Cheat Code to AI Brainwave Decoding with Muse (5 new tips)

Рет қаралды 1,970

Күн бұрын

Techforpsych Custom AI BrainWave Decoder: www.techforpsych.com/aibrainw...
►►►Muse Headband 15% Dr. Cody Discount (applied at checkout): choosemuse.com/codyrallmd
►►►RentMD for Muse Rentals: bit.ly/RentMDco
►►►Myndlift: Want to connect with a Neurofeedback Myndlift coach? $60 off with this link: signup.myndlift.com/subscription. Enter CODYRALL at checkout.
►►►Mendi Affiliate Link (free shipping, 20% off, applied at checkout with code “mendiwithdrcody”): www.mendi.io/?discount=mendiw...
►►►Pulsetto 10% off: Enter "Cody10" at discount checkout: pulsetto.tech/ref/Cody+Rall/
►►►Sens.ai: Enter “drcody” at checkout for a $100 discount sens.ai/drcodyrall
►►►Neurosity Affiliate link (free shipping auto-applied at checkout) neurosity.co/drcody
►►►Oura Affiliate Link: ouraring.com/codyrallmd
►►►FocusCalm $25 off: Enter "DrCody" at discount checkout: www.focuscalm.com
►►► GET YOUR FREE MEDITATION GUIDE HERE: bit.ly/2XIRDNa
►►► INSTAGRAM (Behind The Scenes with Cody Rall MD): / codyrall_techforpsych
Cody Rall, M.D., is a United States Navy trained Psychiatrist who specializes in neurotechnology wearables. He is a co-founder of Stanford Brainstorm, the world’s first academic laboratory dedicated to transforming brain health through entrepreneurship.
Dr. Rall also served as a board member of the psychiatry innovation lab, an annual national competition at the American Psychiatric Association that works as an incubator for groups developing technological solutions to problems in mental health care. He is the founder of Techforpsych, a media and relations company that covers advancements in technology related to neuroscience.
============================================================
**Dr. Cody is not actively taking neurofeedback clients at this time. To connect with an excellent Myndlift coach, head to this link for $60 off, just enter CODYRALL at checkout.: signup.myndlift.com/subscription.
============================================================
Episode Timestamp Notes:
The Best AI Engine: 0:00-3:00
How to load EEG data 3:00-5:02
The best time to use AI : 5:02-5:51
How to Refresh the AI: 5:51-9:27
The Best AI Prompts: 9:27-10:35
A Custom GPT for Techforpsych Audience: 10:35-11:08
Edited with Gling AI: bit.ly/46bGeYv

Пікірлер: 27

@CraigDeServidor Ай бұрын

Fantastic stuff, Doc! However, I respectfully submit that it's data source that is the starting point of the issue. EEG is a 4-dimensional phasor, inherently nonlinear dynamical and chaotic in nature. So, you're getting data from the device that is in the wrong coordinate plane - gaussian and linear in nature and using the wrong mathematical tools to generate. With EEG being a 4-dimensional phasor the correct tools to use are appropriately tuned mathematics, using Joint Time-Frequency Analysis (JTFA) and Nonlinear Dynamical Systems (NDS) Control Procedure Mathematics. JTFA allows for exactly visualizing the 4-Dimensional nature of EEG activity - both traditionally and in novel ways including both frequency AND time. JTFA also offers the ability to perform inline, adaptive JTFA denoising that minimizes the impact of signals that do not conform to the time-frequency signatures of human EEG. This means that simple sinusoids like line current, as well as “flat top” signals like EMG are effectively "denoised" out by NO’s filtering. This means that you do not have to sit still while Training or focus on any task. NDS Control Process Mathematics, allows for capturing the emergent shifting (flutter) of the CNS around its trajectory. Heady stuff, right? I can connect you with my boss, Valdeane W. Brown, Phd, Co-founder of the Zengar Institute, Inc. and the Co-developer of NeurOptimal(R), if you wish to discuss this with him as I know he would love to talk more about this with you. I'm just sorta his "Man Friday." LOL Cheers and be well!

@CodyRallMD Ай бұрын

Thanks for the in-depth look at EEG. How do you think a different way of analyzing EEG would change analysis outcomes?

