Search

Neuromagnetic fields from the left cerebral hemisphere of five healthy, right-handed subjects were investigated under three different experimental conditions: (1) electrical stimulation of the right index finger (task S); (2) voluntary movement of the same finger (M); (3) M+S condition, consisting of voluntary movements of the right index finger triggering the electrical stimulus at the very beginning of the electromyogram. The three conditions were administered in random order every 5-8 s. In addition, the task somatosensory evoked fields (task SEFs) gathered during condition (1) were compared with control SEFs recorded at the beginning of the experiment during rest. In all subjects the overlay of somatosensory stimulation on movement provoked a decrement in brain responsiveness (gating) as determined by the amplitude of gated SEFs. The latter was found as the difference between the neuromagnetic fields during M+S condition (overlaying of movement and sensory stimulation) minus neuromagnetic fields under M condition (M only). The gating effect was found to begin approximately 30 ms after movement onset, and to last for the whole period of the ongoing movement. The theoretical locus of gating was estimated by dipole localisation of the difference between task SEFs and gated SEFS using a moving dipole model. The site of the "early" gating effect (< 40 ms) was found to be more anteriorly located than the "later" (> 40 ms) gating effect. The task SEFs were found to be larger (significant after 30 ms) than the control SEFs elicited under the basal condition. The results are discussed with respect to timing, mechanism (centrifugal and centripetal), locus and selectivity of gating. In addition, the results are discussed with regard to clinical application (measuring attentional deficits in patients with impairments of higher mental functions and measuring gating deficits in patients with disturbed sensorimotor integration.

The magnetic correlate of the P300 wave was registered using a seven-channel detector. The spatial variation of the corresponding field suggests the contribution of outer brain layers to the neural activity of M300.

Upon conventional median nerve stimulation at the wrist early magnetic fields were recorded using a SQUID magnetometer. At the upper arm, mono- and biphasic compound nerve action fields were detected, depending on the subject's distribution of single fiber conduction velocities. At the upper thorax, brachial plexus fields reversed polarity at the level of Erb's point; their distribution was asymmetric, probably due to volume currents. At the upper lateral neck, fields from proximal plexus, spinal cord (P13m) and the primary somatosensory cortex contralateral to the sensor position were detected. The observed P13m field distribution agrees with the electrophysiological concept of a sagittal segmental dorsal horn generator.

After obtaining the measured magnetic field or its components in the form of an isofield map one has to decide which approximation to use in solving the inverse problem. A single current dipole as an equivalent current source is often used. It will be shown when this approximation fails and one possible way to improve the equivalent source description. The expansion of current multipoles is discussed up to the second order. The localization of an equivalent current source in this case is considered. The application of this type of expansion is analysed and discussed.

The neuromagnetic method was applied to the study of photoconvulsive responses. The identification of specific magnetic field distributions over the scalp was achieved by; (a) a stimulation paradigm consisting of series of trains of flicker stimuli randomly presented to the epileptic patient, after eye closure, to get epileptic responses while avoiding seizures; (b) a novel procedure for data analysis, to select consistent responses. These patterns, when sufficiently stable in time and dipolar in shape, were used for source localization in the usual biomagnetic framework of the equivalent current dipole source representation. The results of this approach suggest that different specific cortical areas are repeatedly and randomly activated, involving mainly the frontal, occipital and temporal areas, often with a hemispheric prevalence.