Estimation of localization of neural activity in the spinal cord using a biomagnetometer

Size: px

Start display at page:

Download "Estimation of localization of neural activity in the spinal cord using a biomagnetometer"

Nathan Goodwin
5 years ago
Views:

1 J Med Dent Sci 2003; 50: Original Article Estimation of localization of neural activity in the spinal cord using a biomagnetometer Harunobu Ohkubo, Hiromichi Komori, M.D., Shigenori Kawabata, M.D., Yuko Fukuoka and Kenichi Shinomiya, M.D. The Section of Orthopaedic Spinal Surgery, Department of Frontier Surgical Therapeutics, Division of Advanced Therapeutical Sciences Graduate School. Tokyo Medical and Dental University Spinal cord evoked potentials (SCEPs) measurement is widely used for level diagnosis of spondylotic myelopathy. However, because of the restriction of spatial resolution, SCEPs do not distinguish the neurophysiological activities among tracts in the spinal cord without invasive methods. Magnetic field measurement has the theoretical advantage of high spatial resolution, compared with electric measurement. We recorded spinal cord evoked magnetic fields (SCEFs) in the thoracic spinal cord after stimulation to the motor area in felines, and estimated the source of the magnetic fields. SCEFs showed a quadrupolar pattern, and conducted in a cranial-to-caudal direction at 55 m/sec. According to this result, we estimated that the SCEFs after stimulation to the motor area were generated by the contralateral corticospinal tract. Furthermore, the estimated dipole of the SCEFs after stimulation to the motor area was located on the contralateral side in the spinal cord. These results correspond with the anatomical location of the corticospinal tract of felines, and suggest that magnetic field recording can detect the magnetic source localization of each tract in the spinal cord. Key words: Spinal cord, Spinal cord evoked potentials, Spinal cord evoked magnetic fields, Corticospinal tract Corresponding Author: Harunobu Ohkubo Tokyo Medical and Dental University 5-45, Yushima 1-chome, Bunkyo-ku, Tokyo , JAPAN Received January 31; Accepted March 20, Introduction Spinal cord evoked potentials (SCEPs) measurement is widely used for spinal cord monitoring during operations and level diagnosis of spondylotic myelopathy 1,2,3. The spinal cord contains several neural pathways, and each has different function such as motor and sensory function. SCEPs contain the electrical activities of several tracts in the spinal cord, and show the polyphasic structure. However, the electrical origin of each signal of the SCEPs is not precisely known. In addition, electrical measurement can not distinguish the neurophysiological activities among tracts without invasive methods. On the other hand, electrical activity produces magnetic field. Magnetic field measurement is currently used for recording the neural activity in the brain. Several authors also reported the evoked magnetic fields generated from action currents in the peripheral nerves 4,5,6, the brachial plexus 6,7, the cervical nerve root 6,8 and the lumbosacral nerve root 9,10,11,12. They found that magnetic field generated from action current demonstrated the conductive quadrupolar structure. On the other hand, Kawabata et al. reported that ascending conductive magnetic fields generated by the stimulation of the thoracic spinal cord 13. The magnetic field demonstrated the conductive quadrupolar structure which has been reported as a pattern of the magnetic field generated from peripheral nerves. This report was the first one to demonstrate the spinal cord evoked magnetic fields (SCEFs) elicited by the action current in the spinal cord. They found that magnetic field measurement was less affected by the surrounding tissue of

