Magnetoencephalography

Magnetoencephalography (MEG) is a non-invasive functional neuroimaging and electrophysiological technique that directly measures ultra-weak extracranial magnetic fields (10 to 1,000 femtotesla, 10⁻¹⁴ to 10⁻¹² T) generated by synchronous intracellular postsynaptic currents in cortical pyramidal ensembles. First demonstrated in 1968 by physicist David Cohen using a two-million-turn copper induction coil inside a magnetically shielded room to detect averaged alpha-rhythm fields, the modality achieved single-trial physiological resolution in 1972 when Cohen paired a shielded room at the Massachusetts Institute of Technology with James E. Zimmerman's point-contact superconducting quantum interference device (SQUID) (Cohen, 1968, 1972). Because biological tissues share the magnetic permeability of free space, neuromagnetic fields traverse cerebrospinal fluid, skull, and scalp without spatial smearing, uniting sub-millisecond timing with millimeter-scale cortical localization across cryogenic SQUID helmets and wearable cryogen-free optically pumped magnetometers (OPMs).

Intracellular Primary Currents and the Biot–Savart Law

During neocortical Synaptic Transmission, dendritic ion influx establishes an axial voltage gradient along the apical shaft of each cortical pyramidal Neuron. In neuromagnetic forward theory (Hämäläinen et al., 1993), total macroscopic current density J(r, t) separates into two terms:

J(r, t) = J^p(r, t) + σ(r) E(r, t)

Here J^p(r, t) is the impressed intracellular primary current flowing along aligned apical dendrites, while σ(r) E(r, t) = −σ(r) ∇Φ(r, t) is the passive ohmic volume return current in the extracellular space. Under quasi-static electrodynamics (<1 kHz), magnetic flux density B(r, t) at external sensor coordinate r obeys the Biot–Savart law:

B(r, t) = (μ₀ / 4π) ∫ [J(r′, t) × (r − r′) / |r − r′|³] d³r′

where μ₀ = 4π × 10⁻⁷ T·m/A. A single postsynaptic potential along a ~0.5 mm dendrite carries a dipole moment q ≈ 0.02 pA·m. Producing a measurable net cortical dipole Q ≈ 10 nA·m (~100 fT at 4 cm standoff) requires 5 × 10⁴ to 10⁵ parallel dendrites active synchronously across 1 to 3 cm² of cortex. Even large epileptiform spikes (~1 pT) remain eight to nine orders of magnitude weaker than Earth's ~50 μT geomagnetic field.

Sulcal Tangential Selectivity and Cranial Magnetic Transparency

Two biophysical properties distinguish MEG from scalp Electroencephalography:

1. Tangential Sulcal Selectivity: In a spherically symmetric conductor (Sarvas, 1987), a radial current dipole on a gyral crown produces zero external magnetic field because symmetric volume return currents cancel its primary magnetic field outside the head. Conversely, tangential dipoles along vertical sulcal walls—housing two-thirds of human neocortex, including Heschl's auditory, Rolandic area 3b somatosensory, and calcarine visual cortices—project crisp dipolar fields outside the scalp without an electrical reference electrode.
2. Magnetic Transparency of the Skull: Whereas cranial bone conductivity varies 3- to 5-fold across thickness and sutures, relative magnetic permeability μ_r across brain, CSF, skull, and scalp differs from unity by less than 10⁻⁵ (μ ≈ μ₀). Neuromagnetic fields therefore cross the skull with zero refraction or blurring, allowing single-shell T1 MRI forward models to localize focal dipoles within 2 to 5 mm (Baillet, 2017).

Josephson Junctions, Flux Transformers, and Gradiometer Geometries

Conventional whole-head systems immerse 275 to 306 niobium (Nb) DC SQUIDs inside a liquid helium-4 Dewar at 4.2 K (−268.95 °C). Harnessing macroscopic Superconductivity, a DC SQUID forms a superconducting ring interrupted by two Josephson tunnel junctions; external flux coupled from a niobium pickup coil modulates junction voltage periodically with the magnetic flux quantum Φ₀ = h / (2e) ≈ 2.068 × 10⁻¹⁵ Wb, reaching white-noise floors of 2 to 5 fT/√Hz.

Three superconducting pickup-coil geometries shape spatial sensitivity:

• Pure Magnetometers: Single loops maximizing deep subcortical sensitivity while remaining vulnerable to ambient interference.
• First-Order Axial Gradiometers: Coaxially opposed loops (4–5 cm baseline) canceling uniform far-field noise while retaining 1/r³ cortical signals.
• Orthogonal Planar Gradiometers: Side-by-side figure-eight coils sensing tangential derivatives (∂B_r/∂x, ∂B_r/∂y), peaking directly above the active sulcus.

