ISIViolations can be imaginary

[RoddyMGrieves]

I think the ISIViolations.m function can still output imaginary values.

From Llobet et al. 2022 (pp 11):
"The equations for C and C' may not have a real solution if nv is large; this
can arise if the activities of reference and contaminant spikes are strongly
correlated, thus causing a departure from the assumption of independent
firing, and/or as a result of stochastic variations. In such cases - almost
invariably removed by a quality threshold - we simply follow the convention
of setting C = 1."

I have observed imaginary values for fpRate, even though the comments at the top of the code say that the expression no longer returns imaginary values.

I suggest adding:

if ~isreal(fpRate)
    fpRate = 1;
end