Matlab code for F-Test for Point Processes???

[Dylan Hirsch-Shell]

I'm trying to analyze some spike data recorded from semicircular canal
afferents during purely sinusoidal rotations of an animal's head. I
need a simple, objective way of assessing whether there is a response,
and I'd like to use the F-Test method for assessing the statistical
significance of spectral content in the point process data at the
frequency of stimulation, as described by Jarvis and Mitra in their
2001 Neural Computation paper. I see that the Chronux toolbox has a
function, ftestc, that implements the F-Test on continuous data, but
there doesn't seem to be a corresponding function for point process
data. Has anyone implemented this yet? I'd hate to re-invent the
wheel, and frankly I don't even know if I'd be able to if I tried.

Thanks!
Dylan Hirsch-Shell
Doctoral Candidate in Neuroscience, UCLA

[Dylan Hirsch-Shell]

Partha Mitra has responded to me personally and suggested that I
should be able to simply smear the point process in time slightly and
then use the function for continuous data in Chronux.

I tried this by convolving the spike train with a Gaussian filter
before passing it into ftestc, and it seems to work pretty well on a
trial where I can see a clear response. Now, my only concern is that
the choice I make for the sampling rate and the filter bandwidth might
contaminate the result and give me a false negative on a trial where I
can't rely on my eyes to tell me whether there was any response. Does
anyone have any intuition on how these parameters might affect things?