Surgically Relevant Localization of the Central Sulcus With High-Density Somatosensory-Evoked Potentials Compared With Functional Magnetic Resonance Imaging
Lascano, Agustina M. MD, PhD*; Grouiller, Frédéric PhD‡,§; Genetti, Mélanie PhD*,‡; Spinelli, Laurent PhD*; Seeck, Margitta MD*; Schaller, Karl MD¶; Michel, Christoph M. PhD*,‡
*Department of Neurology, University Hospital of Geneva, Geneva, Switzerland;
‡Functional Brain Mapping Laboratory, Department of Neurology, University Hospital of Geneva and University Medical Centre, Geneva, Switzerland;
§Department of Radiology and Medical Informatics, University Hospital of Geneva, Geneva, Switzerland;
¶Department of Neurosurgery, University Hospitals of Geneva, Geneva, Switzerland
Correspondence: Christoph M. Michel, PhD, Rue Michel-Servet 1, CP 1211, Geneva, Switzerland. E-mail: Christoph.Michel@unige.ch
Received September 13, 2013
Accepted January 07, 2014
BACKGROUND: Resection of abnormal brain tissue lying near the sensorimotor cortex entails precise localization of the central sulcus. Mapping of this area is achieved by applying invasive direct cortical electrical stimulation. However, noninvasive methods, particularly functional magnetic resonance imaging (fMRI), are also used. As a supplement to fMRI, localization of somatosensory-evoked potentials (SEPs) recorded with an electroencephalogram (EEG) has been proposed, but has not found its place in clinical practice.
OBJECTIVE: To assess localization accuracy of the hand somatosensory cortex with SEP source imaging.
METHODS: We applied electrical source imaging in 49 subjects, recorded with high-density EEG (256 channels). We compared it with fMRI in 18 participants and with direct cortical electrical stimulation in 6 epileptic patients.
RESULTS: Comparison of SEP source imaging with fMRI indicated differences of 3 to 8 mm, with the exception of the mesial-distal orientation, where variances of up to 20 mm were found. This discrepancy is explained by the fact that the source maximum of the first SEP peak is localized deep in the central sulcus (area 3b), where information initially arrives. Conversely, fMRI showed maximal signal change on the lateral surface of the postcentral gyrus (area 1), where sensory information is integrated later in time. Electrical source imaging and fMRI showed mean Euclidean distances of 13 and 14 mm, respectively, from the contacts where electrocorticography elicited sensory phenomena of the contralateral upper limb.
CONCLUSION: SEP source imaging, based on high-density EEG, reliably identifies the depth of the central sulcus. Moreover, it is a simple, flexible, and relatively inexpensive alternative to fMRI.
ABBREVIATIONS: DCES, direct cortical electrical stimulation
ESI, electric source imaging
fMRI, functional magnetic resonance imaging
GFP, global field power
HD, high density
MNI, Montreal Neurological Institute
SEP, somatosensory evoked potential
SI, primary somatosensory cortex
Noninvasive functional techniques, allowing accurate mapping of the somatosensory cortex in the individual subject, are of interest in planning the resection of intracranial lesions or epileptogenic foci.1,2 Accurate localization of the central sulcus reduces the risk of postoperative functional deficits in cases where the lesion is close to the sensorimotor region.3
Identification of the central sulcus is still done by visual search for anatomical landmarks from images on computer tomography (CT) or magnetic resonance imaging (MRI).4-6 However, this approach has proven to be unreliable because of the large variability between observers.1 Moreover, brain anatomy can be severely distorted in patients with large brain lesions.7 For this reason, intra- or extraoperative functional mapping through direct cortical electrical stimulation (DCES), performed in combination with electrocorticography, is considered to be the gold standard.8 However, this procedure is time consuming and/or necessitates preoperative electrode implantation, being, therefore, associated with potential risks.9,10 Also, if DCES is applied only during surgery, these results are not available for presurgical evaluation.
In the quest for noninvasive methods with added localization yield during presurgical planning, several studies investigated the use of functional magnetic resonance imaging (fMRI).11-13 fMRI localization accuracy is considered to be very high, providing a topographical layout of the individual digital representation in the primary somatosensory cortex (SI).14-16 On the other hand, given that fMRI is based on the hemodynamic response, it is thought to be less powerful in lesions that lead to changes in the vascular autoregulation, such as gliomas17,18 or cerebral ischemia.19 Moreover, fMRI is not always feasible, particularly in pediatric patients or in those with claustrophobia or carrying ferromagnetic material.
