Combined EEG and MEG analysis of early somatosensory evoked activity in children and adolescents with focal epilepsies.

OBJECTIVE: The study aimed to evaluate differences between EEG and MEG analysis of early somatosensory evoked activity in patients with focal epilepsies in localizing eloquent areas of the somatosensory cortex. METHODS: Twenty-five patients (12 male, 13 female; age 4-25 years, mean 11.7 years) were included. Syndromes were classified as symptomatic in 17, idiopathic in 2 and cryptogenic in 6 cases. 10 patients presented with malformations of cortical development (MCD). 122 channel MEG and simultaneous 33-channel EEG were recorded during tactile stimulation of the thumb (sampling rate 769 Hz, band-pass 0.3-260 Hz). Forty-four hemispheres were analyzed. Hemispheres were classified as type I: normal (15), II: central structural lesion (16), III: no lesion, but central epileptic discharges (ED, 8), IV: lesion or ED outside the central region (5). Analysis of both sides including one normal and one type II or III hemisphere was possible in 15 patients. Recordings were repeated in 18 hemispheres overall. Averaged data segments were filtered (10-250 Hz) and analyzed off-line with BESA. Latencies and amplitudes of N20 and P30 were analyzed. A regional source was fitted for localizing S1 by MRI co-registration. Orientation of EEG N20 was calculated from a single dipole model. RESULTS: EEG and MEG lead to comparable good results in all normal hemispheres. Only EEG detected N20/P30 in 3 hemispheres of types II/III while MEG showed no signal. N20 dipoles had a more radial orientation in these cases. MEG added information in one hemisphere, when EEG source analysis of a clear N20 was not possible because of a low signal-to-noise ratio. Overall N20 dipoles had a more radial orientation in type II when compared to type I hemispheres (p=0.01). Further N20/P30 parameters (amplitudes, latencies, localization related to central sulcus) showed no significant differences between affected and normal hemispheres. Early somatosensory evoked activity was preserved within the visible lesion in 5 of the 10 patients with MCD. CONCLUSIONS: MEG should be combined with EEG when analyzing tactile evoked activities in hemispheres with a central structural lesion or ED focus. SIGNIFICANCE: At time, MEG analysis is frequently applied without simultaneous EEG. Our results clearly show that EEG may be superior under specific circumstances and combination is necessary when analyzing activity from anatomically altered cortex.

Noninvasive source localization of interictal EEG spikes: effects of signal-to-noise ratio and averaging.

Source localization using single current dipoles estimates equivalent centers of the spiking gray matter. The extent of the active cortex, however, is difficult to assess from scalp EEG because of the unknown individual volume conduction. The spatial scatter of dipole localizations of single spikes has been proposed as a measure of extent. Single spike localization, however, is strongly dependent on the signal-to-noise ratio (SNR), that is, the ratio of spike and background EEG amplitudes. On the other hand, averaging of all spikes yields only the localization of equivalent centers of activity. We investigated the influence of SNR and multiple subaverages on the estimation of spatial extent by comparing the localization scatter of 100 single spikes in 27 spike types of 25 epilepsy patients with 1000 different subaverages computed by random sampling and bootstrapping. Averaging increased SNR and therefore allowed for localization not only at the spike peak but also during spike onset when less cortex is active. In several subjects with known cortical lesions, the single spike scatter considerably exceeded the lesion. Single dipole scatter was highly correlated with SNR (r = -0.83, P < 0.0001) and was greatly reduced when analyzing multiple subaverages of 10, 25, 50, and 100 spikes. Thus, we found a dominant role of the SNR on the estimated extent and improvement by scatterplots based on the dipole localization of randomly sampled subaverages.

Interaction of tactile input in the human primary and secondary somatosensory cortex--a magnetoencephalographic study.

Interaction of simultaneous tactile input at two finger sites in primary (SI) and secondary somatosensory cortex (SII) was studied by whole-head magnetoencephalography. Short pressure pulses were delivered to fingers of the right and left hand at an interstimulus interval of 1.6 s. The first phalanx of the left digit 1 and four other sites were stimulated either separately or simultaneously. We compared four sites with increasing distance: the second phalanx of left digit 1, left digit 5, and digits 1 and 5 of the right hand. The temporal evolution of source activity in the contralateral SI and bilateral SII was calculated using spatiotemporal source analysis. Interaction was assessed by comparing the source activity during simultaneous stimulation with the sum of the source activities elicited by separate stimulation. Significant suppressive interaction was observed in contralateral SI only for stimuli at the same hand, decreasing with distance. In SII, all digits of the same and the opposite hand interacted significantly with left digit 1. When stimulating bilaterally, SII source waveforms closely resembled the time course of the response to separate stimulation of the opposite hand. Thus, in bilateral simultaneous stimulation, the contralateral input arriving first in SII appeared to inhibit the later ipsilateral input. Similarly, the separate response to input at two unilateral finger sites which arrived slightly earlier in SII dominated the simultaneous response. Our results confirm previous findings of considerable overlap in the cortical hand representation in SII and illustrate hemispheric specialization to contralateral input when simultaneous stimuli occur bilaterally.

Magnetic source imaging of tactile input shows task-independent attention effects in SII.

We investigated whether attention to different stimulus attributes (location, intensity) has different effects on the activity of the secondary (SII) somatosensory cortex. Tactile stimuli were applied to the left index finger and somatosensory evoked fields (SEFs) were recorded using a whole-head magnetoencephalography (MEG) system. Two oddball paradigms with stimuli varying in location or intensity were performed in an ignore and an attend condition. Brain sources were estimated by magnetic source imaging. No attention effect was observed for the primary SI area. However, attention enhanced SII activity bilaterally from 55 to 130 ms by 52% in the spatial and 64% in the intensity discrimination task. SII attentional enhancement was very similar in both paradigms and occurred both for deviants and standards.