Spectrastrobe, Kasina & Mind Workstation

Hi All,

I've got the EPOC EEG system pretty well integrated into my lab now. I have five dedicated student interns and with their help I hope to get through some basic experiments in EEG-band classical conditioning this summer. As of now, all of our AVS work has been done using my Procyon, because of the way it can be specifically programmed. We've been thinking about the future of our research program and I am interested in manipulating color displays with an open-eyes display. Lately, I have been thinking about Transparent's Mind Workstation software. It seems a good fit in that it integrates with the EPOC device and can use EPOC EEG for biofeedback. Somewhere on-line, I read CraigT's article on Transparent supporting Spectra strobe in Mind Workstation 5.1. Does anybody know if this is the case? I find the Transparent pages hard to get through for specific information. Any help in this area, with some elaboration, if possible, would be greatly appreciated.

Thanks in advance,

John

P.S. Isn't there a specific research forum here? I know I was posting my research ideas somewhere but I can't seem to find the threads.

Hi John,
The Kasina and Mind Workstation are fully and seamlessly integrated. When you plug in your Kasina, new toolbar items appear in MWS to play the session directly through the Kasina by USB audio. Also, you can save your session directly to the Kasina's SD card without have to convert and export first. One of our goals in designing this new machine was to achieve this level of integration and with the help of Transparent's talented team this has been fully realized.

You will find that there is much more you can do with the Kasina in MWS than with the Procyon as with SpectraStrobe you get complete programming access to all the colors.

BTW, Neuroprogrammer 3 supports SpectraStrobe and the Kasina as well.

Concerning your work with the EPOC, MWS is the ideal platform for using it fully. In fact, the lead software engineer at Transparent worked on some of the EPOC's algorithms. (so I am told) You have full access to all channels and sensors even without a developer's license. When you use the Kasina with this combination for stimulation and visualization, you have a wide ranging research platform.

There is a research thread at the Transparent site but not here yet. That is a good idea though and I will contact Andy the forum admin and ask to have one added.

We are very interested to hear of your research and if you need any assistance in setting up the Kasina with Mind Workstation or otherwise, then we are happy to help.

Best regards,
Scott

With MWS, you can easily control just about any aspect of an AVE session when using the Kasina. Intensity, color levels, even multiple waveforms along a timeline are easily adjusted in the interface. Left and right sides can be programmed independently. It is simply amazing! No other Mind Machine that I know of gives you the level of creativity and control as the Kasina. Check out the videos Scott so graciously produced and posted on this thread http://www.transparentcorp.com/commu...ws-and-kasina/ and I'm sure you will agree.

With the Epoc and MWS, you can graph all data available (Professional), use the data to control any part of an AVE session, including light and color, 'drive' entrainment based on EEG data, as well as create visual neurofeedback games.

Quote:

---

Originally Posted by brewmasher

With the Epoc and MWS, you can graph all data available (Professional), use the data to control any part of an AVE session, including light and color, 'drive' entrainment based on EEG data, as well as create visual neurofeedback games.

---

Here is a question that still hasn't been answered. Does MWS work with the standard Emotiv EPOC? The confusion comes from the use of the term "Emotive EPOC" in several posts here and on the Transparent forums. In Emotiv's parlance, the term "Emotiv EPOC" refers to the $299 headset that is, in actuality, a brain control device. The term "EPOC EEG" refers to the $799 EEG headset that comes with the research and development SDK.

I have the EPOC EEG, but if MWS really does work with the Emotiv EPOC ($299), I could probably afford to purchase a few more headsets that would greatly shorten my experiment cycles.

The reason I asked is that the firmware for the $299 Emotiv EPOC headset was cracked as early as 2012, the publication of which led Emotiv to threaten to stop production of the $299 headset. Of course, that is lunacy because there is no way that Emotiv, given its support system, can be relying on sales of the $799 headset to be 'keeping them in business', as they claimed. What had to have made the EPOC (either version) possible, in the first place, was the prospect of an eventual penetration into the computer gaming market. If I was a prospective investor, that would be what would sell me, not the possible availability of research-grade EEG for under $1k. Emotiv's claims to the contrary are obviously fabrications. I'm a member of the Society for Neuroscience. The total membership of SfN is probably still well less than 90,000 (based on recent meeting statistics, about 40k, and extrapolations). Only a very small fraction of the members of SfN have any interest in EEG. So, even if Emotiv was to completely capture that market, they would probably looking at only a few thousand sales (i.e., not very attractive from a investment prospectus standpoint, in my opinion). Contrast that to the tens of millions of avid gamers worldwide and the potential profit associated by penetrating even a small part (e.g., 5%) of that market. The emperor really does appear to have no clothes.

