epi_00955.tex Epilepsia, 48(3):531–538, 2007 Blackwell Publishing, Inc. C© 2007 International League Against Epilepsy Distribution of Auditory and Visual Naming Sites in Nonlesional Temporal Lobe Epilepsy Patients and Patients with Space-Occupying Temporal Lobe Lesions ∗Marla J. Hamberger, ‡Shearwood McClelland, III, †Guy M. McKhann, II, ∗Alicia C. Williams, and †Robert R. Goodman ∗Department of Neurology, and †Department of Neurological Surgery, College of Physicians and Surgeons, Columbia University, New York, New York; and ‡Department of Neurosurgery, University of Minnesota Medical School, Minneapolis, Minnesota, U.S.A. Summary: Purpose: Current knowledge regarding the topog- raphy of essential language cortex is based primarily on stimula- tion mapping studies of nonlesional epilepsy patients. We sought to determine whether space-occupying temporal lobe lesions are associated with a similar topography of language sites, as this information would be useful in surgical planning. Methods: We retrospectively compared the topography of au- ditory and visual naming sites in 25 nonlesional temporal lobe epilepsy patients (“nonlesional”) and 18 patients with space- occupying lesions (“lesional”) who underwent cortical language mapping before left temporal resection. Results: Both groups exhibited a similar pattern of auditory naming sites anterior to visual and dual (auditory–visual) nam- ing sites; no group differences were specific to auditory or visual naming sites. However, significantly fewer lesional (10 of 20) compared with nonlesional patients (21 of 25) exhibited any naming sites in the temporal region (p = 0.04). Although the proportion of naming sites on the superior temporal gyrus was similar between groups, naming sites were found on the mid- dle temporal gyrus in 13 of 25 nonlesional patients, yet in only one of 18 lesional patients (p = 0.002). Across groups, patients with visual naming sites were older than patients without visual naming sites identified (p = 0.02). Conclusions: The precise location of essential language cortex cannot be reliably inferred from anatomic landmarks or patient- related variables. As time constraints are a common quandary in stimulation mapping, the different patterns reported here for patients with and without space-occupying lesions can be used to guide mapping strategies, thereby increasing the effi- ciency by which positive naming sites are identified. Key Words: Cortical language mapping—Cortical stimulation— Epilepsy—Naming sites—Temporal lobe. The temporal lobes are a common locus for neu- ropathology requiring surgical intervention, including two thirds of medically intractable partial seizures with demonstrable foci, many vascular malformations and nearly 40% of gliomas (Laws et al., 1986; Semah et al., 1998; Moran et al., 1999; Yeh et al., 2005; Zhao et al., 2005). One challenge of temporal lobe surgery is to re- move a sufficient amount of pathologic tissue, without removing or damaging tissue that is critical for normal lan- guage function. Although the Wernicke area, loosely con- sidered to represent the posterior portion of the superior temporal gyrus, serves as a general landmark for eloquent cortex supporting language, research in language local- ization has shown considerable interindividual variability in the location of essential language cortex (Penfield and Accepted September 23, 2006. Address correspondence and reprint requests to Dr. M.J. Hamberger at The Neurological Institute, 710 West 168th Street, Box 100, New York, NY 10032, U.S.A. E-mail: mh61@columbia.edu doi:10.1111/j.1528-1167.2006.00955.x Roberts, 1959; Ojemann, 1983a). Consequently, surgical resection involving the language-dominant temporal re- gion often requires pre-resection, stimulation-based corti- cal language mapping (Black and Ronner, 1987; Ojemann et al., 1989). Mapping is performed either extraopera- tively, by using permanently implanted electrode grids and strips, or intraoperatively, by using electrodes placed on the cortical surface after craniotomy (Luders et al., 1987; Ojemann, 1990). Cortical sites at which electrical stim- ulation impedes performance on one or more language tasks are considered essential for normal language func- tion. These positive sites (and 1 cm of adjacent cortex) are typically then spared from resection, with the goal of preserving postoperative language abilities (Ojemann and Dodrill, 1981; Hermann et al., 1999). The clinical population composing most stimulation- mapping studies has been nonlesional epilepsy patients, and the task most often used to identify essential language cortex has been visual object naming (Ojemann, 1983b; Hermann and Wyler, 1988). Within the parameters of this 531 532 M. J. HAMBERGER ET AL. population and this particular task, naming sites have