EEG hardware possibilities for research

About Brain Computer Interfaces, EEG, neurofeedback, related event announcements, jobs, etc.
Post Reply
markandeya
Posts: 2
Joined: Fri Jul 29, 2011 5:45 am

EEG hardware possibilities for research

Post by markandeya »

Hello Friends at OpenVibe,
I wish to start investigating BCI and OpenVibe for meditation, biofeedback, eventually educational game creating/programming, and research into effects of binaural beats and transcranial magnetic stimulation on the brain and consciousness. soooo...
I would like advice on the hardware/eeg aquisition devices that would be applicable. There is the epoc from emotive which is a dry headset, there are the head caps with gel injection, there is the inexpensive neurosky, etc. But i am not knowledgable on which can do the best in so many areas, what an emotiv can do and not do in this research, and what others are being used and considered good in different cost ranges. I don't want to purchase something that will not do enough of what i want, because they are costly, and have to purchase again later on. I also would like to try a dry headset or saline sensored cap if possible. If you have experience or if you have a link to some information i would very much appreciate any help in this area. Thank you so much, Markandeya
p.s. I saw in an openvibe video research being done with a cap which allows inserting electrodes sensors wherever needed and easy insertion with a thick lubricant from a can. Can anyone refer me to some companies and/or some pricing? Thank you again, Markandeya

ddvlamin
Posts: 160
Joined: Thu Aug 13, 2009 8:39 am
Location: Ghent University
Contact:

Re: EEG hardware possibilities for research

Post by ddvlamin »

Hi markandeya,

I have only experience with BCI, so this post may be biased.

The epoc is indeed a good set, especially because it is quite cheap, but maybe less suited if your "core business" is to do research. A paper by researchers of the University of Leuven "Designing a brain-computer interface controlled video-game using consumer grade EEG hardware" have used and evaluated the hardware and it seems to work well for SSVEP based BCI (they claim it is also suitable for P300). I, personally, don't know of other research experiments that are done with it, but there probably are, so I can't say what it is really capable of. For research purposes, I wouldn't go for the Neurosky set.

There are of course many other amplifiers, but most of the time these are a lot more expensive, certainly if they are certified for medical use. I think prices are then easily in the range of 20000 to 30000 euros. Such devices most of the time also have additional channels for measuring other signals, such as EMG, ECG, GSR, accelerometres, respiration,.... If you do combined TMS and EEG you probably need special equipment that minimizes the impact of TMS artifacts on the EEG measurements. TMSI/ANT (http://www.tmsi.com/ and http://www.ant-neuro.com/) equipment can do that, but probably many others too. You have brainamp http://www.brainproducts.com/productdet ... id=5&tab=1, g.Tec http://www.gtec.at/, Neuroscan http://www.neuroscan.com/landing.cfm,....

Biosemi is also a full fledged system that is maybe cheaper than others, but I don't know if its certified. Prices for this system can be found at http://www.biosemi.com/

Maybe this one is the cheapest one I came across: http://www.contecmed.com/paypal/product ... G%20Series It seems that some people are developing drivers for this one so that it will work with OpenViBE

Best regards,
Dieter Devlaminck

markandeya
Posts: 2
Joined: Fri Jul 29, 2011 5:45 am

Re: EEG hardware possibilities for research

Post by markandeya »

Dieter, Thanks so much. It is a great help. I am looking up all the links you sent.
I am learning a lot about what is out there and the many possibilities.
I would like to start with a less expensive set up say under 1000 euros, that can get eeg signals from different places on the head. Then i would like to use openvibe to do biofeedback and make software to allow users to control on-screen displays and movements. I would also like to investigate meditation practices to gather data about brain states and to use biofeedback to assist in getting to those desired brain states.
Thus i need help learning which eeg equiptment might give this functionality and can it be done with a "dry" headset or saline headset/cap. Do i need to get something more than what the emotiv can do? Does having a cap with configurable electrode setups make a big difference? Does more than 2 channels also make a big difference in getting proper and usable eeg data?
Thanks much for the help. Oh and research means my own personal research not for medical use or a university degree. Yet i would like to assist a project by using my testing and data results. Open source philosophy is my way of thinking. Markandeya

ddvlamin
Posts: 160
Joined: Thu Aug 13, 2009 8:39 am
Location: Ghent University
Contact:

Re: EEG hardware possibilities for research

Post by ddvlamin »

Hi,

If 1000euro is your budget, then I think the emotiv is the best consumer grade hardware available, at least to my knowledge. The contec devices may also be good and are just slightly above your budget. To know what this device is capable of (in conjunction with openvibe), you can maybe contact the author of this blog http://engineuring.wordpress.com/2011/0 ... amplifier/

The main problem I see with the emotiv device is the placement of the sensors, they do not span the motor area, so if you want to do feedback based on this paradigm, it could be difficult. Nevertheless, maybe with some engineering this can also be circumvented.

