Cortical mapping in glioma surgery: correlation of fMRI and direct electrical stimulation with Human Connectome Project parcellations

Carlos BennettDepartment of Neurosurgery, Hospital Carlos van Buren, Valparaíso;
School of Medicine, Universidad de Valparaíso;

Search for other papers by Carlos Bennett in
jns
Google Scholar
PubMed
Close
 MD
,
Matías GonzálezDepartment of Neurosurgery, Hospital Carlos van Buren, Valparaíso;
School of Medicine, Universidad de Valparaíso;

Search for other papers by Matías González in
jns
Google Scholar
PubMed
Close
 MD
,
Gisella TapiaSchool of Medicine, Universidad de Valparaíso;
Department of Neurology, Hospital Carlos van Buren, Valparaíso;

Search for other papers by Gisella Tapia in
jns
Google Scholar
PubMed
Close
 MD
,
Rodrigo RiverosSchool of Medicine, Universidad de Valparaíso;
Department of Radiology, Hospital Carlos van Buren, Valparaíso;

Search for other papers by Rodrigo Riveros in
jns
Google Scholar
PubMed
Close
 MD
,
Francisco TorresSchool of Medicine, Universidad de Valparaíso;
Department of Radiology, Hospital Carlos van Buren, Valparaíso;

Search for other papers by Francisco Torres in
jns
Google Scholar
PubMed
Close
 MD
,
Nicole LoyolaDepartment of Neurosurgery, Hospital Carlos van Buren, Valparaíso;
School of Medicine, Universidad de Valparaíso;

Search for other papers by Nicole Loyola in
jns
Google Scholar
PubMed
Close
 MD
,
Alejandro VelozSchool of Biomedical Engineering, Universidad de Valparaíso;
Centro de Investigación y Desarrollo en Ingeniería en Salud CINGS, Universidad de Valparaíso;

Search for other papers by Alejandro Veloz in
jns
Google Scholar
PubMed
Close
 PhD
, and
Stéren ChabertSchool of Biomedical Engineering, Universidad de Valparaíso;
Instituto Milenio Intelligent Healthcare Engineering, Santiago, Chile

Search for other papers by Stéren Chabert in
jns
Google Scholar
PubMed
Close
 PhD
Free access

OBJECTIVE

Noninvasive brain mapping with functional MRI (fMRI) and mapping with direct electrical stimulation (DES) are important tools in glioma surgery, but the evidence is inconclusive regarding the sensitivity and specificity of fMRI. The Human Connectome Project (HCP) proposed a new cortical parcellation that has not been thoroughly tested in a clinical setting. The main goal of this study was to evaluate the correlation of fMRI and DES mapping with HCP areas in a clinical setting, and to evaluate the performance of fMRI mapping in motor and language tasks in patients with glioma, using DES as the gold standard.

METHODS

Forty patients with supratentorial gliomas were examined using preoperative fMRI and underwent awake craniotomy with DES. Functional activation maps were visualized on a 3D representation of the cortex, classified according to HCP areas, and compared with surgical mapping.

RESULTS

Functional MRI was successful in identifying language and motor HCP areas in most cases, including novel areas such as 55b and the superior longitudinal fasciculus (SLF). Functional MRI had a sensitivity and specificity of 100% and 71%, respectively, for motor function in HCP area 4. Sensitivity and specificity were different according to the area and fMRI protocol; i.e., semantic protocols performed better in Brodmann area (BA) 55b/peri-sylvian language areas with 100% sensitivity and 20% specificity, and word production protocols in BAs 44 and 45 with 70% sensitivity and 80% specificity. Some compensation patterns could be observed, such as motor activation of the postcentral gyrus in precentral gliomas.

CONCLUSIONS

HCP areas can be detected in clinical scenarios of glioma surgery. These areas appear relatively stable across patients, but compensation patterns seem to differ, allowing occasional resection of activating areas. Newly described areas such as 55b and SLF can act as critical areas in language networks. Surgical planning should account for these parcellations.

ABBREVIATIONS

AAL = automated anatomical labeling; BA = Brodmann area; BOLD = blood oxygenation level dependent; DES = direct electrical stimulation; fMRI = functional MRI; HCP = Human Connectome Project; IFG = inferior frontal gyrus; IFS = inferior frontal sulcus; PHT = area parietalis basalis; PSL = peri-sylvian language; SFG = superior frontal gyrus; SFL = superior frontal language; SLF = superior longitudinal fasciculus; STSdp = superior temporal sulcus dorsal posterior; STV = superior temporal visual; TPOJ1 = temporo-parieto-occipital junction 1.

