def test_andb():
# test with a range of dimensions and ensure output dtype is int
for ndim in range(1, 5):
shape = (10,) * ndim
out = utils.andb([rs.randint(10, size=shape) for f in range(5)])
assert out.shape == shape
assert out.dtype == int
# confirm error raised when dimensions are not the same
with pytest.raises(ValueError):
utils.andb([rs.randint(10, size=(10, 10)),
rs.randint(10, size=(20, 20))])
def test_andb():
# test with a range of dimensions and ensure output dtype is int
for ndim in range(1, 5):
shape = (10,) * ndim
out = utils.andb([rs.randint(10, size=shape) for f in range(5)])
assert out.shape == shape
assert out.dtype == int
# confirm error raised when dimensions are not the same
with pytest.raises(ValueError):
utils.andb([rs.randint(10, size=(10, 10)),
rs.randint(10, size=(20, 20))])
def test_smoke_andb():
"""
ensure that andb returns reasonable objects with random inputs
in the correct format
"""
arr = np.random.random((100, 10)).tolist() # 2D list of "arrays"
assert utils.andb(arr) is not None
def kundu_tedpca(comptable, n_echos, kdaw=10., rdaw=1., stabilize=False):
"""
Select PCA components using Kundu's decision tree approach.
Parameters
----------
comptable : :obj:`pandas.DataFrame`
Component table with relevant metrics: kappa, rho, and normalized
variance explained. Component number should be the index.
n_echos : :obj:`int`
Number of echoes in dataset.
kdaw : :obj:`float`, optional
Kappa dimensionality augmentation weight. Must be a non-negative float,
or -1 (a special value). Default is 10.
rdaw : :obj:`float`, optional
Rho dimensionality augmentation weight. Must be a non-negative float,
or -1 (a special value). Default is 1.
stabilize : :obj:`bool`, optional
Whether to stabilize convergence by reducing dimensionality, for low
quality data. Default is False.
Returns
-------
comptable : :obj:`pandas.DataFrame`
Component table with components classified as 'accepted', 'rejected',
or 'ignored'.
metric_metadata : :obj:`dict`
Dictionary with metadata about calculated metrics.
Each entry corresponds to a column in ``comptable``.
"""
LGR.info('Performing PCA component selection with Kundu decision tree')
comptable['classification'] = 'accepted'
comptable['rationale'] = ''
eigenvalue_elbow = getelbow(comptable['normalized variance explained'],
return_val=True)
diff_varex_norm = np.abs(
np.diff(comptable['normalized variance explained']))
lower_diff_varex_norm = diff_varex_norm[(len(diff_varex_norm) // 2):]
varex_norm_thr = np.mean(
[lower_diff_varex_norm.max(),
diff_varex_norm.min()])
varex_norm_min = comptable['normalized variance explained'][
(len(diff_varex_norm) // 2) + np.arange(len(lower_diff_varex_norm))[
lower_diff_varex_norm >= varex_norm_thr][0] + 1]
varex_norm_cum = np.cumsum(comptable['normalized variance explained'])
fmin, fmid, fmax = getfbounds(n_echos)
if int(kdaw) == -1:
lim_idx = utils.andb(
[comptable['kappa'] < fmid, comptable['kappa'] > fmin]) == 2
kappa_lim = comptable.loc[lim_idx, 'kappa'].values
kappa_thr = kappa_lim[getelbow(kappa_lim)]
lim_idx = utils.andb(
[comptable['rho'] < fmid, comptable['rho'] > fmin]) == 2
rho_lim = comptable.loc[lim_idx, 'rho'].values
rho_thr = rho_lim[getelbow(rho_lim)]
stabilize = True
LGR.info('kdaw set to -1. Switching TEDPCA algorithm to '
'kundu-stabilize')
elif int(rdaw) == -1:
lim_idx = utils.andb(
[comptable['rho'] < fmid, comptable['rho'] > fmin]) == 2
rho_lim = comptable.loc[lim_idx, 'rho'].values
rho_thr = rho_lim[getelbow(rho_lim)]
else:
kappa_thr = np.average(sorted(
[fmin, (getelbow(comptable['kappa'], return_val=True) / 2), fmid]),
weights=[kdaw, 1, 1])
rho_thr = np.average(sorted([
fmin, (getelbow_cons(comptable['rho'], return_val=True) / 2), fmid
]),
weights=[rdaw, 1, 1])
# Reject if low Kappa, Rho, and variance explained
is_lowk = comptable['kappa'] <= kappa_thr
is_lowr = comptable['rho'] <= rho_thr
is_lowe = comptable['normalized variance explained'] <= eigenvalue_elbow
is_lowkre = is_lowk & is_lowr & is_lowe
comptable.loc[is_lowkre, 'classification'] = 'rejected'
comptable.loc[is_lowkre, 'rationale'] += 'P001;'
# Reject if low variance explained
is_lows = comptable['normalized variance explained'] <= varex_norm_min
comptable.loc[is_lows, 'classification'] = 'rejected'
comptable.loc[is_lows, 'rationale'] += 'P002;'
# Reject if Kappa over limit
is_fmax1 = comptable['kappa'] == F_MAX
comptable.loc[is_fmax1, 'classification'] = 'rejected'
comptable.loc[is_fmax1, 'rationale'] += 'P003;'
# Reject if Rho over limit
is_fmax2 = comptable['rho'] == F_MAX
comptable.loc[is_fmax2, 'classification'] = 'rejected'
comptable.loc[is_fmax2, 'rationale'] += 'P004;'
if stabilize:
temp7 = varex_norm_cum >= 0.95
comptable.loc[temp7, 'classification'] = 'rejected'
comptable.loc[temp7, 'rationale'] += 'P005;'
under_fmin1 = comptable['kappa'] <= fmin
comptable.loc[under_fmin1, 'classification'] = 'rejected'
comptable.loc[under_fmin1, 'rationale'] += 'P006;'
under_fmin2 = comptable['rho'] <= fmin
comptable.loc[under_fmin2, 'classification'] = 'rejected'
comptable.loc[under_fmin2, 'rationale'] += 'P007;'
n_components = comptable.loc[comptable['classification'] ==
'accepted'].shape[0]
LGR.info('Selected {0} components with Kappa threshold: {1:.02f}, Rho '
'threshold: {2:.02f}'.format(n_components, kappa_thr, rho_thr))
# Move decision columns to end
comptable = clean_dataframe(comptable)
metric_metadata = collect.get_metadata(comptable)
return comptable, metric_metadata
def selcomps(seldict,
mmix,
mask,
ref_img,
manacc,
n_echos,
t2s,
s0,
olevel=2,
oversion=99,
filecsdata=True,
savecsdiag=True,
strict_mode=False):
"""
Labels components in `mmix`
Parameters
----------
seldict : :obj:`dict`
As output from `fitmodels_direct`
mmix : (C x T) array_like
Mixing matrix for converting input data to component space, where `C`
is components and `T` is the number of volumes in the original data
mask : (S,) array_like
Boolean mask array
ref_img : str or img_like
Reference image to dictate how outputs are saved to disk
manacc : list
Comma-separated list of indices of manually accepted components
n_echos : int
Number of echos in original data
t2s : (S,) array_like
s0 : (S,) array_like
olevel : int, optional
Default: 2
oversion : int, optional
Default: 99
filecsdata: bool, optional
Default: False
savecsdiag: bool, optional
Default: True
strict_mode: bool, optional
Default: False
Returns
-------
acc : list
Indices of accepted (BOLD) components in `mmix`
rej : list
Indices of rejected (non-BOLD) components in `mmix`
midk : list
Indices of mid-K (questionable) components in `mmix`
ign : list
Indices of ignored components in `mmix`
"""
if filecsdata:
import bz2
if seldict is not None:
LGR.info('Saving component selection data')
with bz2.BZ2File('compseldata.pklbz', 'wb') as csstate_f:
pickle.dump(seldict, csstate_f)
else:
try:
with bz2.BZ2File('compseldata.pklbz', 'rb') as csstate_f:
seldict = pickle.load(csstate_f)
except FileNotFoundError:
LGR.warning('Failed to load component selection data')
return None
# List of components
midk = []
ign = []
nc = np.arange(len(seldict['Kappas']))
ncl = np.arange(len(seldict['Kappas']))
# If user has specified components to accept manually
if manacc:
acc = sorted([int(vv) for vv in manacc.split(',')])
midk = []
rej = sorted(np.setdiff1d(ncl, acc))
return acc, rej, midk, [] # Add string for ign
"""
Do some tallies for no. of significant voxels
"""
countsigFS0 = seldict['F_S0_clmaps'].sum(0)
countsigFR2 = seldict['F_R2_clmaps'].sum(0)
countnoise = np.zeros(len(nc))
"""
Make table of dice values
"""
dice_tbl = np.zeros([nc.shape[0], 2])
for ii in ncl:
dice_FR2 = utils.dice(
utils.unmask(seldict['Br_clmaps_R2'][:, ii], mask)[t2s != 0],
seldict['F_R2_clmaps'][:, ii])
dice_FS0 = utils.dice(
utils.unmask(seldict['Br_clmaps_S0'][:, ii], mask)[t2s != 0],
seldict['F_S0_clmaps'][:, ii])
dice_tbl[ii, :] = [dice_FR2, dice_FS0] # step 3a here and above
dice_tbl[np.isnan(dice_tbl)] = 0
"""
Make table of noise gain
"""
tt_table = np.zeros([len(nc), 4])
counts_FR2_Z = np.zeros([len(nc), 2])
for ii in nc:
comp_noise_sel = utils.andb([
np.abs(seldict['Z_maps'][:, ii]) > 1.95,
seldict['Z_clmaps'][:, ii] == 0
]) == 2
countnoise[ii] = np.array(comp_noise_sel, dtype=np.int).sum()
noise_FR2_Z_mask = utils.unmask(comp_noise_sel, mask)[t2s != 0]
noise_FR2_Z = np.log10(
np.unique(seldict['F_R2_maps'][noise_FR2_Z_mask, ii]))
signal_FR2_Z_mask = utils.unmask(seldict['Z_clmaps'][:, ii],
mask)[t2s != 0] == 1
signal_FR2_Z = np.log10(
np.unique(seldict['F_R2_maps'][signal_FR2_Z_mask, ii]))
counts_FR2_Z[ii, :] = [len(signal_FR2_Z), len(noise_FR2_Z)]
ttest = stats.ttest_ind(signal_FR2_Z, noise_FR2_Z, equal_var=True)
# avoid DivideByZero RuntimeWarning
if signal_FR2_Z.size > 0 and noise_FR2_Z.size > 0:
mwu = stats.norm.ppf(
stats.mannwhitneyu(signal_FR2_Z, noise_FR2_Z)[1])
else:
mwu = -np.inf
tt_table[ii, 0] = np.abs(mwu) * ttest[0] / np.abs(ttest[0])
tt_table[ii, 1] = ttest[1]
tt_table[np.isnan(tt_table)] = 0
tt_table[np.isinf(tt_table[:, 0]),
0] = np.percentile(tt_table[~np.isinf(tt_table[:, 0]), 0], 98)
# Time series derivative kurtosis
mmix_dt = (mmix[:-1] - mmix[1:])
mmix_kurt = stats.kurtosis(mmix_dt)
mmix_std = np.std(mmix_dt, axis=0)
"""
Step 1: Reject anything that's obviously an artifact
a. Estimate a null variance
"""
LGR.debug(
'Rejecting gross artifacts based on Rho/Kappa values and S0/R2 counts')
rej = ncl[utils.andb(
[seldict['Rhos'] > seldict['Kappas'], countsigFS0 > countsigFR2]) > 0]
ncl = np.setdiff1d(ncl, rej)
"""
Step 2: Compute 3-D spatial FFT of Beta maps to detect high-spatial
frequency artifacts
"""
LGR.debug(
'Computing 3D spatial FFT of beta maps to detect high-spatial frequency artifacts'
)
# spatial information is important so for NIFTI we convert back to 3D space
if utils.get_dtype(ref_img) == 'NIFTI':
dim1 = np.prod(check_niimg(ref_img).shape[:2])
else:
dim1 = mask.shape[0]
fproj_arr = np.zeros([dim1, len(nc)])
fproj_arr_val = np.zeros([dim1, len(nc)])
spr = []
fdist = []
for ii in nc:
# convert data back to 3D array
if utils.get_dtype(ref_img) == 'NIFTI':
tproj = utils.new_nii_like(
ref_img,
utils.unmask(seldict['PSC'], mask)[:, ii]).get_data()
else:
tproj = utils.unmask(seldict['PSC'], mask)[:, ii]
fproj = np.fft.fftshift(np.abs(np.fft.rfftn(tproj)))
fproj_z = fproj.max(axis=-1)
fproj[fproj == fproj.max()] = 0
spr.append(np.array(fproj_z > fproj_z.max() / 4, dtype=np.int).sum())
fproj_arr[:, ii] = stats.rankdata(fproj_z.flatten())
fproj_arr_val[:, ii] = fproj_z.flatten()
if utils.get_dtype(ref_img) == 'NIFTI':
fprojr = np.array([fproj, fproj[:, :, ::-1]]).max(0)
fdist.append(
np.max([
utils.fitgaussian(fproj.max(jj))[3:].max()
for jj in range(fprojr.ndim)
]))
else:
fdist = np.load(os.path.join(RESOURCES, 'fdist.npy'))
if type(fdist) is not np.ndarray:
fdist = np.array(fdist)
spr = np.array(spr)
# import ipdb; ipdb.set_trace()
"""
Step 3: Create feature space of component properties
"""
LGR.debug('Creating feature space of component properties')
fdist_pre = fdist.copy()
fdist_pre[fdist > np.median(fdist) * 3] = np.median(fdist) * 3
fdist_z = (fdist_pre - np.median(fdist_pre)) / fdist_pre.std()
spz = (spr - spr.mean()) / spr.std()
Tz = (tt_table[:, 0] - tt_table[:, 0].mean()) / tt_table[:, 0].std()
varex_ = np.log(seldict['varex'])
Vz = (varex_ - varex_.mean()) / varex_.std()
Rz = (seldict['Rhos'] - seldict['Rhos'].mean()) / seldict['Rhos'].std()
Ktz = np.log(seldict['Kappas']) / 2
Ktz = (Ktz - Ktz.mean()) / Ktz.std()
Rtz = np.log(seldict['Rhos']) / 2
Rtz = (Rtz - Rtz.mean()) / Rtz.std()
KRr = stats.zscore(np.log(seldict['Kappas']) / np.log(seldict['Rhos']))
cnz = (countnoise - countnoise.mean()) / countnoise.std()
Dz = stats.zscore(np.arctanh(dice_tbl[:, 0] + 0.001))
fz = np.array([Tz, Vz, Ktz, KRr, cnz, Rz, mmix_kurt, fdist_z])
"""
Step 3: Make initial guess of where BOLD components are and use DBSCAN
to exclude noise components and find a sample set of 'good' components
"""
LGR.debug('Making initial guess of BOLD components')
# epsmap is [index,level of overlap with dicemask,
# number of high Rho components]
F05, F025, F01 = utils.getfbounds(n_echos)
epsmap = []
Rhos_sorted = np.array(sorted(seldict['Rhos']))[::-1]
# Make an initial guess as to number of good components based on
# consensus of control points across Rhos and Kappas
KRcutguesses = [
getelbow_mod(seldict['Rhos']),
getelbow_cons(seldict['Rhos']),
getelbow_aggr(seldict['Rhos']),
getelbow_mod(seldict['Kappas']),
getelbow_cons(seldict['Kappas']),
getelbow_aggr(seldict['Kappas'])
]
Khighelbowval = stats.scoreatpercentile([
getelbow_mod(seldict['Kappas'], val=True),
getelbow_cons(seldict['Kappas'], val=True),
getelbow_aggr(seldict['Kappas'], val=True)
] + list(utils.getfbounds(n_echos)),
75,
interpolation_method='lower')
KRcut = np.median(KRcutguesses)
# only use exclusive when inclusive is extremely inclusive - double KRcut
cond1 = getelbow_cons(seldict['Kappas']) > KRcut * 2
cond2 = getelbow_mod(seldict['Kappas'], val=True) < F01
if cond1 and cond2:
Kcut = getelbow_mod(seldict['Kappas'], val=True)
else:
Kcut = getelbow_cons(seldict['Kappas'], val=True)
# only use inclusive when exclusive is extremely exclusive - half KRcut
# (remember for Rho inclusive is higher, so want both Kappa and Rho
# to defaut to lower)
if getelbow_cons(seldict['Rhos']) > KRcut * 2:
Rcut = getelbow_mod(seldict['Rhos'], val=True)
# for above, consider something like:
# min([getelbow_mod(Rhos,True),sorted(Rhos)[::-1][KRguess] ])
else:
Rcut = getelbow_cons(seldict['Rhos'], val=True)
if Rcut > Kcut:
Kcut = Rcut # Rcut should never be higher than Kcut
KRelbow = utils.andb([seldict['Kappas'] > Kcut, seldict['Rhos'] < Rcut])
# Make guess of Kundu et al 2011 plus remove high frequencies,
# generally high variance, and high variance given low Kappa
tt_lim = stats.scoreatpercentile(
tt_table[tt_table[:, 0] > 0, 0], 75, interpolation_method='lower') / 3
KRguess = np.setdiff1d(
np.setdiff1d(nc[KRelbow == 2], rej),
np.union1d(
nc[tt_table[:, 0] < tt_lim],
np.union1d(
np.union1d(nc[spz > 1], nc[Vz > 2]), nc[utils.andb([
seldict['varex'] > 0.5 *
sorted(seldict['varex'])[::-1][int(KRcut)],
seldict['Kappas'] < 2 * Kcut
]) == 2])))
guessmask = np.zeros(len(nc))
guessmask[KRguess] = 1
# Throw lower-risk bad components out
rejB = ncl[utils.andb([
tt_table[ncl, 0] < 0,
seldict['varex'][ncl] > np.median(seldict['varex']), ncl > KRcut
]) == 3]
rej = np.union1d(rej, rejB)
ncl = np.setdiff1d(ncl, rej)
LGR.debug('Using DBSCAN to find optimal set of "good" BOLD components')
for ii in range(20000):
eps = .005 + ii * .005
db = DBSCAN(eps=eps, min_samples=3).fit(fz.T)
# it would be great to have descriptive names, here
# DBSCAN found at least three non-noisy clusters
cond1 = db.labels_.max() > 1
# DBSCAN didn't detect more classes than the total # of components / 6
cond2 = db.labels_.max() < len(nc) / 6
# TODO: confirm if 0 is a special label for DBSCAN
# my intuition here is that we're confirming DBSCAN labelled previously
# rejected components as noise (i.e., no overlap between `rej` and
# labelled DBSCAN components)
cond3 = np.intersect1d(rej, nc[db.labels_ == 0]).shape[0] == 0
# DBSCAN labelled less than half of the total components as noisy
cond4 = np.array(db.labels_ == -1, dtype=int).sum() / float(
len(nc)) < .5
if cond1 and cond2 and cond3 and cond4:
epsmap.append([
ii,
utils.dice(guessmask, db.labels_ == 0),
np.intersect1d(
nc[db.labels_ == 0],
nc[seldict['Rhos'] > getelbow_mod(Rhos_sorted, val=True)]).
