4 ms·
I just worked on something similar at a hackathon this weekend with the Muse Headband, it was really fun. We used a support vector machine to maximize the "calm
by PascLeRasc 11y ago
I just worked on something similar at a hackathon this weekend with the Muse Headband, it was really fun. We used a support vector machine to maximize the "calmness" output of the band, which is some weighted sum of gamma and beta waves I believe.
- dmazin 11y agoIs the Muse easy to hack? I was thinking of buying one to play with.
- jdpigeon 11y agoYou can fiddle around with it connected to a PC pretty nicely in Python
- PascLeRasc 11y agoThe hardest part of the weekend was getting real-time data through OSC. It's been done in Python and node before, but none of us were super proficient in getting data from the local server the scripts pushed to. This is the matlab library we eventually settled on: https://github.com/MuSAELab/muse_osc https://github.com/MuSAELab/muse_osc I'd say it was pretty easy compared to Myo, but I wouldn't pay $300 for one. I'd use it again though if it was available for a weekend.
- Hydraulix989 11y agoWhat you've worked on sounds much more scientific than OP.
- OJFord 11y agoI'm not sure if that's fair. They (claim to) have [0] a bunch of PhDs, and have a page with the same kind of words on [1] as parent. [0] https://www.brain.fm/about [1] https://www.brain.fm/#science
- Hydraulix989 11y ago#credentials
- merkahbah 11y agoTheir man researcher has a PhD in Neuroscience and the lead guy has been researching and developing audio-neuro software for over 11 years. https://www.transparentcorp.com/company/staff_details.php https://www.transparentcorp.com/company/staff_details.php
- OJFord 11y agoWell, yes?! Your implied complaint was that their work is not very scientific. The team having PhDs in neuroscience is irrelevant/unsatisfactory/bad? What would you like to see? Would it have seemed more scientific to you if they'd done it at a hackathon?
- Hydraulix989 11y agoObviously irrelevant, which renders moot the other two questions that you asked me. My point is that the work itself is what makes something scientific, not the workers' CVs (note the careful wording here; calling them researchers is another appeal to credentialism). With that in mind, what would I like to see? You should be able to answer that one yourself now.
- wutf 11y ago"We used a support vector machine to maximize the "calmness" output of the band" fyi, this doesn't make sense.
- PascLeRasc 11y agoHow so? We give it different variables and let it change them incrementally. Calmness is a 0.0 - 1.0 floating point output from the band. This was custom-made by the way.
- wutf 11y agoDepends on what exactly you did, but the SVM doesn't maximize calmness - it finds a maximally separating hyperplane that separates calm data from not-calm data. If you had real-time biofeedback then the person wearing the handband might have used the feedback to help them relax, but in that case they are the one maximizing calmness, not the SVM.
- PascLeRasc 11y agoOk. My only experience with SVMs was reading the wikipedia article a year ago. Our implementation evolved pretty organically and I only thought later that it was like a SVM. thanks for the info.
- latenightcoding 11y agoI bought the Muse Headband to do something similar, but the headband sucks tbh
- PascLeRasc 11y agoWhat was your problem with it? It's not very isolated from electrical noise, but it can be pretty good still.