OBJECTIVE

Noninvasive brain mapping with functional MRI (fMRI) and mapping with direct electrical stimulation (DES) are important tools in glioma surgery, but the evidence is inconclusive regarding the sensitivity and specificity of fMRI. The Human Connectome Project (HCP) proposed a new cortical parcellation that has not been thoroughly tested in a clinical setting. The main goal of this study was to evaluate the correlation of fMRI and DES mapping with HCP areas in a clinical setting, and to evaluate the performance of fMRI mapping in motor and language tasks in patients with glioma, using DES as the gold standard.

METHODS

Forty patients with supratentorial gliomas were examined using preoperative fMRI and underwent awake craniotomy with DES. Functional activation maps were visualized on a 3D representation of the cortex, classified according to HCP areas, and compared with surgical mapping.

RESULTS

Functional MRI was successful in identifying language and motor HCP areas in most cases, including novel areas such as 55b and the superior longitudinal fasciculus (SLF). Functional MRI had a sensitivity and specificity of 100% and 71%, respectively, for motor function in HCP area 4. Sensitivity and specificity were different according to the area and fMRI protocol; i.e., semantic protocols performed better in Brodmann area (BA) 55b/peri-sylvian language areas with 100% sensitivity and 20% specificity, and word production protocols in BAs 44 and 45 with 70% sensitivity and 80% specificity. Some compensation patterns could be observed, such as motor activation of the postcentral gyrus in precentral gliomas.

CONCLUSIONS

HCP areas can be detected in clinical scenarios of glioma surgery. These areas appear relatively stable across patients, but compensation patterns seem to differ, allowing occasional resection of activating areas. Newly described areas such as 55b and SLF can act as critical areas in language networks. Surgical planning should account for these parcellations.

Current glioma management usually involves resection,1 and several studies have shown that a greater degree of resection is associated with longer survival and lower risk of anaplastic transformation.24 However, it is necessary to balance the objective of maximal resection with the need to reduce the risk of new neurological deficits. This balance poses a problem, as the precise locations of these "eloquent" areas present with individual variability.5,6

Different forms of brain mapping have been proposed to identify critical areas. Surgical mapping with direct electrical stimulation (DES) is considered the gold standard.57 Noninvasive mapping methods include functional MRI (fMRI), an imaging technique based on the higher oxygen consumption in cortical areas that are active during certain tasks. Several studies have compared fMRI with DES; in language studies, observations have led to a wide range of fMRI sensitivity (22%–93%) and specificity (64%–83%).812 Motor task observations have usually found a narrower range of these parameters. These observations have some limitations; for example, many studies include different pathologies,8,9,11,13,14 and potentially each pathology implies different ways of affecting the neurovascular coupling, and therefore the blood oxygenation level–dependent (BOLD) signal. There is limited knowledge on fMRI mapping performance in gliomas.12,15 Most comparison studies are based on identifying Euclidean distances between activated areas on fMRI and the locus identified on DES. This approach suffers from loss of the anatomical meaning of localization: a difference of 1 cm can have different implications if within the same gyrus or if separated by a sulcus.17

Clinical brain mapping requires a working model of cortical parcellations. One of the most well-accepted models comes from the Human Connectome Project (HCP), which generated a multimodal map of 180 distinct areas.16 We used the anatomical section of this parcellation to evaluate the sensitivity and specificity of fMRI in patients with supratentorial low- and high-grade gliomas. Each of the areas described has different roles within the language network (Table 1). We used 3 different fMRI protocols and hypothesized that each protocol would have different sensitivities and specificities according to the cortical area being mapped.

TABLE 1.

