Spherical harmonic based noise rejection and neuronal sampling with multi-axis OPMs

Abstract

In this study we explore the interference rejection and spatial sampling properties of multi-axis Optically Pumped Magnetometer (OPM) data. We use both vector spherical harmonics and eigenspectra to quantify how well an array can separate neuronal signal from environmental interference while adequately sampling the entire cortex. We found that triaxial OPMs have superb noise rejection properties allowing for very high orders of interference (L=6) to be accounted for while minimally affecting the neural space (2dB attenuation for a 60-sensor triaxial system). To adequately model the signals arising from the cortex, we show that at least 11th order (143 spatial degrees of freedom) irregular solid harmonics or 95 eigenvectors of the lead field are needed to model the neural space for OPM data (regardless of number of axes measured). This can be adequately sampled with 75-100 equidistant triaxial sensors (225-300 channels) or 200 equidistant radial channels. In other words, ordering the same number of channels in triaxial (rather than purely radial) configuration gives significant advantages not only in terms of external noise rejection but also minimizes cost, weight and cross-talk.

Test-Retest Reliability of the Human Connectome: An OPM-MEG study

“…In addition, our triaxial design measures the tangential and radial components of the magnetic field, whereas conventional MEG only measures radially. Whilst the use of triaxial sensors has proven to be an excellent means to reduce the effects of non-brain sources of magnetic field (Brookes et al, 2021; Rea et al, 2022; Tierney et al, 2022), the tangential field components are smaller in amplitude and consequently, in terms of absolute signal, OPM-MEG remains disadvantaged compared to cryogenic instrumentation. It is encouraging that, despite the lower channel count, we achieve approximate parity with conventional MEG in terms of repeatability of connectivity measurement.…”

“…They also allow increased total signal strength (i.e. each sensor makes three measurements of field)(Brookes et al, 2021; Rea et al, 2022), improved ability to differentiate brain activity from background fields (Brookes et al, 2021; Rea et al, 2022; Tierney et al, 2022), more uniform coverage in infants (Boto et al, 2022) and optimised calibration. In addition to the triaxial array, our system includes magnetic shielding which operates in active feedback configuration (Rea et al, 2021).…”

Concurrent spinal and brain imaging with optically pumped magnetometers

“…We expect that with a larger number of sensors, especially by using triaxial rather than dual-axis devices 51,52 , we will be able to better characterize both the signal and environmental noise space. Also, although OPM sensitivity does begin to decline at 130Hz, it does so only as a first order filter (i.e., a halving of amplitude for every doubling of frequency).…”

Miniature biplanar coils for alkali-metal-vapor magnetometry

“…In MEG, a few to a hundred zero-field OPM sensor modules are mounted in an array around a subject's head with the aim of mapping the magnetic field normal to the scalp surface. Dipole source location based on inverse modeling of the surface field map requires each OPM's output to be proportional to the normal field component 27,28 , although a small time-independent component at each sensor is sometimes allowed, overall motivating the magnetic field over the entire array to be nulled as far as possible. Presently, magnetically shielded rooms 13,29 or smaller, person-sized magnetic shields 30,31 are used to passively attenuate background magnetic fields down to tens of nT and then shim coils inside the shield may further cancel residual fields at the head position into a range where arrayed OPMs are operable.…”