been considered to be located in the posterior portion of the temporal region, primarily on the superior temporal gyrus (STG) and middle temporal gyrus (MTG), with some suprasylvian representation. Recent work in stimulation mapping using auditory description naming (e.g., “a pet that purrs”), in addition to visual naming, has shown that auditory naming sites (i.e., sites at which stimulation im- pairs auditory but not visual naming) are generally located anterior to visual naming sites or “dual” sites (i.e., sites at which stimulation disrupts both auditory and visual nam- ing) (Hamberger et al., 2001). Similar to that found for visual naming sites, removal or encroachment on auditory naming sites is associated with postoperative word-finding decline (Hamberger et al., 2005). Given this history of stimulation language mapping, relatively little is known regarding the topographic rep- resentation of visual naming sites in patients with space- occupying lesions, and no published studies address the topography of auditory naming sites in this population. Such information would be useful in guiding surgical plan- ning and mapping strategies, particularly given the time constraints inherent in intraoperative mapping. In this ret- rospective study, we sought to examine and compare the cortical distribution of auditory and visual naming sites in nonlesional temporal lobe epilepsy patients and patients with space-occupying temporal lobe lesions. We hypoth- esized that space-occupying lesions might displace both auditory and visual naming sites outside the lateral tempo- ral region, where language sites have been most frequently identified in nonlesional epilepsy patients. Specifically, we reasoned that lesional patients would have fewer posi- tive naming sites in lateral temporal cortex compared with those demonstrated in nonlesional patients. To address this hypothesis, we retrospectively analyzed the topography of auditory and visual naming sites in 25 nonlesional and 18 lesional patients who underwent cortical language map- ping before left temporal surgical resection. METHODS Subjects A series of 43 consecutive patients who underwent cor- tical language mapping before left temporal surgical re- section and met inclusion criteria was included in this study. Subjects were required to be left hemisphere lan- guage dominant, and either native English speakers or to have learned English before age 5 years. Left hemisphere language dominance was identified by intracarotid amo- barbital testing, functional imaging, and/or intraoperative identification of language sites. Twenty-five of these pa- tients had temporal lobe epilepsy and were categorized as “nonlesional,” in that they had no space-occupying le- sions (13 with medial temporal sclerosis as defined by MRI, 12 with no abnormality on MRI), and 18 patients, categorized as “lesional,” had space-occupying temporal lobe lesions. Lesions consisted of 11 temporal tumors (two oligodendrogliomas, three gangliogliomas, one astrocy- toma, one glioblastoma multiforme, two dysembryoplas- tic neuroepithelial tumors, one mixed oligoastrocytoma, and one low-grade glioma), five cavernous malformations, and two arteriovenous malformations. All patients with nonlesional temporal lobe pathology had medically in- tractable seizures, as did eight of 18 patients with temporal lobe lesions. Twenty-six patients underwent intraoperative language mapping before resection, all at Columbia Uni- versity Medical Center (CUMC). The remaining 17 pa- tients (16 nonlesional) underwent extraoperative language mapping via subdural electrodes, nine at CUMC and seven at New York University Medical Center (NYU). Demo- graphic and clinical information for the two groups was as follows: age at surgery (nonlesional, 35.6, SD, 11.5; lesional, 30.9, SD, 12.6; p = 0.21), years of education (nonlesional, 14.6; SD, 3.0; lesional, 15.0; SD, 3.1; p = 0.69), and age at onset (onset of recurrent seizures or detec- tion of lesion; nonlesional, 16.8 years; SD, 11.3; lesional, 27.8 years; SD, 13.9; p = 0.007). The group difference in onset age is addressed later. Preoperative scores for Ver- bal IQ (VIQ) (Wechsler, 1997), Visual naming (Boston Naming Test, Kaplan et al., 1983), and Auditory naming (Hamberger and Seidel, 2003) were available for 23 non- lesional and 13 lesional patients (see Results). Electrodes For patients evaluated intraoperatively (17 lesional, nine nonlesional), perisylvian sites were stimulated by using a bipolar stimulator with 2-mm-diameter ball con- tacts separated by 5 mm (Ojemann Cortical Stimulator; Radionics, Burlington, MA, U.S.A.). The sites were cho- sen based on gyral/vascular anatomy and spaced <10 mm apart. Electrode positions were documented by using dig- ital photography and schematic diagrams. For patients who