Mao, unstable baseline (run-ups) are pretty common in whole-cell LTP experiments, and there are ways to avoid that as well. This is important because you might have to induce LTP in less than 10 mins of achieving whole cell to avoid 'wash-out'. So, the idea is to record atleast 5 mins of a "stable baseline" before you induce LTP. One solution is to pre-stimulate the afferents before you patch a cell. For this, you can run your baseline protocol for 10 mins before you patch a cell, then patch it to follow up with a baseline recording and the induction.

For AMPAR I-V, you are right- wait till you get a stable baseline, then record 15-20 sweeps of -70mV, then change the clamp voltage to -60mV and so forth in 10 mV intervals until +50mV. Then bring it down to -70mV to make sure the amplitude of the AMPAR current hasn't changed from it original size. Include GABAA and NMDAR antagonists when you are doing this.

Jai, thank you for your help.

It is great.

I have another question. Do you do Input-output before LTP experiment to see the amplitude of AMPAR current? What's the amplitude of stimulus intensity when you do I-O? And also, record several sweep of the fixed stimulus intensity and next

Maybe you have a good reason of doing the WC LTP that I am not aware of but in my opinion, doing LTP in Whole-cell, might only be good to get a quick preliminary result. I think you should seriously think about doing LTP using filed potentials instead. This will allow you to do reasonable Long duration experiments (> 90 minutes) and not worry about the overall health of your cell when you patch for more than 30 minutes or so and I am not talking about the obvious dialysis of important components of the cell that do inevitably happen in the first 5 minutes or less.

To answer your question about when to do the I/O? Yes, you do that before you start your baseline. The amplitude of the stimulus will depend on your preparation (usually few volts suffice to see a response) then increase your stimulation with adequate incrementation until you see some saturation of the response.

There are valid reasons for using whole-cell recording to study LTP, and there are also reasons to avoid it. The well-known (and discussed) dialysis of some LTP-critical factor is one reason to avoid, as is the tendency of AMPA currents to run down. I tend to avoid using whole cell (and we use it a lot) when I don't need to use it. Field potentials are another approach with advantages, but also with limitations, mostly that you are always looking at a population response, you can't record the membrane potential/current of individual neurons, and you also have no access to the intracellular space for injecting things like drugs. We tend to arrange our experiments in a technique hierarchy. We use field potentials for those things we can, and to give a broad overview of what is occurring in the variables we are studying. We use sharp microelectrode intracellular recording as our next technique or choice, for the things that you can use it for. Lastly, we go to whole cell, when we can't study what we want with Fields of Sharps. The reasons that one has to resort to whole-cell are typically: 1) You need to voltage clamp, 2) You are looking at voltages or currents that are too small to be detected in sharp-electrode recording, or 3) you are trying to inject something large into the cell. As far as your rundown problem - we see that too - we generally solve it with fresh internal solution, and that contains fresh ATP and GTP.

Someone also mentioned input/output curves. I'm not saying they are not a good idea - they can be very useful, but in LTP studies, people tend to do them for the wrong reason. I have from time-to-time seen people use the I/O to set the size of their baseline EPSP on the rationale that they want to set it 'below saturation' for getting subsequent LTP. This is wrong. You will, if anything, get MORE LTP if you set your baseline near the top of your I/O curve. Using this 'saturation' rationale indicates nothing so much as that the person has not completely thought through what it means to turn the stimulus up (hint:, you are increasing the *number* of presynaptic axons activated, not stimulating synapses at the ends of those axons *harder*).

Professor Madison, thank you very much for your suggestion. But why seting the size of baseline EPSP near the top of I/V curve will get more LTP?

can some body provide me the details how you are recording and analyzing the LTP, even how you make the protocol for the recording LTP, after going the whole cell. we are recording, but We could not able to get the proper recording, also give the external and internal solution.