Passive Mu-Metal Shielded Rooms and Active Interference Suppression

Systems operate inside magnetically shielded rooms (MSRs) built from concentric high-permeability nickel–iron mu-metal shells (μ_r > 60,000, shunting low-frequency flux) and eddy-current aluminum plates yielding 60 to 85 dB attenuation. Residual urban transit, magnetocardiographic, and cranial Electromyography noise is rejected via synthetic reference gradiometers and temporal Signal Space Separation (tSSS) spherical-harmonic decomposition, permitting recordings even in patients carrying dental hardware or implanted stimulators.

Spin-Exchange Relaxation-Free (SERF) Alkali Vapor Physics

Liquid-helium Dewars enforce fixed adult helmet sizing, strict head stillness, and a 20 to 30 mm thermal vacuum standoff between scalp and pickup coils. Optically pumped magnetometers (OPMs) eliminate cryogenics using microfabricated glass cells containing heated rubidium-87 (⁸⁷Rb, ~150 °C) or metastable helium-4 (⁴He) vapor. Circularly polarized 795 nm laser light spin-polarizes alkali valence electrons; when ambient MSR fields are actively nulled below 1 to 2 nT by bi-planar wall coils at high vapor density, inter-atomic collisions preserve spin phase in the spin-exchange relaxation-free (SERF) regime, detecting femtotesla Larmor spin precession via photodiode transmission at 7 to 15 fT/√Hz (Boto et al., 2018).

On-Scalp Standoff Reduction and Naturalistic Movement Paradigms

Slotting compact triaxial OPMs into subject-specific or flexible helmets cuts sensor standoff to 4 to 6 mm from the scalp. Inverse-square proximity boosts cortical signal amplitude 3- to 5-fold—compensating for OPM intrinsic noise—while doubling spatial-frequency resolution across infant and adult heads during natural head movement.

Presurgical Epilepsy Mapping, Neural Oscillations, and Decoding

Magnetic source imaging (MSI) holds clinical approval for presurgical localization of interictal epileptiform spikes—pinpointing fissural or insular focal cortical dysplasias invisible on scalp EEG to guide Electrocorticography or stereo-EEG trajectories—and non-invasive mapping of eloquent language and sensorimotor cortex. In cognitive neuroscience and non-invasive Brain–Computer Interface research, spectral estimation via the Fast Fourier Transform and source-space beamforming resolve millisecond beta/gamma oscillatory routing for Neural Decoding of imagined kinematics and covert speech.

Why are MEG and scalp EEG considered biophysically complementary rather than redundant?

MEG detects tangential sulcal dipoles without skull conductivity blurring, whereas scalp EEG records both radial gyral-crown dipoles (magnetically invisible to MEG) and tangential sources. Simultaneous 306-channel MEG and high-density EEG constrains all three dipole axes for superior cortical source reconstruction.

How can SERF optically pumped magnetometers sit 5 mm from the scalp when their rubidium cell is heated to 150 °C?

Micro-vacuum insulation and aerogel standoffs isolate the millimeter-scale glass cell inside a polymer housing so the outer casing contacting the cap stays below 37–40 °C during prolonged wear.

Why do MEG systems require magnetically shielded rooms when EEG systems operate at the hospital bedside?

Power lines, building vibrations in Earth's 50 μT field, and elevators produce 10⁵–10⁸ fT fluctuations—up to a million times stronger than a 50 fT cortical response—which would saturate SQUID flux-locked loops or push SERF OPMs outside their <2 nT zero-field operating window without mu-metal and active coil shielding.

1. Cohen, D. (1968). Magnetoencephalography: evidence of magnetic fields produced by alpha-rhythm currents. Science, 161(3843), 784–786. https://doi.org/10.1126/science.161.3843.784
2. Cohen, D. (1972). Magnetoencephalography: detection of the brain's electrical activity with a superconducting magnetometer. Science, 175(4022), 664–666. https://doi.org/10.1126/science.175.4022.664
3. Hämäläinen, M., Hari, R., Ilmoniemi, R. J., Knuutila, J., & Lounasmaa, O. V. (1993). Magnetoencephalography—theory, instrumentation, and applications to noninvasive studies of the working human brain. Reviews of Modern Physics, 65(2), 413–497. https://doi.org/10.1103/RevModPhys.65.413
4. Baillet, S. (2017). Magnetoencephalography for brain electrophysiology and imaging. Nature Neuroscience, 20(3), 327–339. https://doi.org/10.1038/nn.4504
5. Boto, E., Holmes, N., Leggett, J., Roberts, G., Shah, V., Meyer, S. S., ... & Brookes, M. J. (2018). Moving magnetoencephalography towards real-world applications with a wearable system. Nature, 555(7698), 657–661. https://doi.org/10.1038/nature26147