def search_all_vs_one(key, plot=False, **kwargs):
"""For each pair of variables, look for a significant regression slope."""
sess = PINGDataSession() # username/password stored in an env variable for me.
sess.login()
all_data = compute_all_asymmetries(prefix=['MRI_cort_area', 'MRI_cort_thick',
'MRI_subcort_vol', 'DTI_fiber_vol'])
results = []
keys = list(set(all_data.keys()) - set(('SubjID',)))
for loop_key in keys:
results.append(sess.regress(key, loop_key, plot=plot, **kwargs))
if plot:
plt.show()
def search_all_vs_itself(covariates, plot=False, **kwargs):
"""For each pair of variables, look for a significant regression slope."""
def add_subplot(fh, *args):
return fh.add_subplot(*args) if plot else None
sess = PINGDataSession() # username/password stored in an env variable for me.
sess.login()
all_data = compute_all_asymmetries(prefix=['MRI_cort_area', 'MRI_cort_thick',
'MRI_subcort_vol', 'DTI_fiber_vol'])
results = []
keys = list(set(all_data.keys()) - set(('SubjID',)))
for X in keys:
# Get relevant covariates
try:
added_covariates = sess.AI2flds(X)
# Then regress on each side
fh2 = plt.figure(figsize=(18, 6)) if plot else None
sess.regress(X.replace('_AI', '_LH_PLUS_RH'),
X,
covariates=covariates, plot=add_subplot(fh2, 1, 3, 1),
**kwargs)
sess.regress(added_covariates[1],
X,
covariates=covariates + ['MRI_cort_area_ctx_total_LH_PLUS_RH'],
plot=add_subplot(fh2, 1, 3, 2),
**kwargs)
sess.regress('MRI_cort_area_ctx_total_LH_PLUS_RH',
X,
covariates=covariates,
plot=add_subplot(fh2, 1, 3, 3),
**kwargs)
except Exception as e:
print("Failed for %s (%s); moving on..." % (X, e))
if plot:
plt.show()
