Cortical language localization in left, dominant hemisphere

An electrical stimulation mapping investigation in 117 patients

Full access

✓ The localization of cortical sites essential for language was assessed by stimulation mapping in the left, dominant hemispheres of 117 patients. Sites were related to language when stimulation at a current below the threshold for afterdischarge evoked repeated statistically significant errors in object naming. The language center was highly localized in many patients to form several mosaics of 1 to 2 sq cm, usually one in the frontal and one or more in the temporoparietal lobe. The area of individual mosaics, and the total area related to language was usually much smaller than the traditional Broca-Wernicke areas. There was substantial individual variability in the exact location of language function, some of which correlated with the patient's sex and verbal intelligence. These features were present for patients as young as 4 years and as old as 80 years, and for those with lesions acquired in early life or adulthood. These findings indicate a need for revision of the classical model of language localization. The combination of discrete localization in individual patients but substantial individual variability between patients also has major clinical implications for cortical resections of the dominant hemisphere, for it means that language cannot be reliably localized on anatomic criteria alone. A maximal resection with minimal risk of postoperative aphasia requires individual localization of language with a technique like stimulation mapping.

Abstract

✓ The localization of cortical sites essential for language was assessed by stimulation mapping in the left, dominant hemispheres of 117 patients. Sites were related to language when stimulation at a current below the threshold for afterdischarge evoked repeated statistically significant errors in object naming. The language center was highly localized in many patients to form several mosaics of 1 to 2 sq cm, usually one in the frontal and one or more in the temporoparietal lobe. The area of individual mosaics, and the total area related to language was usually much smaller than the traditional Broca-Wernicke areas. There was substantial individual variability in the exact location of language function, some of which correlated with the patient's sex and verbal intelligence. These features were present for patients as young as 4 years and as old as 80 years, and for those with lesions acquired in early life or adulthood. These findings indicate a need for revision of the classical model of language localization. The combination of discrete localization in individual patients but substantial individual variability between patients also has major clinical implications for cortical resections of the dominant hemisphere, for it means that language cannot be reliably localized on anatomic criteria alone. A maximal resection with minimal risk of postoperative aphasia requires individual localization of language with a technique like stimulation mapping.

The generally accepted model of language localization in cortex, with a posterior inferior frontal Broca area and a temporoparietal Wernicke area, was developed over a century ago. Originally based on the location of strokes that altered language,3,41 this approach has subsequently been supported by many but not all reports of aphasia with cortical lesions.2,6,20 The investigation of language localization with electrical stimulation mapping during neurosurgical operations under local anesthesia provides a different perspective on this model. Devised by Penfield and Roberts,33 stimulation mapping was based on the observation that applying a current to some cortical sites blocked ongoing object-naming, although no effect of stimulating these sites was reported by the quiet patient. Penfield and Roberts presented the results of stimulation mapping of the right or left hemispheres of a total of 110 patients; those authors reported that sites with evoked errors in naming were located in the rolandic cortex of both hemispheres and in the supplementary motor and classical Broca and Wernicke areas of the left hemisphere. However, a rather large number of left hemisphere sites outside those areas were associated with naming errors. Based on this work, stimulation mapping for language localization became an accepted part of the resective surgical technique for epilepsy32 and has been the subject of a number of subsequent reports.21,24,31,38

The perspective on language localization provided by stimulation mapping differs from that derived from the effect of lesions. With the stimulation technique, language localization can be mapped in an individual subject, limited only by the surgical exposure; lesions generally damage only one cortical region in a patient, thus providing information on function of only a single area in that individual. The brief duration of stimulation makes any functional reorganization during the time it is applied unlikely; some degree of functional recovery usually occurs after lesions. Both stimulation mapping and lesions provide a different perspective on language localization from that derived from neuronal activity recording or blood flow or metabolic measurement techniques (such as positron emission tomography). The latter techniques indicate where neurons participate in language but not whether those neurons are essential for it. Stimulation and lesions indicate only areas essential for language, for the link with behavior is made only when the behavior fails.

In the present report, the cortical localization of language function as determined by stimulation mapping during naming was investigated in the left, dominant hemisphere of 117 patients. This is the largest experience with dominant hemisphere language localization with this technique reported to date. Several issues were investigated. Does the language area in an individual patient actually occupy all of the language cortex of the classical model? What is the variability in this localization between patients? Are there any demographic characteristics that correlate with the variability?

Clinical Material and Methods
Subjects

The study group included 117 patients undergoing left language-dominant frontal or frontotemporoparietal craniotomies. In two children, electrical stimulation mapping was accomplished through chronically implanted subdural electrode arrays.15 In the remaining patients, stimulation mapping was performed during craniotomy under local anesthesia. The 117 patients were selected from a consecutive series of 129 patients undergoing stimulation mapping of the left languagedominant hemisphere as part of resective surgery for medically intractable epilepsy undertaken at the University of Washington Epilepsy Center through January, 1988. Twelve patients were excluded because of previous resections (four cases), inadequate data recording (two cases), preoperative aphasia (two cases), no statistically significant errors on stimulation (two cases), and congenital malformations of rolandic cortex that precluded use of the method for comparing cases (two cases). Thus, no subjects in the present study had previously experienced resection or a preoperative aphasia. All patients had evidence of left brain dominance for language. In 99 patients the evidence was based on language assessment with bilateral intracarotid amobarbital perfusion;40,44 of these, 11 exhibited some language function in the right hemisphere as well. Patients for whom the amobarbital perfusion test showed exclusively or predominantly right language dominance were excluded.

The 117 patients included 56 males and 61 females, ranging in age from 4 to 80 years (mean 30.2 years). Preoperative verbal intelligence quotient (VIQ) data were available for 93 subjects (79%), and ranged from 70 to 126 (mean 99.1). There was no significant difference in VIQ between male and female subjects. In 37 patients the seizure disorder was related to adult-acquired lesions, 31 of which were tumors (most often a low-grade glioma). The sites of stimulation mapping were determined by the surgical exposure: this included both frontal and temporoparietal cortex in 90 subjects, only posterior frontal cortex in 11, and only temporoparietal cortex in 16.

Technique of Intraoperative Stimulation Mapping

Stimulation mapping is a technique for localizing language function within a hemisphere after lateralization has been determined preoperatively by the intracarotid amobarbital perfusion test. In order to insure that sites without evoked naming errors could be resected with a low risk of a postoperative language deficit, stimulation mapping must indicate both where language function is located and where it is not. In this series, the extent of the craniotomy was determined in part by this consideration, covering both the area of the proposed resection and also likely language locations.

Local anesthesia for the craniotomy was induced by a mixture of 0.5% lidocaine and 0.25% Marcaine (bupivacaine hydrochloride) given as a scalp field block and as an intradural injection on each side of the middle meningeal artery. Up to 1 cc of Innovar (droperidol and fentanyl citrate) neuroleptoanalgesia was administered shortly before the operation. No other medications were administered until after stimulation mapping, which was usually performed 3 to 4 hours after the operation was begun. Prior to language mapping, rolandic cortex was identified by stimulation and the threshold for afterdischarge in the electrocorticogram (ECoG) was established for the area of association cortex to be sampled with language mapping. Language mapping used the largest current that did not evoke afterdischarges. For the patients in this study, this ranged from 1.5 to 10 mA, measured between peaks of biphasic square-wave pulses with a total duration of 2.5 msec (1.25 msec for each phase). This was delivered from a constant-current stimulator in 4-second trains at 60 Hz across 1-mm bipolar electrodes separated by 5 mm. Sites for stimulation mapping were randomly selected to cover the exposed cortical surface, including areas where language function was likely to be located as well as the proposed resection. The stimulation sites, usually 10 to 20 per subject, were identified with sterile numbered tickets and the location of these tickets was recorded photographically.

