MAGNETIC LOCALIZATION AND ORGANIZATION OF SOMATOSENSORY CORTEX IN CHILDREN

J. Xiang¹, S. Chuang¹, H. Otsubo² , S. Holowka¹, R. Sharma², A. Hunjan^2
1Department of Diagnostic Imaging,  ²Department of Pediatrics, The Hospital for Sick Children,
555 University Avenue, Toronto, ON, CANADA, M5G 1X8

Abstract: Twenty children have been studied using 151-channel Magnetoencephalography(MEG) system with electrical stimulation applied to left or right median nerve.   The sources represented by the neuromagnetic signals were estimated using single dipole modeling and synthetic aperture magnetometry (SAM).  Frequency-related changes were localized, and three-dimensional neuromagnetic activity was reconstructed. The averaged waveforms demonstrated that the first deflection (M1) was clearly identified in most of children (90%, 18/20). The latency of M1 was 19.5±1.8 ms for right stimulation and 19.4±1.7 ms for left stimulation. Two frequency bands, 30 – 60 Hz and 60 – 120 Hz were observed to be related to somatosensory cortex. Dipole and SAM data overlapped on individual MRI indicated that they were in the same area. The extent of the reconstructed neuromagnetic activity in the left hemisphere was larger than that in the right hemisphere.  Furthermore, the neurons related to 30 – 60 Hz and 60-120 Hz bands were clearly separated in left hemisphere in six subjects, but not in the right hemisphere. We conclude that MEG is not only very useful for localizing somatosensory cortex, but also could be useful in the investigation of the development of the brain function in children.

INTRODUCTION

Magnetoencephalography (MEG) measurements can be used to noninvasively localize functional regions of the brain, and can be used routinely for presurgical planning and imaging guided surgery.  Its temporal resolution is less than 1 millisecond (ms), significantly shorter than that of functional magnetic resonance imaging (fMRI). Its spatial discrimination is 2-3 mm for sources in the cerebral cortex, much better than electroencephalography (EEG) [1,2,3] . Magnetic signals are generally described using an equivalent current dipole (ECD).  The dipole model can indicate the location and orientation of the center of activity, but is unable to describe the organization of the functional area.  In the present study, we applied a new method, synthetic aperture magnetometry (SAM), to estimate and localize event-related changes in rhythm in somatosensory cortex in three-dimension. The objective of this study was to localize somatosensory cortex and reconstruct the three-dimensional neuromagnetic activity in children using MEG and magnetic resonance imaging (MRI).

METHODS

Twenty children (13 girls and 7 boys, aged 6 – 17 years, with a mean age of 9 years) were studied. A Grass electrical stimulation system (Astro-Med Inc., West Warwick, U.S.A) was used in present study. The electric stimulus was a constant voltage square-wave pulse at a rate of 3 Hz. Its duration was 0.3 ms.  The stimulus intensity was adjusted to a level at which a clear thumb twitch was observed.  Left and right median nerves were stimulated sequentially.

A 151-channel whole cortex CTF OMEGA system was used for recordings (CTF Systems Inc., Port Coquitlam, Canada). Data were recorded with noise cancellation of third order gradients. The sampling rate of data acquisition was 1250 Hz.  Four hundred epochs were averaged for one session. Three-dimensional magnetic resonance imaging, 3D-SPGR sequence, was obtained for subjects using a Signa Advantage (GE Medical System, Milwaukee, USA).  Three fiduciary points were marked on nasion, left and right pre-auricular points on the subject’s head with MRI markers.  The positions of the three fiduciary points were identical to the positions of the three coils used in MEG. The dipole source location corresponding to each peak latency was estimated individually using the CTF DipoleFit program with the single dipole model. Six frequency bands including 8-15 Hz, 15-30 Hz, 30 – 60 Hz, 60-120 Hz, 120 – 200 Hz and 200 – 300 Hz were analyzed using SAM.   A real head model was produced from each subject’s MRI, and multiple spheres were used in SAM analysis.  The region of interest (ROI) was set to include the whole brain with a 2.5 mm voxel resolution.    Pseudo-T (related to Student’s t) values were calculated from the changes of spectral power in each voxel between the active states and the control states, normalized by the variance using SAM.   The distributions of spectral power were displayed on the individual MR images [4] .

