notebook.community

Edit and run



In [254]:

    
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns
import subprocess



In [2]:

    
!ls









    



2016-09-06-neurovault_metadata.csv all_tasks.rdf
2016-09-06-pyneurovault.ipynb      all_tasks_parsed.txt
Untitled.ipynb                     cogatlas_explore.ipynb
all_concepts.rdf



In [156]:

    
df = pd.read_csv('2016-09-06-neurovault_metadata.csv')
df.shape









    Out[156]:





(14228, 132)



In [7]:

    
wh_cogatlas = [c for c in df.columns if 'cogatlas' in c]



In [221]:

    
[(n,c) for n,c in enumerate(df.columns) if not str.isupper(c[0])]









    Out[221]:





[(87, 'add_date'),
 (88, 'analysis_level'),
 (89, 'brain_coverage'),
 (90, 'brainpedia_labels'),
 (91, 'brainpedia_study'),
 (92, 'cognitive_contrast_cogatlas'),
 (93, 'cognitive_contrast_cogatlas_id'),
 (94, 'cognitive_paradigm_cogatlas'),
 (95, 'cognitive_paradigm_cogatlas_id'),
 (96, 'cognitive_paradigm_description_url'),
 (97, 'collection'),
 (98, 'collection_id'),
 (99, 'contrast_definition'),
 (100, 'contrast_definition_cogatlas'),
 (101, 'custom_prop'),
 (102, 'data'),
 (103, 'description'),
 (104, 'figure'),
 (105, 'file'),
 (106, 'file_size'),
 (107, 'id'),
 (108, 'image_type'),
 (109, 'is_thresholded'),
 (110, 'is_valid'),
 (111, 'label_description_file'),
 (112, 'map_type'),
 (113, 'modality'),
 (114, 'modify_date'),
 (115, 'n-back'),
 (116, 'name'),
 (117, 'nidm_results'),
 (118, 'nidm_results_ttl'),
 (119, 'not_mni'),
 (120, 'number_of_subjects'),
 (121, 'perc_bad_voxels'),
 (122, 'perc_voxels_outside'),
 (123, 'random_property_1'),
 (124, 'random_property_2'),
 (125, 'reduced_representation'),
 (126, 'smoothness_fwhm'),
 (127, 'statistic_parameters'),
 (128, 'task'),
 (129, 'thumbnail'),
 (130, 'type'),
 (131, 'url')]



In [42]:

    
_ = sns.distplot(df.groupby('cognitive_contrast_cogatlas_id')['id'].count(), kde=False)



In [48]:

    
for i in df['modality'].unique():
    print i









    



nan
fMRI-BOLD
Structural MRI
Other
fMRI-CBV
MEG
Diffusion MRI
fMRI-CBF
EEG
PET other
PET FDG



In [66]:

    
df = df.drop(np.where(df['modality'] != 'fMRI-BOLD')[0])



In [67]:

    
df.shape









    Out[67]:





(9640, 132)



In [80]:

    
df.groupby('map_type')['modality'].count()









    Out[80]:





map_type
F map                              17
P map (given null hypothesis)     146
ROI/mask                          238
T map                            1505
Z map                            2519
anatomical                         25
multivariate-beta map              56
other                            1044
parcellation                       66
univariate-beta map              4914
Name: modality, dtype: int64



In [82]:

    
xt = pd.crosstab(df['modality'], df['map_type'])



In [322]:

    
nv = pd.read_csv('2016-09-06-neurovault_metadata.csv')
nv.shape









    Out[322]:





(14228, 132)



In [329]:

    
for i in nv.map_type.unique():
    print i









    



T map
Z map
other
anatomical
parcellation
P map (given null hypothesis)
F map
ROI/mask
nan
univariate-beta map
multivariate-beta map



In [ ]:

    
nv['map_type'].isin(['T map', 'Z map'])



In [313]:

    
wh_labeled = []
wh_labeled.append(nv['cognitive_contrast_cogatlas'].isnull())
wh_labeled.append(nv['cognitive_contrast_cogatlas'] == 'Other')
wh_labeled.append(nv['cognitive_contrast_cogatlas'] == 'None / Other')
wh_labeled.append(nv['modality'] != 'fMRI-BOLD')
wh_labeled.append(nv['map_type'].str.contains('map') == False)
wh_labeled.append(nv['map_type'])