shape[0]
])
db = None
epsmap = np.array(epsmap)
LGR.debug('Found DBSCAN solutions for {}/20000 eps resolutions'.format(
len(epsmap)))
group0 = []
dbscanfailed = False
if len(epsmap) != 0:
# Select index that maximizes Dice with guessmask but first
# minimizes number of higher Rho components
ii = int(
epsmap[np.argmax(epsmap[epsmap[:, 2] == np.min(epsmap[:, 2]),
1], 0), 0])
LGR.debug('Component selection tuning: {:.05f}'.format(
epsmap[:, 1].max()))
db = DBSCAN(eps=.005 + ii * .005, min_samples=3).fit(fz.T)
ncl = nc[db.labels_ == 0]
ncl = np.setdiff1d(ncl, rej)
ncl = np.setdiff1d(ncl, ncl[ncl > len(nc) - len(rej)])
group0 = ncl.copy()
group_n1 = nc[db.labels_ == -1]
to_clf = np.setdiff1d(nc, np.union1d(ncl, rej))
if len(group0) == 0 or len(group0) < len(KRguess) * .5:
dbscanfailed = True
LGR.debug('DBSCAN guess failed; using elbow guess method instead')
ncl = np.setdiff1d(
np.setdiff1d(nc[KRelbow == 2], rej),
np.union1d(
nc[tt_table[:, 0] < tt_lim],
np.union1d(
np.union1d(nc[spz > 1], nc[Vz > 2]), nc[utils.andb([
seldict['varex'] > 0.5 *
sorted(seldict['varex'])[::-1][int(KRcut)],
seldict['Kappas'] < 2 * Kcut
]) == 2])))
group0 = ncl.copy()
group_n1 = []
to_clf = np.setdiff1d(nc, np.union1d(group0, rej))
if len(group0) < 2 or (len(group0) < 4
and float(len(rej)) / len(group0) > 3):
LGR.warning('Extremely limited reliable BOLD signal space! '
'Not filtering components beyond BOLD/non-BOLD guesses.')
midkfailed = True
min_acc = np.array([])
if len(group0) != 0:
# For extremes, building in a 20% tolerance
toacc_hi = np.setdiff1d(
nc[utils.andb([
fdist <= np.max(fdist[group0]), seldict['Rhos'] < F025,
Vz > -2
]) == 3], np.union1d(group0, rej))
min_acc = np.union1d(group0, toacc_hi)
to_clf = np.setdiff1d(nc, np.union1d(min_acc, rej))
else:
toacc_hi = []
min_acc = []
diagstep_keys = [
'Rejected components', 'Kappa-Rho cut point', 'Kappa cut point',
'Rho cut point', 'DBSCAN failed to converge',
'Mid-Kappa failed (limited BOLD signal)', 'Kappa-Rho guess',
'min_acc', 'toacc_hi'
]
diagstep_vals = [
list(rej), KRcut, Kcut, Rcut, dbscanfailed, midkfailed,
list(KRguess),
list(min_acc),
list(toacc_hi)
]
with open('csstepdata.json', 'w') as ofh:
json.dump(dict(zip(diagstep_keys, diagstep_vals)),
ofh,
indent=4,
sort_keys=True,
default=str)
return list(sorted(min_acc)), list(sorted(rej)), [], list(
sorted(to_clf))
# Find additional components to reject based on Dice - doing this here
# since Dice is a little unstable, need to reference group0
rej_supp = []
dice_rej = False
if not dbscanfailed and len(rej) + len(group0) < 0.75 * len(nc):
dice_rej = True
rej_supp = np.setdiff1d(
np.setdiff1d(
np.union1d(rej, nc[dice_tbl[nc, 0] <= dice_tbl[nc, 1]]),
group0), group_n1)
rej = np.union1d(rej, rej_supp)
# Temporal features
# larger is worse - spike
mmix_kurt_z = (mmix_kurt -
mmix_kurt[group0].mean()) / mmix_kurt[group0].std()
# smaller is worse - drift
mmix_std_z = -1 * (
(mmix_std - mmix_std[group0].mean()) / mmix_std[group0].std())
mmix_kurt_z_max = np.max([mmix_kurt_z, mmix_std_z], 0)
"""
Step 2: Classifiy midk and ignore using separte SVMs for
different variance regimes
# To render hyperplane:
min_x = np.min(spz2);max_x=np.max(spz2)
# plotting separating hyperplane
ww = clf_.coef_[0]
aa = -ww[0] / ww[1]
# make sure the next line is long enough
xx = np.linspace(min_x - 2, max_x + 2)
yy = aa * xx - (clf_.intercept_[0]) / ww[1]
plt.plot(xx, yy, '-')
"""
LGR.debug('Attempting to classify midk components')
# Tried getting rid of accepting based on SVM altogether,
# now using only rejecting
toacc_hi = np.setdiff1d(
nc[utils.andb([
fdist <= np.max(fdist[group0]), seldict['Rhos'] < F025, Vz > -2
]) == 3], np.union1d(group0, rej))
toacc_lo = np.intersect1d(
to_clf, nc[utils.andb([
spz < 1, Rz < 0, mmix_kurt_z_max < 5, Dz > -1, Tz > -1, Vz < 0,
seldict['Kappas'] >= F025, fdist < 3 *
np.percentile(fdist[group0], 98)
]) == 8])
midk_clf, clf_ = do_svm(fproj_arr_val[:, np.union1d(group0, rej)].T,
[0] * len(group0) + [1] * len(rej),
fproj_arr_val[:, to_clf].T,
svmtype=2)
midk = np.setdiff1d(
to_clf[utils.andb([
midk_clf == 1,
seldict['varex'][to_clf] > np.median(seldict['varex'][group0])
]) == 2], np.union1d(toacc_hi, toacc_lo))
# only use SVM to augment toacc_hi only if toacc_hi isn't already
# conflicting with SVM choice
if len(
np.intersect1d(
to_clf[utils.andb([midk_clf == 1, Vz[to_clf] > 0]) == 2],
toacc_hi)) == 0:
svm_acc_fail = True
toacc_hi = np.union1d(toacc_hi, to_clf[midk_clf == 0])
else:
svm_acc_fail = False
"""
Step 3: Compute variance associated with low T2* areas
(e.g. draining veins and low T2* areas)
# To write out veinmask
veinout = np.zeros(t2s.shape)
veinout[t2s!=0] = veinmaskf
utils.filewrite(veinout, 'veinmaskf', ref_img)
veinBout = utils.unmask(veinmaskB, mask)
utils.filewrite(veinBout, 'veins50', ref_img)
"""
LGR.debug(
'Computing variance associated with low T2* areas (e.g., draining veins)'
)
tsoc_B_Zcl = np.zeros(seldict['tsoc_B'].shape)
tsoc_B_Zcl[seldict['Z_clmaps'] != 0] = np.abs(
seldict['tsoc_B'])[seldict['Z_clmaps'] != 0]
sig_B = [
stats.scoreatpercentile(tsoc_B_Zcl[tsoc_B_Zcl[:, ii] != 0, ii], 25)
if len(tsoc_B_Zcl[tsoc_B_Zcl[:, ii] != 0, ii]) != 0 else 0 for ii in nc
]
sig_B = np.abs(seldict['tsoc_B']) > np.tile(
sig_B, [seldict['tsoc_B'].shape[0], 1])
veinmask = utils.andb([
t2s < stats.scoreatpercentile(
t2s[t2s != 0], 15, interpolation_method='lower'), t2s != 0
]) == 2
veinmaskf = veinmask[mask]
veinR = np.array(sig_B[veinmaskf].sum(0),
dtype=float) / sig_B[~veinmaskf].sum(0)
veinR[np.isnan(veinR)] = 0
veinc = np.union1d(rej, midk)
rej_veinRZ = ((veinR - veinR[veinc].mean()) / veinR[veinc].std())[veinc]
rej_veinRZ[rej_veinRZ < 0] = 0
rej_veinRZ[countsigFR2[veinc] > np.array(veinmaskf, dtype=int).sum()] = 0
t2s_lim = [
stats.scoreatpercentile(t2s[t2s != 0],
50,
interpolation_method='lower'),
stats.scoreatpercentile(
t2s[t2s != 0], 80, interpolation_method='lower') / 2
]
phys_var_zs = []
for t2sl_i in range(len(t2s_lim)):
t2sl = t2s_lim[t2sl_i]
veinW = sig_B[:, veinc] * np.tile(rej_veinRZ, [sig_B.shape[0], 1])
veincand = utils.unmask(
utils.andb([
s0[t2s != 0] < np.median(s0[t2s != 0]), t2s[t2s != 0] < t2sl
]) >= 1, t2s != 0)[mask]
veinW[~veincand] = 0
invein = veinW.sum(
axis=1)[(utils.unmask(veinmaskf, mask) *
utils.unmask(veinW.sum(axis=1) > 1, mask))[mask]]
minW = 10 * (np.log10(invein).mean()) - 1 * 10**(
np.log10(invein).std())
veinmaskB = veinW.sum(axis=1) > minW
tsoc_Bp = seldict['tsoc_B'].copy()
tsoc_Bp[tsoc_Bp < 0] = 0
vvex = np.array([
(tsoc_Bp[veinmaskB, ii]**2.).sum() / (tsoc_Bp[:, ii]**2.).sum()
for ii in nc
])
group0_res = np.intersect1d(KRguess, group0)
phys_var_zs.append(
(vvex - vvex[group0_res].mean()) / vvex[group0_res].std())
veinBout = utils.unmask(veinmaskB, mask)
utils.filewrite(veinBout.astype(float), 'veins_l%i' % t2sl_i, ref_img)
# Mask to sample veins
phys_var_z = np.array(phys_var_zs).max(0)
Vz2 = (varex_ - varex_[group0].mean()) / varex_[group0].std()
"""
Step 4: Learn joint TE-dependence spatial and temporal models to move
remaining artifacts to ignore class
"""
LGR.debug(
'Learning joint TE-dependence spatial/temporal models to ignore remaining artifacts'
)
to_ign = []
minK_ign = np.max([F05, getelbow_cons(seldict['Kappas'], val=True)])
newcest = len(group0) + len(
toacc_hi[seldict['Kappas'][toacc_hi] > minK_ign])
phys_art = np.setdiff1d(
nc[utils.andb([phys_var_z > 3.5, seldict['Kappas'] < minK_ign]) == 2],
group0)
rank_diff = stats.rankdata(phys_var_z) - stats.rankdata(seldict['Kappas'])
phys_art = np.union1d(
np.setdiff1d(
nc[utils.andb([phys_var_z > 2, rank_diff > newcest /
2, Vz2 > -1]) == 3], group0), phys_art)
# Want to replace field_art with an acf/SVM based approach
# instead of a kurtosis/filter one
field_art = np.setdiff1d(
nc[utils.andb([mmix_kurt_z_max > 5, seldict['Kappas'] < minK_ign]) ==
2], group0)
field_art = np.union1d(
np.setdiff1d(
nc[utils.andb([
mmix_kurt_z_max > 2,
(stats.rankdata(mmix_kurt_z_max) -
stats.rankdata(seldict['Kappas'])) > newcest /
2, Vz2 > 1, seldict['Kappas'] < F01
]) == 4], group0), field_art)
field_art = np.union1d(
np.setdiff1d(
nc[utils.andb([
mmix_kurt_z_max > 3, Vz2 > 3,
seldict['Rhos'] > np.percentile(seldict['Rhos'][group0], 75)
]) == 3], group0), field_art)
field_art = np.union1d(
np.setdiff1d(nc[utils.andb([mmix_kurt_z_max > 5, Vz2 > 5]) == 2],
group0), field_art)
misc_art = np.setdiff1d(
nc[utils.andb([(stats.rankdata(Vz) - stats.rankdata(Ktz)) > newcest /
2, seldict['Kappas'] < Khighelbowval]) == 2], group0)
ign_cand = np.unique(list(field_art) + list(phys_art) + list(misc_art))
midkrej = np.union1d(midk, rej)
to_ign = np.setdiff1d(list(ign_cand), midkrej)
toacc = np.union1d(toacc_hi, toacc_lo)
ncl = np.setdiff1d(np.union1d(ncl, toacc), np.union1d(to_ign, midkrej))
ign = np.setdiff1d(nc, list(ncl) + list(midk) + list(rej))
orphan = np.setdiff1d(nc,
list(ncl) + list(to_ign) + list(midk) + list(rej))
# Last ditch effort to save some transient components
if not strict_mode:
Vz3 = (varex_ - varex_[ncl].mean()) / varex_[ncl].std()
ncl = np.union1d(
ncl,
np.intersect1d(
orphan, nc[utils.andb([
seldict['Kappas'] > F05, seldict['Rhos'] < F025,
seldict['Kappas'] > seldict['Rhos'], Vz3 <= -1, Vz3 > -3,
mmix_kurt_z_max < 2.5
]) == 6]))
ign = np.setdiff1d(nc, list(ncl) + list(midk) + list(rej))
orphan = np.setdiff1d(
nc,
list(ncl) + list(to_ign) + list(midk) + list(rej))
if savecsdiag:
diagstep_keys = [
'Rejected components', 'Kappa-Rho cut point', 'Kappa cut',
'Rho cut', 'DBSCAN failed to converge', 'Kappa-Rho guess',
'Dice rejected', 'rej_supp', 'to_clf', 'Mid-kappa components',
'svm_acc_fail', 'toacc_hi', 'toacc_lo', 'Field artifacts',
'Physiological artifacts', 'Miscellaneous artifacts', 'ncl',
'Ignored components'
]
diagstep_vals = [
list(rej),
KRcut.item(),
Kcut.item(),
Rcut.item(), dbscanfailed,
list(KRguess), dice_rej,
list(rej_supp),
list(to_clf),
list(midk), svm_acc_fail,
list(toacc_hi),
list(toacc_lo),
list(field_art),
list(phys_art),
list(misc_art),
list(ncl),
list(ign)
]
with open('csstepdata.json', 'w') as ofh:
json.dump(dict(zip(diagstep_keys, diagstep_vals)),
ofh,
indent=4,
sort_keys=True,
default=str)
allfz = np.array([Tz, Vz, Ktz, KRr, cnz, Rz, mmix_kurt, fdist_z])
np.savetxt('csdata.txt', allfz)
return list(sorted(ncl)), list(sorted(rej)), list(sorted(midk)), list(
sorted(ign))
def tedpca(catd,
OCcatd,
combmode,
mask,
t2s,
t2sG,
stabilize,
ref_img,
tes,
kdaw,
rdaw,
ste=0,
wvpca=False):
"""
Use principal components analysis (PCA) to identify and remove thermal
noise from multi-echo data.
