LEADFIELD ATLAS

Installation

Docker

sudo apt-get install docker docker.io Reference

First go

docker run -ti -v $(pwd):/home/fenics/shared --name fenics-container quay.io/fenicsproject/stable:2016.2.0

Subsequent access

docker start fenics-container docker exec -ti -u fenics fenics-container /bin/bash -l

To stop

docker stop fenics-container

List containers

docker ps -a

Remove a container

docker rm -f CONTAINER

Build from source

Required to access the barycenters of the meshes. Critical step - heavily unstable and unreliable. Only used to obtain the bary points at this stage.

See ./build_script.sh for full details on building

This can be also avoided by moving def compute_barycenters to elsewhere [TODO - not crucial]

Mesh generation

This step requires gmsh and dolfin. - do not have to work together Can be skipped if the .xml files are already provided

ITKSnap to stl

to export the rough surface of the brain

Meshlab

stl smoothing and surface fixing

blender

boolean subtracting of the ground electrode.

GMSH

version 3.0+ (installed via sudo apt-get install) see /mesh/mesh_it.sh

Use setup_refined.geo to obtain setup_refined.msh dolfin then converts it setup_refined.msh into mesh_setup*.xml

Probe points

This requires vtk 6.3.0 version installed correctly access to numpy and nifti file This ALSO requires the right fenics version installed along side. Hence the need to mesh.

See ./probe_points.py

Goes through the meshes, fetches the barycenters, goes though the imgage files and fills the necessary fields corresponding to those barycenters and dumps these files inside /points folder

Can skip this step if the points are already provided.

Conductivity

This requires the output from the probepoinst to be in place Can be avoided if simgas are already provided.

See ./conductivity.py

Populates ./sigmas//.xml.gz Which correspond to the sigmas at the mesh barycentes populated via conductivity_c.py file C-code

Forward method

Currently works with point sources sent in as a list of points corresponding to the nifti world coordinates. Specifying the electrode positions is not implemented. Currently dumps as hdf5 files for the phi vector of the potential space.

Critical dependency is with HDF5 - works smooth with docker.

Works only with hypre_amg preconditioner via PETSCL with hypre libraries - exists in docker image luckily.