Series Resistance. Why it’s bad.

At the very start of my Masters, my first experiments appeared to show that the histamine H3 receptor inhibited the release of GABA in the neocortex. It turns out, this was all lies. It was all lies because of series resistance, a concept I had vaguely heard of, but didn’t understand. If you’re just starting electrophysiology this post is for you. The hope is that by the end of this post, you will understand series resistance, and you’ll understand why it is extremely important to monitor it religiously, whether you’re performing voltage clamp or current clamp recordings.

Voltage Clamp

Take Home Points

  • The smaller your series resistance is relative to your membrane resistance, the closer the membrane potential will get to your command potential, given enough time. You should aim for your series resistance (after Rs compensation) to be less than 1/10th of your membrane resistance
  • The membrane capacitor and the series resistances sets up a low-pass filter for your command voltage. Hence you will be able to voltage clamp faster signals when Rs and Cm are small
  • Voltage clamp is never perfect, but it can be good enough. However, if Rs changes during your experiment, all currents you record will change with it, hence you must monitor it, and reject cells when it changes

series resistance openingIn standard whole-cell voltage clamp, the goal is simple: hold the membrane potential of the cell at the command voltage (Vcmd) and tell the experimenter how much current that required, because it should be equal and opposite to any transmembrane current. However, there is one major thing that is stopping things going to plan: series resistance (Rs). Series resistance is the sum of all the resistances between the electronics in the amplifier and the inside of your cell, but in practice is made up of the resistance in the last dozen microns of the pipette and the junk that blocks up the hole between the pipette and the cell (good rule of thumb, your series resistance should be around 3x the open pipette resistance).

But why does series resistance matter? Because together with the resistance (Rm) and capacitance (Cm) of the membrane, it creates a voltage divider, that is, the voltage you are trying to apply to the cell is divided across Rs and the impedance presented by the combination of Rm and Cm. Even simpler yet, you can think of Rs as limiting the amount of current you can use to charge up the membrane. This means that the actual voltage you get inside the cell (Vout) is given by

Vout = \frac{ Vcmd Rm + Vm Rs}{ Rs + Rm  }

Where Vm is the resting membrane potential. As you can see, if Rs is much smaller than Rm, then Vout approaches Vcmd, likewise, if Rs is massive, then Vout approaches Vm. This is where we can see the how the Rs < Rm/10 rule of thumb comes about. If we say Rs = Rm/10, then the above equation becomes:

Vout = \frac{ Vcmd Rm + Vm \frac{Rm}{10}}{ \frac{Rm}{10} + Rm  }

Which simplifies to:

Vout = \frac{ Vcmd +  \frac{Vm}{10}}{ 1+ \frac{1}{10} }

Which basically says Vout = Vcmd plus a tiny fraction of Vm. As Rs becomes a smaller fraction of Rm, the less Vm contributes to the final membrane potential.

But this voltage divider is a frequency dependent voltage divider, because capacitors let more current through when the voltage across them changes rapidly. And do you know the other word for a frequency dependent voltage divider? A filter, in this case a low-pass filter. The corner frequency of this filter is given by a pretty awful looking equation:

f_{c} = \frac{ \sqrt{ Rm^{2}-2 Rm Rs-Rs^{2} } }{2 \pi Cm Rm Rs}

But if Rs is much smaller (about 10x less) than Rm, you can approximate the cut-off frequency of the filter with the much simpler

f_{c} = \frac{ 1 }{2 \pi Rs Cm}

(A little helpful note, if Rs is in MΩ and Cm is in pF then the answer is in MHz).

