Describe the new feature or enhancement
This is Tomasz Piotrowski from Nicolaus Copernicus University in
Toruń, Poland. I am copying Julia Jurkowska from the University of
Warsaw, Poland.
Together with our collaborators, we recently completed work on a
source localization method that extends the classical LCMV-based
approach. The method first identifies candidate active sources using
multi-source neural activity indices and then reconstructs their
activity using a spatial filter. We chose LCMV for the initial
implementation because it is widely used and already well supported in
MNE-Python.
The paper is available here:
https://www.sciencedirect.com/science/article/pii/S1053811926004416
The accompanying code is available at:
https://github.com/julia-jurkowska/mvpure-tools
and the tutorials at:
https://julia-jurkowska.github.io/mvpure-tools/index.html
We developed the implementation specifically for use within the
MNE-Python ecosystem. We would therefore like to ask whether it would
make sense to prepare a pull request with the aim of eventually
including the method in MNE-Python.
Describe your proposed implementation
The first stage, localization using multi-source neural activity indices, could be implemented as an optional step before running the existing LCMV procedure.
The proposed two-stage localization-reconstruction approach would also allow any spatial filter to be used in the second stage for source reconstruction.
Describe possible alternatives
We aimed to align the proposed approach as closely as possible with the existing MNE-Python implementation of spatial filtering and to integrate it naturally into the current framework. It is not intended as an alternative, but rather as a nontrivial extension of the existing functionality.
Additional context
No response
Describe the new feature or enhancement
This is Tomasz Piotrowski from Nicolaus Copernicus University in
Toruń, Poland. I am copying Julia Jurkowska from the University of
Warsaw, Poland.
Together with our collaborators, we recently completed work on a
source localization method that extends the classical LCMV-based
approach. The method first identifies candidate active sources using
multi-source neural activity indices and then reconstructs their
activity using a spatial filter. We chose LCMV for the initial
implementation because it is widely used and already well supported in
MNE-Python.
The paper is available here:
https://www.sciencedirect.com/science/article/pii/S1053811926004416
The accompanying code is available at:
https://github.com/julia-jurkowska/mvpure-tools
and the tutorials at:
https://julia-jurkowska.github.io/mvpure-tools/index.html
We developed the implementation specifically for use within the
MNE-Python ecosystem. We would therefore like to ask whether it would
make sense to prepare a pull request with the aim of eventually
including the method in MNE-Python.
Describe your proposed implementation
The first stage, localization using multi-source neural activity indices, could be implemented as an optional step before running the existing LCMV procedure.
The proposed two-stage localization-reconstruction approach would also allow any spatial filter to be used in the second stage for source reconstruction.
Describe possible alternatives
We aimed to align the proposed approach as closely as possible with the existing MNE-Python implementation of spatial filtering and to integrate it naturally into the current framework. It is not intended as an alternative, but rather as a nontrivial extension of the existing functionality.
Additional context
No response