Automated localization of magnetoencephalographic interictal spikes by adaptive spatial filtering
H E Kirsch1, S E Robinson, M Mantle
1UCSF Epilepsy Center, Department of Neurology, University of California, San Francisco, USA. heidi.kirsch@ucsf.edu
Summary
The software SAM(g(2)) shows promise for localizing epilepsy spikes in magnetoencephalography (MEG) data, especially with single foci and high signal-to-noise ratio (SNR). Further research is needed for cases with multiple foci or poor SNR.
Area of Science:
- Neuroscience
- Medical Imaging
- Epilepsy Research
Background:
- Automated adaptive spatial filtering enhances magnetoencephalography (MEG) data.
- This improves signal-to-noise ratio (SNR) for identifying interictal spikes in epilepsy.
Purpose of the Study:
- To evaluate the sensitivity and specificity of the SAM(g(2)) software tool for analyzing interictal MEG data in epilepsy patients.
- To compare SAM(g(2)) performance against established equivalent current dipole (ECD) fit procedures.
Main Methods:
- Analyzed interictal MEG data from 36 patients with intractable epilepsy using the SAM(g(2)) adaptive spatial filtering algorithm.
- Compared SAM(g(2)) results with those obtained from standard equivalent current dipole (ECD) fit methods.
Main Results:
- Good agreement between SAM(g(2)) and ECD was observed for interictal spikes with high SNR and a single focus.
- Imperfect concordance and overlap were noted between SAM(g(2)) and ECD when multiple epileptic foci were present.
Conclusions:
- SAM(g(2)) may be a viable alternative to manual ECD fitting for localizing single-focus interictal spikes with adequate SNR.
- Additional studies are necessary to validate SAM(g(2)) for complex cases involving multiple foci or low SNR.