Emotiv, like many small companies with innovative technology, became overly protective of a market that has to be almost insignificant to them by now. So, it seems to me that it would be in their best interest to open up their $299 Emotiv EPOC headset by making legitimate agreements with related companies (e.g., Transparent). My guess is that the Transparent MWS users (especially those using the Standard version of MWS) outnumber research EEG neuroscientists by a considerable amount. Additionally, my guess is that the average $199 - $699 MWS user is probably not all that interested in acquiring a $799 "add-on". So, if the $299 headset doesn't work with MWS, I think that Emotiv is seriously neglecting what could be a considerable market for them.

John

Yes it works with the Consumer headset. Fully functional as mentioned above in the 'developer's license' not required. That refers to the Research SDK. Transparent had to pay a princely sum for this ability. I am often confounded by Emotiv's business practices and philosophy. Contrast it by example with Neurosky who does exactly the opposite, meaning everything is open source as far as development, although the actual algorithms are still a black box. But anyway whatever...

I am waiting to see the release of the Emotiv Insight and if it rectifies some of these deficiencies. I definitely am not one to pre-order from them though, being a past customer with a bit of bad taste. (dry sensors anyone?)

Quote:

---

Originally Posted by neuroasis

Yes it works with the Consumer headset. Fully functional as mentioned above in the 'developer's license' not required. That refers to the Research SDK. Transparent had to pay a princely sum for this ability. I am often confounded by Emotiv's business practices and philosophy. Contrast it by example with Neurosky who does exactly the opposite, meaning everything is open source as far as development, although the actual algorithms are still a black box. But anyway whatever...

I am waiting to see the release of the Emotiv Insight and if it rectifies some of these deficiencies. I definitely am not one to pre-order from them though, being a past customer with a bit of bad taste. (dry sensors anyone?)

---

Thanks, Scott! Apparently saner minds are gaining control at Emotiv. Maybe I should approach them as a marketing director :icon_wink:

John

They are very good at marketing, excellent in fact. Product design, reliability, customer service and communication... not so much. The EPOC was their first foray though, surely in this next product line, all of that will improve.

With my EPOC: I have a broken arm hinge, 2 sensor pods where the plastic teeth broke off, and my battery swelled and stopped charging, so I had to replace it. I have had massive green copper crystal growth which is disgusting and needs to be cleaned all the time.. pretty much every problem reported on the forums, I have had. Still it limps along and I use it. Still the best choice for MWS without BioExplorer.

Joe's system for wearing it backwards gives a much more useful placement of electrodes for entrainment. And also at least you can lie down with it on. These early ideas of using headsets like this for gaming, just are not feasible yet. Emotiv may not have anticipated the degree to which their customer base would want these units for 'quantified self' use. Gaming got them all the press I suppose and also might have gotten around some regulation.

Generally though I have switched to an Olimex EEG-SMT OpenEEG system with active electrodes. I pipe this into MWS with BioExplorer.

I agree with you Scott. I am on my second Epoc also. It does need quite a bit of fussing, not only to get it to work, but to keep it maintained. There is only a 90 day warranty, that is if you can get customer service in the first place. However, if treated correctly, it works great for the money. I learned many things with the first headset. It would have been nice to have known that before. The info they give on use and practice is not always the best course of action.

I am hoping that this will improve with the Insight, which is their answer to the dry sensor debacle. Doing away with the saline sensors will solve 90% of the problems.

Same here. Many first time user mistakes that could have been prevented. For instance, do not take off the felt pads after each use. You are just going to be frustrated all the time, before you use the thing. And you will lose and break them. Store the EPOC with them on and then take them off every couple of weeks or so to clean them. Set up this way is actually pretty easy. Also putting the headset on backwards is easier to line up too I think. The only problem is that there 3D visualization software display is not very useful this way as everything is reversed.

Yea, I leave the pads in when I know I am going to be using it often. If you put it up, remove just the felt pads- not the whole assembly. That way the corrosion will not build up between the pad and the gold electroplate. It is easier to just put the pads back, and there is no chance of stripping the sensor assembly. The only disadvantage is that it tears the pads up quicker. I found a bit of cotton stuffed in the cup works just as well.

If you have MWS Enterprise, export EEG driven sessions as it plays, and use those until you can put the headset back on. Maybe not as effective at real time EEG driving, but the next best thing.

I have used that technique before, even to make sessions that more closely model the flow of EEG readings in meditation. A valuable feature in the Enterprise version and a good one for research.