Anatomical descriptions of the HCP parcellation

Anatomical LocationHCP ParcellationBA EquivalenceFunction
Central lobe
 Anterior bank of central sulcus/posterior part of precentral gyrusArea 4Part of area 4Primary motor cortex
 Fundus of the central sulcusArea 3aPart of area 3Sensory information from deep body tissues, proprioception
 Posterior bank central sulcusArea 3bPart of area 3Sensation of tactile stimuli, localization of sensation on the skin
 Postcentral gyrusArea 2Area 2Processing of deep tissue sensations
Frontal dorsomedial cortex
 Posterior medial SFG straddling over the interhemispheric cleftArea SFLPart of area 8Language-related functions
Frontal dorsolateral cortex
 Anterior portion of the IFS, superior to the pars triangularis portion of the IFGArea IFSpPart of area 45Retrieval of specific auditory memories; creating procedural representations in working memory from verbal instructions
 Area IFJp is at the posterior-most part of the IFS, approximately superior to the pars opercularis of the IFGArea IFJpPart of area 44Retrieval of specific auditory memories
 Surface of the pars opercularisArea 44Area 44Production of verbal actions
 Surface of the pars triangularisArea 45Area 45Production of verbal actions
 Anterior part of precentral gyrus/posterior part of middle frontal gyrusArea 55bNot previously describedLanguage processing
Temporoparietal cortex
 Lower portion of the posterior part of the supramarginal gyrusArea PSLPart of area 40Information processing, motional control, & control of cognitive functions; generation of language, visuospatial attention, & assimilation of audiovisual information
 Posterior superior temporal sulcus, just as it angles upward to indent the angular gyrusArea TPOJ1Part of areas 22 & 39Information processing, motional control, & control of cognitive functions; generation of language, visuospatial attention, & assimilation of audiovisual information
 Posterior half of the lateral face of the STG & the posterior half of the superior bank of the STSArea STSdpPart of area 22Motion processing, audiovisual integration, facial processing, & language comprehension
 Inferior posterior STG, inferior to PSLArea STVPart of area 22Information processing, motional control, & control of cognitive functions; generation of language, visuospatial attention, & assimilation of audiovisual information
 Posterior middle temporal gyrusArea PHTPart of area 21Controlled retrieval of conceptual knowledge, automatic retrieval of specific semantic information

IFJp = inferior frontal junction posterior; IFSp = IFS posterior; STG = superior temporal gyrus.

The main goal of the study was to evaluate the correlation of fMRI and DES mapping with HCP areas in a clinical setting, and to evaluate the performance of fMRI mapping in the context of motor and language tasks in patients with glioma, using DES as the gold standard.

Methods

Study Population

We recruited 40 patients presenting with a high- or low-grade supratentorial glioma and requiring excision surgery using awake craniotomy, as indicated by the neuro-oncological committee. Written informed consent was obtained and this study received approval from our local ethics committee (Servicio de Salud Valparaíso San Antonio, Chile).

Image Acquisition

Images were acquired with a 1.5-T General Electric MR machine (Signa HDxt). Three-dimensional T1-weighted images served as anatomical images, obtained with a fast spoiled gradient echo sequence (TE/TR = 1.9/6.1 msec, 256 × 256 matrix, 24-cm field of view, 1.2-mm slice thickness). Functional images were obtained with gradient echo planar imaging sequence (TE/TR = 60/3000 msec, 5-mm slice thickness, 1.875 × 1.875 mm2 pixel size). All functional images were obtained with stimuli presented in blocks, consisting of 5 runs of 30-second OFF/30-second ON periods.

Motor Acquisition

Motor acquisitions consisted of thumb-finger tapping or tongue motion along the palate or feet motion during the ON period, versus no motion to be performed (OFF period). Three different language paradigms were used: 1) naming, 2) reading comprehension, and 3) verb generation. The naming task consisted of presenting images from the Boston test17 for 3 seconds per image (ON period). The control period consisted of a blank screen with a fixation dot. The reading comprehension task consisted of a series of 10 words forming a meaningful sentence. During the control period, a series of pseudowords were shown. Each word or pseudoword was presented for 3 seconds. In the verb generation paradigm, a series of nouns were shown to the patient (ON period), with the instruction to silently think of a verb that could be associated with each noun. The control period consisted of a blank screen with a fixation dot. All stimuli were presented using OpenSesame (version 3.1).18

Image Processing

Images were processed in MATLAB (MathWorks) using SPM12 (UCL; https://www.fil.ion.ucl.ac.uk/spm/). In preprocessing, the slice timing was adjusted, head motion was estimated and corrected, functional and anatomical images were coregistered, normalization was performed to represent activations in a common space, and spatial smoothing was obtained using a 6-mm Gaussian kernel. Statistical parametric mapping was obtained using a general linear model with a classical Bayesian approach and canonical expression of hemodynamic response function, using family-wise error and a threshold fixed at 5%. Blood vessels were segmented using in-house software implemented in Python. Acquisitions were excluded from the analysis if patient motion exceeded 3 mm or if the task was not performed correctly by the patient.

