-
Notifications
You must be signed in to change notification settings - Fork 0
/
GradDistCorrFullHCP.py
323 lines (229 loc) · 9.29 KB
/
GradDistCorrFullHCP.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
#!/usr/bin/env python
import os
import sys
import pathlib
import argparse
import nibabel as nib
import numpy as np
import brainspace as bs
import subprocess as sp
from nilearn import signal
from scipy import spatial
from sklearn.metrics import pairwise_distances
from sklearn.metrics.pairwise import cosine_similarity
from brainspace.gradient import GradientMaps
from brainspace.gradient.embedding import diffusion_mapping
from scipy.spatial.distance import pdist, squareform
import sys
from mapalign import embed
from utils import *
print('imports are good')
parser = argparse.ArgumentParser(description='Embeds functional Connectivity Matrix using PCA and diffusion map embedding, and then determines how far peaks are across sessions of the same subject',\
usage='pkReliability.py --subj <HCP subject> --odir <output directory> ',\
epilog=("Example usage: "+"pkReliability.py --subj 100206 --odir /well/margulies/projects/pks"),add_help=True)
if len(sys.argv) < 2:
parser.print_help()
sys.exit(1)
req_grp = parser.add_argument_group(title='Required arguments')
req_grp.add_argument('--subj',type=str,metavar='',required=True,help='HCP subject to run on')
req_grp.add_argument('--odir',type=str,metavar='',required=True,help='Output directory base. Output will be saved as odir/subj/...')
req_grp.add_argument('--kernel',type=str,metavar='',required=True,help='specify smoothing kernel of time series')
op_grp= parser.add_argument_group(title='Optional arguments')
args=parser.parse_args()
#get arguments into variables
subj=args.subj
odir=args.odir
kernel=args.kernel
### set up subjects output directory
odir=f'{odir}/{subj}'
print(odir)
print(kernel)
os.makedirs(odir,exist_ok=True)
# cluster_path='/well/margulies/users/mnk884/data20/'
cluster_path='/well/win-hcp/HCP-YA/subjectsAll'
###### set up files
subjdir=f'{cluster_path}/{subj}' ### the folder containing the Structural and Rest folders. Change to match cluster when access given
fdir=f'{subjdir}/MNINonLinear/Results/'
anatdirNat=f'{subjdir}/T1w/Native/'
anatdir32=f'{subjdir}/T1w/fsaverage_LR32k/'
MNINonlinearDir=f'{subjdir}/MNINonLinear'
##### resting state time series paths
ses1LR=f'{fdir}/rfMRI_REST1_LR/rfMRI_REST1_LR_Atlas_MSMAll_hp2000_clean.dtseries.nii'
ses2LR=f'{fdir}/rfMRI_REST2_LR/rfMRI_REST2_LR_Atlas_MSMAll_hp2000_clean.dtseries.nii'
ses1RL=f'{fdir}/rfMRI_REST1_RL/rfMRI_REST1_RL_Atlas_MSMAll_hp2000_clean.dtseries.nii'
ses2RL=f'{fdir}/rfMRI_REST2_RL/rfMRI_REST2_RL_Atlas_MSMAll_hp2000_clean.dtseries.nii'
func_ses=[ses1LR,ses2RL,ses1RL,ses2LR]
##### cifti indices
lcort=slice(0,29696)
rcort=slice(29696, 59412)
cortAll=slice(0,59412)
#### left anatomical files
Lsrf32=f'{anatdir32}/{subj}.L.midthickness_MSMAll.32k_fs_LR.surf.gii'
LsrfNative=f'{anatdirNat}/{subj}.L.midthickness.native.surf.gii'
# Lsphere32=f'{anatdir32}/{subj}.L.sphere_MSMAll.32k_fs_LR.surf.gii'
# LsphereNat=f'{anatdirNat}/{subj}.L.sphere.reg.reg_LR.native.surf.gii'
# Laparc=f'{MNINonlinearDir}/fsaverage_LR32k/{subj}.L.aparc.a2009s.32k_fs_LR.label.gii'
#### right anatomical files
Rsrf32=f'{anatdir32}/{subj}.R.midthickness_MSMAll.32k_fs_LR.surf.gii'
RsrfNative=f'{anatdirNat}/{subj}.R.midthickness.native.surf.gii'
# Rsphere32=f'{anatdir32}/{subj}.R.sphere.32k_fs_LR.surf.gii'
# RsphereNat=f'{anatdirNat}/{subj}.R.sphere.reg.reg_LR.native.surf.gii'
# Raparc=f'{MNINonlinearDir}/fsaverage_LR32k/{subj}.R.aparc.a2009s.32k_fs_LR.label.gii'
##### start doing the things ####
print('getting cortex info')
func_ses411=[]
for data in range(len(func_ses)):
#### start by getting the indices of cortical vertices
func_ses411.append(get_corticalVertices(func_ses[data]))
##### smooth and clean the funcitonal time series
print(f'using kernel {kernel}')
### save the cifti encoded brainmodel information
np.save(f'{odir}/{subj}.cifti.info.npy', func_ses411[0])
#### to load dict back do as follows tst=np.load('.cifti.info.npy',allow_pickle=True)
print('smooth and clean the functional time series')
for data in range(len(func_ses)):
##### smooth and clean the funcitonal time series
print(func_ses[data])
func_ses[data]=wb_smoothCleanTs(subj,func_ses[data],kernel,Lsrf32,Rsrf32)
print('concatenating time series')
data=np.vstack(func_ses).T
data1=np.vstack([func_ses[0],func_ses[2]]).T
data2=np.vstack([func_ses[1],func_ses[3]]).T
del func_ses
print(data.shape)
print('the data type of the input is')
print(data.dtype)
print('######################')
print("generating cortical connectome")
print('full cort shape')
print(data[cortAll].shape)
rmat=np.corrcoef(data[cortAll])
print(rmat.shape)
print('correlation matrix done')
#np.save(f'{odir}/{subj}rmat.npy',rmat)
thresh=95
#### the first version of gradients is 90. this one is 95 and will be reflected in file output
thr=threshMat(rmat,thresh)
print('thresholding conn matrix to top 10% connectivity')
del rmat
# Check for minimum value
print("Minimum value is %f" % thr.min())
# The negative values are very small, but we need to know how many nodes have negative values
# Count negative values per row
N = thr.shape[0]
neg_values = np.array([sum(thr[i,:] < 0) for i in range(N)])
print("Negative values occur in %d rows" % sum(neg_values > 0))
thr[thr < 0] = 0
aff=cosine_similarity(thr)
#aff = 1 -squareform(pdist(thr, metric='cosine'))
del thr
print('is the affinity matrix symmetric?')
