InvGammaMap

Namespace: INVLIB  ·  Library: Inverse Library

Python equivalent

mne.inverse_sparse.gamma_map in MNE-Python.

#include <inv/inv_gamma_map.h>

class INVLIB::InvGammaMap

Gamma-MAP sparse inverse solver (Sparse Bayesian Learning).

Iteratively estimates source variance hyperparameters (gamma) and prunes sources whose gamma falls below a threshold, yielding a sparse solution.

Gamma-MAP sparse inverse solver.

src/examples/ex_inv_api/main.cpp

// Gamma-MAP: sparse Bayesian learning with unit (whitened) noise, MacKay updates as in mne.inverse_sparse.gamma_map.
const InvGammaMapResult gammaMap = InvGammaMap::compute(gain, measured, MatrixXd::Identity(gain.rows(), gain.rows()), 1000, 1e-8);

---

Static Methods

compute(matGain, matData, matNoiseCov, nIterations, tolerance, gammaThreshold)

Compute the Gamma-MAP inverse solution.

Parameters:

• matGain : const Eigen::MatrixXd & Forward gain matrix (n_channels x n_sources).

• matData : const Eigen::MatrixXd & Measurement data (n_channels x n_times).

• matNoiseCov : const Eigen::MatrixXd & Noise covariance matrix (n_channels x n_channels).

• nIterations : int Maximum number of fixed-point iterations.

• tolerance : double Convergence tolerance on relative gamma change.

• gammaThreshold : double Threshold below which sources are pruned.

Returns:

• InvGammaMapResult — The Gamma-MAP result containing the sparse source estimate.

---

Example

• Full source: src/examples/ex_inv_api/main.cpp
• Build target: ex_inv_api (runs under CTest)
• Data: mne-cpp-test-data

Used in examples

ex_inv_api

Inverse library: minimum-norm estimates, resolution matrices, label time courses, sparse solvers.

src/examples/ex_inv_api/main.cpp, line 250

// Gamma-MAP: sparse Bayesian learning with unit (whitened) noise, MacKay updates as in mne.inverse_sparse.gamma_map.
const InvGammaMapResult gammaMap = InvGammaMap::compute(gain, measured, MatrixXd::Identity(gain.rows(), gain.rows()), 1000, 1e-8);
ok &= expect(!gammaMap.activeVertices.isEmpty() && gammaMap.activeVertices.size() < 70 && gammaMap.vecGamma.size() == 70,
             QString("InvGammaMap: %1 of 70 V1 sources keep a nonzero variance").arg(gammaMap.activeVertices.size()));

Authors of this file

• Christoph Dinh <christoph.dinh@mne-cpp.org>