Magnetospinography visualizes electrophysiological activity in the cervical spinal cord

Abstract

Diagnosis of nervous system disease is greatly aided by functional assessments and imaging techniques that localize neural activity abnormalities. Electrophysiological methods are helpful but often insufficient to locate neural lesions precisely. One proposed noninvasive alternative is magnetoneurography (MNG); we have developed MNG of the spinal cord (magnetospinography, MSG). Using a 120-channel superconducting quantum interference device biomagnetometer system in a magnetically shielded room, cervical spinal cord evoked magnetic fields (SCEFs) were recorded after stimulation of the lower thoracic cord in healthy subjects and a patient with cervical spondylotic myelopathy and after median nerve stimulation in healthy subjects. Electrophysiological activities in the spinal cord were reconstructed from SCEFs and visualized by a spatial filter, a recursive null-steering beamformer. Here, we show for the first time that MSG with high spatial and temporal resolution can be used to map electrophysiological activities in the cervical spinal cord and spinal nerve.

Figure 1

Spinal cord evoked magnetic fields (SCEFs) of a healthy subject measured after stimulation of the lower thoracic spinal cord. (a) Positions of the sensors superimposed on an X-ray image of a subject. (b) The three-directional magnetic fields recorded by each sensor. Black traces are magnetic fields in the ventral–dorsal direction relative to the cervical spinal cord (dorsal is upwards in the graphs). Red traces are magnetic fields in the left–right direction (right is upwards). Blue traces are magnetic fields parallel to the spinal cord (cranial is upwards). Some malfunctioning pickup coils show a flat line. The red signals, mainly generated from intra-axonal currents, are highest above the spinal cord. The polarity of the black and blue signals is reversed on each side of the spinal cord.

Figure 2

Reconstructed currents of a healthy subject measured after stimulation of the lower thoracic spinal cord. (a) Reconstructed current map. Currents at the level of the spinal canal were reconstructed by a recursive null-steering (RENS) beamformer and superimposed on an X-ray image. Current intensity is shown by a colour scale (red is higher). Small white arrows indicate current vectors. The leading component of the currents (large upwards-pointing red arrow) appeared 4.3 ms after stimulation and propagated cranially along the spinal canal. At 5.8 ms, the trailing component (downwards-pointing red arrow) appeared from the caudal side. Perpendicular currents (blue arrows) flowing towards the spine between the leading and the trailing components are also observed on both sides. (b) Waveforms of reconstructed currents at the “virtual electrodes”. Red traces are currents at the midline of the cervical spinal canal (the caudal to cranial direction is upwards). Blue traces are currents 20 mm lateral from the midline (upwards is towards the spinal canal).

Figure 3

Preoperative MR image of a patient with cervical spondylotic myelopathy showing spinal stenosis. The patient was a 67-year-old male with C4/5 disc herniation and a small indentation of the C5/6 disc. The patient exhibited clumsiness of bilateral hands and gait disturbance.

Figure 4

Reconstructed currents of the cervical spondylotic myelopathy patient. Currents are illustrated in a similar manner to Fig. 2. The current map is in the anterior–posterior direction, and the X-ray image of the lateral view shows the corresponding levels of the cervical spine. (a) In the reconstructed current map, the leading component is not illustrated because its intensity is much lower than that of other currents. When the perpendicular components stopped and disappeared caudal to C4/5, the trailing component simultaneously disappeared caudal to C5/6. (b) The left graphs are the ascending spinal cord evoked potentials (SCEPs) by stimulation of the lower thoracic cord showing conduction block at the C4/5 disc level. The right graphs are the reconstructed currents at the midline of the cervical spinal canal (red) and 2 cm lateral (blue). The leading component (the first waveform in red) attenuated and disappeared through C4–6, and the trailing component (the second waveform in red) disappeared at C5/6. The perpendicular inflow components greatly attenuated at C4/5 (the second waveform in blue).

Figure 5

Spinal cord evoked magnetic fields (SCEFs) in response to stimulation of the median nerve at the elbow. (a) Positions of the sensors. (b) SCEFs illustrated in a manner similar to Fig. 1. SCEFs were mainly recorded from the right side of the spinal cord.

Figure 6

Reconstructed currents after stimulation of the right median nerve at the elbow. (a) Reconstructed current map. The currents flowed into the spinal canal from the right side and propagated cranially 5.6–6.4 ms post-stimulus. Rotated currents were observed 6.6–7.6 ms post-stimulus. Trailing currents propagated cranially 7.8–8.4 ms post-stimulus. (b) Reconstructed currents at the midline of the cervical spinal canal (C3 to C7) and at the adjacent intervertebral foramina (C3/4 to Th1/2). The first peak of the reconstructed currents at the midline of the cervical spinal canal did not propagate, but subsequent currents did (left graphs). Currents flowing into the intervertebral foramina were larger at C5/6 to C7/Th1 (right graphs).

Figure 7

Individual variations in the peak currents flowing into the intervertebral foramina. In most subjects, the largest currents flowed through C6/7 or C7/Th1. Currents passing through the C4/5 intervertebral foramen (outlet of root C5) were observed in two of the ten subjects, a 27-year-old man (upper one) and a 44-year-old man.

Figure 8

The 120-channel superconducting quantum interference device (SQUID) biomagnetometer system used for the magnetospinography. (a) Individual SQUID sensors are arranged in an eight-by-five configuration at 20-mm intervals. Each sensor has three perpendicular pickup coils to detect magnetic fields in three orthogonal directions. (b) Overhead view of the sensor and subject as shown in Figs 1 and 5. (c) Cross-sectional image of the cryostat protrusion. The central grey area indicates the sensor array shown in (a). The sensors are arranged in the vertical direction and positioned in close contact with the upper inwall of the protrusion. The subject was in the supine position on a table with the posterior neck on the protrusion holding the SQUID sensor array inside. The upper surface of the protrusion is curved to fit the lordosis of the cervical spine. (d) Both median nerves were alternatively stimulated at the anterior of the elbow joint. Volar splints to suppress evoked movements are not shown here.