Parameters
----------
catd : (S x E x T) array_like
Input functional data
OCcatd : (S x T) array_like
Optimally-combined time series data
combmode : {'t2s', 'ste'} str
How optimal combination of echos should be made, where 't2s' indicates
using the method of Posse 1999 and 'ste' indicates using the method of
Poser 2006
mask : (S,) array_like
Boolean mask array
stabilize : :obj:`bool`
Whether to attempt to stabilize convergence of ICA by returning
dimensionally-reduced data from PCA and component selection.
ref_img : :obj:`str` or img_like
Reference image to dictate how outputs are saved to disk
tes : :obj:`list`
List of echo times associated with `catd`, in milliseconds
kdaw : :obj:`float`
Dimensionality augmentation weight for Kappa calculations
rdaw : :obj:`float`
Dimensionality augmentation weight for Rho calculations
ste : :obj:`int` or :obj:`list` of :obj:`int`, optional
Which echos to use in PCA. Values -1 and 0 are special, where a value
of -1 will indicate using all the echos and 0 will indicate using the
optimal combination of the echos. A list can be provided to indicate
a subset of echos. Default: 0
wvpca : :obj:`bool`, optional
Whether to apply wavelet denoising to data. Default: False
Returns
-------
n_components : :obj:`int`
Number of components retained from PCA decomposition
dd : (S x T) :obj:`numpy.ndarray`
Dimensionally reduced optimally combined functional data
Notes
-----
====================== =================================================
Notation Meaning
====================== =================================================
:math:`\\kappa` Component pseudo-F statistic for TE-dependent
(BOLD) model.
:math:`\\rho` Component pseudo-F statistic for TE-independent
(artifact) model.
:math:`v` Voxel
:math:`V` Total number of voxels in mask
:math:`\\zeta` Something
:math:`c` Component
:math:`p` Something else
====================== =================================================
Steps:
1. Variance normalize either multi-echo or optimally combined data,
depending on settings.
2. Decompose normalized data using PCA or SVD.
3. Compute :math:`{\\kappa}` and :math:`{\\rho}`:
.. math::
{\\kappa}_c = \\frac{\sum_{v}^V {\\zeta}_{c,v}^p * \
F_{c,v,R_2^*}}{\sum {\\zeta}_{c,v}^p}
{\\rho}_c = \\frac{\sum_{v}^V {\\zeta}_{c,v}^p * \
F_{c,v,S_0}}{\sum {\\zeta}_{c,v}^p}
4. Some other stuff. Something about elbows.
5. Classify components as thermal noise if they meet both of the
following criteria:
- Nonsignificant :math:`{\\kappa}` and :math:`{\\rho}`.
- Nonsignificant variance explained.
Outputs:
This function writes out several files:
====================== =================================================
Filename Content
====================== =================================================
pcastate.pkl Values from PCA results.
comp_table_pca.txt PCA component table.
mepca_mix.1D PCA mixing matrix.
====================== =================================================
"""
n_samp, n_echos, n_vols = catd.shape
ste = np.array([int(ee) for ee in str(ste).split(',')])
if len(ste) == 1 and ste[0] == -1:
LGR.info('Computing PCA of optimally combined multi-echo data')
d = OCcatd[utils.make_min_mask(OCcatd[:,
np.newaxis, :])][:,
np.newaxis, :]
elif len(ste) == 1 and ste[0] == 0:
LGR.info('Computing PCA of spatially concatenated multi-echo data')
d = catd[mask].astype('float64')
else:
LGR.info('Computing PCA of echo #%s' %
','.join([str(ee) for ee in ste]))
d = np.stack([catd[mask, ee] for ee in ste - 1],
axis=1).astype('float64')
eim = np.squeeze(eimask(d))
d = np.squeeze(d[eim])
dz = ((d.T - d.T.mean(axis=0)) / d.T.std(axis=0)).T # var normalize ts
dz = (dz - dz.mean()) / dz.std() # var normalize everything
if wvpca:
dz, cAl = dwtmat(dz)
if not op.exists('pcastate.pkl'):
voxel_comp_weights, varex, comp_ts = run_svd(dz)
# actual variance explained (normalized)
varex_norm = varex / varex.sum()
eigenvalue_elbow = getelbow(varex_norm, return_val=True)
diff_varex_norm = np.abs(np.diff(varex_norm))
lower_diff_varex_norm = diff_varex_norm[(len(diff_varex_norm) // 2):]
varex_norm_thr = np.mean(
[lower_diff_varex_norm.max(),
diff_varex_norm.min()])
varex_norm_min = varex_norm[
(len(diff_varex_norm) // 2) + np.arange(len(lower_diff_varex_norm))
[lower_diff_varex_norm >= varex_norm_thr][0] + 1]
varex_norm_cum = np.cumsum(varex_norm)
# Compute K and Rho for PCA comps
eimum = np.atleast_2d(eim)
eimum = np.transpose(eimum, np.argsort(eimum.shape)[::-1])
eimum = eimum.prod(axis=1)
o = np.zeros((mask.shape[0], *eimum.shape[1:]))
o[mask] = eimum
eimum = np.squeeze(o).astype(bool)
vTmix = comp_ts.T
vTmixN = ((vTmix.T - vTmix.T.mean(0)) / vTmix.T.std(0)).T
LGR.info('Making initial component selection guess from PCA results')
_, ct_df, betasv, v_T = model.fitmodels_direct(catd,
comp_ts.T,
eimum,
t2s,
t2sG,
tes,
combmode,
ref_img,
mmixN=vTmixN,
full_sel=False)
# varex_norm overrides normalized varex computed by fitmodels_direct
ct_df['normalized variance explained'] = varex_norm
# Save state
fname = op.abspath('pcastate.pkl')
LGR.info('Saving PCA results to: {}'.format(fname))
pcastate = {
'voxel_comp_weights': voxel_comp_weights,
'varex': varex,
'comp_ts': comp_ts,
'comptable': ct_df,
'eigenvalue_elbow': eigenvalue_elbow,
'varex_norm_min': varex_norm_min,
'varex_norm_cum': varex_norm_cum
}
try:
with open(fname, 'wb') as handle:
pickle.dump(pcastate, handle)
except TypeError:
LGR.warning('Could not save PCA solution')
else: # if loading existing state
LGR.info('Loading PCA from: pcastate.pkl')
with open('pcastate.pkl', 'rb') as handle:
pcastate = pickle.load(handle)
voxel_comp_weights, varex = pcastate['voxel_comp_weights'], pcastate[
'varex']
comp_ts = pcastate['comp_ts']
ct_df = pcastate['comptable']
eigenvalue_elbow = pcastate['eigenvalue_elbow']
varex_norm_min = pcastate['varex_norm_min']
varex_norm_cum = pcastate['varex_norm_cum']
np.savetxt('mepca_mix.1D', comp_ts.T)
# write component maps to 4D image
comp_maps = np.zeros((OCcatd.shape[0], comp_ts.shape[0]))
for i_comp in range(comp_ts.shape[0]):
temp_comp_ts = comp_ts[i_comp, :][:, None]
comp_map = utils.unmask(
model.computefeats2(OCcatd, temp_comp_ts, mask), mask)
comp_maps[:, i_comp] = np.squeeze(comp_map)
io.filewrite(comp_maps, 'mepca_OC_components.nii', ref_img)
fmin, fmid, fmax = utils.getfbounds(n_echos)
kappa_thr = np.average(sorted(
[fmin, getelbow(ct_df['kappa'], return_val=True) / 2, fmid]),
weights=[kdaw, 1, 1])
rho_thr = np.average(sorted(
[fmin, getelbow_cons(ct_df['rho'], return_val=True) / 2, fmid]),
weights=[rdaw, 1, 1])
if int(kdaw) == -1:
lim_idx = utils.andb([ct_df['kappa'] < fmid,
ct_df['kappa'] > fmin]) == 2
kappa_lim = ct_df.loc[lim_idx, 'kappa'].values
kappa_thr = kappa_lim[getelbow(kappa_lim)]
lim_idx = utils.andb([ct_df['rho'] < fmid, ct_df['rho'] > fmin]) == 2
rho_lim = ct_df.loc[lim_idx, 'rho'].values
rho_thr = rho_lim[getelbow(rho_lim)]
stabilize = True
elif int(rdaw) == -1:
lim_idx = utils.andb([ct_df['rho'] < fmid, ct_df['rho'] > fmin]) == 2
rho_lim = ct_df.loc[lim_idx, 'rho'].values
rho_thr = rho_lim[getelbow(rho_lim)]
# Add new columns to comptable for classification
ct_df['classification'] = 'accepted'
ct_df['rationale'] = ''
# Reject if low Kappa, Rho, and variance explained
is_lowk = ct_df['kappa'] <= kappa_thr
is_lowr = ct_df['rho'] <= rho_thr
is_lowe = ct_df['normalized variance explained'] <= eigenvalue_elbow
is_lowkre = is_lowk & is_lowr & is_lowe
ct_df.loc[is_lowkre, 'classification'] = 'rejected'
ct_df.loc[is_lowkre, 'rationale'] += 'low rho, kappa, and varex;'
# Reject if low variance explained
is_lows = ct_df['normalized variance explained'] <= varex_norm_min
ct_df.loc[is_lows, 'classification'] = 'rejected'
ct_df.loc[is_lows, 'rationale'] += 'low variance explained;'
# Reject if Kappa over limit
is_fmax1 = ct_df['kappa'] == F_MAX
ct_df.loc[is_fmax1, 'classification'] = 'rejected'
ct_df.loc[is_fmax1, 'rationale'] += 'kappa equals fmax;'
# Reject if Rho over limit
is_fmax2 = ct_df['rho'] == F_MAX
ct_df.loc[is_fmax2, 'classification'] = 'rejected'
ct_df.loc[is_fmax2, 'rationale'] += 'rho equals fmax;'
if stabilize:
temp7 = varex_norm_cum >= 0.95
ct_df.loc[temp7, 'classification'] = 'rejected'
ct_df.loc[temp7, 'rationale'] += 'cumulative var. explained above 95%;'
under_fmin1 = ct_df['kappa'] <= fmin
ct_df.loc[under_fmin1, 'classification'] = 'rejected'
ct_df.loc[under_fmin1, 'rationale'] += 'kappa below fmin;'
under_fmin2 = ct_df['rho'] <= fmin
ct_df.loc[under_fmin2, 'classification'] = 'rejected'
ct_df.loc[under_fmin2, 'rationale'] += 'rho below fmin;'
ct_df.to_csv('comp_table_pca.txt',
sep='\t',
index=True,
index_label='component',
float_format='%.6f')
sel_idx = ct_df['classification'] == 'accepted'
n_components = np.sum(sel_idx)
voxel_kept_comp_weighted = (voxel_comp_weights[:, sel_idx] *
varex[None, sel_idx])
kept_data = np.dot(voxel_kept_comp_weighted, comp_ts[sel_idx, :])
if wvpca:
kept_data = idwtmat(kept_data, cAl)
LGR.info('Selected {0} components with Kappa threshold: {1:.02f}, '
'Rho threshold: {2:.02f}'.format(n_components, kappa_thr,
rho_thr))
kept_data = stats.zscore(kept_data,
axis=1) # variance normalize timeseries
kept_data = stats.zscore(kept_data,
axis=None) # variance normalize everything
return n_components, kept_data
def tedpca(catd, OCcatd, combmode, mask, t2s, t2sG, stabilize,
ref_img, tes, kdaw, rdaw, ste=0, mlepca=True):
"""
Use principal components analysis (PCA) to identify and remove thermal
noise from multi-echo data.
Parameters
----------
catd : (S x E x T) array_like
Input functional data
OCcatd : (S x T) array_like
Optimally-combined time series data
combmode : {'t2s', 'ste'} str
How optimal combination of echos should be made, where 't2s' indicates
using the method of Posse 1999 and 'ste' indicates using the method of
Poser 2006
mask : (S,) array_like
Boolean mask array
stabilize : bool
Whether to attempt to stabilize convergence of ICA by returning
dimensionally-reduced data from PCA and component selection.
ref_img : str or img_like
Reference image to dictate how outputs are saved to disk
tes : list
List of echo times associated with `catd`, in milliseconds
kdaw : float
Dimensionality augmentation weight for Kappa calculations
rdaw : float
Dimensionality augmentation weight for Rho calculations
ste : int or list-of-int, optional
Which echos to use in PCA. Values -1 and 0 are special, where a value
of -1 will indicate using all the echos and 0 will indicate using the
optimal combination of the echos. A list can be provided to indicate
a subset of echos. Default: 0
mlepca : bool, optional
Whether to use the method originally explained in Minka, NIPS 2000 for
guessing PCA dimensionality instead of a traditional SVD. Default: True
Returns
-------
n_components : int
Number of components retained from PCA decomposition
dd : (S x E x T) :obj:`numpy.ndarray`
Dimensionally-reduced functional data
Notes
-----
====================== =================================================
Notation Meaning
====================== =================================================
:math:`\\kappa` Component pseudo-F statistic for TE-dependent
(BOLD) model.
:math:`\\rho` Component pseudo-F statistic for TE-independent
(artifact) model.
:math:`v` Voxel
:math:`V` Total number of voxels in mask
:math:`\\zeta` Something
:math:`c` Component
:math:`p` Something else
====================== =================================================
Steps:
1. Variance normalize either multi-echo or optimally combined data,
depending on settings.
2. Decompose normalized data using PCA or SVD.
3. Compute :math:`{\\kappa}` and :math:`{\\rho}`:
.. math::
{\\kappa}_c = \\frac{\sum_{v}^V {\\zeta}_{c,v}^p * \
F_{c,v,R_2^*}}{\sum {\\zeta}_{c,v}^p}
{\\rho}_c = \\frac{\sum_{v}^V {\\zeta}_{c,v}^p * \
F_{c,v,S_0}}{\sum {\\zeta}_{c,v}^p}
4. Some other stuff. Something about elbows.
5. Classify components as thermal noise if they meet both of the
following criteria:
- Nonsignificant :math:`{\\kappa}` and :math:`{\\rho}`.
- Nonsignificant variance explained.