underwent extraoperative mapping (one lesional, 16 nonlesional), a 64-contact (i.e., eight-by-eight) grid ar- ray, with 5-mm-diameter electrodes embedded in Silas- tic with center-to-center interelectrode distances of 1 cm (Ad-Tech, Racine, WI, U.S.A.), was positioned over the temporal/perisylvian region (trimmed as needed to con- form to the covered area). The exposed cortical surface and grid position were documented by digital photography and schematic diagrams. Initial schematics were drawn by the surgeon intraoperatively, while looking directly at the brain surface. Digital photographs were then used post- operatively to refine the diagrams. Additionally, subdu- ral electrode positions were verified by skull radiographs postoperatively. The STG and MTG, >3.5 cm to ∼8.5 cm from the temporal pole, were reliably mapped in all 43 patients. This was determined empirically by segment- ing the STG, MTG, inferior temporal gyrus (ITG), and suprasylvian region into centimeter-wide sections from Epilepsia, Vol. 48, No. 3, 2007 TEMPORAL LOBE LANGUAGE CORTEX 533 the temporal pole and comparing the number of sites tested within each section between groups. Although many pa- tients also had anterior temporal (i.e., ≤3.5 cm from the temporal pole) and suprasylvian mapping, only the areas that were comparably mapped in all patients were included in the analyses described later. Mapping procedures All patients performed both auditory and visual naming tasks. All auditory and visual naming stimuli were admin- istered to patients within 1 to 4 months before surgery. Auditory and visual items, selected from previously pub- lished stimuli, were equated for word frequency and were similar in difficulty level (Hamberger and Seidel, 2003). Only items that patients successfully completed at base- line were administered during cortical mapping (items associated with word-retrieval errors at baseline could not be used to identify stimulation-related errors during mapping). For epilepsy patients, mapping was performed while antiepileptic drug (AED) levels were in the thera- peutic range to minimize afterdischarges and seizure ac- tivity. Extraoperative language mapping was conducted after video-EEG monitoring to identify the seizure-onset zone. Testing was conducted during electrical stimulation ap- plied to adjacent electrodes. When results were positive, each electrode was studied individually and referenced to a remote electrode in “silent cortex.” All available sites along lateral temporal cortex, as well as parietal sites in the perisylvian area, were stimulated. Patients who underwent intraoperative mapping were initially anesthetized with propofol. Language mapping began after craniotomy/dural opening, electrocorticogra- phy, and stimulation to determine the threshold for after- discharges. Several practice trials were conducted to en- sure an adequate level of patient responsiveness, and that naming ability was at the level the patient demonstrated at baseline, preoperatively. Stimulation sites were primarily in the vicinity of the anticipated resection, as determined by the presence of a lesion or intracranial EEG evidence of seizure onset. If no visual naming cortex was identified, additional perisylvian sites were tested in an attempt to identify positively the visual naming cortex (rather than relying on negative responses alone). Sites were tested with a bipolar stimulator, as described earlier. Stimulation mapping followed well-established meth- ods (Ojemann, 1983a, 1991). For both intra- and extraop- erative mapping at CUMC, a constant-current stimulator (Ojemann Cortical Stimulator, Radionics, Inc., Burling- ton, MA, U.S.A.) delivered a biphasic square waveform at a frequency of 60 Hz with a 2-ms pulse duration and deliv- ered amperage ranging from 3 to 15 mA during intraopera- tive mapping. Mapping at NYU was conducted by using a Grass Instruments S-12 cortical stimulator (Quincy, MA, U.S.A.) with a biphasic square waveform at frequency 50 Hz with a 0.3-ms pulse duration, with amperage rang- ing from 3 to 15 mA. Afterdischarge levels were deter- mined by increasing amperage until an afterdischarge was elicited, with an upper limit of 15 mA. Amperage for stim- ulation was set at 0.5 to 1 mA below that which elicited an afterdischarge (or 15 mA), which was determined for each site individually. Results reported here are from trials during which no afterdischarges were elicited. At least two trials of both visual and auditory naming were conducted at each site. If results were ambiguous or the patient was temporarily inattentive, additional tri- als were administered. For visual naming, patients were shown line drawings of common items (e.g., bench, heli- copter), and for auditory naming, patients heard oral de- scriptions of concrete items (e.g., “What