Axon Axopatch 200B, Theory Of Operation

Page: 43 / 153

USE OF THE PATCH CLAMP – A TUTORIAL • 27
Chapter 3
At this point, 95% of the approximately 10 M
Ω
series resistance has been compensated; the
residual series resistance is 500 k
Ω
. An ionic current of 2 nA amplitude would now cause only
a 1 mV error in the membrane potential relative to the command potential, i.e. , a 20-fold
reduction from the situation prior to the use of CORRECTION. Moreover, the true membrane
potential is established within about 30 µs after the start of the step command without overshoot
or ringing. In addition, the bandwidth of current measurement has been increased from 480 Hz
to about 9.6 kHz (of course the measurement bandwidth is still restricted to 5 kHz by the output
filter). It is this increase in the bandwidth of current measurement that is responsible for the
increased noise as the CORRECTION percentage is increased.
Turn on and off the WHOLE CELL CAP. switch and observe the improvement in performance.
Set the vertical gain of the oscilloscope to 2 V/div and set the sweep speed to 1 or 2 ms/div.
With the switch ON, the trace should be essentially flat. Recall that turning off this switch not
only eliminates the correction signal applied to the 5.1 pF capacitor in the headstage used to
compensate for whole-cell capacity transients, but also disables PREDICTION; however,
CORRECTION is not disabled. With the WHOLE CELL CAP. switch turned off, series
resistance is still compensated via positive feedback of the measured current. Turning off this
switch will result in a large ringing capacity transient, with a peak-to-peak amplitude of more
than 10 nA. At a higher vertical gain on the oscilloscope (
e.g. , 200 mV/div), 7 or 8 discernible
peaks can be observed in this transient before they disappear into the noise (about 2 ms following
the beginning of the step command). The membrane potential will also ring severely and have a
1% settling time of nearly 2 ms. Turning the switch back ON completely eliminates the transient
and results in a large improvement in stability: the true membrane potential changes smoothly
without ringing to its new value in about 30 µs following the step command. Similar results can
usually be achieved with real cells.
The percentage CORRECTION can be increased beyond 95%; 100% can often be achieved with
the 10 µs LAG. However, with the parameters of the model cell, as CORRECTION is increased
beyond about 95% the current record shows periodic noise (about 200 µs period; which is
essentially the same as the ring frequency observed above) that may interfere with current
measurement. This can be eliminated by increasing the LAG setting. However, increasing the
LAG setting filters the signal used in CORRECTION (
e.g. , 10 µs corresponds to a 16 kHz
1 pole RC filter, 20 µs corresponds to an 8 kHz filter, etc.). This oscillatory "noise" will also
virtually disappear with the output filter set at 1 or 2 kHz. Similarly, also note that at the 95%
level of CORRECTION the LAG control setting can be reduced to about 2-3 µs before the
system becomes unstable. However, once again, noise will increase.