MRI Activation Maps

MRI activation maps were visualized on a 3D representation of the cortical surface, on which blood vessels were superimposed to facilitate spatial referencing. MRI activation maps were also superimposed on the automated anatomical labeling (AAL) template,19 enabling an approximation of a superimposition of activation in Brodmann areas (BAs). Two neurosurgeons independently analyzed the images and represented the activations on a standard space demonstrating the most relevant parcellations for motor function and language identified by the HCP. Fifteen areas were selected for this study. The main sulcal and gyral anatomy was used to define whether the activation zone corresponded to an HCP area, as defined in Table 1, aided by the information provided by AAL superimposition.

Glioma Surgery

All patients underwent excision surgery with awake craniotomy. Cortical mapping was performed using bipolar stimulation at 60 Hz and a biphasic wave starting at 2 mA, without exceeding 8 mA. Neuronavigation (Brainlab AG) was used in all patients. Intraoperative mapping findings were documented with photographs. Afterward, 3D fMRI with superimposed cortical vessels was compared with the intraoperative photographs, and confusion matrices were created. For each HCP area, if activation was detected through fMRI mapping, a positive result of "activation" in this area was registered for this patient. In an independent manner, the activation maps obtained from surgery were written down. The areas tested during surgery were registered as negative if tested with no activation corresponding to this localization, and as positive if tested with activation corresponding to this localization. Considering activation registered during awake surgery as the gold standard, true-positive, true-negative, false-positive, and false-negative numbers were summarized for each HCP area. We grouped some areas that were closely related and were difficult to test independently during surgery in 1 confusion matrix (described below). An example of the different steps used for analysis of 1 patient is depicted in Fig. 1.

FIG. 1.
FIG. 1.

Images showing the methodology used in the study. A: Motor and language fMRI is performed (hand movement depicted in the picture). B: Activation zones are superimposed on a normalized AAL map, which allows an allocation according to BAs, which in turn can be correlated with HCP areas. C: To ensure that displacement caused by the tumor has been considered, sulcal and gyral anatomy is carefully examined by two neurosurgeons. In this case, the IFS (blue), precentral sulcus (red), postcentral sulcus (black), and intraparietal sulcus (green) are marked. D: Of the studied areas, the patient is shown to have activations in motor fMRI in HCP areas 4, 3a, 3b, and 2. E: The patient is operated on using awake craniotomy; an intraoperative photograph is taken to correlate DES findings with fMRI findings. F: Three-dimensional fMRI with superimposition of cortical vessels is compared with the intraoperative photograph (E). The black outline indicates the area exposed during surgery, and the arrow indicates identical localization of activation, labeled as true positive.

Results

We analyzed 40 patients with supratentorial gliomas (29 low grade, 11 high grade). Thirty-seven patients were examined with language and motor fMRI, and 3 patients using only language fMRI. The mean patient age was 40.1 (range 18–72) years. Nineteen tumors (48%) were frontal or fronto-insular, 12 (30%) temporal or temporo-insular, 6 (15%) parietal, and 3 (7%) were multilobar.

Motor Mapping

Thirty-seven patients were examined with motor mapping. Two patient acquisitions were excluded due to excessive motion; another acquisition was excluded because the patient did not understand the instructions and did not perform the assigned tasks. Thirty-four patients were then included in the analysis of motor task mapping. Activation on motor fMRI was classified as anterior (HCP area 4) or posterior (HCP areas 3a, 3b, and 2) to the central sulcus. We detected activation in area 4 in 29 (85%) of the 34 patients, and posterior to the central sulcus in 19 patients (56%). Comparison with surgical mapping was possible when the area of activation was exposed within the craniotomy, which occurred in 18 patients in the precentral gyrus and in 11 patients in the postcentral gyrus. Confusion matrices are depicted in Table 2, and patient examples in Fig. 2.

TABLE 2.

