How to correct for series resistance (and whole cell capacitance) in real cells

I’ve talked before about the problems series resistance [i.e. the resistance formed by the tip of your electrode and the gunk in it] poses when you’re performing whole-cell voltage clamp recordings. Simply put, it limits your ability to hold the cell at the voltages you want to, because every time you pass current, some voltage forms over the series resistance and you’re amplifier can’t tell the difference between the voltage over the membrane and the voltage over your series resistor. It also limits your temporal accuracy, as it helps to set up up a low pass filter for your command potential. Thankfully, because the problem is essentially inescapable during whole-cell voltage clamp, essentially all voltage clamp amplifiers come with series resistance compensation: an electronic procedure where the amplifier estimates how much of a problem your series resistance will be, and works to minimize it. However, in order for that circuitry to work, your amplifier needs to know the whole-cell capacitance of the cell you’re recording from (Cm) and how much series resistance (Rs) you’re up against. Typically your set these values while looking at the current response to a test pulse. However, outside of a few select examples, this process of telling the amplifier what Cm and Rs is (know as “correcting for whole-cell parameters” or “capacitance compensation”) isn’t actually trivial. Let’s have a look at why this is, and how we should be doing it in real cells

Near the start of any instructional manual on voltage clamp, there is a section describing how when you first break into a cell you correct for whole cell capacitance and series resistance, turning the capacitive transient you see in response to a voltage step into a smoothly rising curve (Fig 1).

Fig 1. By correcting for the whole-cell parameters on the amplifier, you remove the capacitve transient associated with charging the membrane.

This doesn’t do anything to improve your voltage clamp, but it does provide a little electrical help to the amplifier (See Note 1). More importantly however, it informs your amplifier what Rs and Cm are so that the series resistance compensation circuit can work.

Because of the fact that the Rs and Cm values you put into your amplifier are used for series resistance compensation, their exact values matter: If you over-estimate Rs, the Vm of the cell will be pushed further from the resting membrane potential than you expect. If you over estimate Cm, the voltage of the cell will oscillate before settling at a final value.

Despite the importance of setting the Rs and Cm values on your amplifier, I’ve never seen a guide on how to actually do it. If you patch HEK-293 cells, or N1E-115s, it’s simple: either slowly increase Rs and Cm together, and you will eventually zero in on the correct values. Or, slightly more scientifically, estimate Rs by dividing the size of the voltage step you are applying by the peak amplitude of the capacitive transient, dial that in and then increase Cm until you are left with a smooth rising curve like that shown in Fig 1. (you’ll probably need to do some slight adjustment of Rs to get it perfect). However, in neuron with dendrites, things are not quite so simple.

You see, when you set the whole-cell parameters, all the amplifier does is subtract off the current needed to charge a capacitor through a resistor with values equal to the Cm and Rs values you have entered on the amplifier. Want to see what I mean?

The current through a capacitor and resistor in series is described by this equation:

I = \frac{ V_{cmd} }{R_{s}}e^{\frac{-t}{R_{s}C_{m}}}

So we can calculate that current, and mathematically subtract it from the response to out test pulse, and we should end up with what we saw in figure 1. Look at figure 2 below, that’s exactly what I’ve done. The blue line is the raw trace, the orange line was calculated with the equation above, and when I take the difference of those two lines I end up with the yellow curve, which looks almost exactly like what we saw in figure 1.

Fig 2. If you mathematically predict the current you should get trying to charge your membrane capacitor through your series resistor, you can subtract that current from your test pulse, and get something that looks almost identical to a corrected response.

But what all of this means is that if your cell isn’t made up of a capacitor with the value you have entered, and/or that capacitor isn’t being charged through a resistor of the size you have entered, then you’re going to be in trouble. And you’re nearly always going to be in trouble if you’re patching a real neuron as they can never be explained by a single capacitor. And that means that no matter how you set Rs and Cm on your amplifier, you can never get the current to look like what you see in Fig 1. Instead, it looks like Fig 3.

Fig 3. When you voltage clamp cells with dendrites, no matter what values you enter, you cannot make the current response look like it does in the textbook. Cm values in parenthesis are whole-cell capacitance values enter into amplifier, not real value of cell.

What is going on here? Why can’t you get a textbook response in your current when you patch a neuron? Well as I said above, when you dial in your whole cell parameters, you are just subtracting off the response of a capacitor being charged through a resistor. However, a real cell with dendrites can’t be described by a single capacitor, it has dendritic compartments with capacitance, and the dendritic capacitance doesn’t just add to the somatic capacitance because the dendrite is hidden behind the resistance of the cytoplasm down the dendrite (which acts exactly like another series resistor). Because of that, it is impossible to get the waveform to look like figure 1. So how should the waveform appear to let you know you have correctly dialed in the whole cell parameters?

