In dpwynne/mmnst: Multiscale Modeling of Neural Spike Trains

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

```r

library(ggplot2)

library(Rcpp)

library(mmnst)

```

Introduction

Write something about

Export the cross-vlidation and optimization results and read them in the background (cheating) so that the vignette does not take long to run However, the included code must be as if the code is being run in real time. Just warn the user that these steps may take long.
The theory behind the package
The data to be used
All setwd commands to be dropped because everything will be in the same directory. Maybe make a temp_vignette_results directory so that this vignette writes output to, but make sure to clear that directory at the end of the vignette.
Fix NeuronNumber='05a' argument in ISIplot below, as well as the y-axis label.
Include an Boolean argument in lambda.J.cross.validation function to print or not to print the J value as the funciton is running.

$\;$

Step 0B: Loading the data into R and initial data cleaning

First, load one file with spike train data. This file should be a list of vectors, e.g. list(c(...), c(...),...), each vector being the spike trains of one trial/experiment.
For each train in the file, first get only the positive-time spike times (in case spike times are not in reference to the beginning of the experiment).
Then, any spike trains in the file that have 5 or fewer spikes will be dropped.

SpikeTrainFileName = 'data.LGN.txt' # For batch mode, this line is commented out because spike train file name is created before running this script
trials.to.drop = c()
spikes = dget(paste(SpikeTrainFileName,sep=''))
for(i in 1:length(spikes)){
  temp = spikes[[i]]
  temp = temp[temp>0]
  spikes[[i]] = temp
  if (length(spikes[[i]]) <= 5) trials.to.drop = c(trials.to.drop , i)
}
if(!is.null(trials.to.drop)) spikes=spikes[-trials.to.drop]

length(spikes) #number of trains
hist(unlist(lapply(spikes, length)))