Hi Moore. I dont have a huge amount of experience with this, but I noticed this can be due to a dirty stimulating electrode, or differences in the depth of placement of the stimulating electrode. You can vary the position once your slice is settled in the chamber and ensure you give it a wipe with sandpaper between slices and /or use a very high reverse polarity series of stimuli into the ACSF above the slice to dislodge any accumulated ions. To get around this, it is possible to analyse the slope of your IO curves, but with practice they should be reasonably similar. Good luck.

Hi Moore,

Before finding an explanation, it is important to have a description as detailed as possible of your experimental protocol.

Thank you for your help. There are some questions.

Do you think the distance between the stimulate electrode and recording electrode is important? To do this experiment, need I to use the same stimulate electrode?

Thank you.

There is no way to obtain the same response with the same stimulus intensity from slice to slice and from animal to animal. The usual way is to use a stimulus intensity that gives 50-75% of the maximum response.

If your fEPSP is not stable or too variable during a given experiment, I would need to know your exact protocol, species, age, perfusion rate, brain region etc

Hi Moore

It is not possible to get the same results from different slices. Don't worry!

If you like to compare the input-output (I/O) curves of different groups with each other, you have to use "arbitrary stimulus intensity" in your I/O curves by the following manner:

In each slice, measure the I/O curve by using at least 5 different stimuli. The weakest intensity must result to a weak response and the strongest one must result to the maximum response. It would be very nice if you can select the intensities so that the outputs changes by the same slope (I mean, it's better to be a linear curve). Then, you have to determine the I/O curve of the next slice by the same manner and so on. In this manner, the intensities used for plotting I/O curves are not exactly the same. Thus, you will consider arbitrary intensities in your I/O curve. For example, in the above mentioned experiment, you will consider the intensity of the weakest stimulus in all of your slices as "1" and the intensity of the strongest stimulus equal to "5". The intensity of "1" may be equal to 50 micro-ampere" in one slice, 80 micro-ampere in another one and so on. Similarly, "5" may be 700 micro-ampere in one slice, 950 micro-ampere in another one and so on. Other stimuli (between the weakest and strongest) will also be considered as "2", "3" and "4". Therefore, if you plot I/O curves in this manner, you can compare them in different groups. It is also necessary to explain in the legend of this kind of plots that each arbitrary intensity is equal to which range of real intensities.

I hope, my explanation be helpful for you.

If you also search some articles in which the I/O curves have been compared, you can easily find more information in this regards.

Best wishes, and have a nice New Year!

I have done a lot of I/O curves over the years, and I am still learning how to do them in an effective way. There is always going to be slice to slice variation and animal to animal variation. As with anything, the trick is to determine what variation is random or due to confounding factors, and what is an actual effect on a dependent variable (if any). As a rule, try to minimize any variation in experimental factors that might give you stray variation in the result (temperature, slice thickness, distance between electrodes, etc.). So I use slices of similar age (since cutting), fix temperature, always place the electrodes similar distance apart, try to use the same stimulation electrode across a series of experiments, keep that electrode clean of tissue buildup, always adjust the depth of the recording electrode to give maximum response, never take an I/O on a non-stable response, etc. To minimize trial to trial variation, deliver multiple stimuli at each strength so they can be averaged (I typically take 5 samples at each stimulus strength). In the end, the most effective and powerful thing you can do is to do a lot of experiments. There is no substitute for a high 'N'. This will average out the random and confounding variability, and should leave your 'effect' intact. How high the N should be depends, of course, on the size of any effect relative to your background variation. But in general, an N of 15-20 (slices, animals, whatever level at which your N is properly placed) seems to work fairly well. I also 'put a lot of points' on the curve, especially where it is changing quickly. If you know you will want to apply ANOVA statistics to your finished result, you need to reproduce the stimuli used the exact same way every time (i.e. Use the same stimulus strengths). Doing so so will not compromise your ability to normalize curves across experiments, if that is your desire (although IMHO, it is difficult to see how this would improve your interpretation of anything except differences in the slope of the curve). Just FYI, this is our typical stimulus array for an I/O curve: (in mAmps): 0.01, 0.025, 0.05, 0.075, 0.1, 0.125, 0.15, 0.175, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6, 0.65, 0.7, 0.8, 0.9.

After electrode placement in the slices, I generally wait for about 10-15 min (sometimes even longer up to 30 mins) before actually beginning with the stimulation. This helps me a lot, and most of the time I get a stable response when I begin with the stimulation. Also, I usually take 3 values of the slope for each stimulation intensity, and average them. There are some slice to slice variations, because as mentioned previously, each slice is different, and has its own excitability characteristics, but they are not extremely different from each other and fall in a particular range. Another thing that I follow is to first give a small stimulating current sufficient enough to generate a response; the response should be a good and stable one- if the response is not good and stable for about 3-5 minutes at this stage, I usually don't use that slice any further. If the response is stable, I go ahead with the I/O curves. The important criteria is to let the slice stabilize properly. A stable healthy slice will then respond well to stimulation and can give you good results. Good luck.

have to make average from 10-20 fEPSP in each intensity and comparing normalized values.

I agree with many of the comments posted. Daniel Madison is right the larger the number you can get the greater the confidence in the data. Stable recording conditions are important including those already mentioned temperature and depth of solution (this can reduce size of artefact). I have tended to use the same simulating electrode. I also made sure all positioning of recording electrodes was similar relative to the anatomy. Anyhow the result is a function of many factors and I would only take the data seriously if there was a big difference between test and controls.

How you normalized data to make a graph

To normalize you divide the measurements to the maximum response. I find that plotting the fEPSP (slope or max rate of rise) as a function of fiber volley amplitude or integral is more interesting than plotting it vs stimulus intensity.