def get_val(sessions=None, masked=True, repeats=False):
"""Retrieves training data for given sessions.
Default: all sessions, mask applied"""
if sessions is None:
sessions = range(3)
if repeats:
if isinstance(masked, bool) and masked:
return tables.openFile(val_file_repeats).getNode("/alldata").read()
elif isinstance(masked, np.ndarray):
return tables.openFile(val_file_repeats).getNode("/alldata").read().reshape(270, 30, 10, 100, 100).transpose(0, 1, 3, 4, 2)[:, masked, :]
else:
raise Exception("Repeats data is masked")
else:
if isinstance(masked, bool) and masked:
#mask = get_mask()
mask = cortex.get_cortical_mask("MLfs",
"20121210ML_auto1", "thick")
return np.concatenate(
[tables.openFile(t).getNode('/data').read()[:, mask]
for t in [val_files[i] for i in sessions]])
elif isinstance(masked, np.ndarray):
mask = masked
return np.concatenate(
[tables.openFile(t).getNode('/data').read()[:, mask]
for t in [val_files[i] for i in sessions]])
else:
raise NotImplementedError("This will exceed 4G of RAM")
def test_volumedata_copy_with_custom_mask():
mask = cortex.get_cortical_mask(subj, xfmname, "thick")
mask[16] = True
nmask = mask.sum()
data = np.random.randn(nmask)
v = cortex.Volume(data, subj, xfmname, mask=mask)
vc = v.copy(v.data)
assert np.allclose(v.data, vc.data)
def test_volumedata_copy_with_custom_mask():
mask = cortex.get_cortical_mask(subj, xfmname, "thick")
mask[16] = True
nmask = mask.sum()
data = np.random.randn(nmask)
v = cortex.Volume(data, subj, xfmname, mask=mask)
vc = v.copy(v.data)
assert np.allclose(v.data, vc.data)
def create_cortical_mask(subject, xfmname, type, epi, temp_dir_cortex):
import cortex
import nibabel as nb
import os#, tempfile
#tempdir = tempfile.gettempdir()
try:
os.makedirs(temp_dir_cortex)
except OSError:
pass
out_file = os.path.join(temp_dir_cortex, 'cortex_%s.nii.gz'%type)
mask = cortex.get_cortical_mask(subject, xfmname, type=type)
epi_nii = nb.load(epi)
dims = epi_nii.shape
mask_img = nb.Nifti1Image(dataobj=mask.transpose((2,1,0)), affine=epi_nii.affine, header=epi_nii.header)
mask_img.to_filename(out_file)
return out_file
def get_train(sessions=None, masked=True):
"""Retrieves training data for given sessions.
Default: all sessions, mask applied"""
if sessions is None:
sessions = range(3)
if isinstance(masked, bool) and masked:
#mask = get_mask()
mask = cortex.get_cortical_mask("MLfs", "20121210ML_auto1", "thick")
return np.concatenate(
[tables.openFile(t).getNode('/data').read()[:, mask]
for t in [train_files[i] for i in sessions]])
elif isinstance(masked, np.ndarray):
mask = masked
return np.concatenate(
[tables.openFile(t).getNode('/data').read()[:, mask]
for t in [train_files[i] for i in sessions]])
else:
# raise NotImplementedError("This will exceed 4G of RAM")
return np.concatenate(
[tables.openFile(t).getNode('/data').read()
for t in [train_files[i] for i in sessions]])
glob.glob("{base_dir}/pp_data/{sub}/func/*.nii.gz".format(
base_dir=base_dir, sub=subject)))
ref_file = file_list[0]
transform = cortex.xfm.Transform(np.identity(4), ref_file)
transform.save(subject, xfm_name, 'magnet')
# Add masks to pycortex transform
# -------------------------------
print('create pycortex transform')
xfm_masks = analysis_info['xfm_masks']
ref = nb.load(ref_file)
for xfm_mask in xfm_masks:
mask = cortex.get_cortical_mask(subject=subject,
xfmname=xfm_name,
type=xfm_mask)
mask_img = nb.Nifti1Image(dataobj=mask.transpose((2, 1, 0)),
affine=ref.affine,
header=ref.header)
mask_file = "{cortex_dir}/transforms/{xfm_name}/mask_{xfm_mask}.nii.gz".format(
cortex_dir=cortex_dir, xfm_name=xfm_name, xfm_mask=xfm_mask)
mask_img.to_filename(mask_file)
# Create participant pycortex overlays
# ------------------------------------
print('create subject pycortex overlays to check')
voxel_vol = cortex.Volume(np.random.randn(mask.shape[0], mask.shape[1],
mask.shape[2]),
subject=subject,
xfmname=xfm_name)
# then, we use the generated flirt matrix to import the subject into the session.
xfm_data = np.loadtxt(mtx_file)
xfm = cortex.xfm.Transform.from_fsl(xfm=xfm_data,
func_nii=example_epi_file,
anat_nii=fs_T1_nii_file)
# # Save as pycortex 'coord' transform
xfm.save('sub-{subject}'.format(subject=subject), 'fmriprep_T1', 'coord')
###################################################################
###
# and, while we're at it, create a cortical mask and save it.
###
###################################################################
mask = cortex.get_cortical_mask('sub-{subject}'.format(subject=subject),
'fmriprep_T1',
type='thick')
epi_nii = nb.load(example_epi_file)
dims = epi_nii.shape
mask_img = nb.Nifti1Image(dataobj=mask.transpose((2, 1, 0)),
affine=epi_nii.affine,
header=epi_nii.header)
cortical_mask_file = os.path.join(
base_dir, settings['result_dir'],
settings['cortical_mask_filename'].format(subject=subject,
session=settings['session']))
mask_img.to_filename(cortical_mask_file)
def __init__(self):
tmp_mni_mask = cortex.get_cortical_mask("MNI", "atlas", "thin")
self.all_non_zero = np.where(tmp_mni_mask != 0)
self.mni_dim = tmp_mni_mask.shape
self.n_mni_voxels = len(self.all_non_zero[0])
valstim = data.get_gabor("val").T
delays = [2, 3, 4]
deltrnstim = np.hstack([np.roll(trnstim, d, 0) for d in delays])[:numtime]
delvalstim = np.hstack([np.roll(valstim, d, 0) for d in delays])
#sdeltrnstim = scipy.sparse.csr_matrix(deltrnstim)
#sdelvalstim = scipy.sparse.csr_matrix(delvalstim)
zs = lambda m: (m - m.mean(0)) / m.std(0)
sdeltrnstim = deltrnstim = np.nan_to_num(zs(deltrnstim))
sdelvalstim = delvalstim = np.nan_to_num(zs(delvalstim))
# Select some voxels
cort_mask = cortex.get_cortical_mask("MLfs", "20121210ML_auto1", "thick")
#rois = ["V1", "V2", "V3"]
rois = ["V1"]
masks = [cortex.get_roi_mask("MLfs",
"20121210ML_auto1",
roi=roi)[roi] > 0 for roi in rois]
roimask = reduce(lambda x, y: (x + y), masks)
wardmask = cort_mask - roimask
# Load training, test fMRI data
trndata_roi = np.nan_to_num(data.get_train(masked=roimask)[:numtime])
trndata_ward = np.nan_to_num(data.get_train(masked=wardmask)[:numtime])
connectivity = image.grid_to_graph(n_x=wardmask.shape[0],
n_y=wardmask.shape[1],
n_z=wardmask.shape[2],