In cases where fMRI is not possible or not useful due to vascular abnormalities, electrophysiological methods based on source localization of stimulus-evoked activity can be considered as an option. Source localization techniques can be based either on magnetoencephalogram (MEG)20,21 or electroencephalogram (EEG) recordings.22,23
MEG-based source localization has demonstrated its usefulness in presurgical assessment24; whereas source localization based on EEG has rarely been promoted as a clinical tool.25 This is rather surprising given that EEG is readily available in clinics and that previous studies have already demonstrated accurate localization of SI with EEG-based source localization methods.26-30 These studies, however, were performed on few subjects and were targeted at validation of new source localization methods and head models. The stability of SEP source localization across a larger cohort of subjects and direct comparison with fMRI and intracranial recordings in the same subjects has not been systematically evaluated.
In the present study, we intended to evaluate the accuracy of high-density (256-channel) EEG source imaging (HD-ESI) in localizing the SI in individual subjects. We also directly compared the localization of the somatosensory cortex between HD-ESI and fMRI in a group of healthy subjects, and between these 2 noninvasive methods and DCES in a group of patients.
The present study is structured in 3 parts:
1. To determine the variability of the localization of HD-ESI SEPs, healthy subjects underwent mechanical air-puff stimulation. Three-dimensional source localization was determined in the average template MRI of the Montreal Neurological Institute (MNI brain).
2. For the purpose of directly comparing the individual SEP source with the conventionally used cortical stimulation, we proposed HD-ESI to 6 patients who afterward underwent intracranial evaluation with subdural electrodes covering the sensorimotor cortex. We then compared the HD-ESI source localization with the localization of the electrodes that evoked sensory responses during DCES.
3. A second group of healthy volunteers underwent HD-ESI and fMRI under identical stimulation conditions, allowing comparison of both noninvasive techniques in the individual subject.
Two groups of healthy controls were studied: (1) 31 subjects (mean age, 29 years; 16 females; 1 left-handed) in whom we recorded high-density scalp SEP (ie, HD-ESI) without MRI; (2) 18 subjects (mean age, 23 years; 7 females; all right-handed) who participated in both high-density scalp SEP and fMRI acquisitions and in whom the individual structural MRI was available. The first group was used to evaluate the correctness of the mean source localization with respect to the known location of the somatosensory cortex in a template brain and to evaluate the variability of this location across healthy subjects, while the second group was used for direct interindividual comparison with fMRI results. None of the subjects presented any previous or current neurological or psychiatric diseases. Before participation, subjects provided written informed consent to procedures that had been approved by the Ethics Committee of the University Hospital of Geneva, Switzerland, in agreement with the Declaration of Helsinki.
We studied 6 patients with medically refractory epilepsy (median age, 12 years; range, 7-35; 3 females and 3 males; all right-handed) during their noninvasive and, subsequently, their presurgical evaluation in which we performed HD-ESI, DCES, and fMRI (in 4/6). Both HD-ESI and fMRI were performed during the presurgical noninvasive evaluation phase.
We selected patients who satisfied the following criteria: (1) implanted electrodes exploring the precentral and/or the postcentral gyrus; (2) lack of paroxysmal interictal activity or early dissemination of spontaneous ictal discharges toward the peri-rolandic region. Table 1 provides clinical information about the patients who were selected for this study.
Experiments were conducted in an electromagnetically shielded, sound-attenuated and darkened room. Finger clips with balloon diaphragms (0.8 cm2 surface) produced nonpainful stimuli (stimulus duration, 50 ms) driven by bursts of compressed air. Stimuli were delivered to the distal phalanx of the thumb at a repetition rate of 2.14 Hz (system built by Christian Wienbruch and Victor Candia; see Wienbruch et al31 for further details). Air pressure was adjusted to produce a well-defined tactile sensation (2.5 bar). A total of 1000 stimuli per thumb were applied in 1 single block. Identical stimulation parameters were used in all participants. We adapted this protocol for fMRI in which a 24-second block of alternative right and left stimulations separated by a pause of 8 seconds was executed.
EEG Data Collection
EEG was collected from 256 silver-chloride-plated carbon-fiber electrodes by using a HydroCel Geodesic Sensor Net (Electrical Geodesics Inc, Eugene, Oregon). One pediatric patient was recorded with a 128-channel net (patient 1). Electrodes are interconnected by thin rubber bands and each contained a small sponge that directly touched the scalp's surface.32 The nets were soaked in saline water before placement. The whole net was applied at once, and no skin abrasion was required. The net was adjusted so that Fpz, Cz, Oz, and the preauricular points were correctly placed according to the international 10/10 system. The geodesic tension structure of the net ensured that the electrodes were evenly distributed across the head and at similar locations across subjects. Electrode-skin impedances were kept below 20 kΩ. EEG was continuously recorded at a sampling rate of 1 kHz and bandpass filtered between 0.1 and 400 Hz. The Vertex (Cz) electrode was used as recording reference, and the data were referenced offline to the average one.
Evoked potentials were computed for each healthy control and each patient, using the free academic software Cartool (D. Brunet, Geneva University Hospital and Medical School, Center for Biomedical Imaging, Geneva, Switzerland; http://sites.google.com/site/fbmlab/cartool). Epochs were selected ranging from 50 ms before to 120 ms after stimulus onset. A high-pass filter of 10 Hz was applied to the ongoing EEG. Epochs with ocular-motor artifacts were determined by voltage thresholds (50 μV) and were excluded after visual inspection.
Spherical spline interpolation was applied on scalp EEG data to interpolate any artifact-contaminated electrode.33 In addition, the electrodes on the cheek area were excluded and a standard electrode array of 204 channels was used (109 channels in the 1 patient with 128-channel net).
Source Localization of Scalp SEP
Each participant's average SEP map corresponding to the global field power (GFP) peak of the most stable early response, appearing at 40 to 60 ms poststimulus onset,34,35 was subjected to source localization. GFP corresponds to a parametric assessment of map strength computed as standard deviation of the potential value.36
A local autoregressive average distributed linear inverse solution37-39 was applied with the purpose of estimating the intracranial 3-D current density distribution of the SEP response without any a priori restriction on the location, number, or orientation of the sources. Noise regularization was based on the L-curve method40 applied to the grand mean SEP and was kept constant for the analysis of the individual subjects. Several simulation studies and application to real data have demonstrated excellent localization performance of this source localization method.39,41
A simplified realistic head model called SMAC39,42 was used for the source localization that restrains the solution space to the grey matter without constraining source orientation. The method extracts the brain surface from the MRI and calculates the best-fitting sphere for this surface. The ratio of the sphere radius and the real surface radius is then determined and the source space is warped accordingly. The source space is then constrained to the gray matter of this warped space and around 5000 solution points are equally distributed within this space. Standard spatial electrode positions were used in all subjects, coregistered to the MRI by adjusting the position of the nasion, inion, Cz, and preauricular landmarks. In the 31 subjects corresponding to group 1, in whom no individual MRI was recorded, the electrodes were coregistered on the scalp surface of the MNI brain (not in the brain or the skull), whereas the individual brain was used for source localization in the 18 subjects of group 2 and in the patients.
An analytical solution was used for the lead field calculation based on a 3-shell spherical head model. The skull relative conductivity was set to 0.05 and adjusted by the estimated skull thickness under each electrode.43 The results of the inverse solution are back-transformed to the original head shape using the same transformation parameter. It has been shown that this simplified realistic head model produces high localization reliability39,41,44 and that it is comparable to boundary element models.45
In the first group of healthy controls, in whom we evaluated location and stability of HD-ESI, MNI coordinates of the source maximum was determined for each subject. In the second group of healthy participants and the epileptic patients, we compared HD-ESI with the other mapping methods by using the x-y-z coordinates of the individual MRI scans for all modalities.
Electrocorticography and DCES
All patients had extensive coverage of the brain surface with subdural grids, strips, and, occasionally, depth electrodes over the region(s) suspected to be involved in seizure onset and early propagation (Table 1). The platinum electrodes (Ad-Tech; Ad-Tech, Racine, Wisconsin) possessed a diameter of 2.3 mm (subdural grids and strips) or 1.1 mm (depth electrodes) and were arranged at an intercontact distance of 10 mm. The anatomical targeting of electrodes was established in each patient, according to available noninvasive information (interictal and ictal scalp EEG, anatomical MRI, nuclear medicine procedures, and seizure semiology). The exact location of depth electrode contacts in the different cortical and subcortical regions was ascertained by postimplantation high-resolution CT scans rigidly coregistered to preoperative 3-D T1-weighted MRI by maximizing the normalized mutual information between these 2 images.
DCES was delivered between 2 adjacent intracranial electrodes (Astro-Med Inc, Grass Technologies, Rockland, Massachusetts) with biphasic currents (frequency = 50 Hz; pulse length = 300 μs; pulse duration = 2 s; intensity = 1-10 mA). We calculated the mean x-y-z coordinates of the electrodes in which a sensory response of the upper limb was elicited.
Structural and functional MRI acquisitions were made with a 3T whole-body MRI scanner (Siemens Magnetom Trio, Erlangen, Germany). Two hundred fifty-six functional images were acquired using a single-shot T2*-weighted gradient-echo Echo-Planar Imaging sequence (repetition time = 1980 ms, echo time = 30 ms, flip angle = 90°, voxel size = 3 × 3 × 3.75 mm3, 32 slices). A magnetization prepared rapid acquisition gradient-echo 3-D high-resolution T1-weighted structural image was acquired for individual anatomical localization (repetition time = 1900 ms, echo time = 2.27 ms, TI = 900 ms, flip angle = 9°, voxel size = 1 × 1 × 1 mm3 acquisition matrix: 256 × 256).
Preprocessing of functional images using SPM8 software (Wellcome Department of Imaging Neuroscience, UCL, London, UK) included the following: (1) realignment of the fMRI time series; (2) rigid-body coregistration of the realigned functional images on the 3-D T1 structural image by maximizing its normalized mutual information with the mean functional image46; (3) spatial smoothing of functional images with an isotropic Gaussian kernel (6 mm full width at the half-maximum).
Finally, fMRI time series were whitened and serial correlations were modeled by using an autoregressive filter of order 1. Low-frequency noise and signal drift were removed by using a discrete cosine transform basis set with a filter cutoff period of 128 s. For each of the 2 conditions (ie, left and right thumb stimulation), regressors of interest were created by convolving each block with a canonical hemodynamic response function. Motion-related parameters derived from the realignment of functional images were also included in the model as covariates. Statistical analyses of fMRI data were performed for each subject individually by using a mass-univariate approach based on the General Linear Model.47
We contrasted the results using a 1-sample t test. The significance level of the resulting SPM[t] maps was set to a threshold of P < .05 corrected for multiple comparisons across the whole brain by using family-wise error, except in 2 patients in whom we modified the threshold to P < .001 uncorrected for multiple comparisons. In these 2 young children, despite the inclusion of the 6 motion parameters in the model, the sensitivity of fMRI was decreased because of moderate motion (<1.5 mm). However, in these 2 cases, we carefully checked that the fMRI maximum was located in the postcentral gyrus, and that no spurious activation suggestive of motion was visible at this threshold.
In order to have group fMRI activation, we also normalized each individual fMRI of our 18 healthy subjects into the MNI space, and we performed a random-effects group analysis on the individual contrast images by using a 1-sample t test (P < .05, family-wise error-corrected).
Comparison Between HD-ESI With fMRI and DCES
Comparison was performed on the basis of visual analysis as well as through x-y-z coordinates of maximum activation of each method. For each subject and pair of techniques (HD-ESI vs DCES, HD-ESI vs fMRI, fMRI vs DCES), the distance between the maxima of activation was computed for each of the axis as well as the mean Euclidean distance between the different methods. Parametric independent paired-sample t tests were performed for each axis in the group of healthy subjects comparing SEP source localization and fMRI.
Consistency of SEP Source Location (HD-ESI)
Figure 1 shows the grand mean averaged SEP of the 31 healthy controls, the map at the peak latency, and the average source localization in the template MNI brain that was used as head model for the source localization. Across subjects, the GFP peak of the first stable SEP component appeared at a mean latency of 47 ± 5 ms for the left and 47 ± 6 ms for the right thumb stimulation. The mean MNI coordinates on the x-y-z axis were as follows: 39 (±5), −22 (±5), 55 (±4) for the left and −32 (±4), −30 (±5), 60 (±2) for the right thumb SEP. These locations correspond within a 5-mm radius to Brodmann areas 3 and 4, ie, the pre- and postcentral gyrus (Talairach Client from http://www.talairach.org/client.html). In order to illustrate the spread of the source localization across subjects, Figure 2 shows the location of the maximum of the source for each individual subject coregistered to the MNI template. Because the source maximum is overlapped across subjects, we color-coded the location according to the number of subjects having the maximum in that location.
Comparison Between HD-ESI and DCES in Epileptic Patients
Table 2 summarizes the results of the comparison between HD-ESI and DCES in the 6 epileptic patients. Additional fMRI was done in 4 patients. DCES was performed on all 6 patients through implanted subdural electrodes, and both sensory and motor responses were obtained. The electrodes from which a sensory response was elicited were always adjacent to each other. The median Euclidean distance between HD-ESI and DCES was 13 mm (range: 5-20 mm). Note that the distance in the medial-lateral axis was always negative because of the DCES electrode lying on the brain surface while the ESI location was located deeper in the sulcus (see Discussion). An example of this comparison is shown in Figure 3.
Comparison of HD-ESI and fMRI in Healthy Controls
In the second control group of 18 healthy subjects, HD-ESI and fMRI were obtained in all subjects showing reliable activation of the somatosensory cortex with both techniques. Mean MNI coordinates of the fMRI maximal activation were the following: 56 (±2), −15 (±4), 50 (±5) for the left and −55 (±3), −20 (±4), 49 (±4) for the right thumb. These locations corresponded to Brodmann areas 2 and 3 within a 5-mm radius.
Table 3 gives the mean and standard deviation of the difference between the SEP source maximum and the fMRI maximum and the result of the statistical comparison for each of the axes. There was a significant difference (P < .001) between both techniques in the x-axis (medial-lateral orientation). The SEP source maximum was identified on both sides more medially than the blood oxygen level-dependent fMRI maximum (difference 16 mm for the left, and 20 mm for the right thumb stimulation). Differences were not significant (P > .01) for the y- and z-axis (ie, anterior-posterior and superior-inferior, respectively). In the 4 implanted patients, who also had preoperative fMRI, the Euclidean distance of the fMRI maximum to DCES was, on average, 14 mm (range, 11-17).
The aim of this study was to assess the capability of SEP source analysis to localize the primary somatosensory area in the individual subject and to evaluate its yield compared with fMRI.
In agreement with earlier findings, scalp SEP source analysis was able to localize in the contralateral SI cortex with low variability between subjects.1,48 The evaluation of 31 subjects by using the MRI template revealed deep central sulcus localization with small variability. In terms of Talairach coordinates, the mean location corresponded to area 3b, with a slight extension to area 4 for the left thumb stimulation. Area 3b and area 4 lie within the central sulcus on the posterior (area 3b) and anterior (area 4) wall. Stimulation of mechanoreceptors reaches area 3b through the ventral posterior lateral nucleus of the thalamus and then projects to area 1 for further integration, which is located on the surface of the postcentral gyrus. Given this projection pathway, the correct localization of the first cortical SEP component is expected to be in area 3b. Because of the proximity of area 3b to area 4 (1-2 mm) and the spatial limitation of distributed inverse solution, a discrete spatial blurring within these areas across subjects cannot be avoided. Compared with “DCES” identification of the sensory cortex of the contralateral upper limb, accuracy appeared reasonable with a distance of only 10 to 20 mm. It should be remembered that DCES from subdural electrodes imperfectly maps deep sources, given that it stimulates mainly the gyri on which electrodes are placed, and only indirectly the cortex in the sulci where area 3b is located. Consequently, a difference between ESI and DCES in the medial-lateral orientation has to be expected and it is, thus, physiologically plausible.
Correct identification of the central sulcus was also possible with fMRI in the 18 healthy subjects. The contrast between SEP source localization and fMRI in these subjects provided differences that were comparable with results from other studies using fewer electrodes. They reported Euclidean distances of 16 to 17 mm1 and even 23.5 mm in a combined EEG-fMRI study.48 The mean distance between the 2 modalities was also comparable to the distances found by Kober and colleagues,35 contrasting MEG source imaging and fMRI with the use of a similar stimulation technique. One investigation found a small discrepancy of 5.1 mm and 11.9 mm while measuring high-density EEG and fMRI responses to electric median nerve stimulation, but these values were based on 2 subjects only.30 Still, it is possible that the air-puff stimulation to the thumb used here leads to less precise localization than electrical stimulation owing due to a lower signal-to-noise ratio. Also, the simplified realistic head model and the assumed electrode locations might have limited the SEP source localization precision. On the other hand, the air-puff stimulation is much more comfortable and easy to apply (particularly in children). Moreover, the simplified head model and the avoidance of long-lasting precise electrode localization measurements reduce the amount of expert effort and thus reduce costs. Another source of errors that needs to be considered in such studies is the coregistration of the different mapping modalities.49 We tried to minimize this error by using the same high-resolution individual 3-D T1 structural image as a reference for each modality. The functional images were all coregistered with this individual structural image; the coordinates of each solution point in the grey matter were defined in the T1 space, and the position of DCES electrodes were also defined in the T1 space after coregistration with the CT.
Interestingly, we observed in all participants that the major and most consistent difference between the 2 techniques lied again in the x-axis (medial-lateral orientation). The SEP source maximum was systematically localized more medial than the fMRI maximum for both stimulation sites. This observation has also been reported in the combined EEG-fMRI study of Christmann and colleagues,48 and in the MEG-fMRI comparison reported by Kober and colleagues.35 In the latter study of 34 patients, the MEG dipole was significantly more mesial (average 8 mm) compared with the fMRI maximum. The authors explained this difference by the fact that the early peak of the magnetic evoked field results from the activation of area 3b (comprising the anterior wall of the central sulcus35), whereas the fMRI maximum was located in area 1 on the postcentral gyrus. As mentioned above, area 1 is a projection area from area 3b, and is thus activated a few milliseconds later in time.50 fMRI, lacking of temporal resolution, is dominated by this more integrative processes in SI and the maximum is rather in area 1 than in area 3b.
An alternative explanation for the difference is the fact that fMRI and SEP measure different phenomena: fMRI is based on secondary metabolic and hemodynamic changes that are coupled to neuronal activity, whereas EEG measures directly the post-synaptic neuronal activity. Thus, fMRI activity follows the topology of the draining veins that might not necessarily exactly overlap with the location of the active neurons measured by SEP.51,52
Several studies directly compared noninvasive functional imaging methods with surgical mapping techniques by using either fMRI, MEG/EEG, or both.1,25,46,52,53 Results were relatively heterogeneous, showing either a slight advantage of fMRI over MEG1 or vice versa.54 In the current study of 6 patients with subdural electrodes overlying the sensory cortex, we found similar differences of HD-ESI and fMRI compared with DCES results (13 mm), indicating that both techniques provide similar yield.
Although the distances between the localization methods were similar to those reported in other studies, they are still rather large from a neurosurgical point of view. Although the systematic difference in the medial-lateral axis of the ESI compared with the DCES can be explained, the discrepancies in the other axis as well as the difference of the fMRI maximum to the DCES are not fully understood. It is evident that the methods evaluated here (including the DCES) can only direct the neurosurgeons toward certain anatomical structures and to presumably functionally relevant regions. During surgery of tumors, or during epilepsy surgical procedures in and around eloquent brain regions, intraoperative neuromonitoring and/or mapping still need to be applied.
We showed that SEP source localization based on HD-ESI reveals accurate localization in the individual patient/subject. The differences with fMRI were similar to those reported by Kober et al35 when comparing MEG with fMRI. Several studies have demonstrated similar localization accuracy of EEG and MEG provided that the same numbers of sensors are used.55-58 EEG systems allowing fast application of >200 electrodes and providing standard EEG localization software are now available.43,59 Thus, HD-ESI is a reasonable alternative to fMRI or MEG for the delineation of the sensory-motor cortex as part of presurgical planning, particularly in patients with limited cooperation.
The authors have no actual or potential conflicts of interest to report. The authors wish to indicate that they benefitted from equipment that the employing institution purchased with a financial concession. This work was supported by the Swiss National Science Foundation (grant 320030-122073 to K.S., grant 33CM30-140332 to M.S. and C.M.M.). Cartool software (http://sites.google.com/site/fbmlab/cartool) is developed by Denis Brunet, from the Functional Brain Mapping Laboratory, and supported by the Center for Biomedical Imaging (CIBM), Geneva and Lausanne, Switzerland.
We thank V. Candia and C. Wienbruch for the installation of the mechanical stimulation device.
1. Towle VL, Khorasani L, Uftring S, et al.. Noninvasive identification of human central sulcus: a comparison of gyral morphology, functional MRI, dipole localization, and direct cortical mapping. Neuroimage. 2003;19(3):684–697.
2. Vitikainen AM, Lioumis P, Paetau R, et al.. Combined use of non-invasive techniques for improved functional localization for a selected group of epilepsy surgery candidates. Neuroimage. 2009;45(2):342–348.
3. Duffau H, Capelle L, Denvil D, et al.. Usefulness of intraoperative electrical subcortical mapping during surgery for low-grade gliomas located within eloquent brain regions: functional results in a consecutive series of 103 patients. J Neurosurg. 2003;98(4):764–778.
4. Kido DK, LeMay M, Levinson AW, Benson WE. Computed tomographic localization of the precentral gyrus. Radiology. 1980;135(2):373–377.
5. Naidich TP, Valavanis AG, Kubik S. Anatomic relationships along the low-middle convexity: part I—normal specimens and magnetic resonance imaging. Neurosurgery. 1995;36(3):517–532.
6. Yousry TA, Schmid UD, Alkadhi H, et al.. Localization of the motor hand area to a knob on the precentral gyrus. A new landmark. Brain. 1997;120(pt 1):141–157.
7. Orrison WW Jr, Rose DF, Hart BL, et al.. Noninvasive preoperative cortical localization by magnetic source imaging. AJNR Am J Neuroradiol. 1992;13(4):1124–1128.
8. Miller KJ, denNijs M, Shenoy P, Miller JW, Rao RP, Ojemann JG. Real-time functional brain mapping using electrocorticography. Neuroimage. 2007;37(2):504–507.
9. Motamedi GK, Okunola O, Kalhorn CG, et al.. Afterdischarges during cortical stimulation at different frequencies and intensities. Epilepsy Res. 2007;77(1):65–69.
10. Hamer HM, Morris HH, Mascha EJ, et al.. Complications of invasive video-EEG monitoring with subdural grid electrodes. Neurology. 2002;58(1):97–103.
11. Maldjian JA, Gottschalk A, Patel RS, Pincus D, Detre JA, Alsop DC. Mapping of secondary somatosensory cortex activation induced by vibrational stimulation: an fMRI study. Brain Res. 1999;824(2):291–295.
12. Stippich C, Hofmann R, Kapfer D, et al.. Somatotopic mapping of the human primary somatosensory cortex by fully automated tactile stimulation using functional magnetic resonance imaging. Neurosci Lett. 1999;277(1):25–28.
13. Hammeke TA, Yetkin FZ, Mueller WM, et al.. Functional magnetic resonance imaging of somatosensory stimulation. Neurosurgery. 1994;35(4):677–681.
14. Schweizer R, Voit D, Frahm J. Finger representations in human primary somatosensory cortex as revealed by high-resolution functional MRI of tactile stimulation. Neuroimage. 2008;42(1):28–35.
15. Francis ST, Kelly EF, Bowtell R, Dunseath WJ, Folger SE, McGlone F. fMRI of the responses to vibratory stimulation of digit tips. Neuroimage. 2000;11(3):188–202.
16. Gelnar PA, Krauss BR, Szeverenyi NM, Apkarian AV. Fingertip representation in the human somatosensory cortex: an fMRI study. Neuroimage. 1998;7(4 pt 1):261–283.
17. Holodny AI, Schulder M, Liu WC, Maldjian JA, Kalnin AJ. Decreased BOLD functional MR activation of the motor and sensory cortices adjacent to a glioblastoma multiforme: implications for image-guided neurosurgery. AJNR Am J Neuroradiol. 1999;20(4):609–612.
18. Jiang Z, Krainik A, David O, et al.. Impaired fMRI activation in patients with primary brain tumors. Neuroimage. 2010;52(2):538–548.
19. Murata Y, Sakatani K, Katayama Y, Fukaya C. Increase in focal concentration of deoxyhaemoglobin during neuronal activity in cerebral ischaemic patients. J Neurol Neurosurg Psychiatry. 2002;73(2):182–184.
20. Hari R, Karhu J, Hämäläinen M, et al.. Functional organization of the human first and second somatosensory cortices: a neuromagnetic study. Eur J Neurosci. 1993;5(6):724–734.
21. Jousmäki V, Hari R. Somatosensory evoked fields to large-area vibrotactile stimuli. Clin Neurophysiol. 1999;110(5):905–909.
22. Baumgartner C, Doppelbauer A, Sutherling WW, et al.. Somatotopy of human hand somatosensory cortex as studied in scalp EEG. Electroencephalogr Clin Neurophysiol. 1993;88(4):271–279.
23. Buchner H, Adams L, Muller A, et al.. Somatotopy of human hand somatosensory cortex revealed by dipole source analysis of early somatosensory evoked potentials and 3D-NMR tomography. Electroencephalogr Clin Neurophysiol. 1995;96(2):121–134.
24. Ganslandt O, Ulbricht D, Kober H, Vieth J, Strauss C, Fahlbusch R. SEF-MEG localization of somatosensory cortex as a method for presurgical assessment of functional brain area. Electroencephalogr Clin Neurophysiol Suppl. 1996;46:209–213.
25. Bast T, Wright T, Boor R, et al.. Combined EEG and MEG analysis of early somatosensory evoked activity in children and adolescents with focal epilepsies. Clin Neurophysiol. 2007;118(8):1721–1735.
26. Schaefer M, Mühlnickel W, Grüsser SM, Flor H. Reliability and validity of neuroelectric source imaging in primary somatosensory cortex of human upper limb amputees. Brain Topogr. 2002;15(2):95–106.
27. Ding L, He B. Sparse source imaging in electroencephalography with accurate field modeling. Hum Brain Mapp. 2008;29(9):1053–1067.
28. Finke S, Gulrajani RM, Gotman J, Savard P. Conventional and reciprocal approaches to the inverse dipole localization problem for N(20)-P (20) somatosensory evoked potentials. Brain Topogr. 2013;26(1):24–34.
29. Kristeva-Feige R, Grimm C, Huppertz HJ, et al.. Reproducibility and validity of electric source localisation with high-resolution electroencephalography. Electroencephalogr Clin Neurophysiol. 1997;103(6):652–660.
30. Grimm C, Schreiber A, Kristeva-Feige R, Mergner T, Hennig J, Lücking CH. A comparison between electric source localisation and fMRI during somatosensory stimulation. Electroencephalogr Clin Neurophysiol. 1998;106(1):22–29.
31. Wienbruch C, Candia V, Svensson J, Kleiser R, Kollias SS. A portable and low-cost fMRI compatible pneumatic system for the investigation of the somatosensensory system in clinical and research environments. Neurosci Lett. 2006;398(3):183–188.
32. Tucker DM. Spatial sampling of head electrical fields: the geodesic sensor net. Electroencephalogr Clin Neurophysiol. 1993;873:154–163.
33. Perrin F, Pernier J, Bertrand O, Giard MH, Echallier JF. Mapping of scalp potentials by surface spline interpolation. Electroencephalogr Clin Neurophysiol. 1987;66(1):75–81.
34. Mertens M, Lütkenhöner B. Efficient neuromagnetic determination of landmarks in the somatosensory cortex. Clin Neurophysiol. 2000;111(8):1478–1487.
35. Kober H, Nimsky C, Moller M, Hastreiter P, Fahlbusch R, Ganslandt O. Correlation of sensorimotor activation with functional magnetic resonance imaging and magnetoencephalography in presurgical functional imaging: a spatial analysis. Neuroimage. 2001;14(5):1214–1228.
36. Lehmann D, Skrandies W. Reference-free identification of components of checkerboard-evoked multichannel potential fields. Electroencephalogr Clin Neurophysiol. 1980;48(6):609–621.
37. Grave de Peralta Menendez R, Gonzalez Andino S, Lantz G, Michel CM, Landis T. Noninvasive localization of electromagnetic epileptic activity. I. Method descriptions and simulations. Brain Topogr. 2001;14(2):131–137.
39. Michel CM, Murray MM, Lantz G, Gonzalez S, Spinelli L, Grave de Peralta R. EEG source imaging. Clin Neurophysiol. 2004;115(10):2195–2222.
40. Hansen C. Analysis of discrete Ill-posed problems by means of the L-curve. SIAM Rev. 1992;34:561–580.
41. Brodbeck V, Spinelli L, Lascano AM, et al.. Electroencephalographic source imaging: a prospective study of 152 operated epileptic patients. Brain. 2011;134(pt 10):2887–2897.
42. Spinelli L, Andino SG, Lantz G, Seeck M, Michel CM. Electromagnetic inverse solutions in anatomically constrained spherical head models. Brain Topogr. 2000;13(2):115–125.
43. Brunet D, Murray MM, Michel CM. Spatiotemporal analysis of multichannel EEG: CARTOOL. Comput Intell Neurosci. 2011;2011:813870.
44. Brodbeck V, Lascano AM, Spinelli L, Seeck M, Michel CM. Accuracy of EEG source imaging of epileptic spikes in patients with large brain lesions. Clin Neurophysiol. 2009;120(4):679–685.
45. Guggisberg AG, Dalal SS, Zumer JM, et al.. Localization of cortico-peripheral coherence with electroencephalography. Neuroimage. 2011;57(4):1348–1357.
46. Ashburner J, Friston K. Multimodal image coregistration and partitioning—a unified framework. Neuroimage. 1997;6(3):209–217.
47. Kiebel SJ, Poline JB, Friston KJ, Holmes AP, Worsley KJ. Robust smoothness estimation in statistical parametric maps using standardized residuals from the general linear model. Neuroimage. 1999;10(6):756–766.
48. Christmann C, Ruf M, Braus DF, Flor H. Simultaneous electroencephalography and functional magnetic resonance imaging of primary and secondary somatosensory cortex in humans after electrical stimulation. Neurosci Lett. 2002;333(1):69–73.
49. Singh KD, Holliday IE, Furlong PL, Harding GF. Evaluation of MRI-MEG/EEG co-registration strategies using Monte Carlo simulation. Electroencephalogr Clin Neurophysiol. 1997;102(2):81–85.
50. Allison T, McCarthy G, Wood CC, Darcey TM, Spencer DD, Williamson PD. Human cortical potentials evoked by stimulation of the median nerve. I. Cytoarchitectonic areas generating short-latency activity. J Neurophysiol. 1989;62(3):694–710.
51. Ogawa S, Menon RS, Tank DW, et al.. Functional brain mapping by blood oxygenation level-dependent contrast magnetic resonance imaging. A comparison of signal characteristics with a biophysical model. Biophys J. 1993;64(3):803–812.
52. Turner R. How much cortex can a vein drain? Downstream dilution of activation-related cerebral blood oxygenation changes. Neuroimage. 2002;16(4):1062–1067.
53. Histed MH, Bonin V, Reid RC. Direct activation of sparse, distributed populations of cortical neurons by electrical microstimulation. Neuron. 2009;63(4):508–522.
54. Korvenoja A, Kirveskari E, Aronen HJ, et al.. Sensorimotor cortex localization: comparison of magnetoencephalography, functional MR imaging, and intraoperative cortical mapping. Radiology. 2006;241(1):213–222.
55. Cohen D, Cuffin BN. EEG versus MEG localization accuracy: theory and experiment. Brain Topogr. 1991;4(2):95–103.
56. Liu AK, Dale AM, Belliveau JW. Monte Carlo simulation studies of EEG and MEG localization accuracy. Hum Brain Mapp. 2002;16(1):47–62.
57. Malmivuo J, Suihko V, Eskola H. Sensitivity distributions of EEG and MEG measurements. IEEE Trans Bio-med Eng. 1997;44(3):196–208.
58. Malmivuo J. Comparison of the properties of EEG and MEG in detecting the electric activity of the brain. Brain Topogr. 2012;25(1):1–19.
59. Baillet S, Friston K, Oostenveld R. Academic software applications for electromagnetic brain mapping using MEG and EEG. Comput Intell Neurosci. 2011;2011:972050.
1. In localizing the central sulcus, in what orientation is the maximum discrepancy between SEP (somatosensory evoked potential) source imaging and fMRI (functional MRI)?
2. What type of lesion does not affect accuracy of fMRI in localizing the motor cortex?
A. Large brain tumor with surrounding edema
B. Ischemic stroke
C. Vascular tumor
D. Small and calcified meningiomas
3. In a brain tumor patient who has a ferromagnetic cardiac pacemaker, which non-invasive technique is indicated for precentral gyrus localization?
A. Functional MRI
C. High density somatosensory evoked potential mapping (SEP)
D. 64-slice CT
Electrical source imaging; Electroencephalography; Eloquent cortex; Functional imaging; Presurgical evaluation; Primary somatosensory cortex
Copyright © by the Congress of Neurological Surgeons
Highlight selected keywords in the article text.