Step 4c - Visualize the binned gradients

Write out of individual Nifti files for each percentile bin of the primary gradient

written by R. Farahibozorg for the Autism Gradients project at Brainhack Cambridge 2017

In [1]:
# first import the input list from the csv file
import pandas as pd
# read in csv
df_phen = pd.read_csv('./data/SelectedSubjects.csv')
selected = list(df_phen.filename_npy)

In [2]:
import matplotlib.pylab as plt
import nilearn
import nilearn.plotting
import os
import numpy as np
import nibabel as nib
##create path to save the output
out_pathbase='./data/Outputs/Bins/'
if not os.path.exists(out_pathbase):
    os.makedirs(out_pathbase)

In [3]:
def rebuild_nii_indv_bins(num,subs,bins):
    thisfile=[selected[ii] for ii in subs]
    
    for sub in thisfile:
        print(sub)
        data = np.load('./data/Outputs/Regs/%s' % sub)
        a = data[:,num].copy()
        steps=int((1/float(bins))*len(a))
        data_argsort=np.argsort(a)[::-1]
        nim = nib.load('./ROIs_Mask/cc400_roi_atlas.nii')
        imdat=nim.get_data().astype('float')
        for thisperc in range(bins):            
            #print a
            #print data_argsort
            abin=np.zeros(a.shape)#a.copy()
            abin[data_argsort[thisperc*steps:(thisperc+1)*steps+1]]=a[data_argsort[thisperc*steps:(thisperc+1)*steps+1]]            
            imdat_new = imdat.copy()
            for n, i in enumerate(np.unique(imdat)):
                if i != 0 and i < 392:
                    imdat_new[imdat == i] = abin[n-1]
                elif i >= 392:
                    imdat_new[imdat == i] = np.nan
            nim_out = nib.Nifti1Image(imdat_new, nim.get_affine(), nim.get_header())
            nim_out.set_data_dtype('float32')
            
            out_path=out_pathbase + '/' + str(thisperc) + '/'
            if not os.path.exists(out_path):
                os.makedirs(out_path)
            
            out_name=out_path+'Bin'+str(thisperc)+'_'+sub+'.nii'
            #print out_name
            # to save:
            nim_out.to_filename(out_name)            
            #nilearn.plotting.plot_epi(nim_out, cut_coords=(0,0,0),colorbar=True)

In [4]:
plt.close("all")
for i in range(1):
    subs=list(range(len(selected)))
    numbins=10
    nims = rebuild_nii_indv_bins(i,subs,numbins)
    #plt.show()


Pitt_0050009_rois_cc400.1D.npy
Pitt_0050010_rois_cc400.1D.npy