Language function was measured by showing the patient slides of line drawings of objects with common names. These slides were projected at 4-second intervals, with the patient trained to name each one as it appeared. This is an easy task, and there were frequently no naming errors on slides presented in the absence of stimulation, with the highest error rate for the subjects of this study being 22%. While the patient named the slides, sites identified by the numbers were successively stimulated, with the current applied as the slide appeared and continued until the appearance of the next slide. At least one slide without stimulation separated each stimulation, and no site was stimulated twice in succession; usually several slides intervened between each stimulation, and all sites were stimulated once before any site was stimulated a second time. Three samples of stimulation effect at each site were usually obtained. Intraoperative manual scoring of errors and their relation to stimulation provided immediate feedback to the surgeon. In addition, the patient's responses and markers indicating when and where stimulation had occurred were recorded on audio tape and used later to check the results before inclusion of the patient in this study.

Data Analysis

The accuracy of naming during the large number of slides presented in the absence of stimulation was compared to the accuracy of naming during stimulation of a given site, using the single sample binomial test.36 A site was determined to be related to language function if the chance probability of errors evoked at that site was less than 0.05. With the low error rates in the absence of stimulation, evoking errors during two of the three stimulations at a site often achieved that level of statistical significance. The location of sites of stimulation in relation to rolandic cortex, the sylvian fissure, and sulci separating the major gyri was determined from the intraoperative photographs. Figures 1 to 7 and 9 to 11 are examples of the maps thus obtained for individual patients.

Fig. 1.
Fig. 1.

Sites essential for naming (filled circles) in a 24-year-old woman with a verbal IQ of 81. Stimulation at 6 mA; control error rate in the absence of stimulation was 3.7%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by a small dot. M and S identify sites with motor (M) or sensory (S) responses. Note the localized posterior language area with closely spaced surrounding stimulation sites without errors.

Fig. 7.
Fig. 7.

Sites essential for naming (filled circles) in a 46-year-old woman with a verbal IQ of 91. Stimulation at 7 mA; control error rate in the absence of stimulation was 6%. Open circles indicate stimulation sites without evoked errors. Note the relatively large posterior language area, but very localized anterior language site, and a language site in the anterior superior temporal gyrus (arrow) in front of rolandic cortex, 4 cm from the temporal tip. M and S indicate sites with motor (M) or sensory (S) responses.

Fig. 9.
Fig. 9.

Sites of significant evoked naming errors (filled circles) in a 4-year-old boy (upper) and a 70-year-old man (lower), both with medial temporal lobe gliomas. Open circles indicate stimulation sites without evoked errors. The 4-year-old boy was stimulated through a chronic subdural grid. Note that both patients show very localized temporal language sites. M and S indicate sites with motor (M) or sensory (S) responses.

Fig. 11.
Fig. 11.

Large dominant hemisphere temporoparietal resections of much of the classical Wernicke area in two patients, with no postoperative worsening of language function. The resections spared the sites of repeated evoked naming errors. Both patients had only left-sided speech areas based on preoperative intracarotid amobarbital perfusion testing. Filled circles indicate sites essential for naming; open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by a small dot. M and S indicate sites with motor (M) or sensory (S) responses. Upper: This patient with superior temporal gyms oligodendroglioma and intractable seizures experienced no language change after the resection delineated by the shaded area. The patient returned to teaching postoperatively. Lower: This patient with widespread lateral temporal epileptic focus had no language deficits preoperatively. Following the resection indicated by the shaded area, oral language returned to normal within a week, although reading remained slow for a longer period.

The variability in the location of sites with significant evoked changes in naming was determined by aligning the individual patient maps to rolandic cortex and sylvian fissure. Sites were then assigned to a zone defined by an arbitrary grid based on the same landmarks, as illustrated in Fig. 8. For frontal cortex, that grid included 1.5-cm segments in each of the inferior, middle and superior frontal gyri, beginning with the most anterior evoked motor response identifying the anterior limit of motor cortex. In addition, the inferior frontal gyrus was divided into superior and inferior zones by a line 1.5 cm above and parallel to the sylvian fissure. For temporoparietal cortex, that grid was based on a line extending from the posterior end of the sylvian fissure to the projection of the foot of the central sulcus onto that fissure. This line was subdivided into fourths, and zones were established in the superior, middle, and inferior temporal gyri and parietal operculum by these lines. Divisions of the retrosylvian portion of these gyri are also indicated in Fig. 8. The number of patients with sites within each of the zones was thus defined, and the percentage of those sites with significant evoked anomia was determined.

Fig. 8.
Fig. 8.

Variability in language localization in 117 patients. Individual maps are aligned as described in the text, and cortex is divided into zones identified by dashed lines. Upper number in each zone is the number of patients with a site in that zone; lower number in circle is the percentage of those patients with sites of significant evoked naming errors in that zone. M and S indicate motor (M) or sensory (S) cortex.

In order to compare patient characteristics with the variability in localization, zones were combined into five groups: superior or middle temporal gyri, inferior parietal lobe, inferior frontal gyrus, or elsewhere in frontal cortex. The proportion of patients with or without errors in each of these groups was compared in terms of age, preoperative VIQ, presence or absence of a lesion with adult onset, or presence of right hemisphere speech. Fisher's exact, chi-square, or Mann-Whitney U statistical tests were used.36 A set of criteria to identify the extent of the cortical surface area with evoked anomia in each subject was also established. By these criteria, areas with evoked anomia were subdivided into: areas with clearly defined boundaries separated from sites without naming errors by 1 cm or less; areas that were of the same dimension but with mapping that did not absolutely define one boundary; and areas where contiguous sites with anomia extended over at least 2.5 cm. The study determined both the total area of mapping sites with evoked errors and areas of individual sites separated from others by sites with no stimulation effect on language.

Results
Language Localization in Individual Patients

In the most common pattern of individual localization encountered in the 117 patients, sites where stimulation evoked errors were separated by less than 1 cm in all directions from sites without errors. This discrete localization is shown for temporoparietal sites in Figs. 1 and 9, and for frontal sites in Fig. 7. Other examples have been presented elsewhere.21,22 Within the limits of the surgical exposure and the number of sites sampled, in only 39% of the patients were errors evoked in any uninterrupted area of cortex larger than 1.5 cm in any dimension (approximately 2.25 sq cm). In only 14% was there an uninterrupted area of language cortex greater than 2.5 cm in the smallest dimension (but see Fig. 4). The populations with large or small extents of individual language areas could not be distinguished based on sex, preoperative VIQ, age, or levels of stimulation current. The margins of these areas were very sharp in many patients (Figs. 1, 3, 7, and 9); in others they were surrounded by cortex where single naming errors were evoked (Figs. 2, 5, and 6), suggesting a gradation between cortex with no role in naming to cortex that is essential for it. In an individual, then, essential areas for language were often organized in a mosaic pattern, 1 to 2 sq cm in extent.

Fig. 2.
Fig. 2.

Sites essential for naming (filled circles) in a 24-year-old woman with a verbal IQ of 94. Stimulation at 2 mA; control error rate in the absence of stimulation was 1.2%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by a small dot. The posterior language area (slightly larger than in the patient illustrated in Fig. 1) is oriented transversely to the superior temporal gyrus. Note also that the intensity of the stimulating current does not determine the size of language area (compare to Fig. 1). M indicates sites with motor response.