I forgot to mention one other important option - as you support the "open" philosophy :) - there is also the OpenEEG project so that you can make (or order a preassembled) an EEG device with two channels http://openeeg.sourceforge.net/doc/ .

In some cases two channels can be enough, for example foot motor imagery feedback has been done with only one channel. However, it certainly limits your options, e.g. you can't apply certain algorithms such as source localization.

Best regards,
Dieter

Stefanj
Posts: 9
Joined: Fri Dec 30, 2011 5:39 pm
Location: Pretoria, South Africa
Contact:

Re: EEG hardware possibilities for research

Post by Stefanj »

Hi markandeya

Sorry for my reply being a few months late… I have only recently joined this forum, but as this is of interest to many people here, it's worth making a quick list.

So here goes, a list of EEG amplifiers with some of my comments. Starting from rock bottom (price-wise):

1. 100 Euros - Olimex SMT openEEG (built in Bulgaria)
2-channel bipolar amp, 256SPS, 10bit resolution, usb interface
Just launched, unbelievable good price. This is about the same (or cheaper) than some gaming devices such as OCZ NIA or Neurosky. But unlike the gaming devices, this is a real, usable amp. However, don't expect user manuals or any such thing, Olimex is actually a PCB manufacturer and not a product house...
http://www.olimex.com/gadgets/eeg-smt.html

2. less than US$300 – self build modEEG amplifiers having 2, 4 or 6 channels
http://openeeg.sourceforge.net/doc/modeeg/modeeg.html
The only way to go if you want a really cheap 4 or 6 channel amplifier. Ready made PCBs can be brought, but you are on your own with mounting all in a suitable enclosure and figuring out how to inter-connect everything. Well it's easy - I am a EE. But it's a daunting task for beginners, the openeeg forum attests to that.

3. US$ 300 – Emotiv EPOC
I should actually not list this device here, at least not at this price, because this $300 version does not allow for raw EEG data. I also have my misgivings about this device’s extremely low 1Mohm input impedance, making it a bit of a toy, not for serious work.
14-channel monopolar, 128SPS, 14 bit resolution, propriety wireless

4. US$600 – USB or wireless EEG pendant (Australian built)
2-channel bipolar amp, 128/256/512SPS, I think it’s 12bit resolution
Good performance but not well build (it’s build into a cheap MP3 player housing)
http://www.pocket-neurobics.com/

5. US$600 – QDS Focus USB (Argentinean built)
2-channel bipolar amp, 256SPS, 16bit resolution
Many happy users, but I have my reservations, it uses a 9V battery which dangles on the OUTSIDE of the enclosure, at one stage the battery connector that was fitted was of such bad quality/poor fit, that new devices where working intermittently…
http://www.qeeg.com.ar/es1024768/defaultENG.html
http://www.brain-trainer.com/cgi-bin/sh ... item_id=60

6. US$710 – Neurobit Optima2
2-channel bipolar amp, up to 2000SPS, 16bit resolution, wireless bluetooth
The Neurobit name I remember from many years back when they sold a silly little one channel device. I don't know the Optima, but on paper it seems like quite a nice amp.
http://www.neurobitsystems.com/neurobitoptima.htm

7. US$1040 - Neurobit Optima4
4-channel bipolar amp, up to 2000SPS, 16bit resolution, wireless bluetooth
http://www.neurobitsystems.com/neurobitoptima.htm

8. US$1200 – Contec KT88-1016 (Far east)
16-channel monopolar amp, I think they use a 12-bit ADC
This is the only amp that is shipped with cables and electrodes included, especially when considering that this unit has 16 channels, this is a saving of $100s (!)
Unfortunately I can't give any further comments about this amp, as I don't know much about it. If anyone has color photos of the INSIDE of this or other Contec amps, please send them to me [bio.sens.eng@gmail.com]

9. US$1200 – Alpha200
10. US$1800 – Alpha400
2- and 4-channel bipolar amps, 501(k) FDA approved, 256SPS, 16bits, USB
http://www.telediagnostic.com/Alpha200_Main.htm

11. US$1240 – Brainmaster 2EB
12. US$2695 – Brainmaster Atlantis I
2- and 4-channel bipolar amps, 256SPS, USB, Atlantis has 24bit resolution
http://store.brainmaster.com/browse.cfm/4,254.html

13. US$1995 – J&J Engineering I-330-C2+ 6channel (only 2 EEG channels)
14. US$3195 - J&J Engineering I-330-C2+ 12channel (only 4 EEG channels)
2- and 4-channel bipolar amps, 1024SPS, 16bits, USB
http://www.jjengineering.com/C6.htm