@CraigDeServidor Ай бұрын

@@CodyRallMD Thanks, Doc! Val is really the expert on this but it is Sunday and he is semi-retired but extremely involved and leading the effort in ongoing development of NeurOptimal(R). I messaged him about this and he will guide the answer with his expertise. But, first, always respectful regarding any brain tech out there and their laudable efforts to help us improve and enhance our lived experience. We know neurofeedback can work and this is sort of a “Copernican Revolution” philosophy in regard to capturing and measuring EEG data. So, I would say that it absolutely can offer untainted (by excluding EMF as well as EMG artifact) from the data set and also capture exactly what is going on during training. The current data is incomplete and may be tainted with EMG and EMF data. Does that mean that the current analysis is wrong? I don’t really know but there is so much more data to consider. I doubt chatGPT 4 is ready for that.🙏

@CraigDeServidor Ай бұрын

@@CodyRallMD, Thanks, Doc! Val is really the expert on this but it is Sunday and he is semi-retired but extremely involved and leading the effort in ongoing development of NeurOptimal(R). I messaged him about this and he will guide the answer with his expertise. But, first, always respectful regarding any brain tech out there and their laudable efforts to help us improve and enhance our lived experience. We know neurofeedback can work and this is sort of a “Copernican Revolution” philosophy in regard to capturing and measuring EEG data. So, I would say that it absolutely can offer untainted (by excluding EMF as well as EMG artifact) from the data set and also capture exactly what is going on during training. The current data is incomplete and may be tainted with EMG and EMF data. Does that mean that the current analysis is wrong? I don’t really know but there is so much more data to consider. Just humbly offering food for thought, here.🙏

@philipm3173 Ай бұрын

May I ask you some questions over email?

@david-to5qu Ай бұрын

I'm excited for the neurovizr video(Instagram post).😊Was wondering when we get to see it???

@CodyRallMD Ай бұрын

The first Neurovizr video should be out by 22 June :) . I'm excited about that one because there were some really interesting EEG results that I could analyze with ChatGPT

@dollardun5868 Ай бұрын

Use Anthropic Claude 3. Less hallucinating and much more token friendly.

@CodyRallMD Ай бұрын

I'll try it out!

@fairuzastevenson9896 Ай бұрын

Thank you

@juergenlilien2067 Ай бұрын

@CodyRallMD James Clutterbuck (Mind Monitor) wrote in a forum thread about the Peak Alpha calculation: "To get the data you will need, set the "Recording Interval" in settings from it's default of 1 data point per second to "Constant"." Do you also use the RAW Mode in Mind Monitor with this changed configuration?

@slmille4 Ай бұрын

I really like the enthusiasm for consumer EEG hardware, but the quality of the Muse S data is questionable. Interaxon's scientific validation was done on the original 2016 Muse, and there isn't any evidence showing that the Muse S is accurately measuring brain activity instead of just muscle tension. Running Muse S band data through ChatGPT is likely to result in hallucinations. Just something to keep in mind for future discussions.

@CodyRallMD Ай бұрын

This is false. There has been a bunch of work done with Muse S. In particular a lot of sleep studies with comparison to gold standard Polysomnography. choosemuse.com/blogs/news/the-gold-standard-of-sleep-tracking-polysomnography

@slmille4 Ай бұрын

@@CodyRallMD It is not. EMG contamination of EEG results is a huge problem even with clinically applied wet electrodes, see “EMG contamination of EEG: spectral and topographical characteristics”. Also while the study you cited shows accuracy in sleep stage classification using deep learning, it appears to be a 'black-box' model that finds correlations without identifying the underlying mechanisms. Such results could also be achieved by measuring eye movements and muscle tension (EMG artifacts) rather than brain activity.

@CodyRallMD Ай бұрын

Wrong again Steve. I have seen Muse S data clearly showing Sleep Spindle architecture. That would not be the result of EMG artifacts.

@slmille4 Ай бұрын

@@CodyRallMD Doubtful, sleep spindles mainly occur in the central regions of the scalp, particularly in the vertex (Cz), which the Muse S doesn't even measure.

@slmille4 Ай бұрын

@@CodyRallMD Additionally, as mentioned in the paper I referenced, contractions of the frontalis muscle appear as spindles on an EEG.

@barkingspider2007 Ай бұрын

Q: What is the small white device that displays brain waves in your video? BTW, I used your link to purchase the Muse S 2 and the Premium Subscription. Thanks for the excellent content! 😎👍

@ben.morris Ай бұрын

Great video Cody. Love this new series. Have you explored Gemmi Advanced? "With 1.5 Pro, our next-generation model with a 1M token context window"

@CodyRallMD Ай бұрын

Whoa yeah, thanks Ben. I'll have to try that and see if it processes the data any better. Thanks!

@dollardun5868 Ай бұрын

Please know that a developer is not a person. It’s a team of AI engineers of different vocations. If you had a developer that person would not bird dog you to GPT 4. They would be creating an application. It’s not your fault, few people understand this industry and anyone can say they are a developer. I own a development firm and Muse would be the domain expert. Ask Muse if they are in development

@CodyRallMD Ай бұрын

Thanks for the guidance!

@marcorodriguezhenry9858 Ай бұрын

A video on the different VAGUS NERVE stimulators would be awesome!!! Thank you for your work!!

@Optioning Ай бұрын

This is so cool, have you looked into biocybernaut being able to increase IQ by an avg of 11 points though neurofeedback training? wonder if its possible to get some of that value with this muse chatGPT setup? please lmk what you think, would be so amazing if there is a path to get there as I know AI is getting smarter and would be great to try and keep up lol