Meg information based connectivity coding session: Difference between revisions
Jump to navigation
Jump to search
Content added Content deleted
(Created page with "!Under Construction - will be finalized by Friday! We will be meeting virtually over Zoom from 1pm-3pm with breakout rooms for questions and collaboration There will be a qui...") |
No edit summary |
||
(18 intermediate revisions by the same user not shown) | |||
Line 1:
We will be meeting virtually over Zoom from 1pm-3pm with breakout rooms for questions and collaboration <br>
There will be a quick discussion
==FRITES toolbox
[https://megcore.nih.gov/MEG/FRITES_Tutorial_07132022.mp4 Club MEG Presentation Recording]
[ Background]▼
[https://brainets.github.io/frites/ Background]
[
[https://brainets.github.io/frites/auto_examples/index.html Example Scripts]
===Install in new environment (RECOMMENDED!):===
conda create -n frites conda-forge::mne conda-forge::jupyterlab pip -y
conda activate frites
pip install frites
git clone https://github.com/brainets/CookingFrites.git
cd CookingFrites
jupyter lab
#Once jupyter is open, the notebooks are in the notebooks folder
#For MEG data to use with this --- Just test data, there are many optimizations that have not been performed
pip install git+https://github.com/jstout211/connectivity_tutorial.git
cp /vf/users/MEGmodules/modules/frites_dset.zip ./
unzip frites_dset.zip
#In python, frites_output_root=$(pwd)/frites_dset
ipython
from frites_connectivity_tutorial.load_data import load_dataset
dt = load_dataset(<<FRITES_OUTPUT_ROOT>>)
===Install in current environment (Better to use New environment as listed above!)-===
#Download getting started tutorials
git clone https://github.com/brainets/CookingFrites.git
pip install frites
===Use on biowulf===
sinteractive --mem=12G --cpus-per-task=8
module use --append /data/MEGmodules/modulefiles
module load mne_spyder #loads mne/mne-bids/frites/frites_tutorial
==SAM Symoblic Transfer Entropy Connectivity ==
[https://megcore.nih.gov/MEG/Robinson_InformationTheoryMEG_ClubMEG_03102022.mp4 Club MEG SAM Connectivity Presentation Recording]
====Recommendations====
This process is a trigger-free process that estimates the symbolic transfer entropy (connectivity assessed by delayed propagation of signals) between two ROIs
70-185Hz has been tested and is recommended as a filter band
Use an atlas based input for the bivariate connectivity assessment (AAL atlas, Desikan K., Destreux...)
SAMwts or Patch_wts to create the beamformer
Output is a text file
Must determine an embedding dimension (4 as default) - the higher the embedding the more data needed
Minimal requirements (>=1Khz sampling rate and >=300s of data)
The inputs do not need to be brain regions (for example can compare STE between audio track to brain region)
===Use on Biowulf===
#Log into biowulf
sinteractive --mem=6G --cpus-per-task=4 #You may need more mem/cpus
#Make sure modules are loadable -- this can also be added to ~/.bashrc file
module use --append /data/MEGmodules/modulefiles
module load SAMsrcDev
#Make the SAM beamformers and apply the symbolic transfer entropy
sam_cov ...
sam_wts ...
STEdelay ....
Usage: STEdelay [options] Version 5.0 (64-bit) rev-1, Aug 10 2022
Options:
If a parameter begins with '--', it is allowed on the command line,
otherwise it is only allowed in a parameter file (see -m).
All times are in seconds.
-h, --help show this help
-r DSNAME, --DataSet DSNAME
MEG dataset name [env_var = ds]
-m PFILE, --param PFILE parameter file name (optionally ending
in ".param") [env_var = param]
-v, --verbose verbose output
--CovType GLOBAL|SUM|ALL which covariance matrix to use for the analysis
--CovBand LO HI covariance bandwidth limits in Hz
--ImageBand LO HI imaging covariance bandwidth limits in Hz
--OrientBand LO HI band to use for orientation instead of Global (Hz)
--SmoothBand LO HI used to smooth Hilbert envelope, kurtosis, or RVE (Hz)
--Notch apply a powerline notch
(including harmonics)
--Hz HZ frequency for powerline (Hz, default 60)
ImageMetric METRICSPEC imaging metric
--Extent DIST radial extent from ROI voxels or centroid (mm)
-t TARGETFILE, --TargetName TARGETFILE
file containing target coordinates
-i MRIDIR, --MRIDirectory MRIDIR
MRI directory root [env_var = mridir]
-o IMAGEDIR, --ImageDirectory IMAGEDIR
image output directory [env_var = imagedir]
--PrefixLength N DataSet prefix may be specified as a number
or as a prefix delimiter (default "_")
|
Latest revision as of 13:58, 12 August 2022
We will be meeting virtually over Zoom from 1pm-3pm with breakout rooms for questions and collaboration
There will be a quick discussion and some example data on biowulf.