Fig. 3.
Fig. 3.

Sites essential for naming (filled circles) in an 18-year-old woman with a verbal IQ of 95. Stimulation at 4 mA; control error rate in the absence of stimulation was 0%. Open circles indicate stimulation sites without evoked errors. This patient was left-handed, but had only left language according to the intracarotid amobarbital perfusion test. Language in this unusual parietal location imposes no limits on a temporal resection. M and S indicate sites with motor (M) or sensory

Fig. 4.
Fig. 4.

Sites essential for naming (filled circles) in a 37-year-old woman with a verbal IQ of 99. Stimulation at 6 mA; control error rate in the absence of stimulation was 0%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by small dots. No posterior language sites were identified despite extensive mapping. A posterior temporal resection (dashed line) was associated with no language changes, even acutely. M and S indicate sites with motor (M) or sensory (S) responses.

Fig. 5.
Fig. 5.

Sites essential for naming (filled circles) in an 18-year-old man with a verbal IQ of 91. Stimulation at 5 mA; control error rate in the absence of stimulation was 1.7%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by a small dot. No changes in naming or counting were evoked at frontal sites. M and S indicate sites with motor (M) or sensory (S) responses.

Fig. 6.
Fig. 6.

Sites essential for naming (filled circles) in a 20-year-old man with a verbal IQ of 91. Stimulation at 5 mA; control error rate in the absence of stimulation was 0%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by small dots. Inferior frontal language sites extend nearly to the pterion. The frontal resection came within 1 cm of the anterior language area, and was followed by a significant expression aphasia lasting several weeks. M and S indicate sites with motor (M) or sensory (S) responses.

In 67% of patients two or more such mosaics separated by “nonlanguage” cortex were identified (Figs. 5, 6, 7, 10, and 11); in 24% three or more were identified. Almost all of the patients with multiple mosaics had at least one mosaic essential for language function in frontal cortex and another in temporoparietal cortex. Several separate mosaics in temporoparietal cortex were relatively common (Figs. 5, 7, 10, and 11); more rarely, several separate frontal mosaics were identified (Figs. 6 and 11).

Fig. 10.
Fig. 10.

Site of evoked naming errors (filled circles) in a 45-year-old man following removal of a parietal glioma (shaded area) that exposed the planum temporale. Note the localized site of evoked naming errors on the planum (arrow), adjacent to a similar superior temporal gyrus surface site. Stimulation at 4 mA; control error rate in the absence of stimulation was 0.9%. Open circles indicate stimulation sites without evoked errors. No naming errors were evoked from stimulation of cortex overlying the tumor and no language disturbance followed the resection. M and S indicate sites with motor (M) or sensory (S) responses.

When the total area of such mosaics in cortex available for mapping was estimated for each patient, 50% of the patients had total areas of 2.5 cm or less (Figs. 1, 3, and 9); only 16% had areas of 6 sq cm or larger (Fig. 7). Again, no statistically significant difference between patients with large or small total areas was identified, although patients with larger areas had lower average preoperative VIQ's than those with small areas (93.6 vs. 99.9).

Variability Between Patients

Although often highly localized in a given patient, sites essential for naming had exceedingly diverse anatomic locations throughout the population; these sites were not found uniformly in any given cortical region. Figure 4 illustrates mapping in a patient with a large frontal language area but no posterior sites. Neither extensive mapping nor resection of posterior temporal cortex in this patient had any effect on language. The reverse situation is seen in the patient whose responses are illustrated in Fig. 5, where naming changes were not evoked at frontal sites yet mapping at the same current evoked errors at multiple posterior sites. Figure 3 illustrates an even more unusual situation in which errors were evoked only in the parietal lobe, again in an area not exceeding 1 to 1.5 cm in any of the directions mapped; there were no errors in temporal or frontal sites, the areas where they would be expected given classical models of language localization. Naming sites were identified only in the frontal lobe in 17% and only in the temporoparietal lobe in 15% of the 90 patients with both frontal and temporoparietal mapping.

Figure 8 summarizes the variability in language localization in the 117 patients. Language sites were not uncommonly identified more anterior in the temporal or parietal lobe than might be predicted by the conventional extent of Wernicke's area. They were found far beyond the traditional boundaries of Broca's area. Equally variable was the presence of naming sites in the traditional language areas. No one zone in posterior cortex had errors evoked in more than 36% of patients with mapping there. Indeed, with the sole exception of inferior posterior frontal cortex, no zone exceeded 50% in the proportion of cases with significant naming sites. Even the 21% of patients with stimulation sites in posterior inferior frontal cortex did not have significant evoked naming errors. Thus, neither the location nor absence of language function at a given cortical site can be reliably predicted by anatomical considerations.

Patient Characteristics Correlating With Language Localization

Various characteristics of the population were somewhat predictive of language organization. Although the sex of the patient was not itself significant, preoperative VIQ and sex in some combinations corresponded to specific organizations of naming sites (Table 1). In the population of patients with a VIQ below the mean score, males were significantly more likely than females to have parietal language sites, an effect that was not evident in the half of the population with higher VIQ's. There was a suggestion of a similar effect for frontal zones outside the posterior inferior portion of the inferior frontal gyrus, but this difference did not reach statistical significance.

TABLE 1
TABLE 1

Sex and VIQ differences in language localization*

Significant differences in VIQ's for patients with or without evoked naming errors were present for males with stimulation sites in the superior temporal gyrus (Table 1). In that group, patients with evoked errors had significantly lower VIQ's than those without errors. Suggestive differences showing higher VIQ's for patients of either sex with errors were present for stimulation at sites in inferior frontal cortex outside the posterior inferior frontal gyrus and at sites in the middle temporal gyrus. Comparing patterns of language localization in all patients with the highest VIQ's (≥ 110, 17 cases) to those with the lowest VIQ's (≤ 90, 21 cases) demonstrated a significantly greater proportion of superior temporal gyrus sites with errors in the low-VIQ group (81% errors in the low-VIQ group vs. 41% errors in the high-VIQ group, p < 0.05). Thus, the presence of evoked naming errors in the superior temporal gyrus seems to be associated with poor verbal abilities as measured by the VIQ, particularly in males. Sites in the middle temporal and frontal gyri may be associated with facile verbal abilities.

Several subgroups of language localization showed some correlation with sex and VIQ as well. Males with only temporoparietal and no frontal language sites had significantly lower VIQ's than males with both frontal and temporoparietal language sites (five males with only temporoparietal sites had a mean VIQ of 92; 13 males with both frontal and temporoparietal sites, both of which were mapped, had a mean VIQ of 101, p < 0.05). Females were suggestively but not significantly overrepresented among the patients with only frontal language sites (11 of 15 of that group were female, 0.05 < p < 0.1).

The patterns of language localization in the 37 patients with lesions acquired in adulthood were compared to those in the 80 patients in whom the condition giving rise to epilepsy might have occurred perinatally. No significant differences were identified in the proportion of errors in frontal, parietal, superior, or middle temporal areas, even when the groups were matched by VIQ and sex. There was a single suggestive but not significant difference. Males with acquired lesions had more errors evoked by stimulation in the inferior frontal lobe outside the classical Broca's area (60% errors in 15 males with acquired lesions vs. 28% errors in 25 with possibly early-life lesions (0.05 < p < 0.1)).

Age was not a predictive factor with regard to language organization. No significant differences existed in location or extent of essential language areas between older and younger patients. Figure 9 illustrates this finding in a 4-year-old boy and a 70-year-old man, both showing highly localized temporal naming sites.

