Skip to main content

InvVectorSourceEstimate

Namespace: INVLIB  ·  Library: Inverse Library

Python equivalent

mne.VectorSourceEstimate in MNE-Python.

#include <inv/inv_vector_source_estimate.h>

class INVLIB::InvVectorSourceEstimate

Vector source estimate: each vertex carries a 3D orientation vector across time.

data has shape (n_vertices*3 x n_times). Rows are interleaved: [x0,y0,z0, x1,y1,z1, ...]. Use magnitude() to collapse to scalar (n_vertices x n_times) for visualisation.

src/examples/ex_inv_api/main.cpp
// The MNE kernel keeps x, y and z per source: an apply_inverse(..., pick_ori="vector") estimate.
const MatrixXd goodData = evoked.pick_channels(inverse.noise_cov->names).data;
const InvVectorSourceEstimate vectorStc(kernel * goodData, stc.vertices, stc.tmin, stc.tstep);
const InvSourceEstimate magnitude = vectorStc.magnitude(); // 70 x 421 vector lengths

Inheritance​

Public Methods​

InvVectorSourceEstimate()​

InvVectorSourceEstimate(p_sol, p_vertices, p_tmin, p_tstep)​

Construct from data and vertices.

Parameters:

  • p_sol : const Eigen::MatrixXd & Data (n_vertices*3 x n_times).

  • p_vertices : const Eigen::VectorXi & Vertex indices (n_vertices).

  • p_tmin : float Start time.

  • p_tstep : float Time step.

nVertices()​

Number of source vertices (data.rows() / 3).

Returns:

  • int — Number of vertices, each carrying an x/y/z component row.

magnitude()​

Compute the magnitude (L2 norm) at each vertex across time.

Returns:

vertexData(vertexIdx)​

Extract the (x, y, z) data for a specific vertex.

Parameters:

  • vertexIdx : int Local index into vertices array.

Returns:

  • Eigen::MatrixXd — Matrix (3 x n_times) for the requested vertex.

projectToNormals(normals)​

Project the 3D vectors onto surface normals to produce a signed scalar estimate.

Parameters:

  • normals : const Eigen::MatrixX3f & Surface normals (n_vertices x 3).

Returns:

Example​

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 194
// The MNE kernel keeps x, y and z per source: an apply_inverse(..., pick_ori="vector") estimate.
const MatrixXd goodData = evoked.pick_channels(inverse.noise_cov->names).data;
const InvVectorSourceEstimate vectorStc(kernel * goodData, stc.vertices, stc.tmin, stc.tstep);
const InvSourceEstimate magnitude = vectorStc.magnitude(); // 70 x 421 vector lengths
// mne.minimum_norm.apply_inverse(evoked, inv, 1/9, "MNE", pick_ori="vector"): sum of lengths 4.31370e-4
ok &= expect(vectorStc.nVertices() == 70 && near(magnitude.data.sum(), 0.0004313698431300483, 1e-4) && near(magnitude.data(0, 200), 3.865084798591962e-09, 1e-4),
             "InvVectorSourceEstimate: vector MNE lengths match apply_inverse(pick_ori=\"vector\")");

Authors of this file​