15. US$2230 – QDS Excalibur (Argentinean built)
16. US$3580 – QDS Quantum (Argentinean built)
4-and 8-channel bipolar amp, 256SPS, 16bit resolution, RS232
I have my misgivings about this manufacturer. I have back-engineered an Excalibur (from hi-res, color photos which where sent to me). It uses a 6th order anti-aliasing LPF but with 10% tolerance capacitors. Go figure. Also, if the proprietor claims 0nV noise for his amps, my respect for him goes down the toilet. 0nV noise is a scientific impossibility, just as 0Liters/100km fuel consumption for your car would be.
http://www.qeeg.com.ar/es1024768/defaultENG.html


This is by no means a complete list. You can google for names such as BrainQuery (PET), MindMaster, ProComp, NeuroAmp (EEG Info), Nexus, Nexstim, BioSemi, Braintronics, gTec, Cadwell, Philips etc. for more. The last five mentioned companies make these 32, 64… 256 and more channel equipment.

As you can see, most amplifiers are 2 or 4 channels only and of bipolar design. These amplifiers are actually intended for neurofeedback use. For BCI you ideally want as many channels as possible, say at least 8, preferably 16 or more. Many-channel amplifiers are always monopolar, as a bipolar design would be more expensive and clumsy to use, having to jumper many reference inputs together.

There is a real hole in the market for an entry level 8+ channel amplifier for BCI use. I am thus working on a 8/16 channel entry level amplifier. Just started, so it’s early days, but I think a price bracket of US$500-600 should be achievable.

As tempting as the Emotiv is, I would not recommend it (see my comments elsewhere on this forum). Item (1) the 100 Euro Olimex amplifier, although using these silly little “stereo” 3.5mm jack plugs, is an absolute steal. It has just been launched, but don’t expect the supplier to help you with technical queries. You buy it and are on your own. Although only 10bits, I regard this as a “real” (read usable) amplifier. Other cheap gaming devices such as OCZ NIA and the Neurosky are only single channel and should therefore not even be considered.

FORGET dry electrodes, especially on entry level (gaming) devices. For instance, the OCZ NIA, is a reasonably well engineered amplifier, with very good build quality, but it’s performance is lousy. A well respected friend did tests, and he claims that this is due to the dry, carbon bristle sensor used. There is a lot of talk about dry, and capacitive sensors. They are all problematic or, if you find something that works, will be very expensive. Getting a capacitive sensor to work in a lab is one thing, selling it in the real world, that does not want to hear about electrostatic sensitive devices – and the protection of such, is quite another. Wet saline electrodes are OK – just. But if you want proper reliable signals, you should go with proper prepping and conductive gel applied electrodes.

Regards
Stefan

James__D_
Posts: 2
Joined: Wed Feb 08, 2012 4:14 pm

Re: EEG hardware possibilities for research

Post by James__D_ »

Hi,

With regards to the Olimex SMT openEEG.
I'm considering ordering one of these.
Has anyone here had experience using this device and can a; verify that it works and b; suggest some good software which will work with it for neurofeedback.

Thanks

James

Stefanj
Posts: 9
Joined: Fri Dec 30, 2011 5:39 pm
Location: Pretoria, South Africa
Contact:

Re: EEG hardware possibilities for research

Post by Stefanj »

James,

If you want to do neurofeedback, BioExplorer is the software of choice. It’s a no-brainer – excuse the pun.

BioExplorer has practically no technical support from Larry Janow, the creator of BE. Larry has recently released a new version, something that was long overdue. You are unlikely to get any emails answered and for this reason I would suggest that you buy it through a reputable reseller like brain-trainer or some others, rather than trying to place your order via cyberevolution directly. There is a demo version available, so you can try it out first.

So why my recommendation if the support is non-existing? Well, to do NFB you need a EEG amp and software… but there is a third thing that is required – NFB training protocols. If you understand the brain and new exactly how you wanted to train it, you could create your own protocols. But NFB is a rather complex field that takes months, if not years to master. There are many ready designed training protocols available for BioExplorer. Some, you may get for free but many you unfortunately have to buy. BE ships with some examples. Pieter van Deusen runs the brain-trainer yahoo group and he specifically, but also many of his group members, are very, very helpful. This help is for both the technical BioExplorer issues, as well as training issues. There is also a BE specific yahoo group where you can get technical help. Pieter sells his TLC package that contains many, many protocols and other tools, but I myself don’t do NFB and am thus not privy to the exact contents of this package. Because you want to go with entry level hardware, cost is a obviously a big factor. I think the TLC package costs much more than entry level hardware and BE combined… so it’s something you are unlikely to buy now, but at least these options are open for you and will be cheaper than any other training and protocol packages from elsewhere.

The other alternative is BioEra. It costs about the same as BE and has excellent support from it’s creator, Jarek Foltynski. The Othmers (EEGinfo) use a modified version of BioEra known as Cygnet, and this speaks volumes for this software. BioEra is technically more advanced than BE. For instance, BE cannot handle the ultra low frequencies for SCP training. A highly controversial training protocol done by the Othmers.

