Per-sensor projection matrix that turns BEM node potentials into MEG coil readings or EEG electrode voltages. More...


Go to the source code of this file.
Classes | |
| class | FWDLIB::FwdBemSolution |
Channel-specific projection that contracts a BEM node-potential vector down to one entry per MEG coil or EEG electrode — i.e. the Geselowitz surface-integral weights cached as a dense ncoil × np matrix. More... | |
Namespaces | |
| namespace | FWDLIB |
| Forward modelling — BEM solver, spherical models, sensor/coil definitions and the lead-field assembly that links current dipoles to MEG/EEG sensor readings. | |
Per-sensor projection matrix that turns BEM node potentials into MEG coil readings or EEG electrode voltages.
SPDX-License-Identifier: BSD-3-Clause Copyright (c) 2017-2026 MNE-CPP Authors
Solving a BEM model produces a node-potential vector v of length np (one entry per surface vertex). For a fixed sensor array that vector is contracted with the ncoil × np matrix solution to yield the final lead-field column; the matrix encodes the Geselowitz surface-integral weights for MEG coils, or the trilinear-interpolation coefficients of the electrode position on the scalp triangulation for EEG. Caching this projection per channel set lets the dipole loop skip the full BEM forward substitution for every dipole and reduces the marginal cost of one source position to a single dense matrix-vector multiply.
Refactored from the fwdBemSolutionRec struct of MNE-C fwd_types.h.
Definition in file fwd_bem_solution.h.