mimosa_training: Train MIMoSA model on full training set
In neuroconductor/mimosa: 'MIMoSA': A Method for Inter-Modal Segmentation Analysis

This function trains the MIMoSA model from the data frames produced by mimosa_data on all subjects and determines optimal threshold based on training data

mimosa_training(
  brain_mask,
  FLAIR,
  T1,
  T2 = NULL,
  PD = NULL,
  tissue = FALSE,
  gold_standard,
  normalize = "no",
  slices = NULL,
  orientation = c("axial", "coronal", "sagittal"),
  cores = 1,
  verbose = TRUE,
  outdir = NULL,
  optimal_threshold = NULL
)

`brain_mask`	vector of full path to brain mask
`FLAIR`	vector of full path to FLAIR
`T1`	vector of full path to T1
`T2`	vector of full path to T2 if available. If not use NULL.
`PD`	vector of full path to PD if available. If not use NULL.
`tissue`	is a logical value that determines whether the brain mask is a full brain mask or tissue mask (excludes CSF), should be FALSE unless you provide the tissue mask as the brain_mask object
`gold_standard`	vector of full path to Gold standard segmentations. Typically manually segmented images.
`normalize`	is 'no' by default and will not perform any normalization on data. To normalize data specify 'Z' for z-score normalization or 'WS' for WhiteStripe normalization
`slices`	vector of desired slices to train on, if NULL then train over the entire brain mask
`orientation`	string value telling which orientation the training slices are specified in, can take the values of "axial", "sagittal", or "coronal"
`cores`	numeric indicating the number of cores to be used (no more than 4 is useful for this software implementation)
`verbose`	logical indicating printing diagnostic output
`outdir`	vector of paths/IDs to be pasted to objects that will be saved. NULL if objects are not to be saved
`optimal_threshold`	NULL. To run algorithm provide vector of thresholds