call for collaboration: neurofeedback project to train 'Jhana' concentration-meditation states
Hi there!
My name is Luc, I'm currently studying Human-Computer Interaction (5th semester) in Hamburg. For the past 6 months, I've been spending the majority of my time working on this project.
About the project:
There are deep states of meditation called the Jhanas, that are achieved through means of sustained concentration on an object. Measurements of these states have already been shown.
For starters, these states are described as a centering of awareness, where there is a merging between the subject and the object of concentration. They are often described as blissful, and cognitive abilites like focus remain heightened for hours after the meditation session.
The catch is, it takes a long time and a lot of energy to even come close to these states. This is something that people (e.g. monks) devote their lives to. In my own practice, I have found that the biggest roadblock to moving forward in concentration meditation is the fact that I don't know whether or not I am doing it right. I could have spent months meditating every day, and not know whether I have gotten better or not.
That's where the project starts. I am using an OpenBCI in order to measure my brainwaves, and writing a game that will do complex signal processing and machine learning to steer the subject in the right direction while meditating.
The idea is the creation of a positive feedback loop, as well as session tracking, the same way a powerlifter would track all of his variables in order to lift the highest amount of weight.
My biggest hope would be to have a system that can accurately measure the depth and duration of the concentration, in order to accelerate the process of attaining access concentration and/or jhanas. EEGs are limited in spatial resolution but have excellent temporal resolution, and we don't know the upper limits of machine learning.
If something like this works, it could increase the focus of serious practitioners in a way that no other practice would allow.
What I've already made:
I have made a first Unity project that tracks simple parameters like band power and site coherence, and recorded lots of sessions including control sessions. I managed to determine some kind of difference between a control session and a meditation session, but this was not enough for Neurofeedback purposes. I've also got data visualization of all the parameters.
First prototypes of the "gamified" trainer exist, but are very rudimentary.
What I'm planning to do:
I want to delve into deeper methods of analysis, these include, but are not limited to:
- Microstate analysis
- Wavelet Packet Decomposition
- Fractal Dimension of the Signal
- Other Machine Learning methods
I also want to work on the following things:
- Gamification (Achievements, levels, scores, leaderboards)
- High quality Data visualization (the brain is an aesthetic thing to track)
Who I need
For now, I mostly need people that are motivated to join in on the analysis. I don't care too much about your current skill, much rather about how motivated and enthusiastic you are. If you can invest 10hrs+ a week, we could already make something happen.
This is a project that fascinates me to no end, but I couldn't find people that would be interested in joining in. Now is the perfect time, as I am restarting the project in a clean an organized manner.
Of course, this is a very limited description of the project, but if that has sparked your interest, we could talk more about it over Skype!
Cheers,
Luc
Comments
Zoran Josipovic, PhD, is a Research Associate and Adjunct faculty at the Psychology Department, New York University. He is the director of Contemplative Neuroscience Lab at NYU, the founding director of the Nonduality Institute, and the founding member of MARGAM -- metro-area research group on awareness and meditation. His research focuses on the nature of consciousness and its relation to the brain, and on the functioning of anti-correlated networks in the brain. Zoran is a long-term meditation practitioner in the Tibetan, Zen and Advaita Vedanta traditions.
My name is Gabriel Pappalardo from the USA. I hold a PhD in Industrial/Organizational Psychology and recently spent four months at a yoga ashram deepening my meditative practice. I am also working on a meditation themed game and would love to use a neurofeedback controller suchthat players would occasionally meditate to further progress in the game. I love the nature of your project and would love to keep a dialogue going!
I want to alert you to two projects you may or may not already be aware of that I have investigated for my purposes: MUSE and Neuroplus. MUSE is a well-established product with a simple gamification, and I have been able to feel subjective differences in my meditation state translate to what the feedback from the device, but I worry a bit about the reliability, durability, and floor/ceiling effects. Neuroplus focuses on attention rather than meditation, and while it more ambitiously attempts to make real games that use the device, I haven't gained much confidence the device measures anything of substance, but I'm hoping it may improve with time. I can speak much more to either of these.
My availability is fairly limited at this time but may open more over the next 12 months, so I'd love to keep in touch and offer what support I can. I can especially be helpful on the gamification side of the house. It is my passion.
Hi folks -
I find myself similarly requiring a modern, fast and developer-friendly platform for neuromeditation and neurofeedback R&D. Did this project ever go anywhere?
I've started building my own thing and I'm looking for help with it. I'm an experienced engineer and unity dev but with limited time to develop all the components needed myself.
Daniel
Daniel, hi.
Take a close look at the various VPL (Visual Programming Language) platforms that support OpenBCI devices. These VPL systems allow data flow / feedback programs to be designed 'visually' with flow diagrams connecting 'function blocks' such as filters, feedback, thresholds, expression evaluation, device configuration, etc.
The list is actually a long one, and more details on each of these can be found in other Forum posts:
All of these would support the 2 or 4 channel NeuroMeditation protocols of Jeff Tarrant. (See Youtube's above.) Most neurofeedback uses relatively simple audio or visual feedback paradigms. Elaborate 3D feedback can work, but also consider that it may add layers of 'distraction' for neuromeditation applications. (If one goal is quieting activity.)
Regards, William
Hi William -
Thanks for the suggestions. I do have a BioEra license and use it for basic protocols, and have looked at or used most of those packages and several others over the years. That experience has been ample motivation for developing my own stack which is explicitly NOT a VPL.
VPLs are great for non-programmers, but are highly suboptimal for the kind of r&d workflow I'm interested in and a poor foundational choice for larger projects. My later roadmap includes experiments with larger number of channels (e.g. more sophisticated z-scoring and LORETA techniques), custom filter implementations, and intricate feedback logic for special situations that would be a nightmare to write and maintain in a visual paradigm. I'd also like to eventually be able to deploy and operate in environments where a more constrained framework will fall down due to various dumb things like licensing restrictions, platform restrictions, poor performance, unfriendly source control interactions, dependency on a 3rd party for code bugfixes and features, etc.
Code is king, C# is fast, and unity runs everywhere. Plus, its very pretty, and I like pretty
Daniel