def run_perm_test(row):
row_array = row.split(';')
perm_ind = row_array[0]
perm_y = row_array[1]
perm_y = np.asarray(perm_y.split(',')).astype(int)
networks = [
'visual', 'somatomotor', 'dorsal_attention', 'ventral_attention',
'limbic', 'default_mode', 'fronto_parietal'
]
for network in networks:
analysis_name = 'schaefer_{}_noTiv_perm_'.format(network) + str(
perm_ind)
data_folder = '/scratch/tmp/wintern/iq_frankfurt/'
project_folder = '/scratch/tmp/wintern/iq_frankfurt/results/noTivRescaling/perm/schaefer_{}'.format(
network)
os.makedirs(project_folder, exist_ok=True)
cache_dir = '/scratch/tmp/wintern/cache'
# get data
data = IQData(data_folder=data_folder, tiv_rescaled=False)
covariates = np.asarray([data.age, data.gender, data.handedness]).T
data.load_schaefer_network(network, use_cached=True)
# run analysis
pipe = construct_hyperpipe_schaefer(analysis_name, project_folder,
cache_dir)
pipe.groups = data.fsiq
pipe.fit(data.schaefer_network, perm_y, **{'covariates': covariates})
os.remove(pipe.output_settings.pretrained_model_filename)
Translationale Psychiatrie
Universitaetsklinikum Muenster
"""
import sys
sys.path.append('/scratch/tmp/wintern/iq_frankfurt/photonai')
sys.path.append('/scratch/tmp/wintern/iq_frankfurt/')
from analyses.analysis_base import construct_hyperpipe_schaefer
from data.data import IQData
import numpy as np
import os
networks = ['visual', 'somatomotor', 'dorsal_attention', 'ventral_attention', 'limbic',
'default_mode', 'fronto_parietal']
for network in networks:
analysis_name = 'schaefer_{}'.format(network)
data_folder = '/scratch/tmp/wintern/iq_frankfurt/'
project_folder = '/scratch/tmp/wintern/iq_frankfurt/results/TivRescaling/schaefer_{}'.format(network)
os.makedirs(project_folder, exist_ok=True)
cache_dir = '/scratch/tmp/wintern/cache'
# get data
data = IQData(data_folder=data_folder, tiv_rescaled=True)
covariates = np.asarray([data.age, data.gender, data.handedness]).T
data.load_schaefer_network(network, use_cached=False)
# run analysis
pipe = construct_hyperpipe_schaefer(analysis_name, project_folder, cache_dir)
pipe.fit(data.schaefer_network, data.fsiq, **{'covariates': covariates})
os.remove(pipe.output_settings.pretrained_model_filename)