Discussion

Application of an electric current to the cortical surface has both excitatory and inhibitory effects on neuronal populations and en passage fibers, both locally and at a distance.34 Empirically, language responses seem to represent a predominance of inhibitory effects, most likely from temporary inactivation of local populations of neurons by depolarizing blockade. The spatial extent of this blockade is not precisely known. One study using the extent of stimulation-evoked fluorescence with nicotinamide adenine dinucleotide, reduced form, suggested that the extent varied with repeated application of current to the same site.39 However, the frequent uniform behavioral effect of repeated stimulation and the close proximity of sites with and without errors in many patients observed in the present study suggest that the evoked inactivation has little variation between stimulations and remains quite localized.

A second issue in interpreting stimulation effects is whether the behavioral change evoked by this local cortical inactivation represents an effect on the language process of naming, or on other processes such as perception or motor output system. Several observations suggest that usually the language process has been altered. Sites with evoked naming errors were not part of primary motor or sensory cortex, as indicated by the lack of evoked movements or sensations. Slides to be named carried a leader phrase, “this is a —.” Many patients successfully recited the phrase even when naming was incorrect, indicating that both perception and speech production were intact and implying that it was only the language process, naming, that was interrupted. In general, naming errors consisted of omission. Although nonsense words, jargon, and other errors occurred, they were rarely peculiar to one site. Thus, attempts at analyzing localization of different types of naming errors have generally proved to be nonproductive. When the patient did not produce the leader phrase, it was not possible to distinguish by means of naming tasks alone between failures in perception, naming, motor control, or even consciousness. However, in a subset of the present series, the stimulation effect on a memory measure was also assessed, and this was usually intact at sites where naming omissions occurred, indicating that consciousness, perception, and motor output were intact.21 This procedure did not distinguish between an arrest of speech during naming and arrest of motor activity; indeed, at frontal premotor sites, such a distinction may be inappropriate. Lueders, et al.,17 have suggested that posterior frontal premotor sites of speech arrest are really sites of “negative motor response,” with arrest of hand movements as well as speech. Nevertheless, these sites are essential for language, and a motor aphasia often occurs if they are excised; this result is one of several lines of evidence for an intimate relationship between language and motor function in dominant hemisphere.12,16,21

There is evidence to support a correlation between sites with evoked naming errors and sites where lesions are likely to impair language. Penfield and Roberts33 inferred this relationship. Direct evidence for anterior temporal resections was provided by Ojemann and Dodrill.23 In that study a resection that came within 2 cm of a site of significant stimulation-evoked anomia was associated with a subtle but definite general language deficit observed on an aphasia battery administered 1 month after the operation. This deficit was not present when the resection avoided such sites by a wider margin, and could not be accounted for by the size of resection or other patient characteristics. Moreover, naming deficits are associated with all types of aphasias, suggesting that naming changes are likely to arise from manipulation of all cortex essential for language.14 Thus, sites identified as essential for naming by stimulation mapping probably represent essential areas for language, even when assessed by lesion criteria.

In addition to defining the extent of cortical inactivation and the meaning of the naming errors, interpretation of these data requires addressing the type of population examined — one that was necessarily (by virtue of requiring craniotomies) neurologically abnormal. One might expect cortical reorganization or malformation to be related to lesions appearing in early life. However, no differences were identified between patients with lesions of early or adult onset. Thus, the variations seen in this study do not seem to be related to abnormal early development.

The pattern of organization of language in association cortex derived from this study includes localized essential areas (often as small as 1 sq cm), in some cases with very sharp boundaries and in others with a small area of surrounding cortex where single errors were evoked suggesting a more graded border to the area of repeated errors.42 This type of organization would seem most compatible with modules or mosaics of association cortex devoted to language. Each patient usually has several of these, most often at least one frontally and one or more temporoparietally. A minority of patients, however, seem to have only frontal or only temporoparietal language modules. The more common pattern with distributed module-mosaics has become a standard model of cortical organization in animals.9 Based on the finding here, human association cortex is also organized in distributed modules, but on an expanded scale compared to the primate. This localized mosaic pattern was present in the youngest patient investigated (a 4-year-old boy) at a time when language was being acquired.

In some patients the language function mosaics seemed to be oriented transversely to the axis of the gyrus (Fig. 2). Such an alignment suggests that areas essential for language reflect the pattern of connections of these cortical areas; in animals, efferent fibers to temporal association cortex are often oriented in bands perpendicular to the gyral axis (O Creutzfeldt, personal communication). How far the mosaics essential for language extend into the sulci is not entirely known, but several observations suggest that any extension is rather limited. If mosaics essential for language occurred exclusively in buried cortex or extended far from surface sites, then the surface sites would not be the good markers that they are for planning resections extending along a gyrus to avoid language deficits. Moreover, on a few occasions the extent of essential areas for language has been mapped in sulci. In one case (illustrated in Fig. 10), an essential area in the planum temporale, buried in the sylvian fissure, was an extension of a superior temporal surface site. Similarly, in another case, a posterior inferior frontal language area extension into buried frontal opercular cortex was identified.31 No language changes were evoked from the sulcus immediately adjacent to a frontal language site in a third patient (G Ojemann, unpublished observation, Case 8844). To date, no buried sites not represented by adjacent surface sites have been identified.

The mosaics essential for language can also be distinguished from surrounding cortex by physiological changes identified during naming. However, changes in neuronal activity per se do not provide this distinction: recordings of neuronal activity in the human temporal lobe during language have identified changes in activity correlated with naming at sites clearly not essential for that function, based on both stimulation mapping and the effects of excision.5,27 The patterns of those changes in neuronal activity related to language were usually tonic shifts in the level of neuronal firing, most suggestive of mechanisms of selective attention. Because of the invasive nature of microelectrode recording, neuronal activity in essential areas for language has not been investigated. The physiological changes that distinguish essential areas for language from surrounding cortex during naming were identified in the ECoG. During naming, temporoparietal sites essential for language were differentiated from surrounding cortex by local desynchronization,7,29 and frontal sites essential for language were distinguished from surrounding cortex by a slow evoked potential.7 The ECoG changes that differentiate frontal and temporoparietal sites essential for language from surrounding cortex during naming occur simultaneously.7 Thus, frontal and posterior language sites function at least partly in parallel. Physiological evidence for serial processing from posterior to anterior language areas has yet to be obtained. Parallel processing of the separated mosaics has also been a feature of the organization of animal association cortex.9 One mechanism for both a local cortical desynchronization and a frontal slow potential in animals is activity of the thalamocortical activating system,11,37 producing selective attention. A lateral thalamic role in language, also thought to be related to mechanisms of activation-attention, has been described in studies of left thalamic stimulation.25 During language, the mosaics essential for language are likely selected in parallel by a thalamocortical activating system producing intense local selective attention.