a king wears on his head”). For visual naming, patients began with the phrase, “This is a” to enable differentiation between speech arrest and anomia, whereas, for auditory naming, patients were instructed to name the target item. To re- duce differences in duration of cortical stimulation across tasks, the auditory stimuli were limited to those that con- tained a maximum of eight words and could be presented clearly within 4 s. Additionally, the requirement for pa- tients during visual naming to articulate the carrier phase (i.e., This is a ———) before naming the pictured object further balanced the stimulus-processing and stimulation- duration times among tasks. For each task, electrical stim- ulation began immediately before presentation of pictures or auditory descriptions and lasted for a maximum of 10 s, but terminated immediately on the patient’s production of a correct response. For both tasks, patients were in- structed to respond as quickly as possible. Sites were con- sidered critical for task performance if the patient could not name target items during stimulation, but provided correct responses on cessation of stimulation. When one of two trials was performed inaccurately, another two tri- als were administered. Sites were considered critical for task performance only when responses to both of these two trials were incorrect. Sites at which this further test- ing resulted in 50% accuracy were not considered critical for task performance. Additionally, if it became apparent during the procedure that naming abilities deteriorated to a level below that demonstrated at baseline, mapping was discontinued. Data analysis For each patient, the location of electrode sites was determined by intraoperative digitized photographs and schematic drawings, and supplemented by postoperative skull radiographs. Naming sites from each patient were plotted on a schematic of the temporal lobe region and coded to indicate whether auditory, visual, or both au- ditory and visual naming were disrupted by stimulation. The topographic distribution and proportion of auditory and visual naming sites in the STG versus the MTG were Epilepsia, Vol. 48, No. 3, 2007 534 M. J. HAMBERGER ET AL. compared in lesional and nonlesional patients by using χ 2 analysis or Fisher’s exact test, depending on cell sizes. Al- though some patients had more extensive mappings that included anterior and inferior temporal cortex, group com- parisons of the topography of naming sites included only the STG and MTG regions that were reliably mapped in all patients. T tests were used for comparisons of de- mographic and patient-related data. Pearson correlations were performed to assess the relation between the number of naming sites identified per patient and patient-related variables. Data from auditory and visual naming were analyzed separately, and subsequently grouped together when no differences were found in the modality-specific results. RESULTS Overall, patients in the nonlesional group were more likely to have naming sites identified relative to patients in the lesional group (Fig. 1). Of the 25 nonlesional patients, at least one positive naming site was found in 21 patients, whereas, of the 18 lesional patients, only 10 patients exhibited any positive naming sites (p = 0.04). No group differences were noted with respect to the num- ber of patients who exhibited auditory (p = 0.63) or visual (p = 0.10) naming sites. Although fewer sites were identified in the lesional compared with the nonlesional group, the spatial patterns of auditory naming sites anterior to visual naming sites was similar and significant in both groups (nonlesional p = 0.001; lesional p = 0.013). However, the superior/inferior distribution of positive naming sites (both auditory and vi- sual), differed. Specifically, 17 of 25 nonlesional patients (25 sites) and 10 of 18 lesional patients (18 sites) had sites identified on the STG (p = 0.40), whereas 13 of 25 nonlesional patients (23 sites) yet only one of 18 lesional patients (one site) had sites identified on the MTG (p = 0.002). Although most lesional patients were tested intraopera- tively and most nonlesional patients were tested extraoper- atively, no group differences were found in the number of sites tested in lesional (mean, 12.4; SD, 6.5) versus nonle- sional patients (mean, 14.2; SD, 5.7; p = 0.35). As noted, analyses of mapping results were limited to the temporal areas that were reliably mapped in both groups. Thus the number of naming sites identified is not merely a func- tion of the number of sites or the size of the region tested. Additionally, no significant differences were seen in num- ber of naming sites identified in patients with (mean, 2.2; SD, 1.9) versus without epilepsy (mean, 1.7; SD, 3.1; p = 0.53).To determine whether