2 178 H. OHKUBO et al. J Med Dent Sci the spinal cord than the electrical measurement. We can hypothesize that magnetic field measurement has the potential to distinguish the localization of neural activities in the spinal cord without invasive methods, such as partial transaction of the spinal cord or insertion of microelectrodes. To test this hypothesis, it is necessary to record the SCEFs generated by the electrical activity of which localization is already known. The initial negative spike wave of SCEPs after stimulation to the motor cortex of the brain is called the D (direct) wave 14. It is well recognized that the D wave is generated from the corticospinal tract. In the present study, we recorded magnetic fields after stimulation that generated the D wave, and estimated the source of the SCEFs. In addition, we examined whether magnetic field measurement had the potential to distinguish the localization of the neural activity in the spinal cord by comparison of the location of the estimated source with the anatomical location of the corticospinal tract. 2. Materials and Methods 2-1. Preparation Five adult mongrel felines of both sexes ranging from 2.2 to 3.6 kg were used to conduct this study. Initially, anesthesia was induced with ketamine chloride (30 mg/kg). After the right femoral vein was catheterized, anesthesia was maintained by continuous intravenous infusion of pentobarbital (30 mg/kg/hr) and vecuronium bromide (0.2 mg/kg/hr). A tracheotomy was performed and a tracheal tube was inserted to allow ventilation. To expose the dura in the thoracic area, a laminectomy was performed from Th1 to Th8. We performed a craniotomy over the motor cortex in the frontal area and exposed the sigmoid gyrus for electrical stimulation. After the study, the animals were euthanized by intravenous infusion of pentbarbital, 120 mg/kg. This study method was approved by the Ethical Committee of Tokyo Medical and Dental University Stimulation Stimulation was performed using bipolar epidural electrodes placed on the motor cortex, using MEB2200 (Nihon Koden). A monophasic square wave of 0.2 ms duration was applied at 30 Hz with a constant current of 5mA intensity Recording The measurements were performed in a magnetically shielded room. The felines were placed on an X- Y-Z stage, in parallel with the Y-axis of the stage. SCEFs and SCEPs were recorded under identical conditions. Both the SCEFs and the SCEPs were recorded under completely relaxed muscle conditions induced by vecuronium bromide. SCEPs were recorded at five points spaced 10 mm apart on the dural tube (Fig. 1A). We used monopolar recording and an indifferent electrode was placed in the subcutaneous tissue of the buttock. Five hundred trials were averaged by MEB2200 (Nihon Koden). The signals were processed with a 500 Hz to 3000 Hz band pass. The SCEFs were recorded with the biomagnetometer (Yokogawa Electric Corporation, Tokyo, Japan). This system consisted of an 8-channel axial first-order gradiometer model with a 50 mm baseline and sensor of 15mm diameter. Each sensor was arranged along the two orthogonal axes of a rectangular coordinate system (2 rows 4 lines and 21mm distance between neighboring sensor position). The noise level of the entire detection system is below 3fT/ Hz. The SCEFs were measured at 48 different points (4 rows and 12 lines). Measurement points were spaced 7mm apart along the X-axis and 10.5 mm apart along the Y-axis. The map- Fig. 1. Diagram of the experimental setup. A, spinal cord evoked potentials recorded at 5 points on the median of the dural tube in the thoracic level (closed circles). Recording points were spaced 10 mm apart, and monopolar recordings were performed. B, spinal cord evoked magnetic fields measured at 48 different points (12 rows X 4 lines) (closed circles) over the dural tube. Recording points were spaced 7 mm apart along the X-axis and spaced 10.5 mm apart along the Y-axis.

