henhaobio
Sign in

bioRxivNeurosciencePreprint

Measurement Equivalence of On-Scalp OPM-MEG and Cryogenic MEG for Auditory and Somatosensory Cortical Mapping Across Development

William C. GaetziD, George C. O’NeilliD, Charlotte Birnbaum, Teresa CheungiD, Timothy P. L. Robets, K. Jeramy Hughes, Svenja KnappeiD, Orang AlemiD

PDFDOI

In-depth summary

Claude Opus 5.5 · xhigh effort · 7 Nov 2026 · from the full text of version 2 (20 Jul 2026) · CC BY 4.0

In 18 participants aged 10-45, OPM-MEG and SQUID-MEG placed auditory and somatosensory dipoles in the same cortical areas with matched latencies, differing by small systematic offsets and amplitude scaling.

Cryogenic magnetoencephalography (MEG) built on superconducting quantum interference devices (SQUIDs) needs a fixed helmet and liquid helium, which restricts head movement and use in children, whereas optically pumped magnetometers (OPMs) sit on the scalp at room temperature. Earlier OPM versus SQUID studies reported broadly similar evoked fields, but within-subject comparisons across development were scarce, leaving open whether sensor geometry shifts source location or amplitude estimates.

Key results

  1. ShownAuditory M100 dipoles from both systems fell in the superior temporal plane with no significant overall mean displacement. CTF dipoles, however, were consistently placed more medially than OPM (HEDscan) dipoles.

    Paired sign-flip permutation tests (5,000 permutations) found no significant mean signed offset. Absolute separation averaged 11.97 mm (median 11.37 mm; range 3.61 to 22.88 mm). After correcting for hemisphere sign, CTF dipoles were medialized by a mean of 4.22 mm (p = 0.00020); left 4.79 mm (p = 0.048), right 4.76 mm (p = 0.058). · Auditory Evoked Field (AEF) Localization; Overall localization agreement for AEF dipoles; Cross-system AEF source separation; Directional bias (axis-specific behavior)

  2. ShownSomatosensory P50m dipoles localized to the central sulcus on both systems. There was a small but significant offset, mostly CTF sources lying posterior to OPM sources, and it was significant in the right hemisphere only.

    Mean displacement magnitude was 4.10 mm (Δx = −0.75, Δy = −3.97, Δz = −0.71 mm; p = 0.028): right 5.73 mm (p = 0.013), left 3.77 mm (p = 0.31). Posterior shift Δy = −3.97 mm (p = 0.0026). No consistent mediolateral bias. Absolute separation averaged 12.00 mm (median 10.20 mm; range 2.46 to 30.54 mm). · Somatosensory-Evoked Field (SEF) Localization; Overall localization agreement for SEF dipoles; Cross-system SEF source separation; Directional bias analysis

  3. ShownPeak latencies matched across systems for both stimuli, and group source time courses showed no significant between-system difference at any time point. Goodness-of-fit was higher for OPM only in the somatosensory fits.

    Latency difference: auditory +1.86 ms (p = 0.219), somatosensory < 0.001 ms (p = 0.98). Goodness-of-fit (CTF minus HEDscan): auditory −0.025 (p = 0.244), somatosensory −0.092 (p = 0.0002). In one participant, whitened global field power correlated 0.979 (auditory) and 0.993 (somatosensory) between systems. · AEF ‘N100m’ Dipole Latency, Fit Quality (GOF)…; SEF ‘P50m’ Dipole Latency, Fit Quality (GOF)…; Figure legends

  4. SuggestedCTF gave larger dipole moments for both stimuli. For auditory sources the excess grew with the medial displacement of the CTF dipole, which the authors attribute to lead-field geometry rather than physiology. Somatosensory moment differences were unrelated to the offset.

    Mean log|moment| difference (CTF minus HEDscan): auditory +0.380 (p = 0.00080), somatosensory +0.208 (p = 0.008). Auditory moment difference vs mediolateral displacement: Spearman ρ = 0.63 (p = 0.0006). Somatosensory moment did not correlate with displacement magnitude. · Directional bias (axis-specific behavior); AEF ‘N100m’ Dipole Latency, Fit Quality (GOF)…; SEF ‘P50m’ Dipole Latency, Fit Quality (GOF)…; Discussion

  5. SuggestedAuditory dipole moment increased with age on both systems, with no between-system difference in slope. Somatosensory latency, fit and moment showed no age association on either system.