"""
n_samp, n_echos, n_vols = catd.shape
ste = np.array([int(ee) for ee in str(ste).split(',')])
if len(ste) == 1 and ste[0] == -1:
LGR.info('Computing PCA of optimally combined multi-echo data')
d = OCcatd[utils.make_min_mask(OCcatd[:, np.newaxis, :])][:, np.newaxis, :]
elif len(ste) == 1 and ste[0] == 0:
LGR.info('Computing PCA of spatially concatenated multi-echo data')
d = catd[mask].astype('float64')
else:
LGR.info('Computing PCA of echo #%s' % ','.join([str(ee) for ee in ste]))
d = np.stack([catd[mask, ee] for ee in ste - 1], axis=1).astype('float64')
eim = np.squeeze(eimask(d))
d = np.squeeze(d[eim])
dz = ((d.T - d.T.mean(axis=0)) / d.T.std(axis=0)).T # var normalize ts
dz = (dz - dz.mean()) / dz.std() # var normalize everything
if not op.exists('pcastate.pkl'):
# do PC dimension selection and get eigenvalue cutoff
if mlepca:
from sklearn.decomposition import PCA
ppca = PCA(n_components='mle', svd_solver='full')
ppca.fit(dz)
v = ppca.components_
s = ppca.explained_variance_
u = np.dot(np.dot(dz, v.T), np.diag(1. / s))
else:
u, s, v = np.linalg.svd(dz, full_matrices=0)
# actual variance explained (normalized)
sp = s / s.sum()
eigelb = getelbow_mod(sp, val=True)
spdif = np.abs(np.diff(sp))
spdifh = spdif[(len(spdif)//2):]
spdthr = np.mean([spdifh.max(), spdif.min()])
spmin = sp[(len(spdif)//2) + np.arange(len(spdifh))[spdifh >= spdthr][0] + 1]
spcum = np.cumsum(sp)
# Compute K and Rho for PCA comps
eimum = np.atleast_2d(eim)
eimum = np.transpose(eimum, np.argsort(eimum.shape)[::-1])
eimum = eimum.prod(axis=1)
o = np.zeros((mask.shape[0], *eimum.shape[1:]))
o[mask] = eimum
eimum = np.squeeze(o).astype(bool)
vTmix = v.T
vTmixN = ((vTmix.T - vTmix.T.mean(0)) / vTmix.T.std(0)).T
LGR.info('Making initial component selection guess from PCA results')
_, ctb, betasv, v_T = model.fitmodels_direct(catd, v.T, eimum, t2s, t2sG,
tes, combmode, ref_img,
mmixN=vTmixN, full_sel=False)
ctb = ctb[ctb[:, 0].argsort(), :]
ctb = np.vstack([ctb.T[:3], sp]).T
# Save state
fname = op.abspath('pcastate.pkl')
LGR.info('Saving PCA results to: {}'.format(fname))
pcastate = {'u': u, 's': s, 'v': v, 'ctb': ctb,
'eigelb': eigelb, 'spmin': spmin, 'spcum': spcum}
try:
with open(fname, 'wb') as handle:
pickle.dump(pcastate, handle)
except TypeError:
LGR.warning('Could not save PCA solution')
else: # if loading existing state
LGR.info('Loading PCA from: {}'.format('pcastate.pkl'))
with open('pcastate.pkl', 'rb') as handle:
pcastate = pickle.load(handle)
u, s, v = pcastate['u'], pcastate['s'], pcastate['v']
ctb, eigelb = pcastate['ctb'], pcastate['eigelb']
spmin, spcum = pcastate['spmin'], pcastate['spcum']
np.savetxt('comp_table_pca.txt', ctb[ctb[:, 1].argsort(), :][::-1])
np.savetxt('mepca_mix.1D', v[ctb[:, 1].argsort()[::-1], :].T)
kappas = ctb[ctb[:, 1].argsort(), 1]
rhos = ctb[ctb[:, 2].argsort(), 2]
fmin, fmid, fmax = utils.getfbounds(n_echos)
kappa_thr = np.average(sorted([fmin, getelbow_mod(kappas, val=True)/2, fmid]),
weights=[kdaw, 1, 1])
rho_thr = np.average(sorted([fmin, getelbow_cons(rhos, val=True)/2, fmid]),
weights=[rdaw, 1, 1])
if int(kdaw) == -1:
kappas_lim = kappas[utils.andb([kappas < fmid, kappas > fmin]) == 2]
kappa_thr = kappas_lim[getelbow_mod(kappas_lim)]
rhos_lim = rhos[utils.andb([rhos < fmid, rhos > fmin]) == 2]
rho_thr = rhos_lim[getelbow_mod(rhos_lim)]
stabilize = True
if int(kdaw) != -1 and int(rdaw) == -1:
rhos_lim = rhos[utils.andb([rhos < fmid, rhos > fmin]) == 2]
rho_thr = rhos_lim[getelbow_mod(rhos_lim)]
is_hik = np.array(ctb[:, 1] > kappa_thr, dtype=np.int)
is_hir = np.array(ctb[:, 2] > rho_thr, dtype=np.int)
is_hie = np.array(ctb[:, 3] > eigelb, dtype=np.int)
is_his = np.array(ctb[:, 3] > spmin, dtype=np.int)
is_not_fmax1 = np.array(ctb[:, 1] != F_MAX, dtype=np.int)
is_not_fmax2 = np.array(ctb[:, 2] != F_MAX, dtype=np.int)
pcscore = (is_hik + is_hir + is_hie) * is_his * is_not_fmax1 * is_not_fmax2
if stabilize:
temp7 = np.array(spcum < 0.95, dtype=np.int)
temp8 = np.array(ctb[:, 2] > fmin, dtype=np.int)
temp9 = np.array(ctb[:, 1] > fmin, dtype=np.int)
pcscore = pcscore * temp7 * temp8 * temp9
pcsel = pcscore > 0
dd = u.dot(np.diag(s*np.array(pcsel, dtype=np.int))).dot(v)
n_components = s[pcsel].shape[0]
LGR.info('Selected {0} components with Kappa threshold: {1:.02f}, '
'Rho threshold: {2:.02f}'.format(n_components, kappa_thr, rho_thr))
dd = stats.zscore(dd.T, axis=0).T # variance normalize timeseries
dd = stats.zscore(dd, axis=None) # variance normalize everything
return n_components, dd
def selcomps(seldict, mmix, mask, ref_img, manacc, n_echos, t2s, s0, olevel=2,
oversion=99, filecsdata=True, savecsdiag=True, strict_mode=False):
"""
Labels ICA components to keep or remove from denoised data
The selection process uses pre-calculated parameters for each ICA component
inputted into this function in `seldict` such as
Kappa (a T2* weighting metric), Rho (an S0 weighting metric), and variance
explained. Additonal selection metrics are calculated within this function
and then used to classify each component into one of four groups.
Parameters
----------
seldict : :obj:`dict`
As output from `fitmodels_direct`
mmix : (C x T) array_like
Mixing matrix for converting input data to component space, where `C`
is components and `T` is the number of volumes in the original data
mask : (S,) array_like
Boolean mask array
ref_img : :obj:`str` or img_like
Reference image to dictate how outputs are saved to disk
manacc : :obj:`list`
Comma-separated list of indices of manually accepted components
n_echos : :obj:`int`
Number of echos in original data
t2s : (S,) array_like
Estimated T2* map
s0 : (S,) array_like
S0 map
olevel : :obj:`int`, optional
Default: 2
oversion : :obj:`int`, optional
Default: 99
filecsdata: :obj:`bool`, optional
Default: False
savecsdiag: :obj:`bool`, optional
Default: True
strict_mode: :obj:`bool`, optional
Default: False
Returns
-------
acc : :obj:`list`
Indices of accepted (BOLD) components in `mmix`
rej : :obj:`list`
Indices of rejected (non-BOLD) components in `mmix`
midk : :obj:`list`
Indices of mid-K (questionable) components in `mmix`
These components are typically removed from the data during denoising
ign : :obj:`list`
Indices of ignored components in `mmix`
Ignored components are considered to have too low variance to matter.
They are not processed through the accept vs reject decision tree and
are NOT removed during the denoising process
Notes
-----
The selection algorithm used in this function is from work by prantikk
It is from selcomps function in select_model_fft20e.py in
version 3.2 of MEICA at:
https://github.com/ME-ICA/me-ica/blob/b2781dd087ab9de99a2ec3925f04f02ce84f0adc/meica.libs/select_model_fft20e.py
Many of the early publications using and evaulating the MEICA method used a
different selection algorithm by prantikk. The final 2.5 version of that
algorithm in the selcomps function in select_model.py at:
https://github.com/ME-ICA/me-ica/blob/b2781dd087ab9de99a2ec3925f04f02ce84f0adc/meica.libs/select_model.py
In both algorithms, the ICA component selection process uses multiple
metrics that include: kappa, rho, variance explained, compent spatial
weighting maps, noise and spatial frequency metrics, and measures of
spatial overlap across metrics. The precise calculations may vary between
algorithms. The most notable difference is that the v2.5 algorithm is a
fixed decision tree where all sections were made based on whether
combinations of metrics crossed various thresholds. In the v3.5 algorithm,
clustering and support vector machines are also used to classify components
based on how similar metrics in one component are similar to metrics in
other components.
"""
if mmix.ndim != 2:
raise ValueError('Parameter mmix should be 2d, not {0}d'.format(mmix.ndim))
elif t2s.ndim != 1: # FIT not necessarily supported
raise ValueError('Parameter t2s should be 1d, not {0}d'.format(t2s.ndim))
elif s0.ndim != 1: # FIT not necessarily supported
raise ValueError('Parameter s0 should be 1d, not {0}d'.format(s0.ndim))
elif not (t2s.shape[0] == s0.shape[0] == mask.shape[0]):
raise ValueError('First dimensions (number of samples) of t2s ({0}), '
's0 ({1}), and mask ({2}) do not '
'match'.format(t2s.shape[0], s0.shape[0], mask.shape[0]))
"""
handwerkerd and others are working to "hypercomment" this function to
help everyone understand it sufficiently with the goal of eventually
modularizing the algorithm. This is still a work-in-process with later
sections not fully commented, some points of uncertainty are noted, and the
summary of the full algorithm is not yet complete.
There are sections of this code that calculate metrics that are used in
the decision tree for the selection process and other sections that
are part of the decision tree. Certain comments are prefaced with METRIC
and variable names to make clear which are metrics and others are prefaced
with SELECTION to make clear which are for applying metrics. METRICs tend
to be summary values that contain a signal number per component.
Note there are some variables that are calculated in one section of the code
that are later transformed into another metric that is actually part of a
selection criterion. This running list is an attempt to summarize
intermediate metrics vs the metrics that are actually used in decision
steps. For applied metrics that are made up of intermediate metrics defined
in earlier sections of the code, the constituent metrics are noted. More
metrics will be added to the applied metrics section as the commenting of
this function continues.
Intermediate Metrics: seldict['F_S0_clmaps'] seldict['F_R2_clmaps']
seldict['Br_clmaps_S0'] seldict['Br_clmaps_R2'] seldict['Z_maps']
dice_tbl countnoise
counts_FR2_Z tt_table mmix_kurt mmix_std
spr fproj_arr_val fdist
Rtz, Dz
Applied Metrics:
seldict['Rhos']
seldict['Kappas']
seldict['varex']
countsigFS0
countsigFR2
fz (a combination of multiple z-scored metrics: tt_table,
seldict['varex'], seldict['Kappa'], seldict['Rho'], countnoise,
mmix_kurt, fdist)
tt_table[:,0]
spz (z score of spr)
KRcut
"""
"""
If seldict exists, save it into a pickle file called compseldata.pklbz
that can be loaded directly into python for future analyses
If seldict=None, load it from the pre-saved pickle file to use for the
rest of this function
"""
if filecsdata:
import bz2
if seldict is not None:
LGR.info('Saving component selection data')
with bz2.BZ2File('compseldata.pklbz', 'wb') as csstate_f:
pickle.dump(seldict, csstate_f)
else:
try:
with bz2.BZ2File('compseldata.pklbz', 'rb') as csstate_f:
seldict = pickle.load(csstate_f)
except FileNotFoundError:
LGR.warning('Failed to load component selection data')
return None
"""
List of components
all_comps and acc_comps start out as an ordered list of the component numbers
all_comps is constant throughout the function.
acc_comps changes through his function as components are assigned to other
categories (i.e. components that are classified as rejected are removed
from acc_comps)
"""
midk = []
ign = []
all_comps = np.arange(len(seldict['Kappas']))
acc_comps = np.arange(len(seldict['Kappas']))
"""
If user has specified components to accept manually, just assign those
components to the accepted and rejected comp lists and end the function
"""
if manacc:
acc = sorted([int(vv) for vv in manacc.split(',')])
midk = []
rej = sorted(np.setdiff1d(all_comps, acc))
ign = []
return acc, rej, midk, ign # Add string for ign
"""
METRICS: countsigFS0 countsigFR2
F_S0_clmaps & F_R2_clmaps are the thresholded & binarized clustered maps of
significant fits for the separate S0 and R2 cross-echo models per component.
Since the values are 0 or 1, the countsig variables are a count of the
significant voxels per component.
The cluster size is a function of the # of voxels in the mask.
The cluster threshold is based on the # of echos acquired
"""
countsigFS0 = seldict['F_S0_clmaps'].sum(0)
countsigFR2 = seldict['F_R2_clmaps'].sum(0)
countnoise = np.zeros(len(all_comps))
"""
Make table of dice values
METRICS: dice_tbl
dice_FR2, dice_FS0 are calculated for each component and the concatenated
values are in dice_tbl
Br_clmaps_R2 and Br_clmaps_S0 are binarized clustered Z_maps.
The volume being clustered is the rank order indices of the absolute value
of the beta values for the fit between the optimally combined time series
and the mixing matrix (i.e. the lowest beta value is 1 and the highest is
the # of voxels).
The cluster size is a function of the # of voxels in the mask.
The cluster threshold are the voxels with beta ranks greater than
countsigFS0 or countsigFR2 (i.e. roughly the same number of voxels will be
in the countsig clusters as the ICA beta map clusters)
These dice values are the Dice-Sorenson index for the Br_clmap_?? and the
F_??_clmap.
If handwerkerd understands this correctly, if the voxels with the above
threshold F stats are clustered in the same voxels with the highest beta
values, then the dice coefficient will be 1. If the thresholded F or betas
aren't spatially clustered (i.e. the component map is less spatially smooth)
or the clusters are in different locations (i.e. voxels with high betas
are also noiser so they have lower F values), then the dice coefficients
will be lower
"""
dice_tbl = np.zeros([all_comps.shape[0], 2])
for comp_num in all_comps:
dice_FR2 = utils.dice(utils.unmask(seldict['Br_clmaps_R2'][:, comp_num],
mask)[t2s != 0],
seldict['F_R2_clmaps'][:, comp_num])
dice_FS0 = utils.dice(utils.unmask(seldict['Br_clmaps_S0'][:, comp_num],
mask)[t2s != 0],
seldict['F_S0_clmaps'][:, comp_num])
dice_tbl[comp_num, :] = [dice_FR2, dice_FS0] # step 3a here and above
dice_tbl[np.isnan(dice_tbl)] = 0
"""
Make table of noise gain
METRICS: countnoise, counts_FR2_Z, tt_table
(This is a bit confusing & is handwerkerd's attempt at making sense of this)
seldict['Z_maps'] is the Fisher Z normalized beta fits for the optimally
combined time series and the mixing matrix. Z_clmaps is a binarized cluster
of Z_maps with the cluster size based on the # of voxels and the cluster
threshold of 1.95. utils.andb is a sum of the True values in arrays so
comp_noise_sel is true for voxels where the Z values are greater than 1.95
but not part of a cluster of Z values that are greater than 1.95.
Spatially unclustered voxels with high Z values could be considerd noisy.
countnoise is the # of voxels per component where comp_noise_sel is true.
counts_FR2_Z is the number of voxels with Z values above the threshold
that are in clusters (signal) and the number outside of clusters (noise)
tt_table is a bit confusing. For each component, the first index is
some type of normalized, log10, signal/noise t statistic and the second is
the p value for the signal/noise t statistic (for the R2 model).
In general, these should be bigger t or have lower p values when most of
the Z values above threshold are inside clusters.
Because of the log10, values below 1 are negative, which is later used as
a threshold. It doesn't seem like the p values are ever used.
"""
tt_table = np.zeros([len(all_comps), 4])
counts_FR2_Z = np.zeros([len(all_comps), 2])
for comp_num in all_comps:
comp_noise_sel = utils.andb([np.abs(seldict['Z_maps'][:, comp_num]) > 1.95,
seldict['Z_clmaps'][:, comp_num] == 0]) == 2
countnoise[comp_num] = np.array(comp_noise_sel, dtype=np.int).sum()
noise_FR2_Z_mask = utils.unmask(comp_noise_sel, mask)[t2s != 0]
noise_FR2_Z = np.log10(np.unique(seldict['F_R2_maps'][noise_FR2_Z_mask, comp_num]))
signal_FR2_Z_mask = utils.unmask(seldict['Z_clmaps'][:, comp_num], mask)[t2s != 0] == 1
signal_FR2_Z = np.log10(np.unique(seldict['F_R2_maps'][signal_FR2_Z_mask, comp_num]))
counts_FR2_Z[comp_num, :] = [len(signal_FR2_Z), len(noise_FR2_Z)]
ttest = stats.ttest_ind(signal_FR2_Z, noise_FR2_Z, equal_var=True)
# avoid DivideByZero RuntimeWarning
if signal_FR2_Z.size > 0 and noise_FR2_Z.size > 0:
mwu = stats.norm.ppf(stats.mannwhitneyu(signal_FR2_Z, noise_FR2_Z)[1])
else:
mwu = -np.inf
tt_table[comp_num, 0] = np.abs(mwu) * ttest[0] / np.abs(ttest[0])
tt_table[comp_num, 1] = ttest[1]
tt_table[np.isnan(tt_table)] = 0
tt_table[np.isinf(tt_table[:, 0]), 0] = np.percentile(tt_table[~np.isinf(tt_table[:, 0]), 0],
98)
"""
Time series derivative kurtosis
METRICS: mmix_kurt and mmix_std
Take the derivative of the time series for each component in the ICA
mixing matrix and calculate the kurtosis & standard deviation.
handwerkerd thinks these metrics are later used to calculate measures
of time series spikiness and drift in the component time series.
