def inter_subj_cc_sim(subj1_id, subj2_id, subj_dir):
"""Compute inter-subjects CCs similarity."""
subj1_dir = os.path.join(subj_dir, 'vS%s'%(subj1_id))
subj2_dir = os.path.join(subj_dir, 'vS%s'%(subj2_id))
#-- inter-channel similarity
feat_weights_file1 = os.path.join(subj1_dir, 'plscca',
'layer1', 'feat_weights.npy')
feat_weights_file2 = os.path.join(subj2_dir, 'plscca',
'layer1', 'feat_weights.npy')
feat_cc_corr1 = np.load(feat_cc_corr_file1).reshape(96, 121, 10)
feat_cc_corr2 = np.load(feat_cc_corr_file2).reshape(96, 121, 10)
sim_mtx = np.zeros((960, 960))
for i in range(10):
data1 = feat_cc_corr1[..., i]
for j in range(10):
data2 = feat_cc_corr2[..., j]
tmp = corr2_coef(data1, data2)
sim_mtx[i*96:(i+1)*96, j*96:(j+1)*96] = np.abs(tmp)
np.save('feat_cc_weights_sim_subj_%s_%s.npy'%(subj1_id, subj2_id), sim_mtx)
#-- inter-CC similarity
#feat_cc_corr_file1 = os.path.join(subj1_dir, 'plscca',
# 'layer1', 'feat_cc_corr.npy')
#feat_cc_corr_file2 = os.path.join(subj2_dir, 'plscca',
# 'layer1', 'feat_cc_corr.npy')
#feat_cc_corr1 = np.load(feat_cc_corr_file1).reshape(96, 11, 11, 10)
#feat_cc_corr2 = np.load(feat_cc_corr_file2).reshape(96, 11, 11, 10)
#avg_weights1 = vutil.fweights_top_mean(feat_cc_corr1, 0.2)
#avg_weights2 = vutil.fweights_top_mean(feat_cc_corr2, 0.2)
#sim_mtx = corr2_coef(avg_weights1, avg_weights2)
#np.save('feat_cc_sim_subj_%s_%s.npy'%(subj1_id, subj2_id), sim_mtx)
pass
def plscorr_eval(train_fmri_ts, train_feat_ts, val_fmri_ts, val_feat_ts,
out_dir, mask_file):
"""Compute PLS correlation between brain activity and CNN activation."""
train_feat_ts = train_feat_ts.reshape(-1, train_feat_ts.shape[3]).T
val_feat_ts = val_feat_ts.reshape(-1, val_feat_ts.shape[3]).T
train_fmri_ts = train_fmri_ts.T
val_fmri_ts = val_fmri_ts.T
# Iteration loop for different component number
#for n in range(5, 19):
# print '--- Components number %s ---' %(n)
# plsca = PLSCanonical(n_components=n)
# plsca.fit(train_feat_ts, train_fmri_ts)
# pred_feat_c, pred_fmri_c = plsca.transform(val_feat_ts, val_fmri_ts)
# pred_fmri_ts = plsca.predict(val_feat_ts)
# # calculate correlation coefficient between truth and prediction
# r = corr2_coef(val_fmri_ts.T, pred_fmri_ts.T, mode='pair')
# # get top 20% corrcoef for model evaluation
# vsample = int(np.rint(0.2*len(r)))
# print 'Sample size for evaluation : %s' % (vsample)
# r.sort()
# meanr = np.mean(r[-1*vsample:])
# print 'Mean prediction corrcoef : %s' %(meanr)
# model generation based on optimized CC number
cc_num = 10
plsca = PLSCanonical(n_components=cc_num)
plsca.fit(train_feat_ts, train_fmri_ts)
from sklearn.externals import joblib
joblib.dump(plsca, os.path.join(out_dir, 'plsca_model.pkl'))
plsca = joblib.load(os.path.join(out_dir, 'plsca_model.pkl'))
# calculate correlation coefficient between truth and prediction
pred_fmri_ts = plsca.predict(val_feat_ts)
fmri_pred_r = corr2_coef(val_fmri_ts.T, pred_fmri_ts.T, mode='pair')
mask = vutil.data_swap(mask_file)
vxl_idx = np.nonzero(mask.flatten()==1)[0]
tmp = np.zeros_like(mask.flatten(), dtype=np.float64)
tmp[vxl_idx] = fmri_pred_r
tmp = tmp.reshape(mask.shape)
vutil.save2nifti(tmp, os.path.join(out_dir, 'pred_fmri_r.nii.gz'))
pred_feat_ts = pls_y_pred_x(plsca, val_fmri_ts)
pred_feat_ts = pred_feat_ts.T.reshape(96, 14, 14, 540)
np.save(os.path.join(out_dir, 'pred_feat.npy'), pred_feat_ts)
# get PLS-CCA weights
feat_cc, fmri_cc = plsca.transform(train_feat_ts, train_fmri_ts)
np.save(os.path.join(out_dir, 'feat_cc.npy'), feat_cc)
np.save(os.path.join(out_dir, 'fmri_cc.npy'), fmri_cc)
feat_weight = plsca.x_weights_.reshape(96, 14, 14, cc_num)
#feat_weight = plsca.x_weights_.reshape(96, 11, 11, cc_num)
fmri_weight = plsca.y_weights_
np.save(os.path.join(out_dir, 'feat_weights.npy'), feat_weight)
np.save(os.path.join(out_dir, 'fmri_weights.npy'), fmri_weight)
fmri_orig_ccs = get_pls_components(plsca.y_scores_, plsca.y_loadings_)
np.save(os.path.join(out_dir, 'fmri_orig_ccs.npy'), fmri_orig_ccs)
def channel_sim(feat_file):
"""Compute similarity between each pair of channels."""
feat = np.load(feat_file)
print feat.shape
feat = feat.reshape(96, 55, 55, 540)
simmtx = np.zeros((feat.shape[0], feat.shape[0]))
for i in range(feat.shape[0]):
for j in range(i+1, feat.shape[0]):
print '%s - %s' %(i, j)
x = feat[i, :].reshape(-1, feat.shape[3])
y = feat[j, :].reshape(-1, feat.shape[3])
tmp = corr2_coef(x, y)
tmp = tmp.diagonal()
simmtx[i, j] = tmp.mean()
np.save('sim_mtx.npy', simmtx)
im = plt.imshow(simmtx, interpolation='nearest', cmap=plt.cm.ocean)
plt.colorbar(im)
plt.show()
def random_cross_modal_corr(fmri_ts, feat_ts, voxel_num, iter_num, filename):
"""Generate a random distribution of correlation corfficient."""
corr_mtx = np.memmap(filename,
dtype='float16',
mode='w+',
shape=(voxel_num, iter_num))
print 'Compute cross-modality correlation ...'
fmri_size = fmri_ts.shape[0]
feat_size = feat_ts.shape[0]
# select voxels and features randomly
vxl_idx = np.random.choice(fmri_size, voxel_num, replace=False)
feat_idx = np.random.choice(feat_size, voxel_num, replace=False)
for i in range(voxel_num):
print 'voxel index %s' % (vxl_idx[i])
print 'feature index %s' % (feat_idx[i])
feat_data = feat_ts[feat_idx[i], :].reshape(1, -1)
fmri_data = np.zeros((iter_num, fmri_ts.shape[1]))
for j in range(iter_num):
fmri_data[j, :] = np.random.permutation(fmri_ts[vxl_idx[i]])
corr_mtx[i, :] = corr2_coef(feat_data, fmri_data)
narray = np.array(corr_mtx)
np.save(filename, narray)