Fig 4. The dark orange trace shows the properly corrected response. Do not be tempted to over set Rs to get a fast rising transient (yellow trace) or under set Cm to get the transient looking squarer. The response should rise smoothly from when the test pulse starts.

Figure 4 shows a properly corrected response along with two bad examples. I’ve seen people incorrectly dial in Cm and Rs in a futile attempt to get the waveform to look like a textbook response, and it usually looks like the purple and yellow traces. People think that they want a fast rising response, so over set Rs to get a squarish looking step. Yes, you do want a fast rising response, but you achieve this by ACTUALLY lowering Rs, not by mis-setting Rs on the amplifier. I’ve also seen people underset Cm in order to make the response look more square. This can be especially easy to do if you have aggressive low pass filtering (e.g. 1 kHz) as it will clip off where the current goes in the wrong direction (i.e. the positive going current at the start of the step). What you’re looking for is a smooth rise from the holding current when your test pulse starts, and you shouldn’t worry yourself about the departure from squareness as there is nothing you can do about it. The approach of calculating Rs from the peak size of the capacative transient, and then dialing in Cm until the transient disappears is still the best approach.

So do you want to see what this looks like in real life? [All the above graphs were generated in NEURON]. Thankfully I made some recordings where I made dual whole-cell recordings from a single soma. One electrode was in voltage clamp and the other was in current clamp. This allows the electrode in current clamp to measure exactly how our voltage clamp was performing. In the below waveforms, the current trace from the electrode in voltage clamp is on top (Ihold) and the membrane potential is shown on the bottom. The voltage clamp is attempting to make a voltage step from -70 mV to – 80mV. I have turned on 70% series resistance compensation which will then demonstrate the effect of incorrectly applying whole-cell correction.

Fig 5. Incorrectly set capacitance leads to oscillations in the membrane voltage.

The above shows us a very common mistake. Everything about the current response looks right, apart from one crucial detail: the current starts to rise several 100 microseconds after the step truly begins. This means that the whole cell capacitance is over set, and it leads to a little oscillation in the voltage response.

Fig 6. Oversetting Rs leads to Vm being further from rest than intended.

The above is what it looks like when you overset Rs. The only time I’ve seen this in the wild is when the true series resistance into a cell randomly drops without anyone noticing. . You can see the current response shows an instantaneous rise. By oversetting Rs, the series resistance compensation is over doing things, and hence despite met telling the amplifier to clamp the cell at -70 mV, we’re actually at about -72 mV. Likewise, the step amplitude is larger than -10 mV. A 1 mV error is really nothing to worry about, but it’s better to avoid it if you can.

Fig 7. Ah… just right.

Finally, we have an approximately correctly set example. The important thing to note is that if you extrapolate backwards from when the rising phase of the transient, it intersects with the baseline current at the moment the voltage step begins (red dashed line). In this particular instance, there is a small instantaneous rise, but this is probably the remnants of uncompensated pipette capacitance. When you are there doing it for real, and you’re looking for this point where the extrapolated current intersects with baseline, it’s pretty easy to tell the difference between a little bit of pipette capacitance and big instantaneous rise shown in figure 6.

So, in conclusion:

  • When you perform whole-cell correction, i.e. dial in the series resistance and whole-cell capacitance, you are NOT improving your voltage clamp, you are just subtracting off the current needed to charge the membrane capacitor.
  • However, it is important to perform whole-cell correction properly as the values you determine are used to perform series resistance compensation.
  • Your best approach is to estimate the series resistance by dividing the size of your test pulse in mV by the peak amplitude of the capactive transient. Then dial that in and start increasing the whole-cell capacitance until the extrapolation of the remaining transient looks like it intersects the holding current at the moment the test pulse starts.

Note 1. When you are performing voltage clamp, you often need to step the cell you’re recording from to a new voltage. When you do this, you need to charge up the capacitance of the cell membrane. Your ability to charge up this membrane capacitor is limited by the resistance of your electrode and the junk in its tip from when you went whole-cell, i.e. your series resistance (Rs). The lower the Rs, the more current your can pump into your cell, and the faster the membrane potential (Vm) approaches your command potential (Vcmd). While you essentially always want as low Rs as possible, it a low Rs can come with a downside: that initial inrush of current when you try to step the potential can saturate certain electronics inside the amplifier (the primary operational amplifier responsible for voltage clamp). When you correct for whole cell parameters, you pass the job of charging up the membrane capacitor to another operational amp, freeing up the primarily operational amplifier to pass current in response to ion channels. Not only does this minimize the risk of saturating the primarily amplifier, but the values are used to perform series resistance compensation.

Leave a Reply

Your email address will not be published. Required fields are marked *