GUI Focus Widget, algorithm question
NeuroReaper
Pakistan
Hello,
I am relatively new to EEG and related stuff but from my understanding a higher Beta corresponds to a higher attention (focus which is the same thing, right?), while the widget relies on a higher alpha voltage (power) for focused state. I might be mixing somethings up here but i would like to understand the working behind this widget.
Regards
Comments
https://docs.openbci.com/docs/06Software/01-OpenBCISoftware/GUIWidgets#focus-widget
https://github.com/OpenBCI/OpenBCI_GUI/blob/master/OpenBCI_GUI/W_Focus.pde
wjcroft thank you for your reply, a comment in the code mentions "focus detection algorithm based on Jordan's clean mind: focus == high alpha average && low beta average", could you provide some insight into this? Additionally my original question remains, is the 'focused mental state' related to a high meditative state or some event related attentive state?
I sent an email to Wangshu, the widget author. Will update the thread when I hear back from her. For general papers on mental states, suggest some searches, such as,
https://www.google.com/search?q=eeg+correlates+of+mental+focus
https://www.google.com/search?q=eeg+correlates+of+focus
As you can see, you can substitute other states for the final term, such as flow, attention, meditation, etc. All of these and many others have been researched. My guess on why the alpha and beta amplitudes are correlated (for Jordan's approach), is that you want not only mental / cognitive (beta) engagement -- but also a certain degree of relaxation and 'flow' which is more associated with alpha. Most likely this Jordan metric is just one of the focus measurement techniques. Certainly others are out there.
Regards, William
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0001459
Deconstructing Insight: EEG Correlates of Insightful Problem Solving
OK, it looks like her reference to 'Jordan' is from this other project,
https://www.sunwangshu.com/portfolio/dark-maze/
https://github.com/sunwangshu/DarkMaze
I would suggest your continued searching on the more specific type of EEG correlates you are interested in.
I am extremely grateful for your help, i will look into these leads and see where it takes me. "not only mental / cognitive (beta) engagement -- but also a certain degree of relaxation and 'flow'" i suppose i didn't take this into account. I was more focused on Beta as a sole marker for attentiveness to a task at hand.
Here is the email reply from Wangshu below, I don't believe she has a forum login. If you have further questions you can email her at the address shown.
---------- Forwarded message ---------
From: Wangshu Sun ws1108@nyu.edu
Date: Fri, Jan 3, 2020 at 12:45 AM
Subject: RE: forum question on Focus Widget
To: William Croft
Cc: Richard Waltman
Hello William,
Happy New Year! I’m very happy to hear from you.
Rethinking about it, “research” was not an accurate word to use there, as it was not referring to specific papers or studies, but rather empirical.
The basic idea behind the algorithm was a combination of “high in alpha” and “low on the other noises” (beta band), forming a nice spike in alpha band for the simple algorithm to detect. When users are not focusing, energy levels are generally higher and scattered across all bands; when they are blinking their eyes, similar high energy defects (ocular artifacts) would be introduced in beta bands too. So “low on beta” could roughly rule out those cases.
They are only my observations, I haven’t done further study to prove the exact reasons behind them, but they worked to a basic extent.
For the exact values “0.7-2.0uv” or “0.0-0.7uv”, they were also actually referring to this “high alpha low beta” pattern. That’s why later versions of the focus widget would have tweak-able thresholds compared to set values.
The word “focus” was a generic term. While the original data was derived from my classmate Jordan’s brainwave when he’s doing “meditation”, it also mostly worked when I told users to “focus”. I would suggest “focus of visual attention” or “meditation” as more accurate to describe the desired “focused” state.
As per channel selection, channel 1 itself would be actually sufficient. It had some historical reasons (researchers using Neurosky or Emotiv as their BCI tools, some models only had frontal lobes), glimpses in some paper[ref 2] that suggested AF3 and F7 (haven’t vetted in the logistics behind them), etc.
I haven’t tried the other channels either, due to the time limit. If my assumptions were correct, it may work too (or even better) in parietal and occipital areas after tweaking the thresholds of the focus widget, as long as it catches the “high alpha peak” pattern and rules out the noises. I haven’t tried though.
References / Papers I found helpful:
General guidebooks:
Best, Wangshu
PS Salutations! I am always feeling honored to talk with professional researchers. And to be honest, this project was done in a more hobbyist (and even artistic way), but I tried my best to catch up some standards.
===
From: William Croft
Sent: Wednesday, January 1, 2020 9:46 PM
To: Wangshu Sun; [email protected] Cc: Richard Waltman
Subject: forum question on Focus Widget
Wangshu, hi.
The Focus doc at,
https://docs.openbci.com/docs/06Software/01-OpenBCISoftware/GUIWidgets#focus-widget
States:
"The widget recognizes a focused mental state by looking at alpha and beta wave levels on channel 1 & 2. It is based on research supporting focused states aligning with alpha levels between 0.7-2.0 μV, and the beta levels between 0.0- 0.7 μV. If your data is outside of this ratio, the algorithm states that you are not focused."
We had a recent user ask about the algorithm background.
https://openbci.com/forum/index.php?p=/discussion/2418/gui-focus-widget-algorithm-question#latest
Do you have any links or paper references to the research mentioned in the above quote? Also, what head 10-20 sites should the two channels be attached to? We should likely include this in the doc. Because as you know, there is markedly more alpha in the parietal and occipital lobes. Whereas frontal sites are more involved with executive functions.
Feel free to add any info to the forum post, or if not I can do a followup post. Regards,
William
PS hey congrats on your position in San Francisco. I lived for 36 years on the SF peninsula, near Stanford. And worked for both Stanford Medical Center and Stanford Research Institute. Moved in 2017 spring very far north to Mount Shasta.