the presence of naming sites in the lesional group was related to lesion location, lesion lo- cations were classified as “anterior” if the posterior margin of the lesion was <5 cm from the temporal pole, and as “posterior” if the anterior margin of the lesion was >5 cm from the temporal pole. Of the six patients with anterior lesions, three had positive naming sites, and of the 12 pa- tients with posterior lesions, seven had positive naming sites (p = 1.0). Thus to the extent we were able to quan- tify lesion location, this did not account for the presence or absence of positive naming sites. Analysis of lesion type also showed no consistent relation with presence or absence of naming sites (p = 0.67). As noted earlier, age at onset was significantly earlier in the nonlesional group. Although onset age showed no sys- tematic correlation with number of naming sites identified (r = −0.04; p = 0.82), it should be noted that correlations involving number of naming sites are likely limited by the restricted range. The number of naming sites identified per patient ranged from none to 10, yet with a mean of 2.1 and a median of 2. T tests comparing patients with and without any naming sites identified (likely a more valid means of analyzing these data) showed no significant differences in onset age (p = 0.79), VIQ (p = 0.23), or education level (p = 0.26). However, patients with sites identified were older at the time of surgery (mean age, 36.8 years; SD, 12.1) than were patients with no sites identified (mean age, 25.8; SD, 7.8; p = 0.006). This difference was spe- cific to the presence of visual (not auditory) naming sites (visual naming sites found: mean age, 37.7; SD, 10.7; vi- sual naming sites not found: mean age, 29.2; SD, 12.2; p = 0.02). As noted earlier, age at the time of surgery was comparable between lesional and nonlesional groups (p = 0.23). Preoperative scores on verbal measures were avail- able for 23 nonlesional patients and 13 lesional patients (Table 1). As shown, VIQ and visual naming were sig- nificantly stronger in the lesional group; auditory naming was similar in direction, but the group difference was not significant. Neither VIQ (r = 0.04; p = 0.80) nor Visual naming (r = 0.17; p = 0.33) correlated with the number of auditory or visual naming sites identified. Addition- ally, no differences in verbal scores were noted between patients who did and did not have naming sites identified (VIQ: p = 0.23; Visual naming: p = 0.66; Auditory nam- ing: p = 0.24). Of note, VIQ scores for both lesional and nonlesional patients were well within the average range, whereas visual and auditory naming scores were below the average range in both groups with impaired naming performance in the nonlesional group. TABLE 1. Verbal test scores Lesional Nonlesional p Value VIQ 106.2 (10.0) 94.7 (14.1) 0.021 Visual Naming (BNT) 52.9 (8.6) 46.2 (8.8) 0.040 Auditory Naming 47.6 (3.9) 45.2 (4.4) 0.133 Values expressed as mean (SD). VIQ, verbal IQ; Visual Naming maximum score, 60; Auditory Naming maximum score, 50. Epilepsia, Vol. 48, No. 3, 2007 TEMPORAL LOBE LANGUAGE CORTEX 535 FIG. 1. Topographic distribution of audi- tory (solid circles) and visual (open cir- cles) naming sites in nonlesional epilepsy patients (A) and patients with space- occupying temporal lobe lesions (B). DISCUSSION Previous work in stimulation language mapping de- scribes, primarily, the topography of visual naming sites in nonlesional epilepsy patients. In this study, we sought to elucidate the cortical distribution of both auditory and vi- sual naming sites in both nonlesional epilepsy patients and in patients with space-occupying lesions in the temporal lobe. Consistent with our hypothesis, we found that le- sional patients were less likely to have any naming sites identified in lateral temporal cortex. Additionally, whereas naming sites in nonlesional patients were scattered equally across the STG and MTG, naming sites were rarely found on the MTG in lesional patients, with most naming sites clustered in the posterior portion of the STG. In both groups, however, auditory naming sites were generally anterior to visual naming sites, concordant with previous findings (Hamberger et al., 2001). Epilepsia, Vol. 48, No. 3, 2007 536 M. J. HAMBERGER ET AL. The consistency in the anterior/posterior distribution of auditory and visual naming sites across groups is not surprising, as this maintains the spatial relation between auditory and visual association areas. The lower propor- tion of lesional patients with positive naming sites is con- sistent with results of a previously reported series that included 40 patients with temporal gliomas and 83 nonle- sional epilepsy patients (Haglund et al., 1994). However, unlike our findings, this previous series found a lower per- centage of