Quote:

---

Originally Posted by mosspa

P.S. Isn't there a specific research forum here? I know I was posting my research ideas somewhere but I can't seem to find the threads.

---

Hi John;

There are some research threads here if that works for you.

Quote:

---

Originally Posted by neuroasis

Same here. Many first time user mistakes that could have been prevented. For instance, do not take off the felt pads after each use. You are just going to be frustrated all the time, before you use the thing. And you will lose and break them. Store the EPOC with them on and then take them off every couple of weeks or so to clean them. Set up this way is actually pretty easy. Also putting the headset on backwards is easier to line up too I think. The only problem is that there 3D visualization software display is not very useful this way as everything is reversed.

---

Well, when I first looked at the actual construction of the EPOC I realized that several things needed to be addressed. The contact materials on the headset and electrodes appear to be some copper/brass combination. I recognized that this would ultimately be a recipe for trouble in the long run. When we developed our standard operating procedures for the lab I specified that we would remove the electrodes from the headset after each use. Given the saline environment and leakage that seemed to make the most sense. What I did determine, however, was that the locking mechanism on the rehydrator cases is ridiculous, so we filed them down, and then some. Second, in order to deal with the green tarnish and crystallization after every use, we dry the conductors of both the headset and electrodes (if necessary), then we use a Q-tip to apply a bit of Tarn-X (one of those miracle products I could never live without), after which we dry the headset connectors with a dry Q-tip and the electrodes with a cotton cloth. This, effectively deals with the oxidation problem. After playing for a couple of weeks we determined that the electrical interface between the electrodes and the scalp are a lot better if we abrade the scalp a bit, use NuPrep on the frontal electrode sites, and saturate the area around the electrode contact with the contact saline. After we started doing this we have been getting between 11-14 green sites and no orange or red sites. One thing we found early on about the green corrosion is that if it is allowed to get too bad, the electrode ceases to work, almost altogether. That is the electrode will give a sognal that is mostly black with some bursts of red. Since we have been using the Tar-X protocol we have never seen anyything like that again.

Also, after I saw the construction of the system I purchased two bags of felts and two additional hydratror kits with electrodes. In the three, or so, months we have been working with the EPOC, we have only had to replace four electrode assemblies. I would recommend that anybody who buys any EPOC system also purchase an extra hydrator kit that includes 16 extra electrodes.

John

Quote:

---

Originally Posted by neuroasis

I have used that technique before, even to make sessions that more closely model the flow of EEG readings in meditation. A valuable feature in the Enterprise version and a good one for research.

---

I've been in contact with Transparent and they have been very generous with me on their upgrade policy. That would not make the Enterprise edition of MWS out of the question for me. However, I am intrigued to learn what the difference in the EEG functions of the various versions are. For example, what can you do with the Enterprise edition that you can't due with Pro? Also, is there actually any real access to the raw (or filtered) EEG in any of the versions?

Thanks in advance,

John

Thanks, Andy!

I think in general people in a lab environment, where personnel is doing the setup, prep and breakdown, the EPOC probably fares better. This of course is not who it is marketed to. For individual hobbyists who are responsible for all these tasks and are the subject of the study, it is too much of an undertaking and alters one's mood (and increases stress levels I think) It is ridiculous at its face to think that any gamer would undergo this prior to play.

I have also used Ten20 with it and it did improve the readings a bit, maybe, but really made the sponges messy. Now, I have a shaven head so usually EEG setup for me is really simple, but I find that actually the EPOC might not work as well in this case. It makes the contact of the electrodes very important as there is no other channel (like wet hair) for conductance. I have sometimes resorted to surgical pads and tape. Again, after all this, it takes longer to settle down to meditate and often I might even have to rehydrate the pads in the middle of a long meditation.

Active electrodes are where it is at... simple hassle free, with only the crazy chain of cables being the annoying factor.

Quote:

---

I've been in contact with Transparent and they have been very generous with me on their upgrade policy. That would not make the Enterprise edition of MWS out of the question for me. However, I am intrigued to learn what the difference in the EEG functions of the various versions are. For example, what can you do with the Enterprise edition that you can't due with Pro? Also, is there actually any real access to the raw (or filtered) EEG in any of the versions?

---

The Transparent staff are good people..

I have to say though that I am not the best person to compare MWS versions. I have used Enterprise for so long, I just take it for granted. But here is the feature matrix: http://www.transparentcorp.com/produ...comparison.php

From reading it, I would say that the realtime export we have been discussing and the inclusion of external EEG software like BioEra and BioExplorer are the big ones in that context.