How series resistance creates a low-pass filter. Behaviour of voltage clamp on a cell with Rm=100MΩ, Cm=100pF

How series resistance creates a low-pass filter. Behaviour of voltage clamp on a cell with Rm=100MΩ, Cm=100pF

To the right you see what this means. If we compare the ratio of the voltage across the cell membrane to the command voltage (Vout/Vin), we can see the effect of Rs on how good out voltage clamp is. Perfect voltage clamp would produce a horizontal line at Vout/Vin = 1, meaning that no matter how fast you tried to change the command voltage, the membrane potential would follow it perfectly. But instead what you see is what I described above: even at very low frequencies (e.g. constant voltages) when the series resistance is high the membrane potential does not even come close to the command voltage. You also see that as Rs increases, the filter becomes more aggressive, meaning that even relative slow voltage changes cannot be achieved. In the case illustrated, of an 100MΩ/100pF cell, when Rs is 40MΩ, you are beginning to fail to voltage clamp at the relatively glacial frequency of 10 Hz.

Trying to voltage clamp a synaptic current in a 100MΩ/100pF cell

Trying to voltage clamp a synaptic current in a 100MΩ/100pF cell

This may still seem relatively academic at this point, so let me really try to show you what this means. To the right you can see what happens when you try to voltage clamp a modeled synaptic current as Rs increases (the synaptic event is current based, rather than conductance, so it is the same no matter how bad the voltage clamp is). I have flipped the synaptic current, so they’re easy to compare. You see that even in the relatively ideal case of Rs = 10MΩ the measured current is a lot smaller than the real current. In part this is unfair because this current is instantaneously rising (and as the previous figure showed, fast changes are the enemy of any voltage clamp), so note how the decay of the current is relatively well captured in all cases. However, the point is, if you had been performing voltage clamp, and your Rs had changed from 10MΩ to 20MΩ, the current you would have measured would have dropped by about 15%. If you had been washing on a drug at the time, you might think that the drug caused that reduction (which is exactly what happened to me). Hence why I am always a lot more convinced by any long term voltage clamp experiment that shows Rs as a function of time, along with any other measurement (e.g. Fig 1 here).

The EPSP produced by our synaptic current (100pA peak) with different values of Rs

The EPSP produced by our synaptic current (100pA peak) with different values of Rs

Of course, not only does Rs effect the current you end up recording, but it has big effects on the amount of unclamped voltage generated by any transmembrane current. To the right you see the size of the EPSP our synaptic current would produce in the absence of voltage clamp. In a perfect world this would be a flat line, i.e. the voltage would be clamped. But when Rs = 40MΩ, you can see that the EPSP produced is nearly half the size of the unclamped potential, i.e. this is not voltage clamp at all, but really some weird thing half way between current clamp and voltage clamp.

Finally, it’s worth mentioning that series resistance compensation allows us to minimize all of these effects (but not remove them). Specifically, if you have a real Rs of 20 MΩ, and then you perform series resistance compensation to 75%, then your series resistance will appear to be 5 MΩ. It’s also important to remember that in large cells, you run into space clamp problems, where you are unable to voltage clamp distal parts of the cell (you can think about the distal sites being behind another series resistor, this time formed by the resistance down the cytoplasm of the dendrite), but that is a whole other story.

For me, voltage clamp is no different that most science: you are not measuring perfectly and your measurements are just an approximation of the real thing. The important thing is to know when the approximation is good, and when it is bad, e.g. Voltage clamp recordings from cereballar granule cell (Cm = 5 pF, Rm = 1000MΩ) give you a good approximation of transmembrane current. But even if those good cases, when Rs changes, even a constant transmembrane current will appear to change, and that is really bad.

Below you can adjust the properties of your voltage clamp system, and see how good you voltage clamp should be.