"""
mmix_dt = (mmix[:-1, :] - mmix[1:, :])
mmix_kurt = stats.kurtosis(mmix_dt)
mmix_std = np.std(mmix_dt, axis=0)
"""
SELECTION #1 (prantikk labeled "Step 1")
Reject anything that is obviously an artifact
Obvious artifacts are components with Rho>Kappa or with more clustered,
significant voxels for the S0 model than the R2 model
"""
LGR.debug('Rejecting gross artifacts based on Rho/Kappa values and S0/R2 '
'counts')
rej = acc_comps[utils.andb([seldict['Rhos'] > seldict['Kappas'],
countsigFS0 > countsigFR2]) > 0]
acc_comps = np.setdiff1d(acc_comps, rej)
"""
prantikk labeled "Step 2"
Compute 3-D spatial FFT of Beta maps to detect high-spatial
frequency artifacts
METRIC spr, fproj_arr_val, fdist
PSC is the mean centered beta map for each ICA component
The FFT is sequentially calculated across each dimension of PSC & the max
value is removed (probably the 0Hz bin). The maximum remaining frequency
magnitude along the z dimenions is calculated leaving a 2D matrix.
spr contains a count of the number of frequency bins in the 2D matrix where
the frequency magnitude is greater than 4* the maximum freq in the matrix.
spr is later z-normed across components into spz and this is actually used
as a selection metric.
handwerkerd interpretation: spr is bigger the more values of the fft are
>1/4 the max. Thus, if you assume the highest mag bins are low frequency, &
all components have roughly the same low freq power (i.e. a brain-shaped
blob), then spr will be bigger the more high frequency bins have magnitudes
that are more than 1/4 of the low frequency bins.
fproj_arr_val is a flattened 1D vector of the 2D max projection fft
of each component. This seems to be later used in an SVM to train on
this value for rejected components to classify some remaining n_components
as midk
Note: fproj_arr is created here and is a ranked list of FFT values, but is
not used anywhere in the code. Was fproj_arr_val supposed to contain this
ranking?
fdist isn't completely clear to handwerkerd yet but it looks like the fit of
the fft of the spatial map to a Gaussian distribution. If so, then the
worse the fit, the more high frequency power would be in the component
"""
LGR.debug('Computing 3D spatial FFT of beta maps to detect high-spatial frequency artifacts')
# spatial information is important so for NIFTI we convert back to 3D space
if utils.get_dtype(ref_img) == 'NIFTI':
dim1 = np.prod(check_niimg(ref_img).shape[:2])
else:
dim1 = mask.shape[0]
fproj_arr = np.zeros([dim1, len(all_comps)])
fproj_arr_val = np.zeros([dim1, len(all_comps)])
spr = []
fdist = []
for comp_num in all_comps:
# convert data back to 3D array
if utils.get_dtype(ref_img) == 'NIFTI':
tproj = utils.new_nii_like(ref_img, utils.unmask(seldict['PSC'],
mask)[:, comp_num]).get_data()
else:
tproj = utils.unmask(seldict['PSC'], mask)[:, comp_num]
fproj = np.fft.fftshift(np.abs(np.fft.rfftn(tproj)))
fproj_z = fproj.max(axis=-1)
fproj[fproj == fproj.max()] = 0
spr.append(np.array(fproj_z > fproj_z.max() / 4, dtype=np.int).sum())
fproj_arr[:, comp_num] = stats.rankdata(fproj_z.flatten())
fproj_arr_val[:, comp_num] = fproj_z.flatten()
if utils.get_dtype(ref_img) == 'NIFTI':
fprojr = np.array([fproj, fproj[:, :, ::-1]]).max(0)
fdist.append(np.max([utils.fitgaussian(fproj.max(jj))[3:].max() for
jj in range(fprojr.ndim)]))
else:
fdist = np.load(os.path.join(RESOURCES, 'fdist.npy'))
if type(fdist) is not np.ndarray:
fdist = np.array(fdist)
spr = np.array(spr)
# import ipdb; ipdb.set_trace()
"""
prantikk labelled Step 3
Create feature space of component properties
METRIC fz, spz, Rtz, Dz
fz is matrix of multiple other metrics described above and calculated
in this section. Most are all of these have one number per component and
they are z-scored across components
Attempted explanations in order:
Tz: The z-scored t statistics of the spatial noisiness metric in tt_table
Vz: The z-scored the natural log of the non-normalized variance explained
of each component
Ktz: The z-scored natural log of the Kappa values
(the '/ 2' does not seem necessary beacuse it will be removed by z-scoring)
KRr: The z-scored ratio of the natural log of Kappa / nat log of Rho
(unclear why sometimes using stats.zcore and other times writing the eq out)
cnz: The z-scored measure of a noisy voxel count where the noisy voxels are
the voxels with large beta estimates, but aren't part of clusters
Rz: z-scored rho values (why aren't this log scaled, like kappa in Ktz?)
mmix_kurt: Probably a rough measure of the spikiness of each component's
time series in the ICA mixing matrix
fdist_z: z-score of fdist, which is probably a measure of high freq info
in the spatial FFT of the components (with lower being more high freq?)
NOT in fz:
spz: Z-scored measure probably of how much high freq is in the data. Larger
values mean more bins of the FFT have over 1/4 the power of the maximum
bin (read about spr above for more info)
Rtz: Z-scored natural log of the Rho values
Dz: Z-scored Fisher Z transformed dice values of the overlap between
clusters for the F stats and clusters of the ICA spatial beta maps with
roughly the same number of voxels as in the clustered F maps.
Dz saves this for the R2 model, there are also Dice coefs for the S0
model in dice_tbl
"""
LGR.debug('Creating feature space of component properties')
fdist_pre = fdist.copy()
fdist_pre[fdist > np.median(fdist) * 3] = np.median(fdist) * 3
fdist_z = (fdist_pre - np.median(fdist_pre)) / fdist_pre.std() # not z
spz = stats.zscore(spr)
Tz = stats.zscore(tt_table[:, 0])
varex_log = np.log(seldict['varex'])
Vz = stats.zscore(varex_log)
Rz = stats.zscore(seldict['Rhos'])
Ktz = stats.zscore(np.log(seldict['Kappas']) / 2)
# Rtz = stats.zscore(np.log(seldict['Rhos']) / 2)
KRr = stats.zscore(np.log(seldict['Kappas']) / np.log(seldict['Rhos']))
cnz = stats.zscore(countnoise)
Dz = stats.zscore(np.arctanh(dice_tbl[:, 0] + 0.001))
fz = np.array([Tz, Vz, Ktz, KRr, cnz, Rz, mmix_kurt, fdist_z])
"""
METRICS Kcut, Rcut, KRcut, KRcutguesses, Khighelbowval
Step 3: Make initial guess of where BOLD components are and use DBSCAN
to exclude noise components and find a sample set of 'good' components
"""
LGR.debug('Making initial guess of BOLD components')
# The F threshold for the echo fit (based on the # of echos) for p<0.05
# p<0.025, and p<0.001 (Confirm this is accurate since the function
# contains a lookup table rather than a calculation)
F05, F025, F01 = utils.getfbounds(n_echos)
# epsmap is [index,level of overlap with dicemask,
# number of high Rho components]
epsmap = []
Rhos_sorted = np.array(sorted(seldict['Rhos']))[::-1]
"""
Make an initial guess as to number of good components based on
consensus of control points across Rhos and Kappas
For terminology later, typically getelbow _aggr > _mod > _cons
though this might not be universally true. A more "inclusive" threshold
has a lower kappa since that means more components are above that thresh
and are likely to be accepted. For Rho, a more "inclusive" threshold is
higher since that means fewer components will be rejected based on rho.
KRcut seems weird to handwerkerd. I see that the thresholds are slightly
shifted for kappa & rho later in the code, but why would we ever want to
set a common threhsold reference point for both? These are two different
elbows on two different data sets.
"""
KRcutguesses = [getelbow_mod(seldict['Rhos']),
getelbow_cons(seldict['Rhos']),
getelbow_aggr(seldict['Rhos']),
getelbow_mod(seldict['Kappas']),
getelbow_cons(seldict['Kappas']),
getelbow_aggr(seldict['Kappas'])]
KRcut = np.median(KRcutguesses)
"""
Also a bit weird to handwerkerd. This is the 75th percentile of Kappa F
stats of the components with the 3 elbow selection criteria and the
F states for 3 significance thresholds based on the # of echos.
This is some type of way to get a significance criterion for a component
fit, but it's include why this specific criterion is useful.
"""
Khighelbowval = stats.scoreatpercentile([getelbow_mod(seldict['Kappas'],
return_val=True),
getelbow_cons(seldict['Kappas'],
return_val=True),
getelbow_aggr(seldict['Kappas'],
return_val=True)] +
list(utils.getfbounds(n_echos)),
75, interpolation_method='lower')
"""
Default to the most inclusive kappa threshold (_cons) unless:
1. That threshold is more than twice the median of Kappa & Rho thresholds
2. and the moderate elbow is more inclusive than a p=0.01
handwerkerd: This actually seems like a way to avoid using the theoretically
most liberal threshold only when there was a bad estimate and _mod is
is more inclusive. My one concern is that it's an odd way to test that
the _mod elbow is any better. Why not at least see if _mod < _cons?
prantikk's orig comment for this section is:
"only use exclusive when inclusive is extremely inclusive - double KRcut"
"""
cond1 = getelbow_cons(seldict['Kappas']) > KRcut * 2
cond2 = getelbow_mod(seldict['Kappas'], return_val=True) < F01
if cond1 and cond2:
Kcut = getelbow_mod(seldict['Kappas'], return_val=True)
else:
Kcut = getelbow_cons(seldict['Kappas'], return_val=True)
"""
handwerkerd: The goal seems to be to maximize the rejected components
based on the rho cut by defaulting to a lower Rcut value. Again, if
that is the goal, why not just test if _mod < _cons?
prantikk's orig comment for this section is:
only use inclusive when exclusive is extremely exclusive - half KRcut
(remember for Rho inclusive is higher, so want both Kappa and Rho
to defaut to lower)
"""
if getelbow_cons(seldict['Rhos']) > KRcut * 2:
Rcut = getelbow_mod(seldict['Rhos'], return_val=True)
# for above, consider something like:
# min([getelbow_mod(Rhos,True),sorted(Rhos)[::-1][KRguess] ])
else:
Rcut = getelbow_cons(seldict['Rhos'], return_val=True)
# Rcut should never be higher than Kcut (handwerkerd: not sure why)
if Rcut > Kcut:
Kcut = Rcut
# KRelbow has a 2 for components that are above the Kappa accept threshold
# and below the rho reject threshold
KRelbow = utils.andb([seldict['Kappas'] > Kcut,
seldict['Rhos'] < Rcut])
"""
Make guess of Kundu et al 2011 plus remove high frequencies,
generally high variance, and high variance given low Kappa
the first index of tt_table is a t static of a what handwerkerd thinks
is a spatial noise metric. Since log10 of these values are taken the >0
threshold means the metric is >1. tt_lim seems to be a fairly aggressive
percentile that is then divided by 3.
"""
tt_lim = stats.scoreatpercentile(tt_table[tt_table[:, 0] > 0, 0],
75, interpolation_method='lower') / 3
"""
KRguess is a list of components to potentially accept. it starts with a
list of components that cross the Kcut and Rcut threshold and weren't
previously rejected for other reasons. From that list, it removes more
components based on several additional criteria:
1. tt_table less than the tt_lim threshold (spatial noisiness metric)
2. spz (a z-scored probably high spatial freq metric) >1
3. Vz (a z-scored variance explained metric) >2
4. If both (seems to be if a component has a relatively high variance
the acceptance threshold for Kappa values is doubled):
A. The variance explained is greater than half the KRcut highest
variance component
B. Kappa is less than twice Kcut
"""
temp = all_comps[utils.andb([seldict['varex'] > 0.5 *
sorted(seldict['varex'])[::-1][int(KRcut)],
seldict['Kappas'] < 2*Kcut]) == 2]
KRguess = np.setdiff1d(np.setdiff1d(all_comps[KRelbow == 2], rej),
np.union1d(all_comps[tt_table[:, 0] < tt_lim],
np.union1d(np.union1d(all_comps[spz > 1],
all_comps[Vz > 2]),
temp)))
guessmask = np.zeros(len(all_comps))
guessmask[KRguess] = 1
"""
Throw lower-risk bad components out based on 3 criteria all being true:
1. tt_table (a spatial noisiness metric) <0
2. A components variance explains is greater than the median variance
explained
3. The component index is greater than the KRcut index. Since the
components are sorted by kappa, this is another kappa thresholding)
"""
rejB = acc_comps[utils.andb([tt_table[acc_comps, 0] < 0,
seldict['varex'][acc_comps] > np.median(seldict['varex']),
acc_comps > KRcut]) == 3]
rej = np.union1d(rej, rejB)
# adjust acc_comps again to only contain the remaining non-rejected components
acc_comps = np.setdiff1d(acc_comps, rej)
"""
This is where handwerkerd has paused in hypercommenting the function.
"""
LGR.debug('Using DBSCAN to find optimal set of "good" BOLD components')
for ii in range(20000):
eps = .005 + ii * .005
db = DBSCAN(eps=eps, min_samples=3).fit(fz.T)
# it would be great to have descriptive names, here
# DBSCAN found at least three non-noisy clusters
cond1 = db.labels_.max() > 1
# DBSCAN didn't detect more classes than the total # of components / 6
cond2 = db.labels_.max() < len(all_comps) / 6
# TODO: confirm if 0 is a special label for DBSCAN
# my intuition here is that we're confirming DBSCAN labelled previously
# rejected components as noise (i.e., no overlap between `rej` and
# labelled DBSCAN components)
cond3 = np.intersect1d(rej, all_comps[db.labels_ == 0]).shape[0] == 0
# DBSCAN labelled less than half of the total components as noisy
cond4 = np.array(db.labels_ == -1, dtype=int).sum() / float(len(all_comps)) < .5
if cond1 and cond2 and cond3 and cond4:
epsmap.append([ii, utils.dice(guessmask, db.labels_ == 0),
np.intersect1d(all_comps[db.labels_ == 0],
all_comps[seldict['Rhos'] > getelbow_mod(Rhos_sorted,
return_val=True)]).shape[0]])
db = None
epsmap = np.array(epsmap)
LGR.debug('Found DBSCAN solutions for {}/20000 eps resolutions'.format(len(epsmap)))
group0 = []
dbscanfailed = False
if len(epsmap) != 0:
# Select index that maximizes Dice with guessmask but first
# minimizes number of higher Rho components
ii = int(epsmap[np.argmax(epsmap[epsmap[:, 2] == np.min(epsmap[:, 2]), 1], 0), 0])
LGR.debug('Component selection tuning: {:.05f}'.format(epsmap[:, 1].max()))
db = DBSCAN(eps=.005+ii*.005, min_samples=3).fit(fz.T)
acc_comps = all_comps[db.labels_ == 0]
acc_comps = np.setdiff1d(acc_comps, rej)
acc_comps = np.setdiff1d(acc_comps, acc_comps[acc_comps > len(all_comps) - len(rej)])
group0 = acc_comps.copy()
group_n1 = all_comps[db.labels_ == -1]
to_clf = np.setdiff1d(all_comps, np.union1d(acc_comps, rej))
if len(group0) == 0 or len(group0) < len(KRguess) * .5:
dbscanfailed = True
LGR.debug('DBSCAN guess failed; using elbow guess method instead')
temp = all_comps[utils.andb([seldict['varex'] > 0.5 *
sorted(seldict['varex'])[::-1][int(KRcut)],
seldict['Kappas'] < 2 * Kcut]) == 2]
acc_comps = np.setdiff1d(np.setdiff1d(all_comps[KRelbow == 2], rej),
np.union1d(all_comps[tt_table[:, 0] < tt_lim],
np.union1d(np.union1d(all_comps[spz > 1],
all_comps[Vz > 2]),
temp)))
group0 = acc_comps.copy()
group_n1 = []
to_clf = np.setdiff1d(all_comps, np.union1d(group0, rej))
if len(group0) < 2 or (len(group0) < 4 and float(len(rej))/len(group0) > 3):
LGR.warning('Extremely limited reliable BOLD signal space! '
'Not filtering components beyond BOLD/non-BOLD guesses.')
