Variability and bias between magnetoencephalography systems in non-invasive localization of the primary somatosensory cortex

Abstract

Objectives: Magnetoencephalography (MEG) provides functional neuroimaging data for pre-surgical planning in patients with epilepsy or brain tumour. For mapping the primary somatosensory cortex (S1), MEG data are acquired while a patient undergoes median nerve stimulation (MNS) to localize components of the somatosensory evoked field (SEF). In clinical settings, only one MEG imaging session is usually possible due to limited resources. As such, it is important to have an a priori estimate of the expected variability in localization. Variability in S1 localization between mapping sessions using the same MEG system has been previously measured as 8 mm. There are different types of MEG systems available with varied hardware and software, and it is not known how using a different MEG system will impact on S1 localization.

Patients and methods: In our study, healthy participants underwent the MNS procedure with two different MEG systems (Vector View and CTF). We compared the location, amplitude and latency of SEF components between data from each system to quantify variability and bias between MEG systems.

Results: We found 8-11 mm variability in S1 localization between the two MEG systems, and no evidence for a systematic bias in location, amplitude or latency between the two systems.

Conclusion: These findings suggest that S1 localization is not biased by the type of MEG system used, and that differences between the two systems are not a major contributor to variability in localization.

Keywords: Bias; Magnetoencephalography; Pre-surgical mapping; Primary Somatosensory cortex; Variability.