Let's do some back of the envelope maths.

Your induced current looks like a potassium current. So lets say, Erev = -110 mV. You are holding at -50mV, and you get an induced current of 20 pA.

So with a driving force of 60 mV, you get 20pA. If the current is linear, (which is appears to roughly be) then with a delta of 10 mV driving force, you should expect a delta of 3 pA. I dont think you have enough resolution (due to noise) to pick that out of a single trace.

Let's double check this rough math.

I = 1/R*(Vm-Ve), so your drug removed current behaves like a shunt with a resistance of 3GOhm (in parellel with your membrane resistor)

Your Rm is about 500 Meg, so if you remove a 3gig resistor in parellel what should happen to the apparent membrane resistance? 500Meg = (R1*3Gig)/(R1+3Gig), R1 = 600Meg. You can see the estimate of your membrane resistance is jumping around by close to 100 Meg, so it wouldn't be hard for a 100 Meg change in Rm to hide either.

So let me ask you a question? How are you calculating Rm? Are you averaging your current over time during the pulse? Or just picking a single point? You'd be wise to average over the last 50ms or so.

But ultimately, you're looking for a pretty small shift. Perhaps try large voltage steps, or average 3 steps in a row, to try to increase you resolution/decrease noise.

May be you can try doing a current clamp experiment with a constant current injection to hold the membrane potential at - 60 mV for both the cases ie baseline and drug. Now run a series of step current (maybe -25 to +25 pA with 5pA step size, given that the baselinr Rm is ~ 500 Mohm). Now get the voltage responses for them. Fit a linear regression for all the V-I plots to see the slope as the Rin of the plots. Just make sure that your membrane potential for all the cases ie drugged or not remains nearly same (maybe -60 mV) else the driving current for the causal conducatnce will change and might not show up in the Rm estimate.

The shift is very (too) small and will be covered by the noise.

If the holding current changes and the membrane resistance stays constant I would look whether any membrane ion pump could be involved in this. However, I would also check if there is a washout of the drug's effect to test any possible artifacts. Cheers, Paolo

Hummmm I am not sure that it is correct to compare membrane resistance measured in a Ramp experiment and that obtained by steps. During a ramp the change of Vm  produces a dynamics of channel far different than that given by a step. Some channel if opened at -20 are inactivated when membrane voltage in the ramp arrive to zero for example. This is why IV in ramp and step protocols are never the same  and much depends on the slope of the ramp :)

As the shift is very small it is advised to do the constant current injection. did you saw any wash effects of the drugs. As William suggested I also agreed with that.  you should inject the large step with average value otherwise it is difficult to reach in conclusion. I could able to see the noise level is more. so try to reduce the noise also 

Your discussion of the results are legitimate. From the figure though it is clear you measure the membrane resistance at the voltage (-50 mV) away from the RMP (resting membrane potential).  Try to measure the Rm at a holding voltage close to the RMP (which is  between -60 mV and -65 mV in the cell in Fig. E). 

I agree with William's answer although with a larger negative voltage step the difference between no drug vs drug (Fig A) is relatively smaller (e.g.compare difference between -100mV currents vs -20mV currents), maybe due to the mild outward rectification. So you might find it doesn't improve things but worth a shot. Perhaps a large positive voltage step is your answer but at those potentials the slope is not linear but exponential.

The idea of using current clamp is probably worth a try with hyperpolarizing current steps. As Debanjan indicated you need to make sure that you use the constant current injection to correct the resting Vm to the pre-drug value since inhibiting a leak K conductance will depolarize the cell. As an alternate to plotting I/V values to get the the slope (Rm) you can calculate the Rm for each current step (e.g. -100pA, -90A, -80pA etc) and plot the Rm against the current step value. This will give you a range of Rms for control vs drug.

Good luck.

Hi Scott!

(1) If you combine Fig A and Fig B, you will get I-V curves, with steeper dI/dV (e.i., larger total membrane conductance) in the base line compare to the drug. Specifically, just looking at the Fig A, I would say approximately g_base = 400pA/100mV = 4nS, which corresponds to 250 MOhm. While the drug reduces membrane conductance by about 40pA/400pA*100% = 10%, which translates to about 25 MOhm decrease in the resistance.

