A tool to perform Freesurfer volume Harminization in unseen scanner.
Project description
Neuroharmony: A tool for harmonizing volumetric MRI data from unseen scanners
The model presented in Garcia-Dias, et al. (2020).
Documentation
Install Neuroharmony.
pip install neuroharmony
Example of use:
Pre-trained Neuroharmony model
An example plot of how to load and apply pre-trained a Neuroharmony model.
import matplotlib.pyplot as plt
from neuroharmony.models.harmonization import fetch_trained_model, fetch_sample
import seaborn as sns
X = fetch_sample()
neuroharmony = fetch_trained_model()
x_harmonized = neuroharmony.transform(X)
rois = ['Left-Hippocampus',
'lh_bankssts_volume',
'lh_posteriorcingulate_volume',
'lh_superiorfrontal_volume',
'rh_frontalpole_volume',
'rh_parsopercularis_volume',
'rh_parstriangularis_volume',
'rh_superiorfrontal_volume',
'Right-Cerebellum-White-Matter',
]
fig, axes = plt.subplots(3, 3, figsize=(10, 10))
for roi, ax in zip(rois, axes.flatten()):
ax.plot(neuroharmony.kde_data_[roi]['x'], neuroharmony.kde_data_[roi]['y'],
color='#fcb85b', ls='--', label='ComBat harmonized training set')
sns.kdeplot(X[roi], color='#f47376', ls=':', legend=False, ax=ax, label='Original test set')
sns.kdeplot(x_harmonized[roi], color='#00bcab', ls='-', legend=False, ax=ax, label='Harmonized test set')
ax.set_xlabel(roi, fontsize=13)
axes.flatten()[2].legend(ncol=3, bbox_to_anchor=(0.8, 1.175), fontsize=13)
axes.flatten()[3].set_ylabel('Density', fontsize=15)
plt.subplots_adjust(left=0.07, right=0.99,
bottom=0.05, top=0.96,
hspace=0.20, wspace=0.20)
plt.savefig('test.png', dpi=200)
plt.show()
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Neuroharmony-0.1.3.tar.gz
(2.9 kB
view hashes)
Built Distribution
Close
Hashes for Neuroharmony-0.1.3-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 28737474fbd9164af432cee0b4e3e80c37fe65fb9b39c3a9c2d33e81bbf25dfb |
|
MD5 | 487b82497a29f85b6cda652fcc46da20 |
|
BLAKE2b-256 | 8170b7e8163da8bf551a871eb08190774c112a4a9e204c4017656b627a705ba9 |