Volume-conduction-robust imaginary coherence (Nolte et al., 2004) between every channel pair. More...
#include "../connectivity_global.h"#include "abstractmetric.h"#include "../connectivitysettings.h"#include <QSharedPointer>#include <Eigen/Core>

Go to the source code of this file.
Classes | |
| class | CONNECTIVITYLIB::ImagCoherence |
| Imaginary-coherence estimator (Nolte et al. 2004); rejects zero-lag volume-conduction mixing. More... | |
Namespaces | |
| namespace | CONNECTIVITYLIB |
| Functional connectivity metrics (coherence, PLV, cross-correlation, etc.). | |
Volume-conduction-robust imaginary coherence (Nolte et al., 2004) between every channel pair.
SPDX-License-Identifier: BSD-3-Clause Copyright (c) 2018-2026 MNE-CPP Authors
Imaginary coherence is the imaginary part of complex coherency,
ImCoh_{xy}(f) = Im( S_{xy}(f) / sqrt(S_{xx}(f) * S_{yy}(f)) )
with output range [-1, 1] - or [0, 1] when reported as |Im(Coh)|, as done here. Its key property, established by Nolte, Bai, Wheaton, Mari, Vorbach and Hallett (NeuroImage 2004), is that any instantaneous mixing of independent sources (volume conduction, common-reference, common-pickup) projects purely onto the real axis and therefore contributes 0 to Im(Coh). Only interactions with a non-zero phase lag - the signature of finite conduction delay between two true neural generators - survive, which makes this estimator a standard choice for sensor-space MEG/EEG connectivity where volume conduction would otherwise dominate ordinary coherence.
The trade-off is reduced sensitivity to genuine but short-lag interactions and a non-trivial dependence on the relative source orientations, so ImCoh is usually reported alongside the phase-based estimators (PLI, wPLI) which share its anti-leakage property but have different bias/variance profiles.
Definition in file imagcoherence.h.