midkfailed = True
min_acc = np.array([])
if len(group0) != 0:
# For extremes, building in a 20% tolerance
toacc_hi = np.setdiff1d(all_comps[utils.andb([fdist <= np.max(fdist[group0]),
seldict['Rhos'] < F025,
Vz > -2]) == 3],
np.union1d(group0, rej))
min_acc = np.union1d(group0, toacc_hi)
to_clf = np.setdiff1d(all_comps, np.union1d(min_acc, rej))
else:
toacc_hi = []
min_acc = []
diagstep_keys = ['Rejected components', 'Kappa-Rho cut point',
'Kappa cut point', 'Rho cut point',
'DBSCAN failed to converge',
'Mid-Kappa failed (limited BOLD signal)',
'Kappa-Rho guess',
'min_acc', 'toacc_hi']
diagstep_vals = [list(rej), KRcut, Kcut, Rcut, dbscanfailed,
midkfailed, list(KRguess), list(min_acc), list(toacc_hi)]
with open('csstepdata.json', 'w') as ofh:
json.dump(dict(zip(diagstep_keys, diagstep_vals)), ofh,
indent=4, sort_keys=True, default=str)
return list(sorted(min_acc)), list(sorted(rej)), [], list(sorted(to_clf))
# Find additional components to reject based on Dice - doing this here
# since Dice is a little unstable, need to reference group0
rej_supp = []
dice_rej = False
if not dbscanfailed and len(rej) + len(group0) < 0.75 * len(all_comps):
dice_rej = True
temp = all_comps[dice_tbl[all_comps, 0] <= dice_tbl[all_comps, 1]]
rej_supp = np.setdiff1d(np.setdiff1d(np.union1d(rej, temp),
group0), group_n1)
rej = np.union1d(rej, rej_supp)
# Temporal features
# larger is worse - spike
mmix_kurt_z = (mmix_kurt-mmix_kurt[group0].mean()) / mmix_kurt[group0].std()
# smaller is worse - drift
mmix_std_z = -1 * ((mmix_std-mmix_std[group0].mean()) / mmix_std[group0].std())
mmix_kurt_z_max = np.max([mmix_kurt_z, mmix_std_z], 0)
"""
Step 2: Classifiy midk and ignore using separate SVMs for
different variance regimes
# To render hyperplane:
min_x = np.min(spz2);max_x=np.max(spz2)
# plotting separating hyperplane
ww = clf_.coef_[0]
aa = -ww[0] / ww[1]
# make sure the next line is long enough
xx = np.linspace(min_x - 2, max_x + 2)
yy = aa * xx - (clf_.intercept_[0]) / ww[1]
plt.plot(xx, yy, '-')
"""
LGR.debug('Attempting to classify midk components')
# Tried getting rid of accepting based on SVM altogether,
# now using only rejecting
toacc_hi = np.setdiff1d(all_comps[utils.andb([fdist <= np.max(fdist[group0]),
seldict['Rhos'] < F025, Vz > -2]) == 3],
np.union1d(group0, rej))
temp = utils.andb([spz < 1, Rz < 0,
mmix_kurt_z_max < 5,
Dz > -1, Tz > -1, Vz < 0,
seldict['Kappas'] >= F025,
fdist < 3 * np.percentile(fdist[group0], 98)]) == 8
toacc_lo = np.intersect1d(to_clf, all_comps[temp])
midk_clf, clf_ = do_svm(fproj_arr_val[:, np.union1d(group0, rej)].T,
[0] * len(group0) + [1] * len(rej),
fproj_arr_val[:, to_clf].T,
svmtype=2)
midk = np.setdiff1d(to_clf[utils.andb([midk_clf == 1, seldict['varex'][to_clf] >
np.median(seldict['varex'][group0])]) == 2],
np.union1d(toacc_hi, toacc_lo))
# only use SVM to augment toacc_hi only if toacc_hi isn't already
# conflicting with SVM choice
if len(np.intersect1d(to_clf[utils.andb([midk_clf == 1,
Vz[to_clf] > 0]) == 2],
toacc_hi)) == 0:
svm_acc_fail = True
toacc_hi = np.union1d(toacc_hi, to_clf[midk_clf == 0])
else:
svm_acc_fail = False
"""
Step 3: Compute variance associated with low T2* areas
(e.g. draining veins and low T2* areas)
# To write out veinmask
veinout = np.zeros(t2s.shape)
veinout[t2s!=0] = veinmaskf
utils.filewrite(veinout, 'veinmaskf', ref_img)
veinBout = utils.unmask(veinmaskB, mask)
utils.filewrite(veinBout, 'veins50', ref_img)
"""
LGR.debug('Computing variance associated with low T2* areas (e.g., '
'draining veins)')
tsoc_B_Zcl = np.zeros(seldict['tsoc_B'].shape)
tsoc_B_Zcl[seldict['Z_clmaps'] != 0] = np.abs(seldict['tsoc_B'])[seldict['Z_clmaps'] != 0]
sig_B = [stats.scoreatpercentile(tsoc_B_Zcl[tsoc_B_Zcl[:, ii] != 0, ii], 25)
if len(tsoc_B_Zcl[tsoc_B_Zcl[:, ii] != 0, ii]) != 0
else 0 for ii in all_comps]
sig_B = np.abs(seldict['tsoc_B']) > np.tile(sig_B, [seldict['tsoc_B'].shape[0], 1])
veinmask = utils.andb([t2s < stats.scoreatpercentile(t2s[t2s != 0], 15,
interpolation_method='lower'),
t2s != 0]) == 2
veinmaskf = veinmask[mask]
veinR = np.array(sig_B[veinmaskf].sum(0),
dtype=float) / sig_B[~veinmaskf].sum(0)
veinR[np.isnan(veinR)] = 0
veinc = np.union1d(rej, midk)
rej_veinRZ = ((veinR-veinR[veinc].mean())/veinR[veinc].std())[veinc]
rej_veinRZ[rej_veinRZ < 0] = 0
rej_veinRZ[countsigFR2[veinc] > np.array(veinmaskf, dtype=int).sum()] = 0
t2s_lim = [stats.scoreatpercentile(t2s[t2s != 0], 50,
interpolation_method='lower'),
stats.scoreatpercentile(t2s[t2s != 0], 80,
interpolation_method='lower') / 2]
phys_var_zs = []
for t2sl_i in range(len(t2s_lim)):
t2sl = t2s_lim[t2sl_i]
veinW = sig_B[:, veinc]*np.tile(rej_veinRZ, [sig_B.shape[0], 1])
veincand = utils.unmask(utils.andb([s0[t2s != 0] < np.median(s0[t2s != 0]),
t2s[t2s != 0] < t2sl]) >= 1,
t2s != 0)[mask]
veinW[~veincand] = 0
invein = veinW.sum(axis=1)[(utils.unmask(veinmaskf, mask) *
utils.unmask(veinW.sum(axis=1) > 1, mask))[mask]]
minW = 10 * (np.log10(invein).mean()) - 1 * 10**(np.log10(invein).std())
veinmaskB = veinW.sum(axis=1) > minW
tsoc_Bp = seldict['tsoc_B'].copy()
tsoc_Bp[tsoc_Bp < 0] = 0
vvex = np.array([(tsoc_Bp[veinmaskB, ii]**2.).sum() /
(tsoc_Bp[:, ii]**2.).sum() for ii in all_comps])
group0_res = np.intersect1d(KRguess, group0)
phys_var_zs.append((vvex - vvex[group0_res].mean()) / vvex[group0_res].std())
veinBout = utils.unmask(veinmaskB, mask)
utils.filewrite(veinBout.astype(float), 'veins_l%i' % t2sl_i, ref_img)
# Mask to sample veins
phys_var_z = np.array(phys_var_zs).max(0)
Vz2 = (varex_log - varex_log[group0].mean())/varex_log[group0].std()
"""
Step 4: Learn joint TE-dependence spatial and temporal models to move
remaining artifacts to ignore class
"""
LGR.debug('Learning joint TE-dependence spatial/temporal models to ignore remaining artifacts')
to_ign = []
minK_ign = np.max([F05, getelbow_cons(seldict['Kappas'], return_val=True)])
newcest = len(group0) + len(toacc_hi[seldict['Kappas'][toacc_hi] > minK_ign])
phys_art = np.setdiff1d(all_comps[utils.andb([phys_var_z > 3.5,
seldict['Kappas'] < minK_ign]) == 2], group0)
rank_diff = stats.rankdata(phys_var_z) - stats.rankdata(seldict['Kappas'])
phys_art = np.union1d(np.setdiff1d(all_comps[utils.andb([phys_var_z > 2,
rank_diff > newcest / 2,
Vz2 > -1]) == 3],
group0), phys_art)
# Want to replace field_art with an acf/SVM based approach
# instead of a kurtosis/filter one
field_art = np.setdiff1d(all_comps[utils.andb([mmix_kurt_z_max > 5,
seldict['Kappas'] < minK_ign]) == 2], group0)
temp = (stats.rankdata(mmix_kurt_z_max) - stats.rankdata(seldict['Kappas'])) > newcest / 2
field_art = np.union1d(np.setdiff1d(all_comps[utils.andb([mmix_kurt_z_max > 2,
temp,
Vz2 > 1,
seldict['Kappas'] < F01]) == 4],
group0), field_art)
temp = seldict['Rhos'] > np.percentile(seldict['Rhos'][group0], 75)
field_art = np.union1d(np.setdiff1d(all_comps[utils.andb([mmix_kurt_z_max > 3,
Vz2 > 3,
temp]) == 3],
group0), field_art)
field_art = np.union1d(np.setdiff1d(all_comps[utils.andb([mmix_kurt_z_max > 5, Vz2 > 5]) == 2],
group0), field_art)
misc_art = np.setdiff1d(all_comps[utils.andb([(stats.rankdata(Vz) -
stats.rankdata(Ktz)) > newcest / 2,
seldict['Kappas'] < Khighelbowval]) == 2], group0)
ign_cand = np.unique(list(field_art)+list(phys_art)+list(misc_art))
midkrej = np.union1d(midk, rej)
to_ign = np.setdiff1d(list(ign_cand), midkrej)
toacc = np.union1d(toacc_hi, toacc_lo)
acc_comps = np.setdiff1d(np.union1d(acc_comps, toacc), np.union1d(to_ign, midkrej))
ign = np.setdiff1d(all_comps, list(acc_comps) + list(midk) + list(rej))
orphan = np.setdiff1d(all_comps, list(acc_comps) + list(to_ign) + list(midk) + list(rej))
# Last ditch effort to save some transient components
if not strict_mode:
Vz3 = (varex_log - varex_log[acc_comps].mean()) / varex_log[acc_comps].std()
temp = utils.andb([seldict['Kappas'] > F05,
seldict['Rhos'] < F025,
seldict['Kappas'] > seldict['Rhos'],
Vz3 <= -1,
Vz3 > -3,
mmix_kurt_z_max < 2.5])
acc_comps = np.union1d(acc_comps,
np.intersect1d(orphan, all_comps[temp == 6]))
ign = np.setdiff1d(all_comps, list(acc_comps)+list(midk)+list(rej))
orphan = np.setdiff1d(all_comps, list(acc_comps) + list(to_ign) + list(midk) + list(rej))
if savecsdiag:
diagstep_keys = ['Rejected components', 'Kappa-Rho cut point', 'Kappa cut',
'Rho cut', 'DBSCAN failed to converge', 'Kappa-Rho guess',
'Dice rejected', 'rej_supp', 'to_clf',
'Mid-kappa components', 'svm_acc_fail', 'toacc_hi', 'toacc_lo',
'Field artifacts', 'Physiological artifacts',
'Miscellaneous artifacts', 'acc_comps', 'Ignored components']
diagstep_vals = [list(rej), KRcut.item(), Kcut.item(), Rcut.item(),
dbscanfailed, list(KRguess), dice_rej,
list(rej_supp), list(to_clf), list(midk),
svm_acc_fail, list(toacc_hi), list(toacc_lo),
list(field_art), list(phys_art),
list(misc_art), list(acc_comps), list(ign)]
with open('csstepdata.json', 'w') as ofh:
json.dump(dict(zip(diagstep_keys, diagstep_vals)), ofh,
indent=4, sort_keys=True, default=str)
allfz = np.array([Tz, Vz, Ktz, KRr, cnz, Rz, mmix_kurt, fdist_z])
np.savetxt('csdata.txt', allfz)
return list(sorted(acc_comps)), list(sorted(rej)), list(sorted(midk)), list(sorted(ign))
def selcomps(seldict, mmix, mask, ref_img, manacc, n_echos, t2s, s0, olevel=2,
oversion=99, filecsdata=True, savecsdiag=True, strict_mode=False):
"""
Labels components in `mmix`
Parameters
----------
seldict : :obj:`dict`
As output from `fitmodels_direct`
mmix : (C x T) array_like
Mixing matrix for converting input data to component space, where `C`
is components and `T` is the number of volumes in the original data
mask : (S,) array_like
Boolean mask array
ref_img : str or img_like
Reference image to dictate how outputs are saved to disk
manacc : list
Comma-separated list of indices of manually accepted components
n_echos : int
Number of echos in original data
t2s : (S,) array_like
s0 : (S,) array_like
olevel : int, optional
Default: 2
oversion : int, optional
Default: 99
filecsdata: bool, optional
Default: False
savecsdiag: bool, optional
Default: True
strict_mode: bool, optional
Default: False
Returns
-------
acc : list
Indices of accepted (BOLD) components in `mmix`
rej : list
Indices of rejected (non-BOLD) components in `mmix`
midk : list
Indices of mid-K (questionable) components in `mmix`
ign : list
Indices of ignored components in `mmix`
"""
if filecsdata:
import bz2
if seldict is not None:
LGR.info('Saving component selection data')
with bz2.BZ2File('compseldata.pklbz', 'wb') as csstate_f:
pickle.dump(seldict, csstate_f)
else:
try:
with bz2.BZ2File('compseldata.pklbz', 'rb') as csstate_f:
seldict = pickle.load(csstate_f)
except FileNotFoundError:
LGR.warning('Failed to load component selection data')
return None
# List of components
midk = []
ign = []
nc = np.arange(len(seldict['Kappas']))
ncl = np.arange(len(seldict['Kappas']))
# If user has specified components to accept manually
if manacc:
acc = sorted([int(vv) for vv in manacc.split(',')])
midk = []
rej = sorted(np.setdiff1d(ncl, acc))
return acc, rej, midk, [] # Add string for ign
"""
Do some tallies for no. of significant voxels
"""
countsigFS0 = seldict['F_S0_clmaps'].sum(0)
countsigFR2 = seldict['F_R2_clmaps'].sum(0)
countnoise = np.zeros(len(nc))
"""
Make table of dice values
"""
dice_tbl = np.zeros([nc.shape[0], 2])
for ii in ncl:
dice_FR2 = utils.dice(utils.unmask(seldict['Br_clmaps_R2'][:, ii], mask)[t2s != 0],
seldict['F_R2_clmaps'][:, ii])
dice_FS0 = utils.dice(utils.unmask(seldict['Br_clmaps_S0'][:, ii], mask)[t2s != 0],
seldict['F_S0_clmaps'][:, ii])
dice_tbl[ii, :] = [dice_FR2, dice_FS0] # step 3a here and above
dice_tbl[np.isnan(dice_tbl)] = 0
"""
Make table of noise gain
"""
tt_table = np.zeros([len(nc), 4])
counts_FR2_Z = np.zeros([len(nc), 2])
for ii in nc:
comp_noise_sel = utils.andb([np.abs(seldict['Z_maps'][:, ii]) > 1.95,
seldict['Z_clmaps'][:, ii] == 0]) == 2
countnoise[ii] = np.array(comp_noise_sel, dtype=np.int).sum()
noise_FR2_Z_mask = utils.unmask(comp_noise_sel, mask)[t2s != 0]
noise_FR2_Z = np.log10(np.unique(seldict['F_R2_maps'][noise_FR2_Z_mask, ii]))
signal_FR2_Z_mask = utils.unmask(seldict['Z_clmaps'][:, ii], mask)[t2s != 0] == 1
signal_FR2_Z = np.log10(np.unique(seldict['F_R2_maps'][signal_FR2_Z_mask, ii]))
counts_FR2_Z[ii, :] = [len(signal_FR2_Z), len(noise_FR2_Z)]
try:
ttest = stats.ttest_ind(signal_FR2_Z, noise_FR2_Z, equal_var=True)
# avoid DivideByZero RuntimeWarning
if signal_FR2_Z.size > 0 and noise_FR2_Z.size > 0:
mwu = stats.norm.ppf(stats.mannwhitneyu(signal_FR2_Z, noise_FR2_Z)[1])
else:
mwu = -np.inf
tt_table[ii, 0] = np.abs(mwu) * ttest[0] / np.abs(ttest[0])
tt_table[ii, 1] = ttest[1]
except Exception: # TODO: what is the error that might be caught here?