What has not been appreciated previously is the marked individual variability in the location of the mosaics essential for language. However, previous investigations of language localization with intraoperative stimulation mapping have shown sites in unexpected locations: for example, in midfrontal or midparietal,33 and posterior temporoparietal locations.38 In addition, stimulation mapping of language sites through chronic subdural electrode arrays has also demonstrated variation in the exact location of language.15,18 Substantial variability has been evident in the University of Washington series, even when it consisted of only eight cases.24,31 This variability is quite marked in all areas related to language except the most posterior portion of the inferior frontal gyrus. Even there, enough variability is present so that the classical Broca's area is occasionally not involved in language. This variability is probably the explanation for the difficulty in determining the exact location of the Wernicke language area from the extent of temporoparietal lesions producing aphasia.2 Indeed, in each zone within the classical Wernicke area, sites related to language were present in less than 40% of the patients who had language assessed there. Moreover, this variability in location of posterior language areas is also evident when sites are combined over larger areas: only 65% of patients with sites anywhere in the superior temporal gyrus showed evoked naming errors, with lower percentages for those with middle temporal gyrus or parietal lobe sites (Table 1). The Wernicke language area of the classical model is clearly an artifact of combining the locations of these essential areas in different patients, for rarely if ever are essential language areas covering the entire classical Wernicke area found in an individual patient. Indeed, the entire extent of the classical Broca and Wernicke language areas is seldom essential for language in an individual patient. In the maps presented by Penfield and Roberts,33 those combined areas cover over 8 sq cm — an extent encountered in less than one-sixth of our patients. In over half of our patients the identified extent of essential language cortex was less than one-quarter of the area delineated by Penfield and Roberts.

This variability in language localization outside the inferior frontal gyrus exceeds even the considerable variability in gross anatomy of the human perisylvian gyri.35,43 Considered over the entire population, essential areas for language do not correspond to any described cytoarchitectonic areas of cortex. The individual variability in cytoarchitectonic areas has been little investigated, although one study of the “TP” area in the planum temporale suggested a moderate degree of variability in extent;8 however, this degree of variation is also not nearly enough to account for the observed variability in language localization. Individual variability in gross localization of functions in animal cortex has received little attention, but considerable variability in the extent of monkey rolandic cortex, some of which was related to previous sensorimotor experiences, has been reported.19

There is little evidence in the present study to suggest that experience changes the localization of areas essential for language. The pattern in the patient with the least experience with language, at the age of 4 years, was similar to that in patients with the longest experience, at the ages of 70 and 80 years. In all of these cases the essential areas were localized to mosaics of 1 sq cm. In an earlier study in which the extent of areas essential for naming in different languages was investigated in bilingual patients, the areas essential for the language in which the subjects were least competent were larger than those for the language in which they had greater competence, although the actual items named had similar error rates in the two languages.30 This suggested that essential areas might become smaller with increasing facility with a language. That suggestion received only slight corroboration in the present study, from the lower VIQ's in patients with larger total language area.

Some of the variation in language localization reflects differences based on language ability and sex. Differences between males and females in proportion of naming errors evoked from parietal and frontal sites (outside Broca's area) were noted in an analysis of the first 21 cases of the present series.21 In the present much larger series, the parietal lobe seems to be the location of the major difference based on sex, with effects evident only in the low-VIQ group; males were more likely to have naming errors evoked from that area. In addition, there was suggestive evidence that females were overrepresented in the small subgroup of patients with only frontal language sites. Kimura13 suggested that language deficits after temporoparietal strokes were less likely in females. From this hypothesis, she suggested that the posterior language area differed according to sex, with those language functions located in the left parietal lobe of males subserved by left frontal lobe in females. The data from the present study provide some support for that view.

Differences in language organization based on VIQ seemed to involve primarily the temporal lobe. In the present study, naming errors evoked from the superior temporal gyrus were significantly more likely in subjects with low VIQ's. Subjects with high VIQ's were suggestively more likely to have errors evoked from the middle temporal gyrus. An even clearer relationship between patterns of temporal lobe language organization and VIQ was evident in an investigation of stimulation effects on both naming and simple-sentence reading in the 55 patients in the present series in whom both had been tested.26 In that study, the presence of superior temporal sites where only naming errors were evoked and middle temporal sites with only reading errors was significantly associated with low VIQ's, while the reverse pattern (sites with only reading errors in the superior gyrus and sites with only naming errors in the middle gyrus) was associated with high VIQ's. Sites where both naming and reading errors were evoked did not differ based on VIQ. This probably explains why the present study showed a somewhat less robust relationship between language location and VIQ for, when only naming was mapped, sites with errors included those related to naming alone (related to VIQ) and to more general language function including reading (not related to VIQ). Stimulation mapping studies, then, clearly indicate that the biological substrate for language differs in patients with differing verbal abilities.

The clinical implications of the present study derive from the combination of discrete localization of essential areas for language in individual patients, and the marked variability between patients. Figure 8 indicates the probability of encountering essential areas for language in various portions of perisylvian cortex. As is evident from that figure, no anatomic landmark reliably indicates the presence or absence of language, outside the posterior inferior frontal lobe. Anterior temporal resections in front of rolandic cortex and sparing superior temporal gyrus are only relatively safe: 5% of our sample had essential areas for language in that cortex. The probability of encroaching on such essential areas increases substantially if the superior temporal gyrus anterior to rolandic cortex is included in the resection. Figure 7 illustrates a case with language function in this area. Damage to those language areas probably accounts for cases of persisting aphasia after anterior temporal lobectomy.10 In the frontal lobe, sites essential for language occasionally extend forward to the pterion (Fig. 6). On the other hand, the marked individual variability means that in some patients cortex that is considered essential for language based on anatomic criteria can be safely resected without risk of aphasia: for example, in the patients whose language areas are illustrated in Figs. 3, 4, 10, and 11. These findings indicate that, if the maximum resections are to be undertaken with maximum safety to language areas, then the location of language function should be established in that individual patient with a technique like stimulation mapping.

The question for the surgeon becomes whether lateral cortical surface stimulation mapping with naming is reliable in indicating cortical areas that are and are not essential for language. The evidence that sites of stimulation-evoked errors identify the cortex that must be preserved to prevent aphasia after an anterior temporal resection has already been presented. However, to provide this type of information, stimulation mapping must indicate both where language is located and where it is not. Only when both are known can cortex without evoked naming errors be safely resected. Entirely negative stimulation mapping does not provide the information needed to plan a resection. The choice of stimulation currents is particularly important in this regard, for too low a current may not adequately block local cortical function, while too large a current is likely to evoke seizures. Thus, determining the afterdischarge threshold on ECoG and using a current just below that threshold are important parts of accurately identifying language cortex with stimulation mapping.

Although lateral cortical surface stimulation seems to provide the information needed to plan a resection, there is some evidence that it does not identify all of the cortex essential for language even in the temporal lobe. Sites of repeated stimulation-evoked language errors have been identified in the fusiform gyrus on the undersurface of the temporal lobe.18 Stimulation mapping during reading of simple sentences and naming in second languages has provided some evidence that sites essential for naming in the primary language do not define all cortex essential for all language functions.21,26,30 Nor does the location of sites essential for naming indicate cortex important for recent verbal memory function.28 Thus, when resections are planned very close to language areas, particularly in patients dependent on facile function in these other language-related functions, additional stimulation mapping with reading or memory assessment may be desirable.

For most patients however, localizing language with lateral cortical surface stimulation mapping during naming provides the information needed to plan a cortical resection, providing that there is a margin of about 2 cm around identified sites, at least along a contiguous gyrus. Even with this margin, knowing where essential areas for naming are located often allowed for resections in classical language areas of the dominant hemisphere without postoperative aphasias (Figs. 4, 10, and 11). Although originally developed to allow maximum resection of epileptic foci in the dominant hemisphere with a minimum risk of subsequent language disturbance, the technique is equally useful for safely maximizing the extent of other cortical resections near language cortex, especially those for low-grade gliomas1 and arteriovenous malformations.4

Acknowledgments

Dr. Carl Dodrill provided intracarotid amobarbital and VIQ data. Drs. H. Whitaker, C. Mateer, I. Fried, T. Sandquist, and D. Cawthon assisted in patient testing. The operations on 18 patients seen early in the study were performed by Dr. A. A. Ward, Jr., or Dr. A. Wyler.