    Age vs auditory moment: CTF ρ = 0.38 (p = 0.026), HEDscan ρ = 0.47 (p = 0.0032); within-subject slope difference p = 0.66. Somatosensory measures vs age: all p > 0.12, with a HEDscan goodness-of-fit trend of ρ = 0.32 (p = 0.054). · Age-related effects; Age-related effects; Developmental Implications

Methods, data and reproducibility

Each participant was recorded on the same day on a 128-sensor FieldLine HEDscan OPM array and a 275-channel CTF SQUID system, with binaural tones (M100) and pneumatic index-finger stimulation (P50m) and shared fiducial coils. After each system's own denoising, identical preprocessing and FieldTrip dipole fits were compared with paired sign-flip permutation tests.

  • Auditory evoked fields: 1000-Hz pure tones, 300 trials, variable inter-stimulus interval, recorded on both CTF and HEDscan · human · n = 18 participants (14 female; 10.9 to 45.5 years)
  • Somatosensory evoked fields: pneumatic pulses to left and right index fingers, 500 trials per hand, recorded on both systems · human · n = 18 participants
  • Individual structural MRIs with MRI-visible fiducial markers, used for coregistration and single-shell head models · human
Methods & tools
FieldLine HEDscan OPM-MEG (128 single-axis magnetometers), HEDscan Smart Helmet, CTF 275-channel SQUID gradiometer MEG, FieldTrip, Adaptive Multipole Modelling (AMM), third-order synthetic gradiometry, equivalent current dipole fitting, Nolte single-shell head model, head-position indicator (HPI) coils, paired sign-flip permutation tests
Reproducibility gaps
No code or data availability statement in the text; FieldTrip and AMM software versions are not reported.
Limitations (6)
  • The OPM sensors were held in a rigid helmet fixed to a mount rather than a fully wearable cap, so the results likely represent best-case, motion-free OPM performance. · stated by the authors
  • The sample may be too small to detect sensitivity differences in very young children, for whom adult-sized cryogenic helmets increase the distance between sensors and cortex. · stated by the authors
  • Only equivalent current dipole models of early sensory responses were tested; beamformer and distributed source reconstructions and other paradigms were not compared. · stated by the authors
  • Absolute dipole amplitudes are not directly comparable, because magnetometers and gradiometers have different transfer functions. · stated by the authors
  • Equivalence is inferred largely from non-significant paired permutation tests in 18 participants rather than from formal equivalence tests with prespecified margins. · noted in this summary
  • The recording order was fixed (CTF first, then OPM, on the same day), so system is confounded with session order. Several authors are co-founders, employees or equity holders of FieldLine Medical, the OPM vendor (S27). · noted in this summary

Read this if you are weighing OPM-MEG against cryogenic MEG for pediatric or clinical sensory mapping and want within-subject, dipole-level comparisons of location, latency, amplitude and age effects.

Citations
1

Abstract

Wearable optically pumped magnetometer magnetoencephalography (OPM-MEG) reduces sensor-to-cortex distance compared with conventional cryogenic SQUID-MEG, but whether the two technologies yield equivalent neurophysiological conclusions remains unclear. We recorded auditory and somatosensory evoked fields in 18 participants (10-45 years) using a 128-sensor FieldLine HEDscan OPM-MEG system and a 275-channel CTF SQUID-MEG system within the same individuals. Equivalent current dipole source models were estimated using identical preprocessing and modeling procedures and compared using paired permutation testing. Both systems localized canonical auditory and somatosensory cortical generators with matched peak latencies and modest cross-system spatial differences. Auditory sources showed a consistent medial bias in SQUID-MEG localization, whereas somatosensory sources exhibited a small systematic offset ([~]4 mm), indicating stable coordinate differences rather than localization error. Dipole moments were larger for SQUID-MEG and goodness-of-fit higher for OPM-MEG; however, the increased moment was explained by a medial localization bias, demonstrating inverse-model effects rather than physiological disagreement. Auditory dipole moment increased with age in both systems, whereas somatosensory responses showed no age-related change. Together, these observations indicate preserved developmental physiology across platforms. These findings demonstrate that OPM-MEG and SQUID-MEG recover the same cortical generators and support equivalent biological interpretations despite predictable geometry-dependent coordinate differences. OPM-MEG therefore represents a measurement-equivalent implementation of MEG suitable for sensory functional mapping.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISIivanainen '16An '22Gutteling '25Jas '26Schofield '24Corvilain '25
82% semantic matchNeuroscience
Joonas Iivanainen, Matti Stenroos, Lauri T. Parkkonen·NeuroImage·Dec 19, 2016· 295 citations