pass
tt_table[np.isnan(tt_table)] = 0
tt_table[np.isinf(tt_table[:, 0]), 0] = np.percentile(tt_table[~np.isinf(tt_table[:, 0]), 0],
98)
# Time series derivative kurtosis
mmix_dt = (mmix[:-1] - mmix[1:])
mmix_kurt = stats.kurtosis(mmix_dt)
mmix_std = np.std(mmix_dt, axis=0)
"""
Step 1: Reject anything that's obviously an artifact
a. Estimate a null variance
"""
LGR.debug('Rejecting gross artifacts based on Rho/Kappa values and S0/R2 counts')
rej = ncl[utils.andb([seldict['Rhos'] > seldict['Kappas'], countsigFS0 > countsigFR2]) > 0]
ncl = np.setdiff1d(ncl, rej)
"""
Step 2: Compute 3-D spatial FFT of Beta maps to detect high-spatial
frequency artifacts
"""
LGR.debug('Computing 3D spatial FFT of beta maps to detect high-spatial frequency artifacts')
# spatial information is important so for NIFTI we convert back to 3D space
if utils.get_dtype(ref_img) == 'NIFTI':
dim1 = np.prod(ref_img.shape[:2])
else:
dim1 = mask.shape[0]
fproj_arr = np.zeros([dim1, len(nc)])
fproj_arr_val = np.zeros([dim1, len(nc)])
spr = []
fdist = []
for ii in nc:
# convert data back to 3D array
if utils.get_dtype(ref_img) == 'NIFTI':
tproj = utils.new_nii_like(ref_img, utils.unmask(seldict['PSC'],
mask)[:, ii]).get_data()
else:
tproj = utils.unmask(seldict['PSC'], mask)[:, ii]
fproj = np.fft.fftshift(np.abs(np.fft.rfftn(tproj)))
fproj_z = fproj.max(axis=2)
fproj[fproj == fproj.max()] = 0
fproj_arr[:, ii] = stats.rankdata(fproj_z.flatten())
fproj_arr_val[:, ii] = fproj_z.flatten()
spr.append(np.array(fproj_z > fproj_z.max() / 4, dtype=np.int).sum())
fprojr = np.array([fproj, fproj[:, :, ::-1]]).max(0)
fdist.append(np.max([utils.fitgaussian(fproj.max(jj))[3:].max() for
jj in range(fprojr.ndim)]))
fdist = np.array(fdist)
spr = np.array(spr)
"""
Step 3: Create feature space of component properties
"""
LGR.debug('Creating feature space of component properties')
fdist_pre = fdist.copy()
fdist_pre[fdist > np.median(fdist) * 3] = np.median(fdist) * 3
fdist_z = (fdist_pre - np.median(fdist_pre)) / fdist_pre.std()
spz = (spr-spr.mean())/spr.std()
Tz = (tt_table[:, 0] - tt_table[:, 0].mean()) / tt_table[:, 0].std()
varex_ = np.log(seldict['varex'])
Vz = (varex_-varex_.mean()) / varex_.std()
Rz = (seldict['Rhos'] - seldict['Rhos'].mean()) / seldict['Rhos'].std()
Ktz = np.log(seldict['Kappas']) / 2
Ktz = (Ktz-Ktz.mean()) / Ktz.std()
Rtz = np.log(seldict['Rhos']) / 2
Rtz = (Rtz-Rtz.mean())/Rtz.std()
KRr = stats.zscore(np.log(seldict['Kappas']) / np.log(seldict['Rhos']))
cnz = (countnoise-countnoise.mean()) / countnoise.std()
Dz = stats.zscore(np.arctanh(dice_tbl[:, 0] + 0.001))
fz = np.array([Tz, Vz, Ktz, KRr, cnz, Rz, mmix_kurt, fdist_z])
"""
Step 3: Make initial guess of where BOLD components are and use DBSCAN
to exclude noise components and find a sample set of 'good' components
"""
LGR.debug('Making initial guess of BOLD components')
# epsmap is [index,level of overlap with dicemask,
# number of high Rho components]
F05, F025, F01 = utils.getfbounds(n_echos)
epsmap = []
Rhos_sorted = np.array(sorted(seldict['Rhos']))[::-1]
# Make an initial guess as to number of good components based on
# consensus of control points across Rhos and Kappas
KRcutguesses = [getelbow_mod(seldict['Rhos']), getelbow_cons(seldict['Rhos']),
getelbow_aggr(seldict['Rhos']), getelbow_mod(seldict['Kappas']),
getelbow_cons(seldict['Kappas']), getelbow_aggr(seldict['Kappas'])]
Khighelbowval = stats.scoreatpercentile([getelbow_mod(seldict['Kappas'], val=True),
getelbow_cons(seldict['Kappas'], val=True),
getelbow_aggr(seldict['Kappas'], val=True)] +
list(utils.getfbounds(n_echos)),
75, interpolation_method='lower')
KRcut = np.median(KRcutguesses)
# only use exclusive when inclusive is extremely inclusive - double KRcut
cond1 = getelbow_cons(seldict['Kappas']) > KRcut * 2
cond2 = getelbow_mod(seldict['Kappas'], val=True) < F01
if cond1 and cond2:
Kcut = getelbow_mod(seldict['Kappas'], val=True)
else:
Kcut = getelbow_cons(seldict['Kappas'], val=True)
# only use inclusive when exclusive is extremely exclusive - half KRcut
# (remember for Rho inclusive is higher, so want both Kappa and Rho
# to defaut to lower)
if getelbow_cons(seldict['Rhos']) > KRcut * 2:
Rcut = getelbow_mod(seldict['Rhos'], val=True)
# for above, consider something like:
# min([getelbow_mod(Rhos,True),sorted(Rhos)[::-1][KRguess] ])
else:
Rcut = getelbow_cons(seldict['Rhos'], val=True)
if Rcut > Kcut:
Kcut = Rcut # Rcut should never be higher than Kcut
KRelbow = utils.andb([seldict['Kappas'] > Kcut, seldict['Rhos'] < Rcut])
# Make guess of Kundu et al 2011 plus remove high frequencies,
# generally high variance, and high variance given low Kappa
tt_lim = stats.scoreatpercentile(tt_table[tt_table[:, 0] > 0, 0],
75, interpolation_method='lower') / 3
KRguess = np.setdiff1d(np.setdiff1d(nc[KRelbow == 2], rej),
np.union1d(nc[tt_table[:, 0] < tt_lim],
np.union1d(np.union1d(nc[spz > 1],
nc[Vz > 2]),
nc[utils.andb([seldict['varex'] > 0.5 *
sorted(seldict['varex'])[::-1][int(KRcut)],
seldict['Kappas'] < 2*Kcut]) == 2])))
guessmask = np.zeros(len(nc))
guessmask[KRguess] = 1
# Throw lower-risk bad components out
rejB = ncl[utils.andb([tt_table[ncl, 0] < 0,
seldict['varex'][ncl] > np.median(seldict['varex']), ncl > KRcut]) == 3]
rej = np.union1d(rej, rejB)
ncl = np.setdiff1d(ncl, rej)
LGR.debug('Using DBSCAN to find optimal set of "good" BOLD components')
for ii in range(20000):
eps = .005 + ii * .005
db = DBSCAN(eps=eps, min_samples=3).fit(fz.T)
# it would be great to have descriptive names, here
# DBSCAN found at least three non-noisy clusters
cond1 = db.labels_.max() > 1
# DBSCAN didn't detect more classes than the total # of components / 6
cond2 = db.labels_.max() < len(nc) / 6
# TODO: confirm if 0 is a special label for DBSCAN
# my intuition here is that we're confirming DBSCAN labelled previously
# rejected components as noise (i.e., no overlap between `rej` and
# labelled DBSCAN components)
cond3 = np.intersect1d(rej, nc[db.labels_ == 0]).shape[0] == 0
# DBSCAN labelled less than half of the total components as noisy
cond4 = np.array(db.labels_ == -1, dtype=int).sum() / float(len(nc)) < .5
if cond1 and cond2 and cond3 and cond4:
epsmap.append([ii, utils.dice(guessmask, db.labels_ == 0),
np.intersect1d(nc[db.labels_ == 0],
nc[seldict['Rhos'] > getelbow_mod(Rhos_sorted,
val=True)]).shape[0]])
db = None
epsmap = np.array(epsmap)
LGR.debug('Found DBSCAN solutions for {}/20000 eps resolutions'.format(len(epsmap)))
group0 = []
dbscanfailed = False
if len(epsmap) != 0:
# Select index that maximizes Dice with guessmask but first
# minimizes number of higher Rho components
ii = int(epsmap[np.argmax(epsmap[epsmap[:, 2] == np.min(epsmap[:, 2]), 1], 0), 0])
LGR.debug('Component selection tuning: {:.05f}'.format(epsmap[:, 1].max()))
db = DBSCAN(eps=.005+ii*.005, min_samples=3).fit(fz.T)
ncl = nc[db.labels_ == 0]
ncl = np.setdiff1d(ncl, rej)
ncl = np.setdiff1d(ncl, ncl[ncl > len(nc) - len(rej)])
group0 = ncl.copy()
group_n1 = nc[db.labels_ == -1]
to_clf = np.setdiff1d(nc, np.union1d(ncl, rej))
if len(group0) == 0 or len(group0) < len(KRguess) * .5:
dbscanfailed = True
LGR.debug('DBSCAN guess failed; using elbow guess method instead')
ncl = np.setdiff1d(np.setdiff1d(nc[KRelbow == 2], rej),
np.union1d(nc[tt_table[:, 0] < tt_lim],
np.union1d(np.union1d(nc[spz > 1],
nc[Vz > 2]),
nc[utils.andb([seldict['varex'] > 0.5 *
sorted(seldict['varex'])[::-1][int(KRcut)],
seldict['Kappas'] < 2 * Kcut]) == 2])))
group0 = ncl.copy()
group_n1 = []
to_clf = np.setdiff1d(nc, np.union1d(group0, rej))
if len(group0) < 2 or (len(group0) < 4 and float(len(rej))/len(group0) > 3):
LGR.warning('Extremely limited reliable BOLD signal space! '
'Not filtering components beyond BOLD/non-BOLD guesses.')
midkfailed = True
min_acc = np.array([])
if len(group0) != 0:
# For extremes, building in a 20% tolerance
toacc_hi = np.setdiff1d(nc[utils.andb([fdist <= np.max(fdist[group0]),
seldict['Rhos'] < F025, Vz > -2]) == 3],
np.union1d(group0, rej))
min_acc = np.union1d(group0, toacc_hi)
to_clf = np.setdiff1d(nc, np.union1d(min_acc, rej))
diagstep_keys = ['Rejected components', 'Kappa-Rho cut point',
'Kappa cut point', 'Rho cut point', 'DBSCAN failed to converge',
'Mid-Kappa failed (limited BOLD signal)', 'Kappa-Rho guess',
'min_acc', 'toacc_hi']
diagstep_vals = [rej.tolist(), KRcut, Kcut, Rcut, dbscanfailed,
midkfailed, KRguess.tolist(), min_acc.tolist(), toacc_hi.tolist()]
with open('csstepdata.json', 'w') as ofh:
json.dump(dict(zip(diagstep_keys, diagstep_vals)), ofh, indent=4, sort_keys=True)
return list(sorted(min_acc)), list(sorted(rej)), [], list(sorted(to_clf))
# Find additional components to reject based on Dice - doing this here
# since Dice is a little unstable, need to reference group0
rej_supp = []
dice_rej = False
if not dbscanfailed and len(rej) + len(group0) < 0.75 * len(nc):
dice_rej = True
rej_supp = np.setdiff1d(np.setdiff1d(np.union1d(rej,
nc[dice_tbl[nc, 0] <= dice_tbl[nc, 1]]),
group0), group_n1)
rej = np.union1d(rej, rej_supp)
# Temporal features
# larger is worse - spike
mmix_kurt_z = (mmix_kurt-mmix_kurt[group0].mean()) / mmix_kurt[group0].std()
# smaller is worse - drift
mmix_std_z = -1 * ((mmix_std-mmix_std[group0].mean()) / mmix_std[group0].std())
mmix_kurt_z_max = np.max([mmix_kurt_z, mmix_std_z], 0)
"""
Step 2: Classifiy midk and ignore using separte SVMs for
different variance regimes
# To render hyperplane:
min_x = np.min(spz2);max_x=np.max(spz2)
# plotting separating hyperplane
ww = clf_.coef_[0]
aa = -ww[0] / ww[1]
# make sure the next line is long enough
xx = np.linspace(min_x - 2, max_x + 2)
yy = aa * xx - (clf_.intercept_[0]) / ww[1]
plt.plot(xx, yy, '-')
"""
LGR.debug('Attempting to classify midk components')
# Tried getting rid of accepting based on SVM altogether,
# now using only rejecting
toacc_hi = np.setdiff1d(nc[utils.andb([fdist <= np.max(fdist[group0]),
seldict['Rhos'] < F025, Vz > -2]) == 3],
np.union1d(group0, rej))
toacc_lo = np.intersect1d(to_clf,
nc[utils.andb([spz < 1, Rz < 0, mmix_kurt_z_max < 5,
Dz > -1, Tz > -1, Vz < 0, seldict['Kappas'] >= F025,
fdist < 3 * np.percentile(fdist[group0], 98)]) == 8])
midk_clf, clf_ = do_svm(fproj_arr_val[:, np.union1d(group0, rej)].T,
[0] * len(group0) + [1] * len(rej),
fproj_arr_val[:, to_clf].T,
svmtype=2)
midk = np.setdiff1d(to_clf[utils.andb([midk_clf == 1, seldict['varex'][to_clf] >
np.median(seldict['varex'][group0])]) == 2],
np.union1d(toacc_hi, toacc_lo))
# only use SVM to augment toacc_hi only if toacc_hi isn't already
# conflicting with SVM choice
if len(np.intersect1d(to_clf[utils.andb([midk_clf == 1,
Vz[to_clf] > 0]) == 2], toacc_hi)) == 0:
svm_acc_fail = True
toacc_hi = np.union1d(toacc_hi, to_clf[midk_clf == 0])
else:
svm_acc_fail = False
"""
Step 3: Compute variance associated with low T2* areas
(e.g. draining veins and low T2* areas)
# To write out veinmask
veinout = np.zeros(t2s.shape)
veinout[t2s!=0] = veinmaskf
utils.filewrite(veinout, 'veinmaskf', ref_img)
veinBout = utils.unmask(veinmaskB, mask)
utils.filewrite(veinBout, 'veins50', ref_img)
"""
LGR.debug('Computing variance associated with low T2* areas (e.g., draining veins)')
tsoc_B_Zcl = np.zeros(seldict['tsoc_B'].shape)
tsoc_B_Zcl[seldict['Z_clmaps'] != 0] = np.abs(seldict['tsoc_B'])[seldict['Z_clmaps'] != 0]
sig_B = [stats.scoreatpercentile(tsoc_B_Zcl[tsoc_B_Zcl[:, ii] != 0, ii], 25)
if len(tsoc_B_Zcl[tsoc_B_Zcl[:, ii] != 0, ii]) != 0
else 0 for ii in nc]
sig_B = np.abs(seldict['tsoc_B']) > np.tile(sig_B, [seldict['tsoc_B'].shape[0], 1])
veinmask = utils.andb([t2s < stats.scoreatpercentile(t2s[t2s != 0], 15,
interpolation_method='lower'),
t2s != 0]) == 2
veinmaskf = veinmask[mask]
veinR = np.array(sig_B[veinmaskf].sum(0),
dtype=float) / sig_B[~veinmaskf].sum(0)
veinR[np.isnan(veinR)] = 0
veinc = np.union1d(rej, midk)
rej_veinRZ = ((veinR-veinR[veinc].mean())/veinR[veinc].std())[veinc]
rej_veinRZ[rej_veinRZ < 0] = 0
rej_veinRZ[countsigFR2[veinc] > np.array(veinmaskf, dtype=int).sum()] = 0
t2s_lim = [stats.scoreatpercentile(t2s[t2s != 0], 50,
interpolation_method='lower'),
stats.scoreatpercentile(t2s[t2s != 0], 80,
interpolation_method='lower') / 2]
phys_var_zs = []
for t2sl_i in range(len(t2s_lim)):
t2sl = t2s_lim[t2sl_i]