3 ESTIMATION OF LOCALIZATION OF NEURAL ACTIVITIES 179 ping was carried out on a field plane 5mm above the dura (Fig. 1B). Using the stimulus pulse as a trigger, epochs of 15ms duration (pre-trigger 5ms, post-trigger 10 ms) were recorded epochs were averaged. The signals were processed with a Hz band pass filter, and digitally stored at a Hz sampling rate Calibration system We designed a calibration system to determine the precise location of the sensor in the biomagnetometer. Six coils that emit magnetic fields were implanted in the measurement stand, and the locations of these coils were always exactly known. Before the experiment, a biomagnetometer detected the magnetic fields generated from the coils, and we then estimated the location of the sensors in the biomagnetometer from the intensity of the signals generated from each coil. We also set two marker coils which have 8 mm diameter and emit magnetic fields on the midline of the spinal cord. According to the estimated location of the marker coil, we determined the coordinate of the midline of the spinal cord. The position of the spinal cord was confirmed by an X-ray Estimation of the dipole of the magnetic fields To estimate of the localization of the neural activity in the spinal cord, we estimated the dipole of the initial magnetic fields of each of the SCEFs. Theoretical fields were calculated on the basis of a single moving equivalent current dipole model as described by Sarvas (1987) 15. To simplify the calculation, we used six parameters; the location (x, y and z coordinates) and moment (X,Y and Z components) for the dipole. Dipolar source locations were reconstructed from the field pattern using a halfspace volume conductor model. Each parameter was adjusted for the leastsquares fit. We then compared the midline of the spinal cord with the coordinates of the dipole of the SCEFs. 3. Results 3-1. SCEPs after stimulation to the motor area We were able to record the conductive descending action potential in all of the felines, (Fig. 2). The action potential consisted of an initial negative spike wave and following polyphasic waves. The mean conduction velocity of the initial wave was approximately 54m/sec. The waveform of SCEPs after stimulation to Fig. 2. Descending spinal cord evoked potentials (SCEPs) recorded on the midline of the dural tube. A, stimulation to the left motor area. B, stimulation to the right motor area. Measurement points were spaced 10 mm apart. Initial negative peak and delayed polyphasic waves were recorded. The initial signal propagated caudally. SCEPs after stimulation to the left and the right motor area resembled each other very closely. the left motor area closely resembled the right. These results indicate that the initial spike wave of each SCEPs corresponds to the D wave SCEFs after stimulation to the motor area Conductive descending SCEFs were detected in all felines (Fig. 3). SCEFs propagated in the caudal direction. The initial response of the SCEFs from the right side of the felines was outward current and that from left side was inward current. The second response of the SCEFs showed the reverse pattern of polarity. The intensity of the initial magnetic fields was approximately 50fT. The conduction velocity calculated from the peak latency of the initial wave was approximately 55 m/sec (Table 1). The time course of the isofield maps of the SCEFs demonstrated the quadrupolar structure. The quadrupolar configuration propagated to caudal direction (Fig. 4) Magnetic source localization of the SCEFs after stimulation to the motor area The location of the current dipole of the initial magnetic fields was estimated by the procedure described above. The mean estimated dipole of SCEFs after stimulation to the right motor area was located mm (mean SD) left to the midline of the spinal cord, and that after stimulation to the left motor area was located mm (mean SD) right to the midline of the spinal cord (Table 2).

180 H. OHKUBO et al. J Med Dent Sci Fig. 3. Distribution of the spinal cord evoked magnetic fields (SCEFs). The spinal cord was located under the median of the measured field.

4 180 H. OHKUBO et al. J Med Dent Sci Fig. 3. Distribution of the spinal cord evoked magnetic fields (SCEFs). The spinal cord was located under the median of the measured field. The SCEFs propagated from cranial to caudal direction. The initial response of the SCEFs from the right side of the feline was outward current and that from left side was inward current. The second response of the SCEFs showed the opposite pattern of polarity. The SCEFs propagated from cranial to caudal direction. A, stimulation to the left motor area. B, stimulation to the right motor area. The SCEFs after stimulation to the left motor area and that to the right motor area resembled each other. Table 1. Conduction velocity of the spinal cord evoked magnetic fields (SCEFs) and spinal cord evoked potentials (SCEPs) calculated from the initial signal of each of the waves. Fig. 4. Time course of the iso-field maps of spinal cord magnetic fields (SCEFs). Red indicates the outward current of the magnetic fields from ventral to dorsal direction and blue indicates the inward current of the magnetic fields. The SCEFs demonstrated a quadrupolar pattern and propagated from cranial to caudal direction. A, stimulation to the left motor area. B, stimulation to the right motor area. The SCEFs after stimulation to the left motor area and that to the right motor area resembled each other very closely Comparison of the estimated dipoles of the SCEFs with anatomical structure The estimated dipoles of the SCEFs after stimulation to the left and right motor area were traced on the X-ray

ESTIMATION OF LOCALIZATION OF NEURAL ACTIVITIES 181 Table 2. Distance from the estimated dipole to the median of the spinal cord. - indicates the left side of the spinal cord. Fig. 5.