(2) Now look at Fig G.  When you average all Rm before and after the drug application, you will get approximately 590 MOhm. But, when you average Rm values during the hyper-polarizing voltage pulse, the average value will be 530 MOhm, which is also about 10% smaller (as in A).

(3) The differences in my estimations in (1) and (2) are due to non-linear I-V curve with smallest slope at around -50 mV, where you measured Rm in Figs D, E, and G. The effect of the drug on the resistance would be more visible, where the drug induced current in Fig C has strongest dependence on voltage (maximal negative slope). Perhaps, ii would measuring of Rm in the the holding potential range from -35 to -45 mV, or from -85 to -100 mV. 

Thus, all your experimental observations are in complete harmony!

Anatoli

Perhaps, there are some (small) changes of Rm (see the attached figure). Here (middle panel) I graphically overlapped your original traces.

I would try to:

1) subtract baselines from ‘control’ ,‘drug’ and ‘wash’ traces 

2) average ~5 traces  for each codition (immediately before, during drug application and wash)

3) calculate Rm for averaged traces to see whether the difference is systematic.

On the other hand, as judged from the ‘ramp’ experiment, quite large difference of currents should be expected  in ‘square-pulse’ experiment at –60 mV. The mismatch may be explained (for instance) by voltage-dependent inactivation of ‘difference’ current. Indeed,  in square-pulse experiment Vh is –50 mV, while in ramp experiment (initial) Vh  is more negative.

What is concern to the essence of the question, I totally and absolutely agree with the comments of Dr. William M Connelly and can hardly add something significant to his considerations and maths. One terminological question (non-important). Is it correct to call it a holding current in case of voltage-clamp measurements? I'm just curious.

 Hi! The method to calculate Rm using short steps is used only if you suspect possible significant changes in cellular Ion concentrations dew to  your measurements. In your case (the whole cell configuration)  ΔRm can be determined with reasonable accuracy directly from your measurements at holding potential E and the  ΔI:

 ΔR=E/ΔI.

If you have 20 pA changes at 50 mV, that is ΔR=0.05/0.00000002=2.5MOm

So, you have new Rm that is now 2.5 MOm more  then before

And I completely agree with Dr. Connelly that it will be difficult to measure this small change with short steps method at ΔE=10 mV, and unnecessary to my opinion. 

Hi Scott,

It is true that you should increase the signal(effect)/noise ratio for these experiments. But in any case, another way that you can try to detect changes in the membrane resistance which could be a bit more illustrative and obvious is by giving the steps at a constant frequency during the whole experiment. Moreover, if your drug effect increases at more depolarized potentials, I would try to check these changes at -30 or -20 mV for example, and as already suggested, increase the step to have more "room" for change.

So, summarizing, you could record your current at -30 mV, and start giving small (50 ms) pulses to -45 mV (or -50 mV) at low but constant rate (for example, 0.5 Hz). Then, apply your drug and see whether the step size changes along with the change in current. Once the experiment is finished, you can calculate the Rm change from the average of many of these steps before and after drug addition and get an idea of the changes in conductance (1/R).

For your experiments, it would be interesting also to show the effect of the drug on the resting membrane potential, specially if you are putatively dealing with leakage potassium currents.

I attached a figure with two of our experiments done in this way. In this case, the drug opens potassium channels. (A) Voltage-clamp to measure the change in current simultaneously with the pulses to assess the change in conductivity; and (B) current-clamp (I=0) to check how the drug affects the resting membrane potential.

Hi Scott,

You're trying to calculate the membrane resistance based on the assumption that you have only potassium conductance activated. Looking at your traces, this is apparently not the case, as the current magnitude at the holding voltage of -50 mV and that in response to a step of -10 mV are roughly the same size, and the "step" current is not changed in drug, when the holding current does. If only drug sensitive conductance is there, it doesn't make sense, as in such case the step current should be about 5 times smaller at least.  In my opinion, you have another conductance active that is comparable in size with you potassium drug-sensitive one, i.e. a shunt, hence you sensitivity is low as suggested by Dr. William M Connelly and others. Find out what other conductance could be and a way to reduce it (i.e. channel blockers) or better use the current clamp as suggested  Diego Fernández-Fernández if you are interested in membrane resistance. It is much more sensitive approach than the voltage-clamp.