STG naming sites in the lesional group and a lower proportion of MTG relative to STG naming sites in both groups. Although our use of auditory naming pro- vided more positive sites overall, the use of auditory nam- ing in our study does not appear to account for these differ- ences, as no group differences in topography were specific to auditory versus visual naming. One possible explanation for the overall fewer positive sites in the lesional group is that as a space-occupying lesion expands, existing sites might be altered or elimi- nated (Haglund et al., 1994). It should be noted, however, that stimulation mapping is limited spatially by clinical and patient-related factors (see later) or by the extent of subdural grid coverage. Thus it is possible that in some pa- tients, naming sites may have been located in unexpected areas that were not tested. Additionally, it is infrequent that patients who undergo stimulation mapping require remapping at a subsequent time; however, repeated map- ping would potentially address the question of changes in language site location over time. Although simulation mapping is the standard of care for clinical purposes, func- tional imaging carries the advantage of repeatable test- ing and has been used to demonstrate changes in both intra- and interhemispheric language representation after a stroke (Thulborn et al., 1999). The literature on language reorganization after left hemisphere insult is notable for controversy regarding the nature and location of pathology or injury associ- ated with cortical reorganization. As we limited our study to left hemisphere–dominant patients, we were interested specifically in intrahemispheric language organization in these two groups. Whereas some investigators have con- cluded that epileptogenic tissue does not typically dis- place language cortex (De Vos et al., 1995; Duchowny et al., 1996), others report otherwise (Rausch et al., 1991; Devinsky et al., 1993; Schwartz et al., 1998). In addi- tion to nature and location of insult, patient-related fac- tors such as IQ, age at onset, and educational attainment have been shown to correlate with location and quan- tity of positive naming sites (Ojemann, 1983a; Devinsky et al., 1993). Although both VIQ and age at onset dif- fered significantly between our lesional and nonlesional groups (i.e., higher VIQ and later onset age in lesional group), we found no systematic relation between VIQ or onset age and the presence/absence of positive nam- ing sites, consistent with other reports (Duchowny et al., 1996; Liegeois et al., 2004). The more-superior/posterior representation of naming sites in the lesional group is concordant with theories that the maintenance of left hemisphere dominance in patients with left temporal lesions is accomplished by reorganization of speech to adjacent posterior structures in the left hemisphere (Rasmussen and Milner, 1977) that reach functional mat- uration and commitment later in childhood (Fennell et al., 1977; Satz et al., 1988, 1990). Our findings are also consistent with results of a relatively recent functional imaging study, suggesting that either physical displace- ment or intrahemispheric reorganization of language oc- curs in patients with space-occupying lesions (Stowe et al., 2000). Although it is possible that the development of a space-occupying lesion merely pushes aside normal brain tissue, resulting in fewer sites identified within the region studied, several investigators have reported the presence of functional tissue (including language cortex) within clearly tumor-infiltrated brain regions (Ojemann et al., 1996; Skirboll et al., 1996; Duffau et al., 2004). Thus the relation between functional and abnormal tissue does not appear to be straightforward (Haglund et al., 1994). An alternative explanation to consider, particularly given the lower baseline verbal scores and earlier onset age in the nonlesional group, is that the pattern of positive sites in the lesional group is actually a closer approxima- tion of normal language organization than that observed in the nonlesional group. Accordingly, the location of nam- ing sites in the posterior portion of the STG is consis- tent with the purported location of the Wernicke area in- ferred from studies of stroke patients (Brust et al., 1976; Mazzocchi and Vignolo, 1979), which, although lack- ing in resolution, likely reflects normal language orga- nization. In keeping with this line of thought, the broad distribution of naming sites in nonlesional epilepsy pa- tients might reflect the development of additional naming sites, perhaps in an attempt to compensate for damage to the original language area caused by chronic epileptiform activity. It is also reasonable to take a more theoretic stance and consider what positive naming sites