paperKB
coga / coga-kb
Help
Sign in

Chunk #355 — 4 Performance analysis — 4.2 Monte Carlo performance analysis of non-parametric inverse solutions — 4.2.4 Discussion of results

Source
Review on solving the inverse problem in EEG source analysis.
Embedded
yes

Text

Regularized sLORETA is shown to have the lowest errors both in terms of localization error (ED1) and ghost sources (ED2) (see Figure 6b). This is consistent with [32] where sLORETA was proven to have zero localization error for a single source in all noise-free simulations. Furthermore, it can be observed that the error measures and ghost maxima of sLORETA solutions can be reduced considerably when regularization is used – this improvement is particularly marked at large noise levels for ghost sources – see Figure 6 where the results for ED2 for the 5 dB scenario are displayed.