Hi Martin,

The Polhemus systems I've seen  were all pre-historic so I don't know
how they compare to the products they are currently offering. What you
should pay attention to is that there is a reference sensor you can
mount on the head (e.g. using glasses). I don't think those systems
with just a box and a pen are suitable for EEG. The box is usually
stationary and is mounted for instance at the back of the subject's
chair. I haven't seen anyone put it on the head and I don't know if
that'd make sense.  The idea is that you have the box on the chair,
one sensor on the head (in the glasses) and the other in the pen. So
probably you need a two-sensor system and the software should 'know'
that one of them is the reference. I'm sure if you talk to Polhemus
sales reps and say you need to digitize EEG positions on a moving
child they'll know what to offer you.

Best,

Vladimir

On Wed, Feb 4, 2009 at 5:22 PM, Martin Luessi <[log in to unmask]> wrote:
> Dear Christophe, Vladimir,
>
> Thank you very much for your input so far. I have been looking at the
> Zebris and the Polhemus system. The Zebris one looks very nice, but
> the problems that you mentioned make the system far less attractive.
> Also, the fact that the Zerbris system uses ultrasound and the
> Polhemus a magnetic field makes me think that the Polhemus has less
> problems when the sensors are obstructed by the head etc. In fact,
> according to the data sheets, the Polhemus system has a higher
> accuracy.
>
> Thank you also for giving an estimate on the time, I think 15min is doable.
>
> Right now, I think the Polhemus Patriot may be the best choice for us.
> It is by far the least expensive solution (compared to Zerbis and
> Polhemus Fastrak). The only possible shortcoming that I see is that it
> only has 2 sensor inputs, i.e. we can use one sensor mounted to the
> head (Vladimir: what are those glasses that you mentioned?) plus a
> stylus. However, as the sensor gives the 3D coordinates and
> orientation, one fixed sensor should be enough to correct for head
> movements. Does anyone know why people use more than one sensor for
> this?
>
> Martin
>
>
> On Wed, Feb 4, 2009 at 3:49 AM, Vladimir Litvak
> <[log in to unmask]> wrote:
>> I worked with Zebris before, but I thought they have a better system
>> now. What Christophe describes sounds like what I'm familiar with. I'm
>> not sure that'd be the best thing for working with children because
>> (1) those markers tend to fall off especially if the subject moves
>> around (2) You'll get those messages about the markers being
>> obstructed all the time if the children move and that will prolong
>> your coregistration time. Perhaps Polhemus with glasses is a better
>> choice in this case. The glasses should be suitable for a smaller head
>> though.
>>
>> Vladimir
>>
>> On Wed, Feb 4, 2009 at 9:12 AM, Christophe Phillips
>> <[log in to unmask]> wrote:
>>>
>>> Martin,
>>>
>>> as a follow up to Vladimir's email.
>>> The Zebris system uses active sensors that you've got to stick on the face
>>> of the subject. With them one and as long as they're +/- insight of the main
>>> receiver (placed about 1m in front of the subject), localization of any
>>> locations (electrodes, fiducials, etc.) is possible even if the subjects
>>> moves his head. The GUI let you know when localization isn't possible due to
>>> too large movement (rotation) of the head.
>>> Since you've got to stop at each of the 64 electrodes, plus the fiducials,
>>> it takes about 15min for a trained person to digitize the EEG setup. I did a
>>> 256 setup once in about 45 minutes.
>>>
>>> EGI has a system based on cameras that does the job much faster than any
>>> other system, but it's a bit more pricey... and you must be using their
>>> Geodesic caps.
>>> http://www.egi.com/research-division-research-products/sensor-registration
>>>
>>> Best,
>>> Chris
>>>
>>> PS: I'm not sponsored by any of the companies mentioned. :-)
>>>
>>> Vladimir Litvak a écrit :
>>>
>>> Dear Martin,
>>>
>>>
>>>
>>> Do you have any information regarding how long it takes to
>>> digitize the locations of all electrodes (we are planing to use 64
>>> electrodes)? The time it takes is quite important for us as we are
>>> working with children and they don't like to sit still for too long
>>> ;-)
>>>
>>>
>>> I don't think this depends on your digitizer and you can easily find
>>> out without one. In any of the common digitizers the idea is that you
>>> touch each of your electrodes with a pen and when you are ready press
>>> some button. In addition you should measure at least 3 additional
>>> points for fiducials. This is something you can try on a child without
>>> a digitizer and see if it can be tolerated. I'd say it should take
>>> 10-15 min. Note that newer digitizers allow you to put some markers on
>>> the head (for instance with special glasses) so the subject doesn't
>>> have to keep still the whole time.
>>>
>>> I have also seen somewhere that people take photos of their subject's
>>> head from several angles and then use some software to reconstruct
>>> electrode locations in 3D based on those photos. It might be something
>>> that comes with the EGI system, but I'm not sure. That would be a good
>>> solution for you since it's probably very fast. Here is a related link
>>> I found with Google:
>>> http://linkinghub.elsevier.com/retrieve/pii/S1388245705000076
>>>
>>> Best,
>>>
>>> Vladimir
>>>
>>>
>>>
>>>
>>
>
>
>
> --
> ______________________________________
> Martin Luessi, M.S.
> PhD Student
> Department of Electrical Engineering and Computer Science
> Northwestern University
> http://ivpl.eecs.northwestern.edu/people/mluessi
>