might actually rep- resent. In the current data set, we found that individuals with positive (visual) naming sites were older than in- dividuals without any naming sites identified. Although we tend to infer structure/function relations from posi- tive naming sites, we can state with certainty only that stimulation at a particular location disrupts the task at hand. Thus it is possible that both chronic epilepsy and aging induce changes in cortical function such that nor- mal function is more readily disturbed by stimulation (i.e., increasing the likelihood of finding positive naming sites). This perspective is most consistent with the posi- tion that the lesional group provides a closer representa- tion of “normal” organization than does the nonlesional group. Epilepsia, Vol. 48, No. 3, 2007 TEMPORAL LOBE LANGUAGE CORTEX 537 A limitation of the current study is the relatively re- stricted region of cortex that was compared between groups. Although analysis of a broader region of cortex would have been preferable, practical limitations, such as time constraints, patient fatigue and cooperation, clinical concerns, and IRB regulations, render such data difficult to obtain. The main limitation of stimulation mapping, in gen- eral, is that it is invasive; only pathologic populations can be studied. Thus the patterns identified in the two populations studied here might both represent abnormal variants: one due to chronic, irritative electrophysiologic activity, the other due to structural displacement. In ac- cordance with this, the finding of weak baseline perfor- mance on both visual and auditory naming in both groups suggests that neither group is entirely “normal,” at least from a functional perspective. Perhaps discerning both the consistencies and discrepancies between stimulation mapping data from patient populations and functional imaging data from normal subjects by using similar tasks might help elucidate the normal cortical organization of language. Given the inconsistencies in the literature on language localization after left temporal insult, it is reasonable to conclude that the precise localization of essential language cortex cannot be reliably inferred from anatomic land- marks or from demographic or patient characteristics. The findings reported here demonstrate two general patterns that can be used to guide the search for language cor- tex in patients with and without space-occupying lesions. Stimulation mapping, which remains the gold standard for identifying essential language cortex, is a time-intensive, yet also a time-constrained, procedure. It is hoped that the current results will increase the efficiency by which positive sites are identified. Acknowledgment: We thank Jesse Brand and John Y. Chen for assistance with data management, Dr. Werner Doyle for neu- rosurgical information, Dr. Kenneth Perrine for patient informa- tion, and Drs. William T. Seidel and Frank Gilliam for edito- rial comments. This work was supported by NIH grant R01 NS 35140 (M.J.H., S.M. III). REFERENCES Black P, Ronner S. (1987) Cortical mapping for defining the limits of tumor resection. Neurosurgery 20:914–919. Brust JC, Shafer SQ, Richter RW, Bruun B. (1976) Aphasia in acute stroke. Stroke 7:167–174. De Vos KJ, Wyllie E, Glecker C, Kotogal P, Comair Y. (1995) Lan- guage dominance in patients with early childhood tumors near left hemisphere language areas. Neurology 45:349–356. Devinsky O, Perrine K, Llinas R, Luciano DJ, Dogali M. (1993) Anterior temporal language areas in patients with early onset of TLE. Annals of Neurology 34:727–732. Duchowny M, Jayakar P, Harvey AS, Resnick T, Alvarez L, Dean P, Levin B. (1996) Language cortex representation: effects of develop- mental versus acquired pathology. Annals of Neurology 40:31–38. Duffau H, Velut S, Mitchell MC, Gatignol P, Capelle L. (2004) Intra- operative mapping of the subcortical visual pathways using direct electrical stimulations. Acta Neurochirurgica (Wien) 46:265–269. Fennell EB, Bowers D, Satz P. (1977) Within-modal and cross-modal reliabilities of two laterality tests among left handers. Perceptual Motor Skills 45:451–456. Haglund M, Berger M, Shamseldin M, Lettich E, Ojemann GA. (1994) Cortical localization of temporal lobe language sites in patients with gliomas. Neurosurgery 34:567–576. Hamberger M, Seidel WT. (2003) Auditory and visual naming tests: normative and patient data for accuracy, response time and tip-of- the-tongue. Journal of the International Neuropsychology Society 9:479–489. Hamberger M, Goodman RR, Perrine K, Tammy T. (2001) Anatomical dissociation of auditory and visual naming in the lateral temporal cortex. Neurology 56:56–61. Hamberger M, Seidel WT, McKhann GM, Perrine K, Goodman RR. (2005) Brain stimulation reveals critical auditory naming cortex. Brain 