All of the raw EEG readings from each sensor are available from the EPOC (or maybe somewhat filtered, not absolutely sure about that) Also, Emotiv's sensory scale readings are available too. I would say that all the same data you can get from the Research SDK is available. The main thing is that with Mind Workstation you can USE this data, instead of just viewing it.

That's another thing that annoys me to no end... $80 for a bag of felt pads. That is beyond outrageous and says they are not serious about people actively using the system. Anyway, I don't mean to be so snarky but this is the real facts about the Emotiv EPOC.

Well, I just purchased a NeuroSky unit from Amazon and will receive it on Monday (today is Friday for historical purposes). Before I decide to invest in either the Enterprise edition of MWS or an additional $299 standard EPOC headset, I have to get some idea of what I can do with any system I end up with. I will have 12 days remaining on my MWS trial in which to play with things in the simplest (one-electrode) environment possible. That will give me time to try all the possible permutations among EPOC/NeuroSky and MWS. It would really be great if the NeuroSky would do what I currently need even if I need Enterprise MWS to get there. It would be a lot easier asking for $8K in startup funds (100 NeuroSky units) than $30K (100 EPOC units).

You guys have all been, and continue to be an amazing amount of help. Between you and Aaron at Transparent, I might actually figure out what I need to do before the end of next week.

Thank you all,

John

Neurosky headsets are only marginally useful. I considered them much more of a biofeedback system than a true EEG. It is not only the single electrode, this is actually very convenient, even with limited data most people will get a lot more usage than a more complex system. Mind Workstation supports it really well and again puts the signals to good use.

The problem with the Neurosky setup is that the 'raw' values have no relationship to any real world measure. You cannot extract any type of frequency information and must only treat this data as a relative unit. The eSense values (or whatever they call them) are better viewed as relative biofeedback values. All sessions that I design for Neurosky units take an initial reading and then alter parameters in contrast to that.

Quote:

---

Originally Posted by neuroasis

Neurosky headsets are only marginally useful. I considered them much more of a biofeedback system than a true EEG.

---

We have been obtaining 'research quality' EEG with the Neurosky device.

Quote:

---

The problem with the Neurosky setup is that the 'raw' values have no relationship to any real world measure. You cannot extract any type of frequency information and must only treat this data as a relative unit. The eSense values (or whatever they call them) are better viewed as relative biofeedback values. All sessions that I design for Neurosky units take an initial reading and then alter parameters in contrast to that.

---

Quality EEG recording can be done from the Neurosky device. Here's what we are currently doing. First, we found OpenVibe, a modular brain-machine interface development environment sponsored by the French govt. OpenVibe consists of a series of modules, of which we only use the EEG acquisition client, the draw EEG on screen module, and the write the raw data to disk as a .csv file module. The acquisition client is, indeed, modular. It allows acquisition modules to be independently constructed for particular hardware. A number of clinical and research EEG devices are supported, and OpenVibe currently supports both Neurosky (non-mobile versions only) and Epoc (SDK 1.x only).

As far as collecting 'research quality' EEG. Here is what we are doing with the Neurosky. First, our current interest is in looking at changes in the Alpha band. In this regard the Neurosky headset is ideal. If worn backwards, it is possible to put the electrode over the middle of the occipital cortex at the location that would be o0, if that electrode location was part of the 10/20 system (i.e., midway between o1 and o2). We then collect the EEG data using OpenVibe. When the OpenVibe template is closed it saves the .csv file to the HDD. All that is necessary to prepare the ,csv file for Matlab is to remove the first line of the .csv header file with Excel. Another neuroscientist and myself constructed a Matlab routine that reads the .csv file, subtracts out the DC offset and then performs an FFT on the data. The routine plots a power/frequency chart for the data. It then goes into the FFT and extracts the relative power within the Alpha, Theta, Delta, Beta and Gamma bands and writes them to the screen. We save the data as a .jpg that includes both the plot and the band-fraction data.

As I said, the EEG we get from the Neurosky compares very consistently with the EPOC and a a Brainmaster clinical EEG that we occasionally have access to. So for $89, the Neurosky is definitely the most inexpensive EEG device available.

I will be happy to post the Matlab routine and some instructions in its use if anybody is interested.

Currently, we are setting up Mind Workstation to control the music and Kasina visual stimulation. I anticipate that we will begin collecting experimental data from subjects around January.