print(np.allclose(aff,aff.T))
print('#######################################')
print('')
print('affinity matrix built')
print(aff.shape)
#np.save(f'{odir}/{subj}CosAff.npy',aff)
print('running a quick little PCA')
from sklearn.decomposition import PCA
pca = PCA(n_components=3)
pca.fit(aff)
np.save(f'{odir}/{subj}.pca.{thresh}.s0{kernel}mm.npy',pca.components_)
print('pca output has dimensions')
print(pca.components_.shape)
print('doing embedding with mapalign')
emb= embed.compute_diffusion_map(aff, alpha = 0.5,n_components=3)
np.save(f'{odir}/{subj}.mapalign.{thresh}.diffmaps.0{kernel}mm.npy',emb.T)
print('embedding run wihtout errors')
del aff
#################################################################################
#################################################################################
#################################################################################
#REWRITE THIS WHOLE SECTOIN AS A FUNCTIONAL LATER
print(data1.shape)
print('the data1 type of the input is')
print(data1.dtype)
print('######################')
print("generating cortical connectome SESSION 1")
print('full cort shape')
print(data1[cortAll].shape)
rmat=np.corrcoef(data1[cortAll])
print(rmat.shape)
print('correlation matrix done')
#np.save(f'{odir}/{subj}rmat.npy',rmat)
thr=threshMat(rmat,95)
print('thresholding conn matrix to top 10% connectivity')
del rmat
# Check for minimum value
print("Minimum value is %f" % thr.min())
# The negative values are very small, but we need to know how many nodes have negative values
# Count negative values per row
N = thr.shape[0]
neg_values = np.array([sum(thr[i,:] < 0) for i in range(N)])
print("Negative values occur in %d rows" % sum(neg_values > 0))
thr[thr < 0] = 0
aff=cosine_similarity(thr)
#aff = 1 -squareform(pdist(thr, metric='cosine'))
del thr
print('is the affinity matrix symmetric?')
print(np.allclose(aff,aff.T))
print('#######################################')
print('')
print('affinity matrix built')
print(aff.shape)
#np.save(f'{odir}/{subj}CosAff.npy',aff)
print('running a quick little PCA')
from sklearn.decomposition import PCA
pca = PCA(n_components=3)
pca.fit(aff)
np.save(f'{odir}/{subj}.pca.{thresh}.ses1.s0{kernel}mm.npy',pca.components_)
print('pca output has dimensions')
print(pca.components_.shape)
print('doing embedding with mapalign')
emb= embed.compute_diffusion_map(aff, alpha = 0.5,n_components=3)
np.save(f'{odir}/{subj}.mapalign.{thresh}.ses1.diffmap.s0{kernel}mm.npy',emb.T)
print('embedding run wihtout errors')
del aff
print(data2.shape)
print('the data2 type of the input is')
print(data2.dtype)
print('######################')
print("generating cortical connectome SESSION 2")
print('full cort shape')
print(data2[cortAll].shape)
rmat=np.corrcoef(data2[cortAll])
print(rmat.shape)
print('correlation matrix done')
#np.save(f'{odir}/{subj}rmat.npy',rmat)
thr=threshMat(rmat,95)
print('thresholding conn matrix to top 10% connectivity')
del rmat
# Check for minimum value
print("Minimum value is %f" % thr.min())
# The negative values are very small, but we need to know how many nodes have negative values
# Count negative values per row
N = thr.shape[0]
neg_values = np.array([sum(thr[i,:] < 0) for i in range(N)])
print("Negative values occur in %d rows" % sum(neg_values > 0))
thr[thr < 0] = 0
aff=cosine_similarity(thr)
#aff = 1 -squareform(pdist(thr, metric='cosine'))
del thr
print('is the affinity matrix symmetric?')
print(np.allclose(aff,aff.T))
print('#######################################')
print('')
print('affinity matrix built')
print(aff.shape)
#np.save(f'{odir}/{subj}CosAff.npy',aff)
print('running a quick little PCA')
from sklearn.decomposition import PCA
pca = PCA(n_components=3)
pca.fit(aff)
np.save(f'{odir}/{subj}.pca.{thresh}.ses2.s0{kernel}mm.npy',pca.components_)
print('pca output has dimensions')
print(pca.components_.shape)
print('doing embedding with mapalign')
emb= embed.compute_diffusion_map(aff, alpha = 0.5,n_components=3)
np.save(f'{odir}/{subj}.mapalign.{thresh}.ses2.s0{kernel}mm.diffmap.npy',emb.T)
print('embedding run wihtout errors')