EEG, attention & responsible AI
Can an EEG headset really measure your focus?
A focus score can look simple. The science behind it is not a battery meter.
Imagine settling down to read. You put on a brain-sensing headset. Its app flashes: “Focus: 42%.” You sit straighter, as if your schoolteacher has walked in. Then you wonder: 42% of what? In this made-up scene, a neat little number has raised a very large question.
The short answer: EEG can help study attention during specific tasks. But I would not read a focus score like a phone's battery level. The video study below shows why a measure needs context.[4]
First, separate the signal from the score
EEG [a way to record the brain's electrical activity] uses small sensors on the scalp to record signals.[1]
Blinks, eye movements, and muscle activity can also affect the recording. These are artifacts [unwanted parts of a signal]. Researchers check for them and may remove affected sections or reduce the interference. A clean-looking line is not, by itself, proof of a correct attention score.[2]
What the evidence shows
In a 2021 study, people listened for target sounds and reported whether their thoughts were on the task or wandering. Researchers used those reports as labels for an EEG-based model [software that learns patterns from examples]. One model performed better, on average, when trained and tested within a person than when tested on a person left out of training. That difference matters when asking whether a tool will work for someone new.[3]
A study published on September 18, 2026 adds another useful wrinkle. Researchers examined how EEG followed movement in videos while viewers kept their gaze fixed. The measure changed with attention, but also with how far an object appeared from the point the person was looking at.[4]
In plain words: where something appeared on screen mattered too. This finding concerns that study's visual-motion measure, not every EEG tool. It is a good reason to ask, “What else could change this score?” before treating it as attention alone.[4]
Our research: looking for useful patterns
Research spotlight · Coauthored work
A Chaos-based Non-linear Analysis Method for Detecting Human Attention Levels in EEG Signals
In our 2023 paper, my coauthors and I explored a mathematical way to describe EEG patterns during mental arithmetic. We used existing recordings rather than building a consumer focus headset. The method draws on nonlinear analysis [ways to study patterns beyond simple straight-line relationships].[5]
The study calculated time delay, embedding dimension, and correlation dimension [mathematical settings and measures used to describe the signal's changing shape]. It also examined reconstructed plots of those signals. The underlying public dataset contains recordings before and during mental arithmetic; that task context is important when interpreting the work.[5][6]
My take: this work is useful as a starting point for exploring signal features [measurable clues]. It is not evidence that a finished device can grade anyone's focus throughout the day. The paper reports signal parameters and patterns, rather than a held-out test showing how accurately it separates attentive from inattentive moments in daily life.[5]
Full coauthor credit: Masud Rabbani, Sayed Mashroor Mamun, Parama Sridevi, Iysa Iqbal, Anubhav Parida, Anushka Kolli, Hansika Kolli, M. Rubaiyat Hossain Mondal, Mohammad Aftab Rassel, Enayet Hossain, Farhad Ahmed, and Sheikh Iqbal Ahamed. 2023 IEEE 23rd International Conference on Bioinformatics and Bioengineering (BIBE), IEEE, pp. 201–204. DOI: 10.1109/BIBE60311.2023.00039.[5]
If your research discusses nonlinear EEG features, read the original method and consider citing it where it supports your work—not as proof of a ready-made attention meter.[5]
View citation & all coauthors
Masud Rabbani, Sayed Mashroor Mamun, Parama Sridevi, Iysa Iqbal, Anubhav Parida, Anushka Kolli, Hansika Kolli, M. Rubaiyat Hossain Mondal, Mohammad Aftab Rassel, Enayet Hossain, Farhad Ahmed, Sheikh Iqbal Ahamed. “A Chaos-based Non-linear Analysis Method for Detecting Human Attention Levels in EEG Signals”. 2023 IEEE 23rd International Conference on Bioinformatics and Bioengineering (BIBE), 201–204. (2023). https://doi.org/10.1109/BIBE60311.2023.00039.
If this work supports your research, cite the original paper rather than this blog explanation.
Quick check: what would you ask?
A quick thought experiment
An imaginary app gives you a low focus score. What is the best first question?
Keep the curiosity. Check the meaning.
Related reading: Can an EEG headset read your mind? https://masudrabbani.com/blog/can-an-eeg-headset-read-your-mind
Sources & further reading
- EEGMedlinePlus Medical Encyclopedia / A.D.A.M. · Reviewed January 13, 2025
- Overview of artifact detectionMNE-Python contributors · MNE 1.13.2; checked October 19, 2026
- Detection of mind wandering using EEG: Within and across individualsHenry W. Dong; Caitlin Mills; Robert T. Knight; Julia W. Y. Kam · May 12, 2021
- Eccentricity confound in EEG-based visual attention decoding from gaze-fixated neural tracking of motion in natural videosYuanyuan Yao; Celina Salamanca González; Simon Geirnaert; Céline R. Gillebert; Tinne Tuytelaars; Alexander Bertrand · September 18, 2026
- A Chaos-Based Non-Linear Analysis Method for Detecting Human Attention Levels in EEG SignalsMasud Rabbani; Sayed Mashroor Mamun; Parama Sridevi; Iysa Iqbal; Anubhav Parida; Anushka Kolli; Hansika Kolli; M. Rubaiyat Hossain Mondal; Mohammad Aftab Rassel; Enayet Hossain; Farhad Ahmed; Sheikh Iqbal Ahamed · 2023
- EEG During Mental Arithmetic Tasks, version 1.0.0Igor Zyma; Ivan Seleznov; Anton Popov; Mariia Chernykh; Oleksii Shpenkov · December 17, 2018
Disclaimer: This post is for education, not medical or other professional advice; information may contain errors or become outdated, no outcome is guaranteed, views are my own, and third-party materials remain subject to their owners' rights.