In [ ]:

    
wh_unlabeled = []
wh_unlabeled.append(nv['cognitive_contrast_cogatlas'].isnull())
wh_unlabeled.append(nv['cognitive_contrast_cogatlas'] == 'Other')
wh_unlabeled.append(nv['cognitive_contrast_cogatlas'] == 'None / Other')
wh_unlabeled.append(nv['modality'] == 'fMRI-BOLD')
wh_unlabeled.append(nv['map_type'].str.contains('map'))



In [317]:

    
nv_old = nv.copy()
 = nv.loc[~np.any(np.array(wh_labeled), axis=0), :]
print nv.shape



In [318]:

    
nv = nv.drop_duplicates()
print nv.shape



In [319]:

    
x = nv['file'].iloc[0]
x.split('/')[-1]
# ${x##*/}









    Out[319]:





'task001_cope06_neg_loss_param.nii.gz'



In [320]:

    
# proc = subprocess.Popen(['/bin/bash'], stdin=subprocess.PIPE, stdout=subprocess.PIPE)
!rm -r images
!mkdir images
n=0
for f in nv['file']:
#     n=n+1
#     if n > 20:
#         break
    fout = f.split('/')[-1]
    print "downloading: ", fout
    subprocess.call('curl %s > images/%s' % (f, fout), shell=True)
#     subprocess.call('touch images/%s' % (fout), shell=True)









    



downloading:  task001_cope06_neg_loss_param.nii.gz
downloading:  task001_cope04_distance%20from%20indifference.nii.gz
downloading:  task001_cope02_parametric%20gain.nii.gz
downloading:  task001_cope01_task.nii.gz
downloading:  tstatsCashoutsMinusPumpMasked.nii.gz
downloading:  Fig3A_zstat1_2.nii.gz
downloading:  task001_cope01_pumps_fixed.nii.gz
downloading:  task001_cope02_pumps_demean.nii.gz
downloading:  task001_cope07_explode_fixed.nii.gz
downloading:  task001_cope04_cash_fixed.nii.gz
downloading:  task001_cope03_pumps_fixed_real_RT.nii.gz
downloading:  task001_cope01_task.nii.gz
downloading:  task002_cope02_feedback.nii.gz
downloading:  task002_cope01_task.nii.gz
downloading:  task003_cope01_classification-probabilistic.nii.gz
downloading:  task001_cope02_feedback.nii.gz
downloading:  task003_cope02_classification-deterministic.nii.gz
downloading:  task002_cope10_succstop_minus_go.nii.gz
downloading:  task002_cope08_unsucc_minus_succ_stop.nii.gz
downloading:  task002_cope09_go_crit-go_noncrit.nii.gz



In [321]:

    
!lsi images --h -V









    



IMAGE                                                          DIM1  DIM2  DIM3  DIM4     VOXELS         mm^3
images/Fig3A_zstat1_2.nii.gz                                     91   109    91     1     194758   1558064.00
images/task001_cope01_pumps_fixed.nii.gz                         91   109    91     1     228397   1827176.00
images/task001_cope01_task.nii.gz                                91   109    91     1     228180   1825440.00
images/task001_cope02_feedback.nii.gz                            91   109    91     1     228180   1825440.00
images/task001_cope02_parametric%20gain.nii.gz                   91   109    91     1     228483   1827864.00
images/task001_cope02_pumps_demean.nii.gz                        91   109    91     1     228397   1827176.00
images/task001_cope03_pumps_fixed_real_RT.nii.gz                 91   109    91     1     228397   1827176.00
images/task001_cope04_cash_fixed.nii.gz                          91   109    91     1     228397   1827176.00
images/task001_cope04_distance%20from%20indifference.nii.gz      91   109    91     1     228483   1827864.00
images/task001_cope06_neg_loss_param.nii.gz                      91   109    91     1     228483   1827864.00
images/task001_cope07_explode_fixed.nii.gz                       91   109    91     1     228397   1827176.00
images/task002_cope01_task.nii.gz                                91   109    91     1     228149   1825192.00
images/task002_cope02_feedback.nii.gz                            91   109    91     1     228149   1825192.00
images/task002_cope08_unsucc_minus_succ_stop.nii.gz              91   109    91     1     228316   1826528.00
images/task002_cope09_go_crit-go_noncrit.nii.gz                  91   109    91     1     228316   1826528.00
images/task002_cope10_succstop_minus_go.nii.gz                   91   109    91     1     228316   1826528.00
images/task003_cope01_classification-probabilistic.nii.gz        91   109    91     1     227937   1823496.00
images/task003_cope02_classification-deterministic.nii.gz        91   109    91     1     227937   1823496.00
images/tstatsCashoutsMinusPumpMasked.nii.gz                      91   109    91     1     228483   1827864.00