FRITES toolbox
Club MEG Presentation Recording
Background Getting Started Scripts Example Scripts
Install in new environment (RECOMMENDED!):
conda create -n frites conda-forge::mne conda-forge::jupyterlab pip -y conda activate frites pip install frites
git clone https://github.com/brainets/CookingFrites.git cd CookingFrites jupyter lab #Once jupyter is open, the notebooks are in the notebooks folder
#For MEG data to use with this --- Just test data, there are many optimizations that have not been performed pip install git+https://github.com/jstout211/connectivity_tutorial.git cp /vf/users/MEGmodules/modules/frites_dset.zip ./ unzip frites_dset.zip
#In python, frites_output_root=$(pwd)/frites_dset ipython from frites_connectivity_tutorial.load_data import load_dataset dt = load_dataset(<<FRITES_OUTPUT_ROOT>>)
Install in current environment (Better to use New environment as listed above!)-
#Download getting started tutorials git clone https://github.com/brainets/CookingFrites.git pip install frites
Use on biowulf
sinteractive --mem=12G --cpus-per-task=8 module use --append /data/MEGmodules/modulefiles module load mne_spyder #loads mne/mne-bids/frites/frites_tutorial
SAM Symoblic Transfer Entropy Connectivity
Background
Club MEG SAM Connectivity Presentation Recording
Recommendations
This process is a trigger-free process that estimates the symbolic transfer entropy (connectivity assessed by delayed propagation of signals) between two ROIs 70-185Hz has been tested and is recommended as a filter band Use an atlas based input for the bivariate connectivity assessment (AAL atlas, Desikan K., Destreux...) SAMwts or Patch_wts to create the beamformer Output is a text file Must determine an embedding dimension (4 as default) - the higher the embedding the more data needed Minimal requirements (>=1Khz sampling rate and >=300s of data) The inputs do not need to be brain regions (for example can compare STE between audio track to brain region)
Use on Biowulf
#Log into biowulf sinteractive --mem=6G --cpus-per-task=4 #You may need more mem/cpus #Make sure modules are loadable -- this can also be added to ~/.bashrc file module use --append /data/MEGmodules/modulefiles module load SAMsrcDev
#Make the SAM beamformers and apply the symbolic transfer entropy sam_cov ... sam_wts ... STEdelay ....
Usage: STEdelay [options] Version 5.0 (64-bit) rev-1, Aug 10 2022 Options: If a parameter begins with '--', it is allowed on the command line, otherwise it is only allowed in a parameter file (see -m). All times are in seconds. -h, --help show this help -r DSNAME, --DataSet DSNAME MEG dataset name [env_var = ds] -m PFILE, --param PFILE parameter file name (optionally ending in ".param") [env_var = param] -v, --verbose verbose output --CovType GLOBAL|SUM|ALL which covariance matrix to use for the analysis --CovBand LO HI covariance bandwidth limits in Hz --ImageBand LO HI imaging covariance bandwidth limits in Hz --OrientBand LO HI band to use for orientation instead of Global (Hz) --SmoothBand LO HI used to smooth Hilbert envelope, kurtosis, or RVE (Hz) --Notch apply a powerline notch (including harmonics) --Hz HZ frequency for powerline (Hz, default 60) ImageMetric METRICSPEC imaging metric --Extent DIST radial extent from ROI voxels or centroid (mm) -t TARGETFILE, --TargetName TARGETFILE file containing target coordinates -i MRIDIR, --MRIDirectory MRIDIR MRI directory root [env_var = mridir] -o IMAGEDIR, --ImageDirectory IMAGEDIR image output directory [env_var = imagedir] --PrefixLength N DataSet prefix may be specified as a number or as a prefix delimiter (default "_")