veinW = sig_B[:, veinc]*np.tile(rej_veinRZ, [sig_B.shape[0], 1])
veincand = utils.unmask(utils.andb([s0[t2s != 0] < np.median(s0[t2s != 0]),
t2s[t2s != 0] < t2sl]) >= 1,
t2s != 0)[mask]
veinW[~veincand] = 0
invein = veinW.sum(axis=1)[(utils.unmask(veinmaskf, mask) *
utils.unmask(veinW.sum(axis=1) > 1, mask))[mask]]
minW = 10 * (np.log10(invein).mean()) - 1 * 10**(np.log10(invein).std())
veinmaskB = veinW.sum(axis=1) > minW
tsoc_Bp = seldict['tsoc_B'].copy()
tsoc_Bp[tsoc_Bp < 0] = 0
vvex = np.array([(tsoc_Bp[veinmaskB, ii]**2.).sum() /
(tsoc_Bp[:, ii]**2.).sum() for ii in nc])
group0_res = np.intersect1d(KRguess, group0)
phys_var_zs.append((vvex - vvex[group0_res].mean()) / vvex[group0_res].std())
veinBout = utils.unmask(veinmaskB, mask)
utils.filewrite(veinBout.astype(float), 'veins_l%i' % t2sl_i, ref_img)
# Mask to sample veins
phys_var_z = np.array(phys_var_zs).max(0)
Vz2 = (varex_ - varex_[group0].mean())/varex_[group0].std()
"""
Step 4: Learn joint TE-dependence spatial and temporal models to move
remaining artifacts to ignore class
"""
LGR.debug('Learning joint TE-dependence spatial/temporal models to ignore remaining artifacts')
to_ign = []
minK_ign = np.max([F05, getelbow_cons(seldict['Kappas'], val=True)])
newcest = len(group0) + len(toacc_hi[seldict['Kappas'][toacc_hi] > minK_ign])
phys_art = np.setdiff1d(nc[utils.andb([phys_var_z > 3.5,
seldict['Kappas'] < minK_ign]) == 2], group0)
rank_diff = stats.rankdata(phys_var_z) - stats.rankdata(seldict['Kappas'])
phys_art = np.union1d(np.setdiff1d(nc[utils.andb([phys_var_z > 2, rank_diff > newcest / 2,
Vz2 > -1]) == 3],
group0), phys_art)
# Want to replace field_art with an acf/SVM based approach
# instead of a kurtosis/filter one
field_art = np.setdiff1d(nc[utils.andb([mmix_kurt_z_max > 5,
seldict['Kappas'] < minK_ign]) == 2], group0)
field_art = np.union1d(np.setdiff1d(nc[utils.andb([mmix_kurt_z_max > 2,
(stats.rankdata(mmix_kurt_z_max) -
stats.rankdata(seldict['Kappas'])) > newcest / 2,
Vz2 > 1, seldict['Kappas'] < F01]) == 4],
group0), field_art)
field_art = np.union1d(np.setdiff1d(nc[utils.andb([mmix_kurt_z_max > 3,
Vz2 > 3, seldict['Rhos'] >
np.percentile(seldict['Rhos'][group0],
75)]) == 3],
group0), field_art)
field_art = np.union1d(np.setdiff1d(nc[utils.andb([mmix_kurt_z_max > 5, Vz2 > 5]) == 2],
group0), field_art)
misc_art = np.setdiff1d(nc[utils.andb([(stats.rankdata(Vz) -
stats.rankdata(Ktz)) > newcest / 2,
seldict['Kappas'] < Khighelbowval]) == 2], group0)
ign_cand = np.unique(list(field_art)+list(phys_art)+list(misc_art))
midkrej = np.union1d(midk, rej)
to_ign = np.setdiff1d(list(ign_cand), midkrej)
toacc = np.union1d(toacc_hi, toacc_lo)
ncl = np.setdiff1d(np.union1d(ncl, toacc), np.union1d(to_ign, midkrej))
ign = np.setdiff1d(nc, list(ncl) + list(midk) + list(rej))
orphan = np.setdiff1d(nc, list(ncl) + list(to_ign) + list(midk) + list(rej))
# Last ditch effort to save some transient components
if not strict_mode:
Vz3 = (varex_ - varex_[ncl].mean())/varex_[ncl].std()
ncl = np.union1d(ncl, np.intersect1d(orphan,
nc[utils.andb([seldict['Kappas'] > F05,
seldict['Rhos'] < F025,
seldict['Kappas'] > seldict['Rhos'],
Vz3 <= -1,
Vz3 > -3,
mmix_kurt_z_max < 2.5]) == 6]))
ign = np.setdiff1d(nc, list(ncl)+list(midk)+list(rej))
orphan = np.setdiff1d(nc, list(ncl) + list(to_ign) + list(midk) + list(rej))
if savecsdiag:
diagstep_keys = ['Rejected components', 'Kappa-Rho cut point', 'Kappa cut',
'Rho cut', 'DBSCAN failed to converge', 'Kappa-Rho guess',
'Dice rejected', 'rej_supp', 'to_clf',
'Mid-kappa components', 'svm_acc_fail', 'toacc_hi', 'toacc_lo',
'Field artifacts', 'Physiological artifacts',
'Miscellaneous artifacts', 'ncl', 'Ignored components']
diagstep_vals = [rej.tolist(), KRcut, Kcut, Rcut, dbscanfailed,
KRguess.tolist(), dice_rej, rej_supp.tolist(),
to_clf.tolist(), midk.tolist(), svm_acc_fail,
toacc_hi.tolist(), toacc_lo.tolist(),
field_art.tolist(), phys_art.tolist(),
misc_art.tolist(), ncl.tolist(), ign.tolist()]
with open('csstepdata.json', 'w') as ofh:
json.dump(dict(zip(diagstep_keys, diagstep_vals)), ofh, indent=4, sort_keys=True)
allfz = np.array([Tz, Vz, Ktz, KRr, cnz, Rz, mmix_kurt, fdist_z])
np.savetxt('csdata.txt', allfz)
return list(sorted(ncl)), list(sorted(rej)), list(sorted(midk)), list(sorted(ign))
def tedpca(catd,
OCcatd,
combmode,
mask,
t2s,
t2sG,
stabilize,
ref_img,
tes,
kdaw,
rdaw,
ste=0,
mlepca=True,
wvpca=False):
"""
Use principal components analysis (PCA) to identify and remove thermal
noise from multi-echo data.
Parameters
----------
catd : (S x E x T) array_like
Input functional data
OCcatd : (S x T) array_like
Optimally-combined time series data
combmode : {'t2s', 'ste'} str
How optimal combination of echos should be made, where 't2s' indicates
using the method of Posse 1999 and 'ste' indicates using the method of
Poser 2006
mask : (S,) array_like
Boolean mask array
stabilize : :obj:`bool`
Whether to attempt to stabilize convergence of ICA by returning
dimensionally-reduced data from PCA and component selection.
ref_img : :obj:`str` or img_like
Reference image to dictate how outputs are saved to disk
tes : :obj:`list`
List of echo times associated with `catd`, in milliseconds
kdaw : :obj:`float`
Dimensionality augmentation weight for Kappa calculations
rdaw : :obj:`float`
Dimensionality augmentation weight for Rho calculations
ste : :obj:`int` or :obj:`list` of :obj:`int`, optional
Which echos to use in PCA. Values -1 and 0 are special, where a value
of -1 will indicate using all the echos and 0 will indicate using the
optimal combination of the echos. A list can be provided to indicate
a subset of echos. Default: 0
mlepca : :obj:`bool`, optional
Whether to use the method originally explained in Minka, NIPS 2000 for
guessing PCA dimensionality instead of a traditional SVD. Default: True
wvpca : :obj:`bool`, optional
Whether to apply wavelet denoising to data. Default: False
Returns
-------
n_components : :obj:`int`
Number of components retained from PCA decomposition
dd : (S x E x T) :obj:`numpy.ndarray`
Dimensionally-reduced functional data
Notes
-----
====================== =================================================
Notation Meaning
====================== =================================================
:math:`\\kappa` Component pseudo-F statistic for TE-dependent
(BOLD) model.
:math:`\\rho` Component pseudo-F statistic for TE-independent
(artifact) model.
:math:`v` Voxel
:math:`V` Total number of voxels in mask
:math:`\\zeta` Something
:math:`c` Component
:math:`p` Something else
====================== =================================================
Steps:
1. Variance normalize either multi-echo or optimally combined data,
depending on settings.
2. Decompose normalized data using PCA or SVD.
3. Compute :math:`{\\kappa}` and :math:`{\\rho}`:
.. math::
{\\kappa}_c = \\frac{\sum_{v}^V {\\zeta}_{c,v}^p * \
F_{c,v,R_2^*}}{\sum {\\zeta}_{c,v}^p}
{\\rho}_c = \\frac{\sum_{v}^V {\\zeta}_{c,v}^p * \
F_{c,v,S_0}}{\sum {\\zeta}_{c,v}^p}
4. Some other stuff. Something about elbows.
5. Classify components as thermal noise if they meet both of the
following criteria:
- Nonsignificant :math:`{\\kappa}` and :math:`{\\rho}`.
- Nonsignificant variance explained.
Outputs:
This function writes out several files:
====================== =================================================
Filename Content
====================== =================================================
pcastate.pkl Values from PCA results.
comp_table_pca.txt PCA component table.
mepca_mix.1D PCA mixing matrix.
====================== =================================================
"""
n_samp, n_echos, n_vols = catd.shape
ste = np.array([int(ee) for ee in str(ste).split(',')])
if len(ste) == 1 and ste[0] == -1:
LGR.info('Computing PCA of optimally combined multi-echo data')
d = OCcatd[utils.make_min_mask(OCcatd[:,
np.newaxis, :])][:,
np.newaxis, :]
elif len(ste) == 1 and ste[0] == 0:
LGR.info('Computing PCA of spatially concatenated multi-echo data')
d = catd[mask].astype('float64')
else:
LGR.info('Computing PCA of echo #%s' %
','.join([str(ee) for ee in ste]))
d = np.stack([catd[mask, ee] for ee in ste - 1],
axis=1).astype('float64')
eim = np.squeeze(eimask(d))
d = np.squeeze(d[eim])
dz = ((d.T - d.T.mean(axis=0)) / d.T.std(axis=0)).T # var normalize ts
dz = (dz - dz.mean()) / dz.std() # var normalize everything
if wvpca:
dz, cAl = dwtmat(dz)
if not op.exists('pcastate.pkl'):
# do PC dimension selection and get eigenvalue cutoff
if mlepca:
from sklearn.decomposition import PCA
ppca = PCA(n_components='mle', svd_solver='full')
ppca.fit(dz)
v = ppca.components_
s = ppca.explained_variance_
u = np.dot(np.dot(dz, v.T), np.diag(1. / s))
else:
u, s, v = np.linalg.svd(dz, full_matrices=0)
# actual variance explained (normalized)
sp = s / s.sum()
eigelb = getelbow_mod(sp, return_val=True)
spdif = np.abs(np.diff(sp))
spdifh = spdif[(len(spdif) // 2):]
spdthr = np.mean([spdifh.max(), spdif.min()])
spmin = sp[(len(spdif) // 2) +
np.arange(len(spdifh))[spdifh >= spdthr][0] + 1]
spcum = np.cumsum(sp)
# Compute K and Rho for PCA comps
eimum = np.atleast_2d(eim)
eimum = np.transpose(eimum, np.argsort(eimum.shape)[::-1])
eimum = eimum.prod(axis=1)
o = np.zeros((mask.shape[0], *eimum.shape[1:]))
o[mask] = eimum
eimum = np.squeeze(o).astype(bool)
vTmix = v.T
vTmixN = ((vTmix.T - vTmix.T.mean(0)) / vTmix.T.std(0)).T
LGR.info('Making initial component selection guess from PCA results')
_, ctb, betasv, v_T = model.fitmodels_direct(catd,
v.T,
eimum,
t2s,
t2sG,
tes,
combmode,
ref_img,
mmixN=vTmixN,
full_sel=False)
ctb = ctb[ctb[:, 0].argsort(), :]
ctb = np.vstack([ctb.T[:3], sp]).T
# Save state
fname = op.abspath('pcastate.pkl')
LGR.info('Saving PCA results to: {}'.format(fname))
pcastate = {
'u': u,
's': s,
'v': v,
'ctb': ctb,
'eigelb': eigelb,
'spmin': spmin,
'spcum': spcum
}
try:
with open(fname, 'wb') as handle:
pickle.dump(pcastate, handle)
except TypeError:
LGR.warning('Could not save PCA solution')
else: # if loading existing state
LGR.info('Loading PCA from: pcastate.pkl')
with open('pcastate.pkl', 'rb') as handle:
pcastate = pickle.load(handle)
u, s, v = pcastate['u'], pcastate['s'], pcastate['v']
ctb, eigelb = pcastate['ctb'], pcastate['eigelb']
spmin, spcum = pcastate['spmin'], pcastate['spcum']
np.savetxt('comp_table_pca.txt', ctb[ctb[:, 1].argsort(), :][::-1])
np.savetxt('mepca_mix.1D', v[ctb[:, 1].argsort()[::-1], :].T)
kappas = ctb[ctb[:, 1].argsort(), 1]
rhos = ctb[ctb[:, 2].argsort(), 2]
fmin, fmid, fmax = utils.getfbounds(n_echos)
kappa_thr = np.average(sorted(
[fmin, getelbow_mod(kappas, return_val=True) / 2, fmid]),
weights=[kdaw, 1, 1])
rho_thr = np.average(sorted(
[fmin, getelbow_cons(rhos, return_val=True) / 2, fmid]),
weights=[rdaw, 1, 1])
if int(kdaw) == -1:
kappas_lim = kappas[utils.andb([kappas < fmid, kappas > fmin]) == 2]
kappa_thr = kappas_lim[getelbow_mod(kappas_lim)]
rhos_lim = rhos[utils.andb([rhos < fmid, rhos > fmin]) == 2]
rho_thr = rhos_lim[getelbow_mod(rhos_lim)]
stabilize = True
if int(kdaw) != -1 and int(rdaw) == -1:
rhos_lim = rhos[utils.andb([rhos < fmid, rhos > fmin]) == 2]
rho_thr = rhos_lim[getelbow_mod(rhos_lim)]
is_hik = np.array(ctb[:, 1] > kappa_thr, dtype=np.int)
is_hir = np.array(ctb[:, 2] > rho_thr, dtype=np.int)
is_hie = np.array(ctb[:, 3] > eigelb, dtype=np.int)
is_his = np.array(ctb[:, 3] > spmin, dtype=np.int)
is_not_fmax1 = np.array(ctb[:, 1] != F_MAX, dtype=np.int)
is_not_fmax2 = np.array(ctb[:, 2] != F_MAX, dtype=np.int)
pcscore = (is_hik + is_hir + is_hie) * is_his * is_not_fmax1 * is_not_fmax2
if stabilize:
temp7 = np.array(spcum < 0.95, dtype=np.int)
temp8 = np.array(ctb[:, 2] > fmin, dtype=np.int)
temp9 = np.array(ctb[:, 1] > fmin, dtype=np.int)
pcscore = pcscore * temp7 * temp8 * temp9
pcsel = pcscore > 0
dd = u.dot(np.diag(s * np.array(pcsel, dtype=np.int))).dot(v)
if wvpca:
dd = idwtmat(dd, cAl)
n_components = s[pcsel].shape[0]
LGR.info('Selected {0} components with Kappa threshold: {1:.02f}, '
'Rho threshold: {2:.02f}'.format(n_components, kappa_thr,
rho_thr))
dd = stats.zscore(dd.T, axis=0).T # variance normalize timeseries
dd = stats.zscore(dd, axis=None) # variance normalize everything
return n_components, dd
