain activity during meditation, we recommend measuring Alpha waves: you can learn how to do this in our following experiment. Studies on mindfulness meditation, assessed in a review by Cahn and Polich in 2006, have linked lower frequency alpha waves, as well as theta waves, to meditation, but this review is not open access, so not everybody can read it to see the results or methodology to replicate the experiment. If you decide to do the experiment and want to make it open to the community, please write to us so we can post it in our blog.