Current Clamp

Take Home Points

  • A combination of the pipette capacitance and the series resistance sets up a low-pass filter for membrane voltages
  • This filter works at a much higher frequency that the filter effecting voltage clamp
  • Again, you should monitor it, as it can change the features of the voltages you are trying to measure

series resistance ic circuit2People often forget about the effect of Rs on current clamp recording, outside of bridge balance (which I very much hope you understand, but if you don’t, the simple thing is that when you inject current, it falls across the series resistor, meaning that you record a voltage that is the sum of the transmembrane voltage and the voltage across the series resistance. Bridge balance subtracts that voltage). But in a fashion very similar to in the voltage clamp case, series resistance effects the nature of filter that messes with your ability to record cellular properties. However, in current clamp, the capacitor that is involved is the capacitance formed by the pipette (Cp), and the other resistance is basically the input impedance of the amplifier, which I have labelled Rp. This is due to the fact that the circuitry in your amplifier draws some small amount of current. Modelling it as a resistor to ground is a bit crude, but for our purposes it is fine, as the main result is captured: you amplifier draws some current from your cell.

How series resistance effects current clamp recordings. The set up is modeled with Cp = 5pF and Rp = 5GΩ.

How series resistance effects current clamp recordings. Modeled with Cp = 5pF and Rp = 5GΩ.

Because the resistances in parallel with the capacitor are much much larger (aka the input impedance), and the capacitor is much much smaller, the frequencies where this filter starts to kick in are much higher, but action potentials are pretty fast voltage transients. As you can see in the figure to the right where we plot the ratio of the recorded voltage to the actual membrane potential (Vout/Vm), in the case where Rs = 100 MΩ, even signals around 100 Hz are being filtered, which will really filter action potentials (remember, the frequency components that make up the action potential are much faster than the rate at which action potentials occur). Don’t believe me, need more proof?

The effect of series resistance on action potentials recorded via a 5pF pipette and an amplifier with 5GΩ input resistance.

The effect of series resistance on action potentials recorded via a 5pF pipette and an amplifier with 5GΩ input resistance.

Here we see how series resistances effects the ability to record a simple action potential. Clearly, even a relatively modest series resistance start to effect the ability to record action potentials. Potentially the Rp I have included in the model is a bit low, but it’s not massively too low. Moreover, the point remains, Rs, Cp and how much current your amplifier draws creates a filtering system, and if Rs changes, the filter changes. Your amplifiers capacitance neutralization is there to minimize this effect, but it only works if everything is set correctly (and it still only minimizes the effect).

So Long story short: In voltage clamp, series resistance prevents your amplifier from charging the membrane capacitor, and in current clamp, series resistance stops your cell from being able to charge the capacitance of your pipette. These things are bad, but what is worse is when it changes over time. So monitor your series resistance constantly throughout your experiment.

14 thoughts on “Series Resistance. Why it’s bad.

  1. Hello Bill,
    my name is Riccardo and I am a postdoctoral fellow from Pisa (Italy). I am very grateful to you because your discussion about th series resistance solved most of my doubts about this damned issue. I would like to take advantage from your expertise to ask something about my troubles…
    Actually I always have to fight against Rs and often is not easy to me to obtain very low and good values of it…
    I know that important tricks are a quite low pipette resistance (as you have written), a good balance in osmolarity between external and intenal solution (with the internal a bit less concentrated than external – 280/290 vs 300), healthy cells, good approach to them and right positive/negative pressure.
    Unfortunately what too often happens to me is that I not easily obtain the gigaseal and I have to suck a lot with mouth after releasing the positive pressure, in order to allow resistance to rise. And when I reach the gigaseal it is often hard to break the membrane, and after the fact my Rs is never so low… I noticed that the more easy is the reaching of the gigaseal and breaking of membrane, the higher is the probability to obtain nice Rs.
    Unfortunately my pressure system is not very good, because it has an air leak somewhere, and I have to blow with mouth continuously during the cell approach, and I think that this not helps me to obtain nice whole-cells.. What do you think about that?
    Thank you very much for your attention, I will appreciate any suggestion!


    • Hey Riccardo,

      First thing, you’ve got to fix the positive pressure leak. This probably isn’t the cause of the problem, but that will really not be helping (I’ll tell you why soon). Go and beg, borrow or steal some new tubing. Plus, when you’re patching, you’ve got enough to worry about, trying to keep positive pressure by mouth is adding more stress.

