Tool for computing brain network stability, a biomarker for brain aging.
Project description
BRAin NEtwork STAbility
Introduction
Tool for computing brain network stability, a biomarker for brain aging.
Free software: MIT license
Documentation: https://branesta.readthedocs.io.
Please cite our article:
Mujica-Parodi, Lilianne R., et al. “Diet modulates brain network stability, a biomarker for brain aging, in young adults.” Proceedings of the National Academy of Sciences 117.11 (2020): 6170-6177. link: https://www.pnas.org/content/117/11/6170
Description
Brain network stability measures the extent of temporal reorganization that takes place in brain networks. Brain networks describe inter-regional communication across the brain. Lower network stability (represented by higher values) is related to weaker persistence of brain networks. The terms Network Stability and Network INstability are used interchangibly and they refer to the exact same metric.
The procedure of computing brain network stability is as follows: fMRI time-series that were previously parcelled into ROIs are first binned into time windows (=snapshots) of N timepoints without overlaps. Next, pairwise correlations among all ROIs are computed separately for each time window. For the whole brain, brain network stability (scalar) is quantified by taking the l2 norm of the element-wise differences of correlation matrices corresponding to two different snapshots. τ is the number of steps separating two snapshots from which a given value of brain network stability is calculated from. For instance, if τ=1, two consecutive snapshots snapshots are used (e.g. #4 and #5). If τ=16, then 16 snapshots are separating the two snapshots (e.g. #3 and #19). If the time-series have a length of 720 timepoints, then there will be 24 snapshots (720/30=24, given N=30). At τ=1, there are 23 instability values, whereas at τ=20 4 different instability values are calculated.
For functional networks (labeled as “subnetworks” in our program), the procedure is analog to the above. The only difference is that once correlations are computed for each time window, element-wise differences are calculated only across those ROIs that spatially overlap with the functional network. In order to facilitate comparison of network instability among networks consisting of different number of nodes, network stability is normalized with the number of edges.
Features
computes network stability from parcelled time-series
performs computations at every τ
computes for subnetworks (optional)
allows user-defined time window length
easy to install (pip)
command line tool
Credits
This package was developed within the Laboratory for Computational Neurodiagnostics (LCNeuro).
This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.
History
0.1.7 (2020-12-22)
Logo fix
0.1.6 (2020-12-22)
Added logo
0.1.5 (2020-12-22)
Added unittests.
Updated imports.
0.1.4 (2020-11-29)
Updated logging.
0.1.3 (2020-11-28)
First standalone release.
0.1.2 (2020-11-28)
Updated dependencies.
0.1.1 (2020-11-27)
Added analysis module.
0.1.0 (2020-11-26)
First release on PyPI.
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
Built Distribution
Hashes for branesta-0.1.7-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 78e6d932eeabf14a2d9f94e1e252900d0e0a8f0b50f70d9c378f95bdb1a87134 |
|
MD5 | 18611dc08b92924b7d0d91d1082d9e79 |
|
BLAKE2b-256 | 03eedbc89b178bc129b7014bd4518b1012629c7e2dbb5b308adbcf8faf13d934 |