7b0c84ac43d81f2c0d980645e693df8452ce1cb5,nyu_rest_decomposition.py,,,#,13

Before Change

from sklearn.decomposition import FastICA
// X = dataset.func[0][:, :, :, 0]
X_masked_shape = X_masked.shape
X_masked = np.reshape(X_masked.ravel(), (-1, 1))
ica = FastICA()
S_masked = ica.fit(X_masked).transform(X_masked)
S_masked = np.reshape(S_masked.squeeze(), X_masked_shape)
S = np.zeros(X.shape)
S[mask] = S_masked
plt.imshow(S[:, :, 20, 0], interpolation="nearest", cmap=plt.cm.hot)

After Change

// Mask non brain areas
from nisl import mask
X = np.concatenate((
    dataset.func[0],
    dataset.func[1],
    dataset.func[2]),
    axis=3)

mean_img = np.mean(X, axis=3)
mask = mask.compute_mask(mean_img)
X_masked = X[mask]

////// Apply requested algorithm //////////////////////////////////////////////////////////////////////////////////////////////////

plt.axis("off")

from sklearn.decomposition import FastICA
X_masked_shape = X_masked.shape
ica = FastICA(n_components=n_components)
S_masked = ica.fit(X_masked).transform(X_masked)
(x, y, z) = mean_img.shape
S = np.zeros((x, y, z, n_components))
S[mask] = S_masked

// Threshold

Instances