John

Yes, do post the MatLab files. I use EEGLab with it for analysis of EDF files I am obtaining from OPI TrueSense hardware. So I would be interested in your designs. I agree that you can get good data, it is just not in a form that most people expect. (ie specific frequency information) When treated as relative units and evaluating band power change then yes. I do the same trick as you, that is turning the headset around for O1 data. I have the MindSet so I can also turn it to get CZ,FZ or PZ data. A little surgical tape stabilizes it but with care it is fine.

I wish MWS supported OpenVibe directly and have requested this. I can share some of my Neurosky MWS designs with you if you would like. I have some formula engines that do some relative calculations and with the Graphing engine you can get a useful data set as well.

Thanks so much for sharing your results. I hope my earlier posts didn't leave you a wrong impression. I was simply trying to inform about what the 'raw' stream actually was.

Scott

Sorry this took so long. I have included a Matlab file <plotFFTb.m>. As well as the source (below in text). Have fun if you can use it.
You will need OpenVibe to collect the raw data stream. The only module that works for Neurosky is for the non-mobile units. In OpenVibe
you only need to use the display EEG and write CSV to disk routines. Make sure you specify in OpenVibe that you want it to write
a comma delimited file (it defaults to semicolon delimiters). After you have written the data file, open it with Excel and remove the
first header line (you must delete it). Then save the file with a .CSV extension. Run Matlab and run the plotFFTb function. When it
asks for a file name type the saved data file without the .CSV extension. The program will then plot the power/frequency data as well
as the proportion of the total power in each relevant frequency band.

John

------------------------------------------------------------------------------------

% plotFFTb.m
%
% Christopher DiMattina & John Rossi III
%
%
% Florida Gulf Coast University
%
% Brief Description: This program loads an EEG data file and plots the FFT
% and PSD. It also calculates the signal power in the
% alpha, theta, delta beta and gamma frequency bands
% as a proportion of overall power
%
% Version: 2.0 Beta
%
% Created: 9-24-2014
%
% Copyright (C) 2014 Christopher DiMattina & John Rossi III
%
% This program is free software: you can redistribute it and/or modify
% it under the terms of the GNU General Public License as published by
% the Free Software Foundation, either version 3 of the License, or
% (at your option) any later version.
%
% This program is distributed in the hope that it will be useful,
% but WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
% GNU General Public License for more details.
%
% You should have received a copy of the GNU General Public License
% along with this program. If not, see <http://www.gnu.org/licenses/>.
%

Fs = 512; % Sampling frequency = 512 Hz
fmax = Fs/2; % Nyquist frequency

% Get Filename

filename = input('Filename to load? ', 's');
filename = strcat(filename, '.csv');

M = csvread(filename);

% M = csvread('test.csv');

t = M(:,1);
y = M(:,2);
y = y-mean(y); % Remove DC component

N = 4096;
FFTy = fftshift(abs(fft(y,N)));
faxis = linspace(-fmax,fmax,N-1);

plotInd = (N/2):length(faxis);
FFTy_temp = FFTy(2:end);

faxis = faxis(plotInd);
FFTy = FFTy_temp(plotInd);

FFTpow = (FFTy.^2);

plot(faxis,FFTy);

deltaInd = find( (faxis > 0.1) & (faxis <=3));
alphaInd = find( (faxis > 8 ) & (faxis <= 13) );
betaInd = find( (faxis > 13) & (faxis <= 30) );
gammaInd = find( (faxis > 30) & (faxis <=100) );
thetaInd = find( (faxis > 3 ) & (faxis <= 8) );
junkloInd = find( (faxis < 0.1) );
junkhiInd = find( (faxis > 100));

deltaFrac = sum(FFTpow(deltaInd))/sum(FFTpow);
alphaFrac = sum(FFTpow(alphaInd))/sum(FFTpow);
betaFrac = sum(FFTpow(betaInd))/sum(FFTpow);
gammaFrac = sum(FFTpow(gammaInd))/sum(FFTpow);
thetaFrac = sum(FFTpow(thetaInd))/sum(FFTpow);
junkloFrac = sum(FFTpow(junkloInd))/sum(FFTpow);
junkhiFrac = sum(FFTpow(junkhiInd))/sum(FFTpow);

disp(sprintf('alpha = %.3f',alphaFrac));
disp(sprintf('theta = %.3f',thetaFrac));
disp(sprintf('delta = %.3f',deltaFrac));
disp(sprintf('beta = %.3f',betaFrac));
disp(sprintf('gamma = %.3f',gammaFrac));
disp(sprintf('lowout = %.3f',junkloFrac));
disp(sprintf('hiout = %.3f',junkhiFrac));