I work with electrophysiology on neuronal cells. I apply to each cell a series of currents, in 20 pA steps, ranging from 0 to 200 pA. At each step the cell produces a number of action potential (AP). I’d like to see if the application of certain drugs (treatments) changes the excitability (number of APs for each current) of the cells. One option would be to use multi-level models, but I'm not totally sure how to do this.
A good explanation of the problem, and with my attempts to solve it, is available in the link.
Is there any way we could compare these curves, between control and different treatments?
http://www.talkstats.com/showthread.php/54986-R-%E2%80%93-lme-function-Comparing-curves-in-longitudinal-data-%28growth-curves%29?highlight=
One idea would be to compare the cumulative number of APs, but this way we would lose information of the excitability shifting to the left or right in the graph.
Another would be comparing the APs at each step (0, 20, 40...) but this would result in a lot of calculation and kind of assuming that at each step the values are independent. So, if anyone has another idea, it would be great to know!
One method is to take the average action potential frequency over the last 25-30% of each step, and than graph those frequency values (y-axis) as a function of the injected current (x-axis). The slope of this "f-I" curve is the gain, and the current level associated with spike onset is the threshold. By taking your measures at the end of each step, you avoid the effects of spike-frequency adaptation (i.e., very high frequency at the onset of each step, with a decay in frequency as the step ensues). However, you can also measure the adaptation rate if you think that the drug would effect it. Good luck!
Hello Rodrigo, I include one of our publications which shows ways to compare average firing frequency changes in neurons as a result of the application of a neuromodulator. It also includes an analysis of spike adaptation. Hope this helps. Good Luck!
You can also compare the exact timing of the spikes before and after drug application. In addition, you can look at the height of the spikes and how that changes during the spikes. For example, a fast decrease in height may indicate that there is more sodium channel inactivation in the presence of drugs.
Hello, I do agree with Paul. According to you posted data, I would average frequency for each binned values of current and fit the curve in condition # 1 to a Boltzmann function that would give you the slope factor k (sensitivity of F to the current) and the current at half frequency I 1/2. You would hence get x values (for x neurons) of k and I 1/2 for control and do the same after drug, then compared parameter with a paired t-tests or Willcoxon.
- Badges
- Science topic