Playing Tetris while wearing an EEG shows human like AI is impossible
mindmatters.ai
mindmatters.ai
Your spectrogram shows difference in muscle tension and activity in your scalp. Unfortunately they overlap with most of the EEG spectrum you are trying to measure and are stronger. When you concentrate you become more tense.
You need to do more signal processing and learn more to really measure changes in EEG.
I tested out the idea by taking an EEG while furrowing my brow, but it did not look the same. Here is the furrowed brow EEG:
You can see some blue bands at the bottom, but no where near the degree that showed up while playing Tetris either slow or fast.
So, I don't think it is merely the tensing of my facial muscles that created the Tetris EEG patterns.
Here is something you can try: Lay relaxed in a bed eyes closed. Program a timer to give pleasant audio signal at random intervals 10-60 seconds between. See if you can detect the signal from EEG.
In general you need to process the signal if you want something useful out of it. The real EEG is very weak coming under skull. Noise from the surface has 100x higher amplitude. Most of what you see visually from spectrogram is just signal noise.
Here is good intro: Classification of Artefacts in EEG Signal Recordings and Overview of Removing Techniques https://pdfs.semanticscholar.org/cba7/b93c6810a7bbc5917bbaef...
With my current setup I can easily detect alpha waves when my eyes are closed, so I am definitely getting something out of the brain.