References

  • 1.

    Berger MSOjemann GA: Cortical mapping techniques used to maximize tumor resection and safety in children with brain tumors. Ann Neurol 24:3611988 (Abstract)Ann Neurol 24:

  • 2.

    Bogen JBogen G: Wernicke's region — where is it? Ann NY Acad Sci 280:8348431975Ann NY Acad Sci 280:

  • 3.

    Broca P: Remarques sur le siège de la faculté du language articulé, suivies d'une observation d'aphémie (perte de la parole). Bull Soc Anat 36:3303571861Bull Soc Anat 36:

  • 4.

    Burchiel KJClark HOjemann GAet al: Use of stimulation mapping and corticography in the excision of arteriovenous malformations in sensorimotor and language-related neocortex. Neurosurgery 24:3233271989Neurosurgery 24:

  • 5.

    Cawthon DLettich EOjemann G: Human temporal lobe neuronal activity: inhibition during naming in only one of two languages. Soc Neurosci Abst 13:8391987Soc Neurosci Abst 13:

  • 6.

    Damasio A: Concluding remarks: neuroscience and cognitive science in the study of language and the brainPlum F (ed): New York: Raven Press1988275282

  • 7.

    Fried IOjemann GAFetz EE: Language-related potentials specific to human language cortex. Science 212:3533561981Science 212:

  • 8.

    Galaburda AMSanides FGeschwind N: Human brain. Cytoarchitectonic left-right asymmetries in the temporal speech region. Arch Neurol 35:8128171978Arch Neurol 35:

  • 9.

    Goldman-Rakic PS: Topography of cognition: parallel distributed networks in primate association cortex. Annu Rev Neurosci 11:1371561988Goldman-Rakic PS: Topography of cognition: parallel distributed networks in primate association cortex. Annu Rev Neurosci 11:

  • 10.

    Heilman KMWilder BJMalzone WF: Anomic aphasia following anterior temporal lobectomy. Trans Am Neurol Assoc 97:2912931972Trans Am Neurol Assoc 97:

  • 11.

    Jasper HH: Unspecific thalamocortical relationsField JMagoun HW (eds): Handbook of Physiology. Neurophysiology. Washington, DC: American Physiology Society2Part 2196013071321Handbook of Physiology. Neurophysiology.

  • 12.

    Kimura D: Left-hemisphere control of oral and brachial movements and their relation to communication. Philos Trans R Soc Lond (Biol) 298:1351491982Kimura D: Left-hemisphere control of oral and brachial movements and their relation to communication. Philos Trans R Soc Lond (Biol) 298:

  • 13.

    Kimura D: Sex differences in cerebral organization for speech and praxic functions. Can J Psychol 37:19351983Kimura D: Sex differences in cerebral organization for speech and praxic functions. Can J Psychol 37:

  • 14.

    Kohn SEGoodglass H: Picture-naming in aphasia. Brain Lang 24:2662831985Brain Lang 24:

  • 15.

    Lesser RPLueders HDinner DSet al: The location of speech and writing functions in the frontal language area. Results of extraoperative cortical stimulation. Brain 107:2752911984Brain 107:

  • 16.

    Liberman AMCooper FSShankweiler DPet al: Perception of the speech code. Psychol Rev 74:4314611967Psychol Rev 74:

  • 17.

    Lueders HLesser RPDinner DSet al: Inhibition of motor activity by elicited electrical stimulation of the human cortex. Epilepsia 24:5191983 (Abstract)Epilepsia 24:

  • 18.

    Lueders HLesser RPHahn Jet al: Basal temporal language area demonstrated by electrical stimulation. Neurology 36:5055101986Neurology 36:

  • 19.

    Merzenich MMNelson RJKaas JHet al: Variability in hand surface representations in areas 3b and 1 in adult owl and squirrel monkeys. J Comp Neurol 258:2812961987J Comp Neurol 258:

  • 20.

    Mohr JP: Broca's area and Broca's aphasiaWhitaker HWhitaker HA (eds): Studies in Neurolinguistics. New York: Academic Press19761201236Studies in Neurolinguistics.

  • 21.

    Ojemann GA: Brain organization for language from the perspective of electrical stimulation mapping. Behav Brain Sci 6:1892061983Ojemann GA: Brain organization for language from the perspective of electrical stimulation mapping. Behav Brain Sci 6:

  • 22.

    Ojemann GA: Effect of cortical and subcortical stimulation in human language and verbal memoryPlum F (ed): Language Communication and the Brain. New York: Raven Press1988101115Language Communication and the Brain.

  • 23.

    Ojemann GA: Electrical stimulation and the neurobiology of language. Behav Brain Sci 6:2212301983Ojemann GA: Electrical stimulation and the neurobiology of language. Behav Brain Sci 6:

  • 24.

    Ojemann GA: Individual variability in cortical localization of language. J Neurosurg 50:1641691979Ojemann GA: Individual variability in cortical localization of language. J Neurosurg 50:

  • 25.

    Ojemann GA: Language and the thalamus: object naming and recall during and after thalamic stimulation. Brain Lang 2:1011201975Ojemann GA: Language and the thalamus: object naming and recall during and after thalamic stimulation. Brain Lang 2:

  • 26.

    Ojemann GA: Some brain mechanisms for readingvon Euler C (ed): Brain and Reading. New York: Macmillan19894759Brain and Reading.

  • 27.

    Ojemann GACreutzfeldt OLettich Eet al: Neuronal activity in human lateral temporal cortex related to shortterm verbal memory, naming and reading. Brain 111:138314031988Brain 111:

  • 28.

    Ojemann GADodrill CB: Verbal memory deficits after left temporal lobectomy for epilepsy. Mechanism and intraoperative prediction. J Neurosurg 62:1011071985J Neurosurg 62:

  • 29.

    Ojemann GAFried ILettich E: Electrocorticographic (ECoG) correlates of language: I. Desynchronization in temporal language cortex during object naming. EEG Clin Neurophysiol EEG Clin Neurophysiol

  • 30.

    Ojemann GAWhitaker HA: The bilingual brain. Arch Neurol 35:4094121978Arch Neurol 35:

  • 31.

    Ojemann GAWhitaker HA: Language localization and variability. Brain Lang 6:2392601978Brain Lang 6:

  • 32.

    Penfield WJasper HH: Epilepsy and the Functional Anatomy of the Human Brain. Boston: Little, Brown & Co1954Epilepsy and the Functional Anatomy of the Human Brain.

  • 33.

    Penfield WRoberts L: Speech and Brain Mechanisms. Princeton, NJ: Princeton University Press1959Speech and Brain Mechanisms.

  • 34.

    Ranck JB Jr: Which elements are excited in electrical stimulation of mammalian central nervous system: a review. Brain Res 98:4174401975Ranck JB Jr: Which elements are excited in electrical stimulation of mammalian central nervous system: a review. Brain Res 98:

  • 35.

    Rubens ABMahowald MWHutton JT: Asymmetry of the lateral (sylvian) fissures in man. Neurology 26:6206241976Neurology 26:

  • 36.

    Siegel S: Nonparametric Statistics for the Behavioral Sciences. New York: McGraw-Hill1956Siegel S: Nonparametric Statistics for the Behavioral Sciences.

  • 37.

    Skinner JYingling C: Central grating mechanisms that regulate event-related potentials and behavior: a neural model for attention. Prog Clin Neurophysiol 1:30691977Prog Clin Neurophysiol 1:

  • 38.

