We're currently investigating instances when response times spike, as our max times are orders of magnitude different than our average times. Knowing this information would allow us to better speculate as to root causes of these spikes so we can better mitigate them.
Is it possible to learn when these spikes happen?
Yes, you would want to use a Graph to plot the counters of the Response time for the specific element(s). Keep in mind, the Max Response time might be only an aberration, and not a major occurrence. If it does not happen often or is isolated, then you might not have an issue. You want to make sure the Average Response Duration for the specific request or page request is not high or spike for long periods of time.
You can go to the Graphs section, select the Element(s) you want to graph (expand tree if it is a specific request or page), select Average Response Duration, and drag that to an open graph spot on the right.
You will be able to see the Average Duration for that element, and see when the spikes might have happened during the test.
Here are some interesting links that provide more reference: