AUTHOR(S): Hayashi, Nakamasa, M.D.; Endo, Shunro, M.D.; Kurimoto, Masanori, M.D.; Nishijo, Hisao, Ph.D.; Ono, Taketoshi, Ph.D.; Takaku, Akira, M.D. Departments of Neurosurgery (NH, SE, MK, AT) and Physiology (HN, TO), Toyama Medical and Pharmaceutical University, Toyama, Japan Neurosurgery 37; 694-703, 1995 ABSTRACT: THIS REPORT DESCRIBES a preoperative simulation and intraoperative localization system for use with three-dimensional functional images in intracranial surgery. The system, which produces three-dimensional functional images by superimposition of the generators of somatosensory evoked potentials derived from dipole tracing, was applied in the preoperative localization of central sulci in 12 patients with intracranial disorders. The preoperative localization of central sulci was assessed by intraoperative cortical recording of somatosensory evoked potentials in three patients. In six patients, three-dimensional computer graphics were reconstructed from magnetic resonance images and allowed visualization of the lesions through the semitransparent views of the scalp surface and the brain. The three-dimensional functional images were created by superimposing the generators of somatosensory evoked potentials on three-dimensional computer graphics. This combined technique provided preoperative data regarding the three-dimensional relationship between the sensorimotor cortex and the lesions. We also developed a three-dimensional digitizer that incorporates a mechanical arm and a laser pointer for use in integrating information obtained in the preoperative simulation into the surgical field. Intraoperative localization can be performed in real time on the three-dimensional computer graphics in combination with the functional images created from dipole tracing. This system was applied in five patients who underwent surgical brain tumor resections with minimal damage to the normal brain tissue. The use of this system for stereotactic craniotomy increased the safety of the surgery by affording preoperative simulation and intraoperative localization with the three-dimensional functional images. Two illustrative case reports are presented. KEY WORDS: Dipole tracing; Magnetic resonance imaging; Somatosensory evoked potentials; Stereotaxy; Threedimensional computer graphics A boundary element method incorporating a realistic head model (dipole tracing [DT]) was recently Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 reported as a noninvasive technique of threedimensional localization of current source generators (dipoles) in the brain (9-11,19). This technique, used in combination with magnetic resonance (MR) imaging, provides information regarding functional localization within the brain. Recent advances in computer graphics have facilitated reconstruction of three-dimensional anatomic structures from MR images, contributing to comprehension of the morphological structure of the brain (31) and to preoperative neurosurgical planning (12). In the present study, generators of somatosensory evoked potentials (SEPs) were investigated by DT and the estimated dipoles were superimposed on MR images in patients with intracranial disorders. To preoperatively obtain information regarding the threedimensional relationship between the sensorimotor cortex and the location of lesions, three-dimensional functional images were created by superimposing the generators of SEPs derived from DT on threedimensional computer graphics (3d-CG) reconstructed from MR images. The 3d-CG created by our system enable surgeons to visualize lesions through semitransparent images of the scalp surface and the brain. Using this system, the surgeon, by examining the three-dimensional functional images on the computer display, can accurately estimate the optimal head position, the craniotomy placement, the trajectory to the target lesion, and the extent of resection possible without inflicting damage on the normal areas of the brain. Furthermore, to integrate the information obtained in the preoperative simulation into the surgical field, we developed a three-dimensional digitizer that incorporates a mechanical arm and a laser pointer. The accuracy of the preoperative noninvasive localization of the central sulcus (CS) was assessed by the intraoperative cortical recording of SEP in patients with pararolandic tumors. This preoperative simulation and intraoperative localization system using threedimensional functional images was found to increase the safety of brain surgery. PATIENTS AND METHODS Generators of SEP were investigated by DT in each of 12 patients with various intracranial disorders, including meningioma in 2 patients, metastatic tumor in 2, glioma in 6, recurrent glioma in 1, and arteriovenous fistula in 1. In three patients with pararolandic tumors, intraoperative cortical recordings of SEP were performed to examine the accuracy of the preoperative localization of the CS by DT. In six patients, three-dimensional functional images were created. In five of these patients, the three-dimensional digitizer was intraoperatively used to integrate the information obtained in the preoperative simulation into the surgical field. Dipole tracing The dipole localization method used has been described in detail elsewhere (9,11,19). The DT, by means of a boundary element method, calculates potential distributions generated on a surface with any shape by an assumed dipole. Because the DT is Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. Neurosurgery 1992-98 October 1995, Volume 37, Number 4 694 Functional Image-guided Neurosurgical Simulation System Using Computerized Threedimensional Graphics and Dipole Tracing Technique Assessment MR imaging data MR imaging was performed with an adapter with three small markers (3-mm-diameter cotton balls soaked in oil) placed on the nasion and in the meatus of the bilateral ears, to determine reference coordinates for DT (Fig. 2A). Imaging was conducted using a whole-body 1.5-T MR imaging system with a volumetric gradient echo pulse sequence (FLASH) with a TR of 40 ms, a TE of 5 ms, and a flip angle of 40 degrees. The image acquisition time was approximately 20 minutes, and 120 contiguous sagittal sections (1.5 mm thick; 256 × 256 matrix of Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 1.1-mm pixels) were depicted. In the patients scheduled to undergo stereotactic craniotomies using the three-dimensional digitizer, three markers of the same type were also placed in the region surrounding the surgical field. These sites were labeled with India ink for use in image registration at surgery. Three-dimensional functional imaging The three-dimensional functional images were created by superimposing the generators of SEP on the 3d-CG. The system used to generate the 3d-CG (AdVANS; Tomiki Medical Instrument Co., Kanazawa, Japan) consisted of a graphics workstation (Indy; Silicon Graphics, Mountain View, CA) and an image scanner (GT-6500; Epson, Suwa, Japan). The software was implemented in the C++ programming language under the UNIX operating system. Images were entered into the graphics workstation using an image scanner that reads film. Digital image data from sources such as MR imaging can also be transmitted to the graphics workstation in approximately 30 minutes using magneto-optical disks or floppy disks. The original cross-sectional images were interpolated to obtain image volume with cubic voxels. At the graphics workstation, the data were subjected to image-processing and volume-rendering manipulations to produce three-dimensional views of the skin surface, the brain, and the region of interest traced out of the boundary on serial slices using a mouse-controlled cursor before volume-rendering manipulations. The segmentation of the gross constituents, such as the skin surface and the brain, was semiautomatically conducted based on the minimum and maximum gray values in the eight-bit image. The three-dimensional coordinates of the positions of the dipoles for SEP estimated by DT were based on the plane intersecting the nasion and the bilateral external meatus. Figure 3 shows the threedimensional views of the skin surface generated by the graphics workstation. The markers at the nasion and the meatus of the ear are identified on the display. The user points to these three markers using the cursor, and the line between the markers at the meatus of the ear is then defined as the x axis (interaural line), and the line from the marker at the nasion perpendicular to the x axis is defined as the y axis. The x/y intersect is defined as the origin, O. The line orthogonal to the x and y axes in the direction of the vertex is defined as the z axis. Therefore, the Cartesian coordinates of each cubic voxel composing an image volume can be automatically obtained within this reference coordinate system. The preoperative localization of the CS was performed by the input of the three-dimensional coordinates of the position of the estimated dipole from DT into the graphics workstation. Thus, the three-dimensional functional images were created by superimposing the estimated dipole from DT. Preoperative simulation After the original cross-sectional images are transmitted into the graphics workstation, the user Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. based on a head model that has a realistic shape, the exact three-dimensional head geometry of each patient is reconstructed by measuring the surface of the head using 64 position sensors in the hemispherical helmet (Fig. 1A) (18). This helmet also measures the positions of the electrodes used for recording the electroencephalography data. The computer-reconstructed head model is derived from 608 triangulations connecting 306 nodes, which are intermediate points of 64 positions measured by the hemispherical helmet. Thus, in this method, because the actual geometry of the head is precisely delineated, the anatomic locations of the estimated dipoles obtained by DT can be stereotactically superimposed on the MR images. The computer estimated the position, orientation, and moment of each equivalent dipole to minimize the difference between the measured and the calculated distributions of SEP. The root-mean-square goodness of fit (dipolarity) is defined by {1-∑(Uo-Uc)2/∑Uo2}1/2 × 100, where Uo and Uc are the observed and calculated potentials, respectively, at each electrode (9) . We adopted optimal dipoles that had dipolarities exceeding 97%, because this dipolarity is a good estimate of the measured potential distribution (19). The depth of the dipole calculated from the surface potential distributions is subject to overestimation, because the electrical conductivity of the cranium is much lower than that of the brain tissue, and the scalp potential distributions could be smoothed by the properties of the cranium (18). Therefore, to reduce the influence of the cranium, we used a threedimensional reconstructed model consisting of scalp, cranium, and brain tissue components with relative conductivities (Fig. 1, B, C, and D) (10). SEP recorded through the 21 electrodes placed on the scalp were simultaneously amplified and filtered (1.6-1500 Hz; -3 dB points) through a 32-channel amplifier (Biotop 6R12; NEC San-ei, Tokyo, Japan). The output signals were relayed to a 32-bit workstation (Spark Station 10; Sun Microsystems, Inc., Mountain View, CA), where approximate dipoles were calculated. The data were sampled every 100 µs and averaged (n = 200) on the computer. More than four averages were obtained to assess the reliability of the SEP in each patient. All recordings were in reference to the linked ear lobes. Stimulation was a sequence of constant-current 1-Hz pulses, 0.5 ms in duration, delivered to the median nerve at the wrist with sufficient intensity to produce moderate thumb twitch. Three-dimensional digitizer The three-dimensional digitizer (AdVANS) incorporates a mechanical arm and a laser pointer and is for use in the integration of information obtained in the preoperative simulation into the surgical field (Fig. 2B). The mechanical arm consists of three joints mounted with rotary encoders forming an isocenter system. The arm is interlinked with the base ring of the stereotactic frame by the base table, allowing three-dimensional movement from the origin. The center of the arm is moved in relation to the Cartesian coordinates (x, y, and z) by the superior-inferior movement (y) of the base table and the right-left (x) and anteroposterior (z) movement of the proximal end of the arm along the base table. The center of the arm is aligned with that of the stereotactic frame to fix the patient's head when the coordinates are 0, 0, and 0. The user need only match the center of the arm with a target point on the Cartesian coordinate system. The distal end of the arm has a laser pointer. In this apparatus, the laser beam is always oriented at the center of the orb formed by this arm in composite rotation of the three joints. Therefore, when the center of the arm is aligned with a target point on the Cartesian coordinate system, the position illuminated by the laser pointer indicates the trajectory to the target point. The laser pointer has a charge-coupled device for measuring the distance between the outlet of the laser beam and the position illuminated by the laser pointer. The point at which the distance measured by the charge-coupled device coincides with the radius of the orb formed by this arm (180 mm) corresponds to the target point. The geometric data from the rotary encoders mounted in the joints of the arm and the charge-coupled device in the laser pointer are entered into the graphic workstation, where the position illuminated by the laser pointer is calculated and superimposed on the 3d-CG. Thus, the intraoperative location can be monitored in real time on the computer display. At surgery, this location is Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 used to delineate the surgical trajectory and to assist definition of the lesion boundaries preoperatively simulated using the present system on the computer display. We tested the accuracy of the three-dimensional digitizer used in this study in two ways. The reproducibility of the center of the mechanical arm was tested in isocentric movement of the arm. First, the center of the arm was aligned with the center of the base ring of the stereotactic frame and the joints of the arm were rotated in various randomly chosen angles. The deviation of the center of the arm from that of the base ring of the stereotactic frame was within 1.0 mm. In the examination of the accuracy of position detection using a phantom, the center of the arm was moved in various directions based on the Cartesian coordinate system and aligned in various positions in a latticed phantom. The average error was 1.9 ± 0.7 mm (mean ± standard deviation) in the measurement of 100 points in the phantom. The assessment of accuracy was performed by independent engineers (Tamagawa Seiki Co. Ltd., Iida, Japan). Procedure in the operating room The base ring of the stereotactic frame allows rotation of the patient's head using the adapter fitted to the patient's external meatus and nasion around the x and y axes. The patient's head is rotated along the base ring of the stereotactic frame to the preoperatively simulated head position and fixed to the base ring of the stereotactic frame with steel pins. After fixation of the head, the upper half of this stereotactic frame and the adapter fitted to the head can be removed, so that they will not interfere with the surgical procedures during craniotomy. The three-dimensional digitizer is then operated to read the three markers placed on the region surrounding the surgical field. These three points on the patient's head are illuminated by the laser pointer of the three-dimensional digitizer, which measures their Cartesian coordinates. The coordinates of these markers are automatically entered into the graphic workstation and matched with those on the 3d-CG. Thus, the coordinates of the patient's head are aligned with the coordinate system of the three-dimensional digitizer and the coordinates of each cubic voxel composing an image volume at the graphic workstation. The center of the arm of the three-dimensional digitizer is then matched with a target point on the Cartesian coordinate system. The position illuminated by the laser pointer on the patient's scalp indicates the trajectory to the target point. Because this point can be monitored on the 3d-CG, the surgeon can accurately estimate the craniotomy placement, the trajectory to the target lesion, and the boundaries of the lesion preoperatively on the patient's scalp and intraoperatively on the cranium, the dura mater, and the brain surface (Fig. 6). The preoperatively estimated position of the CS derived by DT can be also assessed by surveying the patient's head and brain surface using the three-dimensional digitizer. Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. can obtain correlated sagittal, coronal, and axial reconstructions using the mouse-controlled cursor on the display, as shown in Figure 4. In the simulation of routine stereotactic surgery, the trajectory from the insertion point on the skin to the target lesion, to which the user points using the cursor, is interactively displayed through the anatomy of the patient on the display. Figure 5 shows three-dimensional views of the region of interest seen through the semitransparent views of the scalp surface and the brain. This three-dimensional view can also be explored from any angle of view using image rotation, and the expected operative field can, therefore, be preoperatively visualized on the computer display. When the user outlines the portion to be uncovered using the mouse-controlled cursor on the opaque skin surface, the brain surface and the region of interest become visible through the semitransparent brain image on the computer display. In simulating stereotactic craniotomy, the serial plane orthogonal to the direction into the target lesion is also displayed. ILLUSTRATIVE CASES The results in two of the five patients who underwent stereotactic craniotomy using the threedimensional digitizer are presented in the following case reports. Patient 1 The patient was a 45-year-old man with a history of seizure. MR imaging demonstrated a mass lesion in the right frontal lobe. Figure 9A shows the threedimensional functional image of this patient, which revealed that the tumor was located anterior to the premotor area. The position of the CS estimated by DT was 3.5 cm posterior to the posterior border of the tumor. The area outlined with red using the mousecontrolled cursor indicates the simulated craniotomy. Thus, in the simulated craniotomy in this patient, the projected operative field, the brain surface, and the tumor after the craniotomy could be preoperatively visualized on the computer display. After the patient's head was fixed using the stereotactic frame, the three-dimensional digitizer was used to integrate information obtained in the preoperative simulation into the surgical field. The operator outlined the portion to be uncovered using the three-dimensional digitizer on the patient's head, in the same manner as with the use of the mousecontrolled cursor, and assessed the simulated craniotomy on the computer display. Figure 9B shows the intraoperative view of the surgical field after craniotomy. The tumor was not visible on the brain surface. The area enclosed with dots indicates the boundaries of the tumor delineated by survey on the brain surface using the threedimensional digitizer. In this patient, the location of the CS estimated by DT corresponded to that identified by the potential reversal of the primary cortical response of SEP (Fig. 9C). A strip recording electrode holding four electrodes (10-mm interelectrode distance) was inserted under the dura mater posterior to the edge of the bone window. Negative evoked potentials of 20 ms from the stimulator (N20) were recorded from Electrode 3, which was located 4 cm posterior to the posterior border of the tumor, whereas positive evoked Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 potentials of 20 ms from the stimulator were recorded from Electrode 2, located 3 cm posterior to the posterior border of the tumor. Thus, it was confirmed that the CS was located 3 to 4 cm posterior to the posterior border of the tumor. This result was consistent with that preoperatively obtained by DT. Resection of the tumor was uneventfully performed. Postoperatively, the patient showed no additional neurological deficit. Patient 2 The patient was a 64-year-old woman with a history of left hemiparesis, in whom MR imaging had demonstrated a mass lesion in the right parietotemporal lobe. The three-dimensional functional image of this patient revealed anterior deviation of the CS resulting from the mass and perifocal edema (Fig. 10). Figure 11A shows the operative view created by the graphics workstation, which enabled the surgeons to visualize the lesion located beneath the superior temporal sulcus. This operative view facilitated the planning for the craniotomy placement, the trajectory to the lesion, and the optimal resection of the lesion. After the patient's head was rotated to the preoperatively simulated head position and fixed using the stereotactic frame, the three-dimensional digitizer was operated to integrate information obtained in the preoperative simulation into the surgical field. The operator outlined the portion to be uncovered using the three-dimensional digitizer on the patient's head, in the same manner as with the use of the mouse-controlled cursor, and assessed the simulated craniotomy on the computer display. Thereafter, the center of the arm of the threedimensional digitizer was aligned with the target point on the Cartesian coordinate system and the trajectory to the target lesion was assessed on the computer display. Figure 11B shows the intraoperative view of the surgical field after stereotactic craniotomy. The position of the laser point on the superior temporal sulcus indicated the optimal trajectory to the tumor. This position was intraoperatively displayed on the three-dimensional view, the correlated sagittal, coronal, and axial reconstructions, and the serial plane orthogonal to the trajectory on the computer display (Fig. 12). Resection of the tumor was uneventfully performed. Postoperatively, the patient showed no additional neurological deficit. DISCUSSION Recently, noninvasive localization of the somatosensory cortex by the combined use of magnetoencephalography and MR imaging has been reported (1-3,6,13,17,21,22,26-28,33,34). Another application of the combination of dipole source analysis of scalprecorded SEPs with MR imaging is in the preoperative localization of the CS (4). There are at least three main sources of error in dipole localization in the brain derived from electroencephalography data recorded from the surface of the scalp: measurement of the electrode position, measurement of the head geometry, and electrical inhomogeneity Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. RESULTS On the normal side of the brain in each patient, the CS was identified on the basis of anatomic landmarks in both axial and sagittal MR reconstructions, and the location of the CS estimated by DT corresponded to its anatomic location as determined with MR images (Fig. 7A). On the abnormal side of the brain, the location of the CS estimated by DT showed variable deviation resulting from the presence of spaceoccupying lesions and perifocal edema (Fig. 7B). In the three patients with pararolandic tumors who underwent preoperative localization of the CS, the location of the CS was intraoperatively assessed by the cortical recording of SEP. The location of the CS estimated by DT corresponded to that identified by the potential reversal of the primary cortical response of SEP (Fig. 8). Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 illuminated by the laser pointer, indicates the trajectory to the target point and is intraoperatively monitored in real time on the computer display. Thus, our system can be used for intraoperative localization. For instance, in patients with intra-axial space-occupying lesions, which frequently have irregular three-dimensional spatial configurations (15) , the three-dimensional reconstructions allow the surgeons to easily comprehend the three-dimensional surgical orientation. CONCLUSION The stereotactic craniotomy using the present system increases the safety of surgery in the brain by affording preoperative simulation and intraoperative localization with the three-dimensional functional images. However, as in any image-based system of neurosurgical localization (5,7,14,16,23-25,29,30,32), the present system does not overcome the potential limiting factor of postcraniotomy movement of intracranial structures. We are currently developing a computational method of correction for such distortions in this system, which will compare the positions shown on the computer display of certain cortical landmarks, such as the sulci, with those observed intraoperatively. Received, November 29, 1994. Accepted, March 29, 1995. Reprint requests: Nakamasa Hayashi, M.D., Department of Neurosurgery, Faculty of Medicine, Toyama Medical and Pharmaceutical University, Sugitani 2630, Toyama 930-01, Japan. REFERENCES: (1-34) 1. 2. 3. 4. 5. 6. Benzel EC, Lewine JD, Bucholz RD, Orrison WW Jr: Magnetic source imaging: A review of the Magnes system of Biomagnetic Technologies Incorporated. Neurosurgery 33:252-259, 1993. Berger MS, Cohen WA, Ojemann GA: Correlation of motor cortex brain mapping data with magnetic source imaging. J Neurosurg 72:383-387, 1990. Berger MS, Kincaid J, Ojemann GA, Lettich E: Brain mapping techniques to maximize resection, safety, and seizure control in children with brain tumors. Neurosurgery 25:786-792, 1989. Buchner H, Adams L, Knepper A, Ruger R, Laborde G, Gilsbach JM, Ludwig I, Reul J, Scherg M: Preoperative localization of the central sulcus by dipole source analysis of early somatosensory evoked potentials and three-dimensional magnetic resonance imaging. J Neurosurg 80:849-856, 1994. Drake JM, Prudencio J, Holowaka S, Rutka JT, Hoffman HJ, Humphreys RP: Frameless stereotaxy in children. Pediatr Neurosurg 20:152-159, 1994. Gallen CC, Sobel DF, Waltz T, Aung M, Copeland B, Schwartz BJ, Hirschkoff EC, Bloom FE: Noninvasive presurgical Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. of the head (20). DT that incorporates a realistic head model consisting of a homogenous medium was recently reported, but in this model, the high resistance of the cranium was disregarded (9,11,19). We previously reported an accurate determination of the dipole by DT with application in a three-dimensional reconstructed monkey brain, by which localization of electric dipoles in the brain was estimated from potentials recorded on the dura mater under the assumption that the brain is electrically homogenous and is bounded by a highly resistant substance, that is, the cranium (8,20). Furthermore, in this study, we applied a three-dimensional reconstructed model consisting of scalp, cranium, and brain tissue components with relative conductivities in relation to the DT (10). Our intraoperative cortical recordings of SEPs established the validity of this method. Thus, DT seems to provide preoperative data regarding the location of the CS that are comparable to those in magnetoencephalography. We have shown that our system enables surgeons to visualize lesions in their three-dimensional relations to the surrounding brain structures. It delineates the optimal trajectory to the target lesion through the anatomy of the patient as well as the sufficient extent of resection of the lesion without damage to the surrounding healthy brain. Hu et al. (12) showed that three-dimensional views of the brain surface derived from MR images could be used in the preoperative assessment of the relationship between a lesion and anatomic landmarks, such as the pre- and postcentral gyri. However, as in the patient in our study with recurrent astrocytoma, there are some patients in whom the location of the CS cannot be reliably established on the basis of MR images alone because of the distortion resulting from spaceoccupying lesions and perifocal edema (1,17). The advantage of our system is that it provides information regarding the relationship between the sensorimotor cortex and the location of the lesion in combination with DT. Furthermore, the preoperative simulation is performed on the three-dimensional functional image in relation to the cortical surface anatomy via the semitransparent view of the scalp surface of individual patients. Using this system, the surgeon can accurately assess the optimal head position, the craniotomy placement, and the trajectory to the target lesion. The integration of preoperative imaging information into the surgical field has conventionally been an intuitive process on the part of the operating surgeons (25). Therefore, we developed the threedimensional digitizer to integrate the information obtained in the preoperative simulation directly into the surgical field. Using the three-dimensional digitizer on the patient's head in the same manner as using the mouse-controlled cursor, the surgeon can assess the simulated craniotomy on the computer display. This allows the preoperative prediction of the intraoperative view of the brain surface and the lesion after craniotomy. As the arm of the three-dimensional digitizer forms an isocenter system, the beam of the laser is always oriented via the trajectory to the lesion. The position, 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. Y, Ackerman RH, Correia JA, Alpert NM: Dipole-tracing of abnormal slow brain potentials after cerebral stroke: EEG, PET, MRI correlations. Neurosci Lett 112:59-64, 1990. Nishijo H, Hayashi N, Fukuda M, Endo S, Musha T, Ono T: Localization of dipole by boundary element method in threedimensional reconstructed monkey brain. Brain Res Bull 33:225-230, 1994. Orrison WW, Davis LE, Sullivan GW, Mettler FA Jr, Flynn ER: Anatomic localization of cerebral cortical function by magnetoencephalography combined with MR imaging and CT. AJNR Am J Neuroradiol 11:713-716, 1990. Orrison WW Jr, Rose DF, Hart BL, Maclin EL, Sanders JA, Willis BK, Marchand EP, Wood CC, Davis LE: Noninvasive preoperative cortical localization by magnetic source imaging. AJNR Am J Neuroradiol 13:1124-1128, 1992. Reinhardt HF, Meyer H, Amrein E: A computer-assisted device for the intraoperative CT-correlated localization of brain tumors. Eur Surg Res 20:51-58, 1988. Reinhardt HF, Zweiful HJ: Interactive sonaroperated device for stereotactic and open surgery. Stereotact Funct Neurosurg 5455:393-397, 1990. Roberts DW, Strohbehn JW, Hatch JF, Murray W, Kettenberger H: A frameless stereotaxic integration of computerized tomographic imaging and the operating microscope. J Neurosurg 65:545-549, 1986. Sobel DF, Gallen CC, Schwartz BJ, Waltz TA, Copeland B, Yamada S, Hirschkoff EC, Bloom FE: Locating the central sulcus: Comparison of MR anatomic and magnetoencephalographic functional methods. AJNR Am J Neuroradiol 14:915925, 1993. Suk J, Ribary U, Cappell J, Yamamoto T, Llinas R: Anatomical localization revealed by MEG recording of the human somatosensory system. Electroencephalogr Clin Neurophysiol 78:185-196, 1991. Sutherling WW, Crandall PH, Darcey TM, Becker DP, Levesque MF, Barth DS: The magnetic and electric fields agree with intracranial localizations of somatosensory cortex. Neurology 38:1705-1714, 1988. Takizawa T: Isocentric stereotactic threedimensional digitizer for neurosurgery. Stereotact Funct Neurosurg 60:175-193, 1993. Tan KK, Grzeszczuk R, Levin DN, Pelizzari CA, Chen GTY, Erickson RK, Johnson D, Dohrmann GJ: A frameless stereotactic approach to neurosurgical planning based on retrospective patient-image registration: Technical note. J Neurosurg 79:296-303, 1993. Tiede U, Bomans M, Hohne KH, Pommert A, Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. 