5 ESTIMATION OF LOCALIZATION OF NEURAL ACTIVITIES 181 Table 2. Distance from the estimated dipole to the median of the spinal cord. - indicates the left side of the spinal cord. Fig. 5. Dipolar source reconstructions of the sequential position of the initial magnetic field of the spinal cord evoked magnetic fields (SCEFs) traced on the X-ray image. Thick gray lines were drawn on the pedicle of the thoracic spine to indicate the lateral boundary of the spinal cord. Solid line traced the sequential movement of the dipole of the SCEFs after stimulation to the right motor area. Broken line traced the sequential movement of the dipole of the SCEFs after stimulation to the left motor area. The dipoles of the SCEFs after stimulation to the left motor area descended through the right side of the spinal cord, and that to the right motor area descended through the left side of the spinal cord. image (Fig. 5). The dipoles of the SCEFs after stimulation to the right motor area descended through the left side of the spinal cord. The dipoles of the SCEFs after stimulation to the left motor area descended through the right side of the spinal cord. 4. Discussion Magnetic fields produced by the synaptic potential is represented by a single dipole, and the iso-field map of the magnetic field exhibits a bipolar structure. On the other hand, when axons are activated, intra-cellular current flows in proximal and distal directions from the depolarized region. Therefore, magnetic field generated by these two anti-parallel intra-cellular currents show a quadrupolar structure 4,6-13. Also the duration of the action potential is much shorter than that of the synaptic potential. According to our results, we can speculate that the SCEFs after stimulation to the motor area are generated by the action current in the spinal cord. SCEPs after stimulation to the motor area are generated by the electrical activity of the pyramidal neurons and propagate in the corticospinal tract. In the present study, the conduction velocity of SCEFs after stimulation to the motor area was equivalent to that of SCEPs. Therefore, we concluded that the SCEFs and the SCEPs were generated by the same electrical activity in the corticospinal tract. It is well known that the majority of the nerve fibers in the corticospinal tracts of the felines descend thorough the opposite side of the spinal cord at the thoracic level 16. To simplify the source model, we hypothesized that the neural activity elicited by the stimulation to the motor area was propagated on a single neural pathway, and we adopted a single dipole model representing the actual generators for each neural activity. However, it is reported that a part of fibers do not cross and descend through ipsilateral medial side of the spinal cord in feline 16. It is possible that the presence of the magnetic fields generated from the ipsilateral medial neural pathway affects the location of the estimated dipole. In the present study, the source of the SCEFs was located 1.4 mm lateral from the midline in the contralateral spinal cord. We measured the diameter of the spinal cord at the thoracic level, and found that corticospinal tract was located approximately 2 mm lateral from the midline of the spinal cord. Thus, the location of the estimated source of the SCEFs was equivalent to the anatomical location of the corticospinal tract, and presence of the ipsilateral medial fibers hardly effected the estimation of the source location. It is possible that the number of the ipsilateral medial fibers was not sufficient to influence the location of the estimate source. Thus, magnetic field measurement can distinguish the estimated dipole from each corticospinal tract in the

