EEG Research Analysis
Contributed to published research on neural signatures in chess puzzle solving and cognitive tasks using consumer-grade EEG devices.
Overview
I had the incredible opportunity to contribute to a published research study that explored how consumer-grade EEG devices could detect different cognitive states. The study was titled "Neural Signatures Within and Between Chess Puzzle Solving and Standard Cognitive Tasks for Brain-Computer Interfaces: A Low-Cost Electroencephalography Study" - quite a mouthful! We used the MUSE 2 headband (a consumer EEG device) to see if we could distinguish between different types of thinking tasks, including chess puzzles, memory games, and visual rotation tasks.
What I Did
- Helped Design the Study: I worked with the team to figure out how to structure the experiments and what tasks would best test different types of thinking
- Collected Data: I helped run participants through the experiments, making sure they wore the MUSE 2 headband correctly and completed all the cognitive tasks
- Did the Math: I contributed to the statistical analysis, using linear mixed-effects models to see if we could reliably distinguish between different cognitive states
- Made Visualizations: I created charts and graphs to help everyone understand what the brain activity data was telling us
- Built ML Models: I helped implement machine learning models that could predict what type of cognitive task someone was doing based on their brain activity
How We Did It
Mixed Up the Tasks
We combined classic cognitive tests (like memory games and color-word matching) with chess puzzles to see how different types of thinking showed up in brain activity
Used a Cool Headband
The MUSE 2 is basically a fancy headband that reads brain waves - much cheaper than lab EEG equipment but surprisingly effective!
Crunched the Numbers
We used linear mixed-effects models to tease out patterns in the brain activity data and see if we could reliably tell different cognitive states apart
Taught Computers to Read Minds
We built machine learning models that could predict what someone was thinking about based on their brain activity patterns
The Technical Stuff
- Cleaned Up the Brain Data: EEG signals are messy! I helped filter out noise and extract the meaningful brain activity patterns from the MUSE 2 recordings
- Found Patterns in the Chaos: Using linear mixed-effects models, we figured out how to reliably distinguish between different types of thinking based on brain activity
- Made Pretty Pictures: I created visualizations that made the complex brain activity data actually understandable and presentable
- Built Brain-Reading AI: We trained machine learning models that could predict whether someone was doing memory tasks, chess puzzles, or other cognitive activities
- Proved It Actually Works: We rigorously tested everything to make sure our findings were solid and reproducible
What We Discovered
The Headband Actually Works!
We were surprised to find that the MUSE 2 (a relatively cheap consumer device) could reliably tell the difference between different types of thinking and cognitive workload levels.
Why This Matters: This means we could potentially build brain-computer interfaces that adapt to your thinking in real-time, without needing expensive lab equipment!
Our AI Could Read Minds (Sort Of)
We successfully built machine learning models that could predict what type of cognitive task someone was doing based on their brain activity patterns.
Why This Matters: This opens up possibilities for adaptive systems that respond to your cognitive state - imagine a computer that knows when you're focused vs. distracted!
Why This Research Matters
- It's one of the first studies to show that affordable EEG devices can actually be useful for serious research
- We figured out how to tell the difference between different types of thinking using brain activity
- We discovered unique brain activity patterns when people solve chess puzzles
- We built AI that can predict what someone is thinking about based on their brain waves
- My visualizations helped make complex brain data understandable to everyone
- This research could lead to better brain-computer interfaces that actually work in real life
About the Publication
Title: "Neural Signatures Within and Between Chess Puzzle Solving and Standard Cognitive Tasks for Brain-Computer Interfaces: A Low-Cost Electroencephalography Study"
Authors: Russell, Matthew; Youkeles, Samuel; Xia, William; Zheng, Kenny; Shah, Aman; Jacob, Robert J. K.
My Role: While I'm not listed as an author (which is common in academic research), I was actively involved in designing experiments, collecting data, running statistical analyses, and creating visualizations that helped make the research findings clear and compelling.
What Could Come Next
- Brain-computer interfaces that actually work in your living room, not just in expensive labs
- Combining EEG with other brain imaging techniques to get an even clearer picture of what's happening in your head
- Computers that adapt to your mental state in real-time - like knowing when you're focused vs. daydreaming
- Using this technology to help doctors monitor cognitive function in patients
- Building even smarter AI that can understand more complex patterns in brain activity