The problem with BioEra is that the user base is small, it is much more difficult to learn and there are not really any training protocols available. (Apparently there are some Cygnet protocols that can be downloaded for free from eeginfo but I am not sure if those are compatible with the standard BioEra software.)

Some companies sell their own software suites, but these are closed to their individual user groups and you have to pay thousands to train with them and normally are also locked into using their specific hardware. Although BioExplorer software is proprietary and needs to be purchased, it gives you a sort of open platform. It is what just about all home-trainers use. I believe this is the cheapest route for getting started and the cheapest way of getting professional training, if you would want to. BE also supports most hardware, but the same is true for BioEra, and then some...

SMT-EEG (Olimex)

Olimex has posted further pictures and a user guide on their website. It seems that the active electrodes are much smaller than what I judged them from an initial picture I’ve seen. Apparently they are 23x13mm and have their bottom side of the PCB gold plated to act as the electrode. The passive electrode has similar dimensions and similar construction. Nothing can beat silver (Ag/AgCl) electrodes, how good, acceptable or mediocre this gold-plated, PCB material is, I can’t say, I don’t have hands-on experience.

Even if these electrodes have acceptable performance, I would not be surprised if they age quickly. The PCB gold plating is normally very thin and one should be very careful not to scratch the surface. Few electrodes are made from a homogeneous material (such as solid silver). Anything like the gold plating or the AgCl layer on Ag/AgCl electrodes when scratched, will present mixed metals to the electrochemically formed half-cell and will then create large DC offset voltages. (It is important that all electrodes are of the same material so as to balance, or cancel the half-cell voltage. Scratches will cause imbalance.)

For those that can solder, I would recommend making your own cables to be used with either reusable Ag/AgCl electrodes or the disposable DSC electrodes (plastic electrodes with an Ag/AgCl coating).

One way would be to buy ready made cables and replace the DIN connector with a 3.5mm jack plug. The disadvantage with this route being high expense and no screened cable.

The alternative is to by “mono” screened cable and jack plugs at any electronics outlet, and make your own from scratch. Going this route you will need to purchase either snap-on sockets for the reusable electrodes, or suitable pin sockets in case you want to use the DSC disposables. The problem is to find a supplier for this latter item.

If SMT-EEG proves itself (as I think it will), I expect Olimex themselves, or some private individual, to see the gap and start selling such cables in due course.

To conclude, I would try SMT-EEG and purchase 4 active and 5 passive electrodes (they are cheap enough). I would then experiment by applying an absorptive pad and apply saline solution such as contact lens cleaner. Then I would try them with conductive paste. If the 50/60 Hz is too high, or the signal quality not good, I would invest in Ag/AgCl electrodes and go through the schlep of making my own cables to interface with those Ag/AgCl.

Regards
Stefan

zackbam
Posts: 5
Joined: Fri Mar 09, 2012 5:54 pm
Location: Barcelona
Contact:

Re: EEG hardware possibilities for research

Post by zackbam »

Has anyone any experience with thisEEG amplifier? I just sent an email to the supplier asking for the price. It might be quite affordable...

hatma
Posts: 12
Joined: Wed May 08, 2013 11:10 pm

Re: EEG hardware possibilities for research

Post by hatma »

Dear all..

I was planning to spend 750USD to buy emotiv eeg (research edition), because I need to upgrade from Neurosky Mindwave Mobile which only has 1 electrode.
I thought it would give a good results since it has 14 electrodes.

I still don't understand why you think Emotiv device is not "good enough".
could you please explain me some more?

PS: I am newbie, please forgive me if I am asking a stupid question

thanks
hatma

duvi
Posts: 17
Joined: Fri Sep 04, 2009 8:20 am

Re: EEG hardware possibilities for research

Post by duvi »

Dear all,

Here is one of our study that compares Emotiv and a medical device. I hope it will be helpful for investment decision.
http://www.biomedical-engineering-onlin ... nt/12/1/56

Best regards,

Duvi

salamandyr
Posts: 4
Joined: Wed Nov 14, 2012 7:37 pm
Contact:

Re: EEG hardware possibilities for research

Post by salamandyr »

Pocket Neurobics has some great new amps coming in - a rugged 2 channel EEG ($700) and a HEG / 4 channel EEG that can sample sets of 4 channels to gather a whole cap ($1K).

I don't have any affiliation with them, other than using a few of the previous generation products. http://www.pocketneurobics.com but I'm going to get an "E-Wiz" at least. Looks like EEG devices are 24 bit, too.

I use them mostly with BioExplorer - any chance of a community-generated driver for OpenVibe? :D Are there sample drivers? I'm sure Bruce at PN would share sample C code to read packets...

Post Reply