Abstracts

Rationale: Excessive head motion, particularly in pediatric epilepsy patients, is known to reduce the localization accuracy of MEG activity sources, by either biasing source localization or decreasing signal-to-noise ratio, with extreme cases requiring discarding of data which may otherwise contain critical epileptiform events. Advanced MEG signal processing includes techniques which can compensate for the adverse effects of head motion on source localization accuracy, thus improving testing yield. Here, we tested the performance of the Movement Compensation (MC) feature implemented in the MEGIN MaxFilter^TM analysis software in a sample of pediatric epilepsy patients referred for MEG interictal mapping, exhibiting natural movement in varying degrees.

Methods: Resting MEG recordings (30-45 minutes in duration) from 5 pediatric patients (8 months-15 years of age) referred for a Phase 1 epilepsy surgery evaluation through the Le Bonheur Comprehensive Epilepsy Program, Le Bonheur Children’s Hospital (Memphis, TN, USA), were reviewed to assess the degree of spontaneous (natural) motion during the course of the testing session and examine the influence of the MC algorithm on source localization integrity. For each patient, two sets of dipole clusters were derived for the same interictal epileptiform event, before (d1) and after (d2) the application of MC, and compared to a reference (ref) cluster derived from dipoles where the head position was < 4 mm from the initial starting position of the recording. Each cluster was characterized by the center of mass, i.e. the mean dipole localization, and radius, which is the mean deviation from the dipole localizations from the center.

Results: In 4/5 patients, the application of MC during instances of natural motion over the course the MEG recording session resulted in effective suppression of motion artifacts which improved the localization accuracy. The localization improvement ranged from 2 to 12.7 mm, and the spatial variance of d2, compared to d1, relative to the ref set, was slightly reduced. The results are summarized in Table 1 and a representative example presented in Figure 1.

Conclusions: The present findings demonstrate that the application of MC to pediatric MEG recordings containing variable degrees of spontaneous, temporary head movements can provide localization of the irritative zone equivalent to that obtained when patients exhibit little-to-no motion. In addition to enhancing proprietary clinical MEG data interpretation, the MC functionality investigated in this study can particularly improve the utility of MEG in infants and children, especially prone to exhibiting movement, thus avoiding the need to prolong or repeat testing, or the need for sedation.

Funding: Please list any funding that was received in support of this abstract.: Le Bonheur Neuroscience Institute.