986aacd1c7ed2f7000cc9816057f96d59701e066,plot_poldrack_space_net.py,,,#,7

Before Change

import os
import sys
sys.path.append(
    os.path.join(os.environ["HOME"], "CODE/FORKED/parietal-python"))
from examples.proximal.load_data import load_gain_poldrack
mem = Memory(cachedir="cache", verbose=3)
X, y, _, mask, affine = mem.cache(load_gain_poldrack)(smooth=0)
img_data = np.zeros(list(mask.shape) + [len(X)])
img_data[mask, :] = X.T

// prepare input data for learner
import nibabel
mask_img = nibabel.Nifti1Image(mask.astype(np.int), affine)
X_train = nibabel.Nifti1Image(img_data, affine)
y_train = y

////// Fit and predict ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
from nilearn.decoding import SpaceNetRegressor
penalty = "TV-L1"
l1_ratio = .3
alpha = None
decoder = SpaceNetRegressor(memory=mem, mask=mask_img, verbose=2,
                            n_jobs=int(os.environ.get("N_JOBS", 1)),
                            l1_ratio=l1_ratio, penalty=penalty, alpha=alpha)
decoder.fit(X_train, y_train)  // fit
coef_niimg = decoder.coef_img_
coef_niimg.to_filename("poldrack_%s(l1_ratio=%g, alpha=%s)_weights.nii" % (
        penalty, l1_ratio, alpha))

After Change

////// Fit and predict ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
from nilearn.decoding import SpaceNetRegressor
penalties = ["Smooth-LASSO", "TV-L1"]
decoders = {}
for penalty in penalties:
    decoder = SpaceNetRegressor(memory=mem, mask=mask_img, verbose=2,
                                penalty=penalty)
    decoder.fit(X_train, y_train)  // fit
    decoders[penalty] = decoder

////// Visualization //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
import matplotlib.pyplot as plt
from nilearn.image import mean_img
from nilearn.plotting import plot_stat_map
background_img = mean_img(X_train)

Instances