6 182 H. OHKUBO et al. J Med Dent Sci intact spinal cord. These results suggest that SCEF recording can differentiate the localization of the electrical activity with intact spinal cord. However, some problems remain to record the SCEFs for clinical use. Magnetic fields generated by the action current are relatively unaffected by the surrounding tissue, compared with the electrical measurement 4,13. However, the magnetic field theoretically reduced in inverse proportion to the square of the distance 17. Therefore, the distance between the sensor and the magnetic source is critical. In the present study, we performed laminectomy to reduce the distance. Without operation, the distance from the spinal cord to the sensor, which is located on the surface of skin, will increase up to 2 cm. In this configuration, it is possible that the signal intensity is reduced and insufficient to obtain meaningful SCEF waveform. To overcome this difficulty and to obtain clear SCEFs, averaging of a large number of trials should be necessary. In conclusion, the present study revealed that magnetic field measurement was able to distinguish the localization of the electrical activity in the spinal cord. Clinical application of magnetic field measurement to diagnose of spinal cord dysfunction in human being will be possible by improving the number and the configuration of sensor and the estimation method. References 1. Shinomiya K, Furuya K, Yamaura I, et al. Spinal cord monitoring of spinal cord function using evoked spinal cord potentials. In: Homma S, Tamaki T, Shimoji K, et al, editors, Fundamentals and clinical application of spinal cord monitoring. Tokyo: Saikon Publishing, 1984: Tamaki T, Noguchi T, Takano H, et al. Spinal cord monitoring as a clinical utilization of the spinal evoked potential. Clin Orthop 1984;184: Levy WJ, York DH, McCaffery M, et al. Motor evoked potentials from transcranial stimulation of the motor cortex in humans. Neurosurgery 1984;15: Hashimoto I, Mashiko T, Mizuta T, et al. Visualization of a moving quadrupole with magnetic measurements of peripheral nerve action fields. Electroenceph Clin Neurophysiol 1994;93: Trahms L, Errne SN, Tronteji Z, et al. Biomagnetic functional localization of a peripheral nerve in man. Biophys J 1989;55: Curio G, Erne SN, Sandfort J, et al. Exploratory mapping of evoked neuromagnetic activity from human peripheral nerve, brachial plexus and spinal cord. Electroenceph Clin Neurophysiol 1991;81: Mackert BM, Burghoff M, Hiss LH, et al. Non-invasive magnetoneurography for 3D-monitoring of human compound action current propagation in deep brachial plexus. Neurosci Lett 2000;289: Mackert BM, Burghoff M, Hiss LH, et al. Magnetoneuroghaphy of evoked compound action currents in human cervical nerve roots. Clin Neurophysiol 2001;112: Mackert BM, Curio G, Burghoff M, et al. Mapping of tibial nerve evoked magnetic fields over the lower spine. Electroenceph Clin Neurophysiol 1997;104: Mackert BM, Curio G, Burghoff M, et al. Magnetoneurographic 3D localization of conduction blocks in patients with unilateral S1 root compression. Electroenceph Clin Neurophysiol 1998;109: Mackert BM, Burghoff M, Hiss LH, et al. Tracing of proximal lumbosacral nerve conduction a comparison of simultaneous magneto and eloctroneurography. Electroenceph Clin Neurophysiol 2001;112: Yasuma M, Mashiko T, Iwase Y, et al. Visualization of signal propagation from sciatic nerve to spinal cord in canine. Neurosci Lett 2001;315: Kawabata S, Komori H, Mochida K, et al. Visualization of conductive spinal cord activity using a biomagnetometer. Spine 2002;27: Patton HD, Amassian VE. Single- and multiple-unit analysis of cortical stage of pyramidal activation. J Neurophysiol 1954;17: Sarvas J. Basic mathematical and elecrtomagnetic concepts of the biomagnetics invers problem. Phys Med Biol 1987;32: Nyberg-Hansen R. Corticospinal fibers from the medial aspect of the cerebral hemisphere in the cat an experimental study with the nauta method. Exp Brain Res 1969;7: Wiskwo JP. Biomagnetic Sources and their models. In: Williamson SJ, Hoke M, Stroink G, et al, editors, Advance in Biomagnetism. New York: Plenum Press, 1990:1-18.

MOTOR EVOKED POTENTIALS AND TRANSCUTANEOUS MAGNETO-ELECTRICAL NERVE STIMULATION

MOTOR EVOKED POTENTIALS AND TRANSCUTANEOUS MAGNETO-ELECTRICAL NERVE STIMULATION MOTOR EVOKED POTENTIAS AND TRANSCUTANEOUS MAGNETO-EECTRICA NERVE STIMUATION Hongguang iu, in Zhou 1 and Dazong Jiang Xian Jiaotong University, Xian, People s Republic of China 1 Shanxi Normal University,