RESULTS

The first three deflections (responses) were identifiable in most of subjects, we termed them as M1, M2 and M3. The first identifiable response, M1, was the most reproducible and robust component, and observed in 18 out of 20 subjects (90%). The latency of the M1 less than 18 ms was only found in children under 10 years old. The latency was more consistent than amplitude across the subjects. There was no significant differences between the M1 and the M2 in terms of amplitude (p > 0.05), however, their latencies were significantly different (p < 0.005). Obviously, the M3 was significantly later and smaller than the M1 and M2 (p < 0.005).

A typical averaged waveform and contour maps elicited by electrical stimuli applied to left median nerve are shown in Fig. 1. The contour maps of the responses indicated one dipole in the right hemisphere.  The orientation of the M1 was constantly anterior, the orientation of the M2 was constantly posterior, and the orientation of the M3 varied among the subjects.

Figure 1. The waveform, contour map, SAM images and dipole overlapped on 3D MRI. Three deflections are clearly identifiable.

The dipoles estimated for M1 were consistently located around the hand area of somatosensory cortex. The dipole locations for M2 were very close to that of M1.  The results of SAM analysis demonstrated that the spectral distributions in the 8 – 15 Hz, 15- 30 Hz, 120 – 200 Hz and 200 – 300 Hz bands varied among subjects.  In contrast, the spectral distributions in frequency bands of 30 –60 Hz and 60 – 120 Hz were reproducible and robust.  One or two local spectral power increases (activations) were clearly identified just around the hand area of the somatosensory cortex (Fig. 1). Comparing dipole location and SAM image indicated that the dipole was located just around the center of the SAM image.  The two results were in agreement with each other.    The reconstructed three-dimensional areas in the right hemisphere which was responsible for electrical stimulation applied to left median nerves. Activity in the 60-120 Hz band was described by a single source in all subjects, while 30-60 Hz activity appeared as two sources in two subjects. The area activated in the left hemisphere was clearly larger than that in the right hemisphere.   The activations estimated by SAM for the 30 – 60 Hz and 60 – 120 Hz bands overlapped one another (Figure 1). Two frequency bands were clearly separated in left hemisphere but not in right hemisphere in 6 subjects.

DISCUSSION

Our study indicated that somatosensory evoked magnetic fields following median nerve stimulation was reliable and reproducible in children. With this robust stimulus, we also demonstrated that dipole and SAM results were in agreement with each other.   The latency and contour map of the somatosensory evoked magnetic fields in children was notably different from that of adults.   The most interesting finding in our study was the three-dimensional organization of somatosensory cortex in children.  There were two frequency bands, 30 – 60 Hz and 60 – 120 Hz, were related to somatosensory cortex.  The extend of the reconstructed functional area was not symmetric between the left and right hemispheres.  The left hemisphere was bigger than the right hemisphere.  Furthermore, the two groups of the neurons responsible to two frequency bands were clearly separated in left hemisphere, but not in the right hemisphere in 6 subjects. It was suggested that there was a dominant side in somatosensory system in humans.  The volumetric characteristics of the three-dimensional reconstruction of neuromagentic activity were considered to be able to indicate the development of brain function in children.

ACKNOWLEDGEMENTS

We thank Dr. Paul Babyn for insightful theoretical discussions and many practical suggestions during the course of these experiments. This study was supported by Seed Grant (77331) from the Research Institute at The Hospital for Sick Children, Toronto, Canada. 

REFERENCES

[1] T.Kawamura, N.Nakasato, K.Seki, et al. Neuromagnetic evidence of pre- and post-central cortical sources of somatosensory evoked responses, Electroencephalogr Clin Neurophysiol, vol.100, pp.44-50,1996.

[2] J.Xiang, M.Hoshiyama, S. Koyama, et al. Somatosensory evoked magnetic fields following passive finger movement, Brain Res Cogn Brain Res, vol. 6,pp.73-82, 1997.

[3] R. Hari, S.L. Joutsiniemi, J.Sarvas. Spatial resolution of neuromagnetic records: theoretical calculations in a spherical model, Electroencephalogr Clin Neurophysiol, vol.71,pp.64-72, 1988.

[4] J. Xiang, D. Wilson, H.Otsubo, et al. Neuromagnetic spectral distribution of implicit processing of words, Neuroreport vol.12,pp.3923-3927, 2001.