128:2742–2749. Hermann BP, Wyler AR. (1988) Comparative results of dominant tempo- ral lobectomy under general or local anesthesia: language outcome. Journal of Epilepsy 1:127–134. Hermann BP, Perrine K, Chelune GJ, Barr W, Loring DW, Strauss E, Trenerry MR, Westerveldt M. (1999) Visual confrontation naming following left ATL: a comparison of surgical approaches. Neuropsy- chology 13:3–9. Kaplan EF, Goodglass H, Weintraub S. (1983) The Boston Naming Test. 2nd ed. Lea & Febiger, Philadelphia. Laws ER Jr, Taylor WF, Bergstralh EJ, Okazaki H, Clifton MB. (1986) The neurosurgical management of low-grade astrocytoma. Clinical Neurosurgery 33:575–588. Liegeois F, Connelly A, Cross JH, Boyd SF, Gadian DG, Vergha- Khadem F, Baldeweg T. (2004) Language reorganization in children with early-onset lesions of the left hemisphere: an fMRI study. Brain 127:1229–1236. Luders H, Lesser R, Dinner R. (1987) Commentary: chronic intracranial recording and stimulation with intracranial electrodes. In Engel J (Ed) Surgical treatment of the epilepsies. Raven Press, New York, pp. 297–321. Mazzocchi F, Vignolo L. (1979) Localisation of lesions in aphasia: clinical-CT scan correlations in stroke patients. Cortex 15:627– 653. Moran NF, Fish DR, Kitchen N, Shorvon S, Kendall BE, Stevens JM. (1999) Supratentorial cavernous haemangiomas and epilepsy: a re- view of the literature and case series. Journal of Neurology, Neuro- surgery and Psychiatry 66:561–568. Ojemann GA. (1983a) Brain organization for language from the perspec- tive of electrical stimulation mapping. Behavioral Brain Research 6:189–230. Ojemann GA. (1983b) The intrahemispheric organization of human lan- guage derived with electrical stimulation techniques. Trends in Neu- roscience 6:184–189. Ojemann GA. (1990) Organization of language cortex derived from in- vestigations during neurosurgery. Seminars in Neuroscience 2:297– 305. Ojemann GA. (1991) Cortical organization of language. Journal of Neu- roscience 11:2281–2287. Ojemann GA, Dodrill CB. (1981) Predicting postoperative language and memory deficits after dominant hemisphere anterior temporal lobec- tomy by intraoperative stimulation mapping. Conference Proceed- ings, American Association of Neurological Surgeons: Paper 38, pp. 76–77. Ojemann GA, Ojemann J, Lettich E, Berger M. (1989) Cortical lan- guage localization in left-dominant hemisphere: an electrical stimu- lation mapping investigation in 117 patients. Journal of Neurosurgery 71:316–326. Ojemann JG, Miller JW, Silbergeld DL. (1996) Preserved function in brain invaded by tumor. Neurosurgery 39:253–258. Penfield W, Roberts L. (1959) Speech and brain mechanisms. Princeton University Press, Princeton, NJ. Rasmussen T, Milner B. (1977) The role of early left-brain injury in determining lateralization of cerebral speech functions. New York Academy of Science 299:355–379. Epilepsia, Vol. 48, No. 3, 2007 538 M. J. HAMBERGER ET AL. Rausch R, Boone K, Ary CM. (1991) Right-hemisphere language dom- inance in temporal lobe epilepsy: clinical and neuropsychological correlates. Journal of Clinical and Experimental Neuropsychology 13:217–231. Satz P, Strauss E, Whitaker H. (1990) The ontogeny of hemispheric specialization: some old hypotheses revisited. Brain and Language 38:596–614. Satz P, Strauss E, Wada J, Orsini DL. (1988) Some correlates of intra- and interhemispheric speech organization after left focal brain injury. Neuropsychologia 26:345–350. Schwartz TH, Devinsky O, Doyle W, Perrine K. (1998) Preoperative pre- dictors of anterior temporal language areas. Journal of Neurosurgery 89:962–970. Semah F, Picot MC, Adam C, Broglin D, Arzimanoglou A, Bazin B, Cavalcanti D, Baulac M. (1998) Is the underlying cause of epilepsy a major prognostic factor for recurrence? Neurology 51:1256–1262. Skirboll SS, Ojemann GA, Berger MS, Lettich E, Winn HR. (1996) Functional cortex and subcortical white matter located within gliomas. Neurosurgery 38:678–684. Stowe LA, Go IG, Pruim J, den Dunnen W, Meiners LC, Paans AMJ. (2000) Language localization in cases of left temporal lobe arachnoid cysts: evidence against interhemispheric reorganization. Brain and Language 75:347–358. Thulborn KR, Carpenter PA, Just M. (1999) Plasticity of language- related brain function during recovery from stroke. Stroke 30:749– 754. Wechsler D. (1997) Wechsler Adult Intelligence Scale–III manual. The Psychological Corporation, New York. Yeh SA, Ho JT, Lui CC, Huang YJ, Hsiung CY, Huang EY. (2005) Treat- ment outcomes and prognostic factors in patients with supratentorial low-grade gliomas. British Journal of Radiology 78:230–235. Zhao J, Wang S, Li J, Qi W, Sui D, Zhao Y. (2005) Clinical charac- teristics and surgical results of patients with cerebral arteriovenous malformations. Surgical Neurology 63:156–161. Epilepsia, Vol. 48, No. 3, 2007