Hello everyone,
I was using CV of EPSCs amplitudes to assess neurotransmission variability and synaptic efficacy. But when I looked at some articles in the litterature they were using CV² and 1/CV². Then, I wondered why using CV² and 1/CV² rather than just CV? What are the conditions for using CV² and 1/CV²? What information do these two measures add that CV does not?
Thank you everyone for your reply.
Basically, reason for using CV2 or 1/ CV2 is that
m ~1/(CV2)
where m - “quantal content” = mean number of quanta that are released to make up postsynaptic potential (current), ~ means proportional.
Given that PSC amplitude is changed by some modulator and there is proportional change of 1/(CV2), there are reasons to think that the effect, at least in part, is presynaptic. (Actually things are more complicated, but the above mentioned is still a reasonable approximation).
In any case, converting CV to 1/ CV2 should not be a problem.
I would say the reason for using 1/CV^2 is because this quantity is proportional to p in the binomial release model. In other words, you can read off normalized p on the y-axis when you plot 1/CV^2 vs the mean. If 1/CV^2 (normalized p) is above the diagonal unity line, you may conclude that LTP was expressed by a change in probability of release presynaptically, otherwise it was expressed postsynaptically (assuming the binomial model holds). Conversely, if normalized p is below the diagonal, LTD is expressed by a change in p, otherwise likely not.
The thing is, you can have a change in CV that does not imply a change in p. By plotting 1/CV^2 vs µ, you now have a criterion for knowing when p changed. Again, assuming the binomial model holds, which may or may not be true.
For details, such as mathematical derivations, see:
Brock JA, Thomazeau A, Watanabe A, Li SSY, & Sjöström PJ: A Practical Guide to Using CV Analysis for Determining the Locus of Synaptic Plasticity. Frontiers in Synaptic Neuroscience (2020) 12:11 DOI: 10.3389/fnsyn.2020.00011
Per Jesper Sjöström Thank you very much for your reply! (I read and appreciated your article while searching the CV analysis literature)
In the meantime, I read a chapter by Zucker et al, 1999, 7.Release of neurotransmitters, in. Fundamental Neuroscience (a very classic book that I read as a student and recently found on my computer). This could be complementary to your answer. Let me know.
By plotting the 1/CV² ratio (ordinates) against the average amplitude of the PSCs ratio (abscissa) during recordings that increase “synaptic strength” (i.e. postsynaptic PSCs, e.g. during a stimulation train), we can appreciate the following hypotheses:
- Horizontal relationship corresponds to a change in quantum amplitude (Q, e.g. the postsynaptic increase in the number of receptors).
- For a Poisson model, the identity line relationship (i.e. the 45° straight line) reveals a change in m (quantum content).
- For a binomial model, the identity line relationship reveals a change in n (number of putative or active release sites).
- For a binomial model, if the relationship is greater or less than the identity line, it reveals a change in p.
Please let me know if I'm wrong, and thanks again for your reply!
- Badges
- Science topic