Improving Localization Accuracy of Neural Sources by Pre-processing: Demonstration With Infant MEG Data

Abstract

We discuss specific challenges and solutions in infant MEG, which is one of the most technically challenging areas of MEG studies. Our results can be generalized to a variety of challenging scenarios for MEG data acquisition, including clinical settings. We cover a wide range of steps in pre-processing, including movement compensation, suppression of magnetic interference from sources inside and outside the magnetically shielded room, suppression of specific physiological artifact components such as cardiac artifacts. In the assessment of the outcome of the pre-processing algorithms, we focus on comparing signal representation before and after pre-processing and discuss the importance of the different components of the main processing steps. We discuss the importance of taking the noise covariance structure into account in inverse modeling and present the proper treatment of the noise covariance matrix to accurately reflect the processing that was applied to the data. Using example cases, we investigate the level of source localization error before and after processing. One of our main findings is that statistical metrics of source reconstruction may erroneously indicate that the results are reliable even in cases where the data are severely distorted by head movements. As a consequence, we stress the importance of proper signal processing in infant MEG.

Keywords: artifact; brain; infant; magnetoencephalography (MEG); movement compensation; signal processing; signal space projection; signal space separation.

Figure 1

Field topographies of bilateral evoked auditory responses (at 150 ms) in a 6-month-old participant. (A) Raw data (no artifact suppression), (B) data processed with tSSS, and (C) data processed with tSSS + movement compensation. Black arrows represent estimation of equivalent current flow at the MEG sensor locations.

Figure 2

In the top row, the mean plus or minus standard error (across subjects) of the position deviation (from each subject's mean head position) shows smaller movements in the more recent recording (published in 2021) compared to an older recording (published in 2014). In the second row, localization bias is reduced by movement compensation, with greater reduction in bias when the covariance is computed from movement-compensated data. There are systematic outward and inward (relative to the head center) localization biases shown for the newer (2021) and older (2014) data, respectively. In the last row, mean goodness of fit values exceed 80% in all cases.

Figure 3

The average position in device coordinates relative to the initial position (x axis; +Z means upward) is strongly correlated with the inward bias of the ECD fit dipoles relative to the true source locations for raw and SSS-processed data (first and second columns), but not either movement-compensated case (p > 0.05, both).

Figure 4

The average QRS complex from exemplar infant (left) and adult (right) MEG recordings.

Figure 5

Topographic maps of cardiac fields for sample infant and adult MEG recordings. (Top) SSP projector fields computed from a 100 ms window surrounding the average QRS peak from this figure was sufficient to capture the artifact. (Bottom) latent sources and field maps for cardiac-related latent components computed with ICA. Both approaches show that fields from heart-generated volume currents in the brain may appear shallow or asymmetrical compared to adult cardiac fields.

Figure 6

Cardiac artifact on a subset of channels after different pre-processing approaches. Top: No processing; Second Panel: SSS only; Third Panel: SSS followed by SSP (2 orthogonal projectors); Bottom Panel: SSS followed by ICA (4 cardiac-related latent sources removed). In the third and bottom panels the cardiac artifacts have been successfully repaired. The spatial distribution of the cardiac artifact stays virtually intact in the SSS process, indicating that most of its signal energy comes from the internal SSS volume where the head is located.