      It would have been helpful to know what cell type you’re trying to patch. Anyone who has patched a lot of different cell types will tell you that different cells seal at different rates. But given that, as you’ve said, good seals generally = good access (low and stable Rs). When you’re sitting there sucking on the tubing trying to get a seal you are drawing large amounts of the cell (and sometimes processes of other cells) into your pipette. When you finally go whole-cell, all this stuff is still stuck up your pipette (indeed, anyone who has done 2-photon imaging during whole-cell recordings will tell you there is more stuff up your pipette than you realize. Check out these images to see what I mean). So I think if you get good seals, your problem of increasing Rs will be reduced.

      At this point, I want to say, can you give me some numbers about your Rs? That is to say, if you start at 10 MOhm, and 20 minutes later, you’re up to 20 MOhm, I would say this is vaguely normally, and sometimes/often you get that. If on the other hand, you start at 20MOhm, and in 3 minutes you’re at 40 MOhm, this is very bad. Also, so long as you’re not trying to voltage clamp sodium currents, for most people, it’s more important to have STABLE Rs, than it has to have LOW Rs. So sometimes actually using a smaller pipette (Rs = 6-7MOhm) can give you more stable access.

      So, how do you get a faster former, more stable seal? First, fix that positive pressure leak! If your pressure drops when you’re in the bath or near tissue, this will hamper your ability to form a seal. Also, in my experience, applying too much positive pressure (>300 mbar) can also make you form bad seals. Secondly, are you pipettes and solutions good? I would find someone who is performing patching with no trouble (perhaps on a different cell type), and give them your pipettes and solutions. If they find their seals are bad using your solutions and pipettes, then that tells you the problem. Conversely, use someone elses pipettes and solutions on your cells. If you can seal perfectly, then you know the problem. Finally, are you SURE you cells are healthy? For instance working from very old animals can cause a lot of seal problems. I often tell people to cut cortical slices or cerebellar spices from P16 rat pups. If you can’t patch a layer V pyramidal cell, or a Purkinje cell (assuming you don’t slices their dendrites off), then there is something seriously wrong. (Also, do you have a way of measuring your positive pressure? Purchasing an manometer can be quite expensive (150-300 Euros), but you can make one yourself for not much at all (60 Euros) using something like a SSCSNBN250MBAA5 and an arduino)

      But generally speaking, with somatic recordings from healthy cells from animals in the normal age range, the moment you turn off the positive pressure, the resistance should increase to ~>500 MOhm pretty much instantly. If it doesn’t, you’re in trouble. Make that happen, and I think your problems will be fixed. If you’re working with cultured cells, cells from really old or really young animals, then … hard to say.

      Hope that helps, but if you want to continue this conversation via email (due to privacy) be my guest.

  2. Hi Bill,

    I enjoy your article. This kind of stuff is rarely explained. I’ve got an elementary question. You mention the importance of observing the series resistance during the course of measurements. I’m wondering if it would be possible to measure Rs simultaneously during measurements of synaptic current. Or would it be necessary to pause synaptic current measurement every once in while and to apply pulses in the command voltage for determining Rs?

    Thank you so much!


    • Hi Daisuke,

      Sorry about the slow reply. When measuring EVOKED synaptic currents, it is trivial to measure Rs during your recordings. About 100 ms before you evoke the synaptic event, do a -5 mV test pulse. Here is a beautiful old waveform (from a MSDOS based version of winLTP) showing exactly what I mean.

      If you have Rs correction/compensation on, then you wont get a nice current response with a capacitive transient, however, you can still monitor changes in your Rs.

      If you are doing spontaneous or miniature synaptic events, then you’re in a bit more trouble. I usually patch a cell, waiting around 10 minutes for everything to stabilize, and then start monitoring Rs constantly. I wait for the Rs to get steady, then perform Rs correction/compensation. I then get my 5/10 minute recording of spontaneous events, and monitor Rs again. If it hasn’t changed (and the holding current has stayed largely stable) then I say that recording is good. Generally speaking, an increase in Rs should be paired with the holding current moving towards zero, the holding current itself is a pretty reliable measure of Rs (when you can’t measure it more directly).

