Research Question:

I have ADHD and I find it very difficult to focus on a singular task as my mind keeps wandering and often processing several thoughts at the same time. I have always wanted to find a way to monitor my attention and go behind the scenes to see what was really going on inside my brain.

And so, I asked myself the question “How often do I lose attention during different times of the day?” and set out to find answers utilizing an Electroencephalogram (EEG) device.

An electroencephalogram (EEG) is a test that measures electrical activity in the brain using small, metal discs (electrodes) attached to the scalp. Brain cells communicate via electrical impulses and this activity shows up as wavy lines on an EEG recording.

For my project, I utilized the Muse 2 – an EEG tracker that is marketed as a consumer grade meditation assisting device that can help achieve better focus and calmness.

Besides the main meditation focused app, the device allows developers to get access to timestamped EEG data which can also be exported to analyze or build applications around.

One issue I faced early on is that it is not very comfortable to wear the device for long periods of time. As a result, I could not track the data I initially planned to over the course of a day but chose to record data during three different activities to see if any patterns emerge.

As a result, I changed my research question to “How often do I lose attention while studying, watching a movie / multitasking and meditating?” – I chose the 3 because they represent the three broader categories of activities I do during a day.

Thanks to Professor McSweeney’s input, I also started tracking when I consciously start realizing that I am losing attention and sought out to ask “whether active realization helps bring attention back“.

Audience:

I am the primary audience for this project as it arises from a personal question. However, it can also be interesting for those who also suffer from ADHD or researchers in the field.

Brief Introduction to Brainwaves:

The Muse 2 device collects a range of data but the ones we are interested in is the timestamp and the five brainwaves (measured in bels).

There are five waves that the human brain emits, and each signify specific states of the brain and are more dominant during the following activities:

Wave Type	Brain State
Gamma (γ)	Deep concentration. Reflects conscious awareness, deep compassion
Beta (β)	Reflects the active / engaged brain like reading books, problem solving etc.
Alpha (α)	Passive attention. Awake but relaxed. Mindfulness meditation.
Theta (θ)	Deeply relaxed, inward focused, dreamy, intuition, deep emotions, creativity
Delta (δ)	Deep sleep (dreamless state) or deep meditation

For my reading, although I have gathered data for all the 5 waves, the Delta one must be dropped because Delta waves show up as dominant on the readings which it should not be. According to some forums, this happens due to muscle movement and other issues that are quite common with the Muse 2 device and affects the Delta waves specifically.

Visualizations Made:

I made 5 visualizations in total:

• 1 bar chart of brainwave averages (alpha, beta, theta & gamma) while performing different activities (studying, meditating & watching a movie / multitasking)
• 1 bar chart of Theta to Beta ratio averages while performing the different activities
• 3 line charts of Theta to Beta ratio over time while performing the different activities (different graphs for different activities)

Data, Findings & Explanation:

The first visualization that naturally came to my mind was to see the averages of each brainwave while performing different activities and see if they are consistent with conventional logic.

For gamma and beta waves, this did follow conventional logic with highest average values for studying followed by meditation and movie / multitasking. These waves are most active during conscious activities and problem solving.

Theta and alpha waves were the highest while studying which it traditionally should not have been. However, some studies show that both high theta and alpha values while performing concentrative tasks can be indicative of ADHD.

Now my goal was to see patterns in attention and absolute values of brainwaves are not the best indicator of that.

After studying multiple papers and sources, I learnt that theta to beta ratio is one of the most commonly sought-after measurement to track attention and diagnose ADHD. In fact higher values of this ratio is related to loss of attention and increased potential of having ADHD.

So, the next visualization I made was a bar chart to see the average theta to beta ratio during different activities. These findings do align with our conventional understanding with highest value while watching movie / multitasking (highest distraction) followed by studying and meditating (lowest distraction).

At this point I realized that I was not really answering the actual question that I initially wanted to find answers to which is tracking attention loss over the course of performing an activity (which is ironic to say the least).

So, for the final analysis, I created the three line charts of theta to beta ratio over time for the three different activities.

Now, theta to beta ratios will not stay consistent (the brain is a complex organ with fluctuating thoughts and processes). However, consistently high values of the ratio indicate to periods of low concentration. But to visualize that better, we need to set a limit for the ratio above which we can say that attention is getting lost.

Unfortunately, this value can vary a lot with different researchers indicating different numbers and in fact it can very much differ from person to person which a small amount of data cannot deduce (a limitation of the experiment).

To make things easier, I went with one researcher’s value of “2” being the magic number. So, we assume that consistent theta to beta ratios (for at least a few minutes) above 2 will indicate periods of inattention.

The value does seem to be represented quite well by the charts applied to conventional wisdom. Throughout the watching movie / multitasking chart, the numbers consistently are higher than 2 which makes sense because I was not concentrating then at all.

While meditating (mindfulness meditation to be more precise), the values are consistently below 2 except sudden and short instances which is normal. This does show that while meditating, especially one that is designed to help reduce distractions, attention is rarely lost.

Finally, we come to the studying chart and while it is overall not as bad, there are 2 periods of time during which the values are consistently above 2.

What I found more interesting is that I recorded the time when I noticed I was losing concentration and once I did notice, I can see that my conscious recognition leads to me forcing myself back into paying attention.

Possible time of attention loss	Recoded time when I noticed I was losing attention
3:22-3:26	3:24
3:49-3:53	3:52

Limitations, Improvements & Potential

The core limitation of this project certainly is the amount of data collected. For a better analysis, a more constant data collection throughout the day as compared to portions would help make the deductions more accurate.

Also, it may vary from day to day so instead of just one day, data from several days would be more ideal.

Additionally, as I mentioned previously, the threshold beyond which attention loss is signified by the theta:beta ratio can be very different for each person, so more data would be ideal to find a better threshold in the first place.

Another issue is that experiment for just one person might not be the best source for analysis. Using the same device on several people both diagnosed and not diagnosed with ADHD might help better understand the broader context of findings.

For all of this, a better product that is more comfortable is required. Muse itself has a newer model, the Muse S which is more suitable for longer use.

One major potential I see based on this experiment is the finding that noticing that you are losing attention itself can help bring back concentration.

So, if we can gather enough data both from a wide range of experiments and then an individualized learning, we might potentially be able to automate an EEG device to notify a person that they are losing attention (maybe through vibrations or a sound) and that can help people with ADHD and others in general to improve their attention span when needed (for studying, working etc.).

Research is being conducted in this field with some experiments showing potential including one I read about that MIT is currently conducting. Utilizing that research to build an affordable consumer grade product can be beneficial to many.