OpenEEG - EEG hardware and software

On Thu, Jan 12, 2012 at 00:38, Joris Dolderer <jolderer@...> wrote:
> 1. From the datasheets, it seems to be that the microcontroller and the
> amplifiers can work with much higher frequencies than 500 Hz. So what
> happens if you only use one
> op-amp cascade and connect the electrodes to this cascade via a
> demultiplexer, then use the microcontroller to switch frequently the
> electrode which is currently mapped to the op-amp cascade and then sort the
> signal the microcontroller receives back to the appropriate electrode in
> software?

Multiplexing signals of such small amplitude is likely to add
significant amounts of noise, to the point that your original signal
may be unrecoverable. Pretty much all uC's with multi-channel ADC DO
multiplex multiple inputs to a single converter, but this happens in
the uC, after amplification and conditioning so noise is less of an
issue.

> 2. If you use active electrodes: What would happen if you set the gain not
> to unity in these electrodes? Could you maybe save one amplification stage
> in the later circuit?

You'll probably overload the frontend of your EEG, that said you can
shift gain around between the active electrode and the rest of the
system if you want to redesign, there have been systems proposed which
have ALL the amplification in active electrodes, pros and cons with
every option, just a matter of striking the best balance for your
application.

First question: It would seem that the data obtained from an EEG assessment would need to be 
"raw" data, in order to analyze it, prior to devising a prescriptive 
training regime. In other words, it would imply that the software design, such as a BrainBay design, would need to simply pass--unaltered--the incoming brain wave information 
directly to the recording targets, such as a virtual oscilloscope, a 
fast-Fourier transform graph, or to a file writer module (for exporting to 
the spreadsheet)...all without the altering/interference of any "processing" 
elements, such as filters or amplifiers, which might change the various brain 
wave characteristics, such as amplitude, wavform, or frequency. Then, after the assessment has been done, analyzed, and a theraputic training program decided upon, it would seem that one would use the various processing elements of BrainBay, in order to provide the desired "uptraining" or "downtraining" of the different frequencies at the 10-20 sites of interest. Am I correct in assuming this? I am asking this question, because I am trying to visualize what a 
good BrainBay assessment design might look like. If I am correct, It would seem that the assessment design would be totally devoid of any filters, threshold elements, magnitude elements, etc.; instead, it would consist of the EEG as a source element, and only "target" elements, without any "processing" elements.  Am I right on this, or totally off-base?   Second question: Are there any existing assessment 
designs available for download, somewhere? It would seem like a forum like this would 
be the perfect place for such a sharing process to take place, among people 
with a like interest. It would be nice to be able to see just how others have 
arranged the various components (filters, reward/inhibit threshold elements, 
etc.), when doing an assessment, as well as later, when designing a neurofeedback 
treatment plans. For one thing, and as the saying goes, "Why reinvent the 
wheel?, and for another thing, I'm afraid that my own attempts to design 
might omit things that should be included, and/or include things that should 
be omitted.  Thank you, in advance, for any help which any of you 
might be willing to share.  --Ray Cole