      Hope that helps.

  3. I really enjoyed the article, thank you very much for sharing knowledge and explain this important subject in a very clear way

  4. Hi Bill,

    I’m a master student approaching electrophysiology for the first time. I’m patching cells in current clamp in the midbrain and I always have a really high Rs (around 30 MOhm, sometimes even 50). I tried to reduce pipette resistance, change IC and apply different pressure but there was no improvement. I’m using one year old mice expressing Tdtomato in specific neurons. Do you think this can be a reason for the high Rs? I have to say that the Rs even if high is pretty stable throughout the recording (around 10 minutes) and the shape of AP and firing frequency looks “normal”. Any idea why this happens?

    Thank you,


    • Hey Petra,

      1 year old mice! Rather you than me. Recordings are never going to be as beautiful from such old as they can be from younger mice (though in my experience the recordings can be more long lived). In my books, if your Rs is 3 times the open tip resistance of your electrode, then you’re doing well: i.e. if your pipette is 5 Meg, then a 15 Meg series resistance is good. Obviously, you’re getting much more than this. But as you can read in my post, in current clamp, this isn’t such a big deal. I don’t have any particular tips for you to get better series resistance, especially if they’re starting off high. If the recordings are stable, and less than some threshold, then keep them.

      As I imply in the post, to know what theoretical “threshold” you should use for your Rs in current clamp, you need to know the input impedance of your amplifier, which you’re not going to be able to find or measure*. So I would just pick a number, (40 Megs sounds good to me).

      *With certain amplifiers, if you pop open the headstage, you can see the op-amp that your signal feeds into, so you could look up the specs, but I suspect the biggest source of current drain isn’t actually the input impedance of that op-amp, but stray capacitances… though don’t quote me on that.

  5. Hi Bill,

    A stupid question, how to do calculate the series resistance change after offline analysis? I mean which one are you using to subject ? average series resistance or the lowest series resistance or the highest series resistance or the initial series resistance?

    Thank you.

    • Not a stupid question at all. In order to calculate series resistance (Rs) you’ve got to be in a position to measure it! That means, every once and a while, you’ve got to apply a small voltage step, and measure the size of the current transient. In a perfect world, this is part of your protocol e.g. you apply a 10 mV step 100 ms before you turn on your LED/activate your stimulator/squirt on your drugs. But in some situations (like recording miniature synaptic events) you don’t want to do this. So in that case I would make a measurement ever couple of minutes. You have to have some threshold where you say “if it increases by X% or ever goes about X MegaOhm then I kill the recording”. If you’re not applying voltage steps, you can’t calculate the resistance offline (as far as I know).

        • Just the same way as you would for anything else. If A is the original Rs, and B is the later measured Rs, then the percentage change is 100*(A-B)/A

          • Sorry, I MAY make things confused. Let me be clear. How do you determine the A? Are you using the average Rs , the maximum or the minimum Rs?


          • Oh I see. It really depends on how you’re measuring it and what kind of experiment you’re doing. If you measure Rs every 10 minutes, then when you measure it the first time, that one value is your beginning Rs. If you’re measuring it every 5 seconds, then you might as well take an average for the first three readings, but honestly, it shouldn’t really change much over the kind of time course. On the other hand, if you’re measuring Rs by reading off the amplifier Dial/display, then whatever that value is when your experiment starts, that is your starting value.

          • But giving advice over the internet is hard. When you comment, because I’m the admin, I can see a few details about where in the world you are. And I’ve got to say, there are some EXTREMELY accomplished electrophysiologists near you (or there were last time I was there), so perhaps you should go knock on a few doors. In my experience, when you come looking for advice, people are usually very helpful.

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>