tt: Time Transformation Using a gssanova Objet
In STAR: Spike Train Analysis with R

Description Usage Arguments Details Value Author(s) References See Also Examples

Performs time transformation using a gssanova fit. If the model is correct, the result of the transformation should be a Poisson process with rate 1.

1	gssObj %tt% dataFrame

`gssObj`	a `gssanova` or a `gssanova0` object.
`dataFrame`	a `data.frame` with variables corresponding to the ones used in the `gssanova` call giving rise to `gssObj`.

The binary operator applies predict.ssanova with the left side as the first argument and the right side as the second argument. The right side (dataFrame) must therefore contain the variables included in the formula used in the call giving rise to gssObj. The result of the predict method call is then transformed with an inverse logistic function or with an exponential (depending on the family argument, "binomial" or "poisson", used in the previous gssanova call). The cumulative sum is computed, that is, the integrated conditional intensity, and its value at the events times is returned as a CountingProcessSamplePath object.

A CountingProcessSamplePath object.

Christophe Pouzat christophe.pouzat@gmail.com

Gu C. (2002) Smoothing Spline ANOVA Models. Springer.

Brillinger, D. R. (1988) Maximum likelihood analysis of spike trains of interacting nerve cells. Biol. Cybern. 59: 189–200.

Brown, E. N., Barbieri, R., Ventura, V., Kass, R. E. and Frank, L. M. (2002) The time-rescaling theorem and its application to neural spike train data analysis. Neural Computation 14: 325-346.

Ogata, Yosihiko (1988) Statistical Models for Earthquake Occurrences and Residual Analysis for Point Processes. Journal of the American Statistical Association 83: 9-27.

gssanova, predict.ssanova, mkGLMdf, mkCPSP, summary.CountingProcessSamplePath

## Not run: 
## load e060517spont data set
data(e060517spont)
## make a data frame using a 2 ms bin width
e060517spontDF <- mkGLMdf(e060517spont,0.002,0,60)
## Keep data relevant to neuron 3
e060517spontDFn3 <- e060517spontDF[e060517spontDF$neuron == "3",]
## Split data in an "early" and a "late" part
e060517spontDFn3e <- e060517spontDFn3[e060517spontDFn3$time <= 30,]
e060517spontDFn3l <- e060517spontDFn3[e060517spontDFn3$time > 30,]
## fit the late part with a nonparametric renewal model
e060517spontDFn3lGF <- gssanova(event ~ lN.3, data=e060517spontDFn3l,family="binomial")
## transform the time of the early part
e060517spont.n3e.tt <- e060517spontDFn3lGF %tt% e060517spontDFn3e
## Test the goodness of fit
e060517spont.n3e.tt
summary(e060517spont.n3e.tt)
plot(summary(e060517spont.n3e.tt))

## End(Not run)