Reproduce the method from this article: Analysing linear multivariate pattern transformations in neuroimaging data, doi: https://doi.org/10.1101/497180

Steps

1. run step1, to calculate 'GOF', 'RDD', 'RDSV' respectively:

[T, GOF, Y_hat, lambda, density_list, RDD, sv, RDSV] =  step1(Y, X, toplot)

% Input X Y should be in the format of (#voxels, #stimuli)
% toplot = 0: do not plot, 1: plot
% please check the fitting curve are correct.

2. run step2, to obtain the GOF-RDD curve using monte carlo simulation approach, and estimate 'Sparsity':

step2(X, Y, GOF, RDD)

% You need to specify the parameters in the 'step2.m' file:
% 	num_rpt :number of repeats, 1000 for final result, 100-200 if you'd like to check the result in half an hour.
% 	noise_list, sparsity_list: test and find the appropriate range so that our true GOF/RDD can fall into that range.

3. run step3, to obtain the GOF-RDSV curve using monte carlo simulation approach, and estimate 'deformation':

step3(X, Y, GOF, RDSV)

% the same as step2