[Neuroscience] Re: Minimizing stimulus artifacts

r norman via neur-sci%40net.bio.net (by r_s_norman from _comcast.net)
Thu Mar 1 19:35:03 EST 2007


On 1 Mar 2007 14:53:24 -0800, "Bill" <connelly.bill from gmail.com> wrote:

>On Mar 2, 8:37 am, "Aric Agmon" <aric.ag... from gmail.com> wrote:
>> Bill (and Christian),
>>
>> Without knowing very much about your recording setup, here are two tips that may help:
>> 1. Always use a separate ground for your stimulus (don't share the headstage ground)
>> 2. The recording electrode and the stimulator ground electrode should be, as much as possible, symmetrically arranged relative to the stimulation electrode (e.g. on opposite sides of it).
>>
>> Aric.
>
>Thanks for the comments, though I'm a bit confused.
>When you're talking about "grounding" your stimulus, do you mean like
>the ground pole on the bipolar stimulator? Or do you mean the ground
>pin on the stimulus isolator? If the later, I have always grounded the
>stimulus isolator to the faraday cage or chassis ground. Are you
>suggesting I should ground it directly to the bath?
>
>Also, when it comes to shielding the electrodes from one another, I'm
>not quite sure how I could do that, as both electrodes poke under a
>water immersion DIC lens. Or should I try wrapping foil around the two
>leads which power the stimulating electrode?

It has been a number of years since I actually did this kind of thing
but there was always one cardinal rule that MUST be followed.  There
must be exactly one ground in the bath.   I actually grounded the bath
through a "virtual ground" that forced the bath to follow my reference
ground potential and, incidentally, measured any current flowing
through the bath (my intracellular electrode current that flowed
through the cell to the bath or my intophoresis current for injecting
synaptic transmitter analogs).  This was in the days before commercial
equipment that would handle all that at the front end.

The entire pathway from the input of the recording microelectrode to
the tip of that electrode where all the series resistance lies is a
huge antenna to pick up capacitatively coupled noise and artifact.  As
much of that pathway must be shielded.  Of course, adding shield
increases the electrode capacitance that must be compensated and
interferes with signal bandwidth.  In any event, I suspect that this
is the source of your problem.  Put the shield however you can fit it
between the stimulating and the recording electrodes, near the
stimulating one is easier and surer and avoids the recording electrode
capacitance problem.  The shield, of course, must not enter the
solution or it will cause a ground loop from having multiple grounds
in the solution.  I made a semi-cylindrical shield from a thin
stainless steel sheet that I could fit around most of the electrode
barrel, leaving the end sticking out to go into the solution.  It fit
around the electrode holder and so was spaced away from the actual
electrode a short distance.  It all depends on how closely packed your
equipment is.

I also had a bunch of 120Mohm resistors that I would use as electrode
simulators.  If I wired up a system using resistors for electrodes (I
used 10K or 100K resistors for extracellular type macroelectrodes)
but keeping the geometry as close as possible to the real experimental
situation, I could work out where the problem was coming from without
messing with tissue or even with the bath.



More information about the Neur-sci mailing list