    Van Buren JMFedio PFrederick GC: Mechanism and localization of speech in the parietotemporal cortex. Neurosurgery 2:2332391978Neurosurgery 2:

  • 39.

    Van Buren JMLewis DVSchuette WHet al: Fluorometric monitoring of NADH levels in cerebral cortex: preliminary observations in human epilepsy. Neurosurgery 2:1141211978Neurosurgery 2:

  • 40.

    Wada JRasmussen T: Intracarotid injections of sodium amytal for the lateralization of cerebral speech dominance. J Neurosurg 17:2662821960J Neurosurg 17:

  • 41.

    Wernicke C: Der Aphasische Symptomen Komplex. Breslau: Cohn & Weigart1874Wernicke C: Der Aphasische Symptomen Komplex.

  • 42.

    Whitaker HSeines O: Anatomic variations in the cortex: individual differences and the problem of the localization of language functions. Ann NY Acad Sci 280:8448541975Ann NY Acad Sci 280:

  • 43.

    Whitaker HAOjemann GA: Graded localisation of naming from electrical stimulation mapping of left cerebral cortex. Nature 270:50511978Nature 270:

  • 44.

    Woods RPDodrill CBOjemann GA: Brain injury, handedness, and speech lateralization in a series of amobarbital studies. Ann Neurol 23:5105181988Ann Neurol 23:

This work was supported by National Institutes of Health Grants NS17111, 21724, and 20482.

Article Information

Address reprint requests to: George Ojemann, M.D., Department of Neurological Surgery, University of Washington, RI-20, Seattle, Washington 98195.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Sites essential for naming (filled circles) in a 24-year-old woman with a verbal IQ of 81. Stimulation at 6 mA; control error rate in the absence of stimulation was 3.7%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by a small dot. M and S identify sites with motor (M) or sensory (S) responses. Note the localized posterior language area with closely spaced surrounding stimulation sites without errors.

  • View in gallery

    Sites essential for naming (filled circles) in a 46-year-old woman with a verbal IQ of 91. Stimulation at 7 mA; control error rate in the absence of stimulation was 6%. Open circles indicate stimulation sites without evoked errors. Note the relatively large posterior language area, but very localized anterior language site, and a language site in the anterior superior temporal gyrus (arrow) in front of rolandic cortex, 4 cm from the temporal tip. M and S indicate sites with motor (M) or sensory (S) responses.

  • View in gallery

    Sites of significant evoked naming errors (filled circles) in a 4-year-old boy (upper) and a 70-year-old man (lower), both with medial temporal lobe gliomas. Open circles indicate stimulation sites without evoked errors. The 4-year-old boy was stimulated through a chronic subdural grid. Note that both patients show very localized temporal language sites. M and S indicate sites with motor (M) or sensory (S) responses.

  • View in gallery

    Large dominant hemisphere temporoparietal resections of much of the classical Wernicke area in two patients, with no postoperative worsening of language function. The resections spared the sites of repeated evoked naming errors. Both patients had only left-sided speech areas based on preoperative intracarotid amobarbital perfusion testing. Filled circles indicate sites essential for naming; open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by a small dot. M and S indicate sites with motor (M) or sensory (S) responses. Upper: This patient with superior temporal gyms oligodendroglioma and intractable seizures experienced no language change after the resection delineated by the shaded area. The patient returned to teaching postoperatively. Lower: This patient with widespread lateral temporal epileptic focus had no language deficits preoperatively. Following the resection indicated by the shaded area, oral language returned to normal within a week, although reading remained slow for a longer period.

  • View in gallery

    Variability in language localization in 117 patients. Individual maps are aligned as described in the text, and cortex is divided into zones identified by dashed lines. Upper number in each zone is the number of patients with a site in that zone; lower number in circle is the percentage of those patients with sites of significant evoked naming errors in that zone. M and S indicate motor (M) or sensory (S) cortex.

  • View in gallery

    Sites essential for naming (filled circles) in a 24-year-old woman with a verbal IQ of 94. Stimulation at 2 mA; control error rate in the absence of stimulation was 1.2%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by a small dot. The posterior language area (slightly larger than in the patient illustrated in Fig. 1) is oriented transversely to the superior temporal gyrus. Note also that the intensity of the stimulating current does not determine the size of language area (compare to Fig. 1). M indicates sites with motor response.

  • View in gallery

    Sites essential for naming (filled circles) in an 18-year-old woman with a verbal IQ of 95. Stimulation at 4 mA; control error rate in the absence of stimulation was 0%. Open circles indicate stimulation sites without evoked errors. This patient was left-handed, but had only left language according to the intracarotid amobarbital perfusion test. Language in this unusual parietal location imposes no limits on a temporal resection. M and S indicate sites with motor (M) or sensory

  • View in gallery

    Sites essential for naming (filled circles) in a 37-year-old woman with a verbal IQ of 99. Stimulation at 6 mA; control error rate in the absence of stimulation was 0%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by small dots. No posterior language sites were identified despite extensive mapping. A posterior temporal resection (dashed line) was associated with no language changes, even acutely. M and S indicate sites with motor (M) or sensory (S) responses.

  • View in gallery

    Sites essential for naming (filled circles) in an 18-year-old man with a verbal IQ of 91. Stimulation at 5 mA; control error rate in the absence of stimulation was 1.7%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by a small dot. No changes in naming or counting were evoked at frontal sites. M and S indicate sites with motor (M) or sensory (S) responses.

  • View in gallery

    Sites essential for naming (filled circles) in a 20-year-old man with a verbal IQ of 91. Stimulation at 5 mA; control error rate in the absence of stimulation was 0%. Open circles indicate stimulation sites without evoked errors; single nonsignificant error shown by small dots. Inferior frontal language sites extend nearly to the pterion. The frontal resection came within 1 cm of the anterior language area, and was followed by a significant expression aphasia lasting several weeks. M and S indicate sites with motor (M) or sensory (S) responses.

  • View in gallery

    Site of evoked naming errors (filled circles) in a 45-year-old man following removal of a parietal glioma (shaded area) that exposed the planum temporale. Note the localized site of evoked naming errors on the planum (arrow), adjacent to a similar superior temporal gyrus surface site. Stimulation at 4 mA; control error rate in the absence of stimulation was 0.9%. Open circles indicate stimulation sites without evoked errors. No naming errors were evoked from stimulation of cortex overlying the tumor and no language disturbance followed the resection. M and S indicate sites with motor (M) or sensory (S) responses.

  • View in gallery

    Sex and VIQ differences in language localization*

References

1.

Berger MSOjemann GA: Cortical mapping techniques used to maximize tumor resection and safety in children with brain tumors. Ann Neurol 24:3611988 (Abstract)Ann Neurol 24:

2.

Bogen JBogen G: Wernicke's region — where is it? Ann NY Acad Sci 280:8348431975Ann NY Acad Sci 280:

3.

Broca P: Remarques sur le siège de la faculté du language articulé, suivies d'une observation d'aphémie (perte de la parole). Bull Soc Anat 36:3303571861Bull Soc Anat 36:

4.

Burchiel KJClark HOjemann GAet al: Use of stimulation mapping and corticography in the excision of arteriovenous malformations in sensorimotor and language-related neocortex. Neurosurgery 24:3233271989Neurosurgery 24:

5.

Cawthon DLettich EOjemann G: Human temporal lobe neuronal activity: inhibition during naming in only one of two languages. Soc Neurosci Abst 13:8391987Soc Neurosci Abst 13:

6.