7. neuromagnetic mapping of somatosensory cortex. Neurosurgery 33:260-268, 1993. Gurthrie BL, Adler JR: Frameless stereotaxy: Computer interactive neurosurgery, in Barrow DL (ed): Perspective in Neurological Surgery. St. Louis, Quality Medical Publishing, 1991, vol 2, pp 1-19. Hayashi N, Nishijo H, Endo S, Fukuda M, Homma S, Ono T: Dipole tracing of monkey somatosensory evoked potentials. Brain Res Bull 33:231-235, 1994. He B, Musha T, Okamoto Y, Homma S, Nakajima Y, Sato T: Electric dipole tracing in the brain by means of the boundary element method and its accuracy. IEEE Trans Biomed Eng 34:406-414, 1987. Homma S, Musha T, Nakajima Y, Okamoto Y, Blom S, Flink R, Hagbarth KE, Mostrom U: Location of electric current sources in the human brain estimated by the dipole tracing method of the scalp-skull-brain (SSB) head model. Electroencephalogr Clin Neurophysiol 91:374-382, 1994. Homma S, Nakajima Y, Musha T, Okamoto Y, He B: Dipole- tracing method applied to human brain potentials. J Neurosci Methods 21:195-200, 1987. Hu XP, Tan KK, Levin DN, Galhotra S, Mullan JF, Hekmatpanah J, Spire JP: Threedimensional magnetic resonance images of the brain: Application to neurosurgical planning. J Neurosurg 72:433-440, 1990. Kamada K, Takeuchi F, Kuriki S, Oshiro O, Houkin K, Abe H: Functional neurosurgical simulation with brain surface magnetic resonance images and magnetoencephalography. Neurosurgery 33:269-273, 1993. Kato A, Yoshimine T, Hayakawa T, Tomita Y, Ikeda T, Mitomo M, Harada K, Mogami H: A frameless, armless navigational system for computer-assisted neurosurgery: Technical note. J Neurosurg 74:845-849, 1991. Kelly PJ, Kall BA, Goerss S: Computer simulation for the stereotactic placement of interstitial radionuclide sources into computed tomography-defined tumor volumes. Neurosurgery 14:442-448, 1984. Laborde G, Gilsbach J, Harders A, Klimek L, Moesges R, Krybus W: Computer assisted localizer for planning of surgery and intraoperative orientation. Acta Neurochir (Wien) 119:166-170, 1992. Morioka T, Yamamoto T, Katsuta T, Fujii K, Fukui M: Presurgical three-dimensional magnetic source imaging of the somatosensory cortex in a patient with a perirolandic lesion: Technical note. Neurosurgery 34:930-934, 1994. Musha T, Homma S: Do optimal dipoles obtained by the dipole tracing method always suggest true source locations? Brain Topogr 3:143-150, 1990. Nakajima Y, Homma S, Musha T, Okamoto 33. 34. COMMENTS Over the past 20 years, we have seen an evolution of stereotactic techniques from radiographically based point-in-space procedures to computer tomography- and magnetic resonance imaging-based point-in-space and, later, volumetric interventions. Surgical planning in the latter depends on the anatomic detail provided by our imaging studies. However, safe approaches to or removal of intra-axial tumors are not really determined by anatomy alone; they relate to our crude knowledge of neurological function inherent in anatomy depicted by our imaging studies. Mapping techniques that establish neurological function for specific anatomic areas have expanded our ability to safely remove intra-axial lesions from important brain areas. Most of these methods are invasive. These include chronic and acute cortical and depth electrode recording and stimulation, cortical stimulation in awake patients undergoing craniotomy, and somatosensory evoked potential phase shifts detected by an array of serial electrodes over the central sulcus. These studies have some risk and can be inconvenient and time-consuming during the surgical procedure. Positron emission tomography and magnetoencephalotomography, both noninvasive methods, have been available for some time. These require complicated and expensive facilities, available at only a few centers. Functional magnetic resonance imaging is now being investigated at some centers and shows promise for noninvasive preoperative mapping, which may be available to most neurosurgeons. The method described in the present article could be made available at a reasonable cost. It may sound complicated, but it is not. It involves recording and amplifying scalp electroencephalography activity from known and reproducible positions on the scalp, performing an averaged somatosensory evoked potential paradigm, and then calculating with a computer program the position of the dipole, which is then translated into stereotactic space. Its use is to Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 preoperatively and noninvasively establish the location of the sensorimotor cortex. This information can then be formatted in a computer-based three dimensional planning matrix that also contains the anatomic information provided by the computed tomography and magnetic resonance imaging database for surgical planning. I think that preoperative noninvasive functional mapping used in the preoperative simulation of intracranial surgical procedures will become an important and widely used concept in future neurosurgery. Patrick J. Kelly New York, New York Hayashi and colleagues describe a preoperative planning and intraoperative localization system with a number of particularly interesting components. The most significant of these is the dipole estimation based upon somatosensory evoked potentials. This has long been recognized as a feasible task that is hampered by the difficulties in adequately modeling tissue heterogeneity and individual variability. The methodology of the authors for generating the model that in turn enables reasonable dipole estimation demonstrates what can be achieved with existing scalp potential recording equipment, a simple measuring helmet, and sophisticated software. It offers a potentially inexpensive alternative to magnetoencephalography for functional somatosensory cortex localization. Quantitative validation of the methodology over a larger number of patients in whom cortical evoked potentials (or other reference standards) are also obtained would have been desirable and will be of great interest as it becomes available. The authors also describe a novel threedimensional digitizer consisting of a three-jointed mechanical arm and an attached laser pointer. The former is designed on an isocenter principle, enabling facile exploration of an infinite number of linear