I calculated the membrane time constant of neurons from the whole-cell capacitance and the input resistance, both values I got from voltage-clamp recordings (application of a pulse of -10 mV from Vh). I wondering, because when I estimated the membrane time constant directly in current clamp recordings by fitting in the voltage after injection of a negative current, the values where different. In VC recordings the membrane time constant was 2x higher (app. 35 ms) compared to the estimation in CC recordings (app. 16 ms). One possible explanation could be that an additional conductance is activated under CC recordings. What do you think, what is the better way to get the membrane time constant from neurons?
Something is wrong. Those should be the same since the time constant is a property of the cell. Is the amplifier a patch clamp amplifier? Could be poor at CCing but difference shouldn't be that big.
Did IR change? How was capacitance measured? Poor electrode compensation in VC?
Actually, your time constant should be slower in current clamp than in voltage clamp. That's not really correct either - the time course of current should be faster than the time constant of voltage. The capacitive current is the simple charging of the capacitor by the injected current. The voltage of the membrane will typically lag behind that.
@Oliver: I agree, from VC (Tau(m)=Cm*RIN) as well as from CC recordings we should get the same Tau(m) (CC fitted directly). I use definitely a patch amplifier - EPC10/2, Raccess was stabil, Cm was calculated from capacitive currents manually (area, not deacy time).
@Damiel: its true in my case, too, that in VC the time measuered constant is longer compared to CC recordings, when I fitted the current/voltage.
I assume, that an additional conductance in CC contaminates my calculations of Tau(m) in CC, I have to do more recordings to find the bug. Under which conditions it is usefuel to calculte Tau(m) in CC, I used Vh=-80 mV (set by current injection) hyperpolarizing currents of 200 pA.
Rene,
Last time I calculated Cm manually by integrating the charge under the transient in VC, I was not getting the right capacitance because I was integrating from the beginning of the transient (including rise time). If I integrated to calculate the charge from the peak of the transient (and this is why the electrode capacitance has to be spot on) the numbers were more accurate. Also, integration was done with the relaxed current at zero (fully decayed), so not to include the added measure caused by the input resistance.
Hope this makes sense.
Rene,
Please be sure not to use slow-capacitance compensation during your measurement. In this case theoretically Cm must be the same by two methods.
thanks, for Cm (without Cm-slow compensation, otherwise Cm should be zero, good tip for beginners) I take the area below the capacitive currents, but I get only negligible differences if I include the area below the rise tise area or not. Calculation of Cm by using the decay time results in more variable results, because sometimes fitting is not the best and sometime I have a tendency for bi-exponetial curves. I do not belive that the Cm calculation is the reason for differences in Tau.
@ Oliver: I tried it on Purkinje cell in slices at postnatal day 15 were Cm is around 200-400 pF, here the bi-exponetial fit becomes clear, and the area from below the rise time is negligible.
What I am saying is perhaps you should try these measures on a fake cell that has no active currents and single exponential decays to see if you methods give you similar numbers in VC and CC.
When you do measurements of capacitative currents (to get time constants) make sure to also turn down your filter settings to a minimum. You do not want to filter out the peak of the capacitative current with too much filtering. This peak value is higher the lower your series resistance.
@Henrique: thanks, for capacitance recordings I used the minimal filter settings, and in addition I applied the highest possible AQ frequency, therefore I assume, this can not be be the reason for the discrepance between VC and CC calculations.
Is it possible that your cells have an V-dependent inward rectifier that would be activated in CC but you have not seen it in VC?
Hi Rene,
I think that it may be useful to apply a long depolarizing/hyperpolazing pulse near resting membrane potential (in first CC and then in VC). Then measure the currents or potentials at the late period after conductances have decayed. You can then get input resistance at steady-state conditions. Near the resting potential there may be many potassium, chloride and I(h) currents, and all of these may be voltage-dependent. The input resistance will also be affected by the putative presence of open gap junctions in your cells.
@Matti: Thanks, do you know, why the inward rectifier may be activated in CC, but not in VC?
@Henrique: Thanks, gap junctions are a good idea. The suggested recording conditions are similar to the conditions I use.
About the possible inward rectifier: the CC and VC situations are so different that I cannot say, why the IR-current may not be seen. One possible explanation possibility is that it can be small, if your V-hold is near its reversal potential. Thus, if the conductance is significant, however, you might be able to see its effect in CC. Other possibilities may exist, I'm sure.
You should also have a look to this article : "What we talk about when we talk about capacitance measured with the voltage-clamp step method
Adam L. Taylor J comput Neuroscience.
The problem in VC is that you have to be very "clean". First fully compensate the capacitance of the pipette (by using bee wax or sylgard or dental wax coating), use a very high sampling rate (40 KHz) during the +/- 5 mv pulse. Compare the pulse with the different polarities for potential active currents. Calculate the charge (Q) that will give the capacitance of the cell, then calculate the series resistance (by measuring the extrapolated peak to time zero of the "instantaneous current" (Rs= U/Ipeak). Calculate tau : tau =( Rs*Rm/RS+Rm)*C but if Rs
Does anyone have a solution for measuring membrane tau in CC when there are both H-currents and T-type rebound currents with a hyperpolarizing step (plus low depolarizing firing threshold)?
Hello Philip,
You should try to use small negative current injection that does not activate these currents; I would guess that a 5 mV (or less 2 mV) hyperpolarization may be enough. Do it 50 times and average the sweeps.
Hi Rene,
I am going back to your old question. Yes, I agree with Fabien.
It is a fundamental issue the membrane time constant and the CC and VC differences. Please take a look at this cool paper: "Membrane Capacitance Measurements Revisited: Dependence of Capacitance
Value on Measurement Method in Nonisopotential Neurons" J Neurophysiol 102: 2161–2175, 2009. by Jorge Golowasch et al. doi:10.1152/jn.00160.2009.
Its points out that neurons are no spheres so one needs at least 2 exponentials to fit the membrane decay in CC and VC because of the soma, dendrites and axon compartments. So using a single exponential fit with one time constant is just a first approximation.