Damasio A: Concluding remarks: neuroscience and cognitive science in the study of language and the brainPlum F (ed): New York: Raven Press1988275282

7.

Fried IOjemann GAFetz EE: Language-related potentials specific to human language cortex. Science 212:3533561981Science 212:

8.

Galaburda AMSanides FGeschwind N: Human brain. Cytoarchitectonic left-right asymmetries in the temporal speech region. Arch Neurol 35:8128171978Arch Neurol 35:

9.

Goldman-Rakic PS: Topography of cognition: parallel distributed networks in primate association cortex. Annu Rev Neurosci 11:1371561988Goldman-Rakic PS: Topography of cognition: parallel distributed networks in primate association cortex. Annu Rev Neurosci 11:

10.

Heilman KMWilder BJMalzone WF: Anomic aphasia following anterior temporal lobectomy. Trans Am Neurol Assoc 97:2912931972Trans Am Neurol Assoc 97:

11.

Jasper HH: Unspecific thalamocortical relationsField JMagoun HW (eds): Handbook of Physiology. Neurophysiology. Washington, DC: American Physiology Society2Part 2196013071321Handbook of Physiology. Neurophysiology.

12.

Kimura D: Left-hemisphere control of oral and brachial movements and their relation to communication. Philos Trans R Soc Lond (Biol) 298:1351491982Kimura D: Left-hemisphere control of oral and brachial movements and their relation to communication. Philos Trans R Soc Lond (Biol) 298:

13.

Kimura D: Sex differences in cerebral organization for speech and praxic functions. Can J Psychol 37:19351983Kimura D: Sex differences in cerebral organization for speech and praxic functions. Can J Psychol 37:

14.

Kohn SEGoodglass H: Picture-naming in aphasia. Brain Lang 24:2662831985Brain Lang 24:

15.

Lesser RPLueders HDinner DSet al: The location of speech and writing functions in the frontal language area. Results of extraoperative cortical stimulation. Brain 107:2752911984Brain 107:

16.

Liberman AMCooper FSShankweiler DPet al: Perception of the speech code. Psychol Rev 74:4314611967Psychol Rev 74:

17.

Lueders HLesser RPDinner DSet al: Inhibition of motor activity by elicited electrical stimulation of the human cortex. Epilepsia 24:5191983 (Abstract)Epilepsia 24:

18.

Lueders HLesser RPHahn Jet al: Basal temporal language area demonstrated by electrical stimulation. Neurology 36:5055101986Neurology 36:

19.

Merzenich MMNelson RJKaas JHet al: Variability in hand surface representations in areas 3b and 1 in adult owl and squirrel monkeys. J Comp Neurol 258:2812961987J Comp Neurol 258:

20.

Mohr JP: Broca's area and Broca's aphasiaWhitaker HWhitaker HA (eds): Studies in Neurolinguistics. New York: Academic Press19761201236Studies in Neurolinguistics.

21.

Ojemann GA: Brain organization for language from the perspective of electrical stimulation mapping. Behav Brain Sci 6:1892061983Ojemann GA: Brain organization for language from the perspective of electrical stimulation mapping. Behav Brain Sci 6:

22.

Ojemann GA: Effect of cortical and subcortical stimulation in human language and verbal memoryPlum F (ed): Language Communication and the Brain. New York: Raven Press1988101115Language Communication and the Brain.

23.

Ojemann GA: Electrical stimulation and the neurobiology of language. Behav Brain Sci 6:2212301983Ojemann GA: Electrical stimulation and the neurobiology of language. Behav Brain Sci 6:

24.

Ojemann GA: Individual variability in cortical localization of language. J Neurosurg 50:1641691979Ojemann GA: Individual variability in cortical localization of language. J Neurosurg 50:

25.

Ojemann GA: Language and the thalamus: object naming and recall during and after thalamic stimulation. Brain Lang 2:1011201975Ojemann GA: Language and the thalamus: object naming and recall during and after thalamic stimulation. Brain Lang 2:

26.

Ojemann GA: Some brain mechanisms for readingvon Euler C (ed): Brain and Reading. New York: Macmillan19894759Brain and Reading.

27.

Ojemann GACreutzfeldt OLettich Eet al: Neuronal activity in human lateral temporal cortex related to shortterm verbal memory, naming and reading. Brain 111:138314031988Brain 111:

28.

Ojemann GADodrill CB: Verbal memory deficits after left temporal lobectomy for epilepsy. Mechanism and intraoperative prediction. J Neurosurg 62:1011071985J Neurosurg 62:

29.

Ojemann GAFried ILettich E: Electrocorticographic (ECoG) correlates of language: I. Desynchronization in temporal language cortex during object naming. EEG Clin Neurophysiol EEG Clin Neurophysiol

30.

Ojemann GAWhitaker HA: The bilingual brain. Arch Neurol 35:4094121978Arch Neurol 35:

31.

Ojemann GAWhitaker HA: Language localization and variability. Brain Lang 6:2392601978Brain Lang 6:

32.

Penfield WJasper HH: Epilepsy and the Functional Anatomy of the Human Brain. Boston: Little, Brown & Co1954Epilepsy and the Functional Anatomy of the Human Brain.

33.

Penfield WRoberts L: Speech and Brain Mechanisms. Princeton, NJ: Princeton University Press1959Speech and Brain Mechanisms.

34.

Ranck JB Jr: Which elements are excited in electrical stimulation of mammalian central nervous system: a review. Brain Res 98:4174401975Ranck JB Jr: Which elements are excited in electrical stimulation of mammalian central nervous system: a review. Brain Res 98:

35.

Rubens ABMahowald MWHutton JT: Asymmetry of the lateral (sylvian) fissures in man. Neurology 26:6206241976Neurology 26:

36.

Siegel S: Nonparametric Statistics for the Behavioral Sciences. New York: McGraw-Hill1956Siegel S: Nonparametric Statistics for the Behavioral Sciences.

37.

Skinner JYingling C: Central grating mechanisms that regulate event-related potentials and behavior: a neural model for attention. Prog Clin Neurophysiol 1:30691977Prog Clin Neurophysiol 1:

38.

Van Buren JMFedio PFrederick GC: Mechanism and localization of speech in the parietotemporal cortex. Neurosurgery 2:2332391978Neurosurgery 2:

39.

Van Buren JMLewis DVSchuette WHet al: Fluorometric monitoring of NADH levels in cerebral cortex: preliminary observations in human epilepsy. Neurosurgery 2:1141211978Neurosurgery 2:

40.

Wada JRasmussen T: Intracarotid injections of sodium amytal for the lateralization of cerebral speech dominance. J Neurosurg 17:2662821960J Neurosurg 17:

41.

Wernicke C: Der Aphasische Symptomen Komplex. Breslau: Cohn & Weigart1874Wernicke C: Der Aphasische Symptomen Komplex.

42.

Whitaker HSeines O: Anatomic variations in the cortex: individual differences and the problem of the localization of language functions. Ann NY Acad Sci 280:8448541975Ann NY Acad Sci 280:

43.

Whitaker HAOjemann GA: Graded localisation of naming from electrical stimulation mapping of left cerebral cortex. Nature 270:50511978Nature 270:

44.

Woods RPDodrill CBOjemann GA: Brain injury, handedness, and speech lateralization in a series of amobarbital studies. Ann Neurol 23:5105181988Ann Neurol 23:

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 8 8 8
Full Text Views 123 123 56
PDF Downloads 98 98 68
EPUB Downloads 0 0 0

PubMed

Google Scholar