trajectories toward a target. The charged coupled device with which the laser pointer is equipped allows determination of the length of the light beam, thereby letting the laser function as a threedimensional digitizer rather than as just an illuminator of a trajectory. Both the somatosensory dipole modeling and the intraoperative digitizer may be useful as independent tools. The computer graphics of the system, however, serve to integrate both of these tools with preoperative imaging studies and with each other. The graphics workstation and the types of display described are similar to those used at a number of centers involved in computer-based operating systems. The system described in this article illustrates the type of preoperative simulation and intraoperative guidance that is rapidly becoming available and reshaping the process of neurosurgery. David W. Roberts Lebanon, New Hampshire Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. 32. Riemer M, Schiemann T, Schubert R, Lierse W: A computerized three-dimensional atlas of the human skull and brain. AJNR Am J Neuroradiol 14:551-561, 1993. Watanabe E, Watanabe T, Manaka S, Mayanagi Y, Takakura K: Three-dimensional digitizer (Neuronavigator): New equipment for computed tomography-guided stereotaxic surgery. Surg Neurol 27:543-547, 1987. Wood CC, Cohen D, Cuffin BN, Yarita M, Allison T: Electrical sources in human somatosensory cortex: Identification by combined magnetic and potential recording. Science 227:1051-1053, 1985. Yamada S, Gallen CC: Biomagnetic Technologies: Magnetic Source Imaging (MSI)--Magnes Biomagnetometer. Neurosurgery 33:166-168, 1993. Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 Figure 2. Apparatus for stereotactic craniotomy. A, adapter fitted to the external meatus and nasion of the patient. B, stereotactic frame and three-dimensional digitizer. Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. Figure 1. Three-dimensional reconstructed model. A, hemispherical helmet with 64 position sensors (18). The three-dimensional reconstructed model consists of scalp (B), cranium (C), and brain tissue (D) components with relative conductivities (12). Figure 4. Triplanar display. Alternatively, a twodimensional triplanar display can show correlated sagittal, coronal, and axial reconstructions with the mouse-controlled cursor. A trajectory from the insert point on the skin to the target lesion is displayed. Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. Figure 3. Three-dimensional views of the head. A, right profile; B, left profile; C, frontal view; D, head position used in the surgical procedure. The markers placed on the nasion, in the meatus of the bilateral ears, and in the region surrounding the surgical field are shown. Figure 6. Three-dimensional digitizer. The mechanical arm consists of three joints forming an isocenter system. The position illuminated by the laser pointer on the patient's scalp indicates the trajectory to the target point, and this point is monitored on the 3d-CG (arrow). Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. Figure 5. Three-dimensional display. Threedimensional views of the region of interest seen through semitransparent views of the scalp surface and the brain (upper left) are shown. The area outlined with red with the use of the mousecontrolled cursor indicates the simulated craniotomy. A serial plane orthogonal to the direction into the target lesion is also displayed (upper right). Figure 8. Preoperative localization of the CS in a patient with a meningioma. A, two-dimensional surface image. The estimated dipole for N20 of SEP is indicated by a closed circle. B, operative photograph demonstrating the positions of the electrodes used to record the SEPs. The interelectrode distance is 10 mm. C, waveforms of SEPs. Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. Figure 7. Three-dimensional functional images generated in a patient with recurrent astrocytoma in the right parietal lobe. A, normal side; B, abnormal side. The estimated dipole for N20 of SEP is indicated by a blue circle. Figure 10. Three-dimensional functional images in Patient 2. A, normal side; B, abnormal side. A blue circle indicates the estimated dipole for N20 of SEP. Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. Figure 9. Three-dimensional functional images in Patient 1. A, three-dimensional functional image. A blue circle indicates the estimated dipole for N20 of SEP. The area outlined with red using the mousecontrolled cursor indicates the simulated craniotomy. B, intraoperative magnified view. The area enclosed by the dots indicates the boundaries of the tumor delineated by survey on the brain surface using the three-dimensional digitizer. C, waveforms of SEPs. Figure 12. Three-dimensional display. The point illuminated by the laser pointer (red cross-line) is displayed on the three-dimensional view (upper left), a serial plane orthogonal to the trajectory (upper right), and the two-dimensional axis view (lower). Downloaded from https://academic.oup.com/neurosurgery/article-abstract/37/4/694/2758079 by East Carolina University user on 25 March 2018 Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. Figure 11. A, three-dimensional operative view; B, intraoperative magnified view. The arrows indicate the superior temporal sulcus, and the asterisks indicate the intraparietal sulcus. The position of the laser point on the superior temporal sulcus delineates the optimal trajectory to the tumor.