Confusion matrices between fMRI and surgical mapping

VariablefMRI (+)fMRI (−)SensitivitySpecificityPPVNPV
Motor mapping
 Precentral gyrus (area 4), n = 18
  DES (+)110100%71%85%100%
  DES (−)25
 Postcentral gyrus (areas 3a, 3b, 2), n = 11
  DES (+)20100%78%50%100%
  DES (−)27
Language mapping
 Whole brain
  Image naming, n = 48
   DES (+)9756%63%43%74%
   DES (−)1220
  Pseudowords, n = 50
   DES (+)10663%85%67%83%
   DES (−)529
  Verb generation, n = 43
   DES (+)10471%72%56%84%
   DES (−)821
  Union of fMRI mapping, n = 50
   DES (+)14288%62%52%91%
   DES (−)1321
 55b mapping
  Image naming, n = 8
   DES (+)30100%20%43%100%
   DES (−)41
  Pseudowords, n = 8
   DES (+)2167%60%50%75%
   DES (−)23
  Verb generation, n = 7
   DES (+)2167%25%40%50%
   DES (−)31
  Union of fMRI mapping, n = 8
   DES (+)30100%20%43%100%
   DES (−)41
 Broca’s complex mapping
  Image naming, n = 20
   DES (+)5550%70%63%58%
   DES (−)37
  Pseudowords, n = 20
   DES (+)6460%90%86%69%
   DES (−)19
  Verb generation, n = 20
   DES (+)7370%80%78%73%
   DES (−)28
  Union of fMRI mapping, n = 20
   DES (+)8280%60%67%75%
   DES (−)46
 Temporo-parieto-occipital junction mapping (PSL, STV, & TPOJ1)
  Image naming, n = 10
   DES (+)0080%0%100%
   DES (−)28
  Pseudowords, n = 10
   DES (+)00100%100%
   DES (−)010
  Verb generation, n = 10
   DES (+)00100%100%
   DES (−)010
  Union of fMRI mapping, n = 10
   DES (+)0080%0%100%
   DES (−)28
 SFL mapping
  Image naming, n = 2
   DES (+)10
   DES (−)10
  Pseudowords, n = 4
   DES (+)10
   DES (−)12
  Verb generation, n = 1
   DES (+)00
   DES (−)10
  Union of fMRI mapping, n = 4
   DES (+)10
   DES (−)12

NPV = negative predictive value; PPV = positive predictive value.

Union of fMRI mapping = at least 1 protocol is considered as positive (+). Image naming = stimulation using the Boston naming test. Pseudowords = stimulation using reading of sentences versus pseudowords. Verb generation = stimulation using verb generation based on substantive reading. See text for more details on stimuli. Differences in the number of patients (n) included between different applied stimuli are due to either difficulty of participation of the patient or images excluded due to motion. SFL mapping is reported here only in observations, as the n value is too low to estimate sensitivity or specificity.

FIG. 2.
FIG. 2.

Examples of motor mapping. A: Functional MRI showing activation (yellow) in the postcentral gyrus (and absence of activation in the precentral gyrus) in a patient with a precentral glioma. Surgical mapping of the precentral gyrus was negative. B: Postoperative MRI obtained in the same patient, showing resection of the precentral gyrus without motor worsening. It was labeled as a true negative in area 4. C: Three-dimensional fMRI of a patient with activation of both the pre- and postcentral gyrus in a patient with a postcentral glioma. The black outline indicates the area exposed during surgery. D: Intraoperative photograph obtained in the same patient in panel C shows the absence of elicited deficit in the postcentral gyrus in the localization (arrow). This area was labeled as a false positive in area 2.

Activation of the precentral gyrus on fMRI was more common in the precentral gyrus compared with the postcentral gyrus (91% of all patients presented with activation in the precentral gyrus vs 53% in the postcentral gyrus). There was no difference in sensitivity (100% in both locations) and only a 7% difference in specificity (71% in the precentral gyrus and 78% in the postcentral gyrus). Notably, the negative predictive value of motor fMRI was 100% (all cases with negative fMRI in the pre- or postcentral gyrus had negative surgical mapping in these locations), while positive fMRI mapping was not as accurate, and was better in the precentral gyrus (85%) than in the postcentral gyrus (50%).

Language Mapping

The extent to which each HCP area was activated in different fMRI language protocols is shown in Fig. 3. Overall, the image naming protocol performed worse than the other protocols in all areas except the superior temporal visual (STV). Some areas showed marked differences in activation percentage according to the fMRI protocol used, with the strongest differences shown in peri-sylvian language (PSL; 82% in the pseudoword protocols vs 43% in verb generation) and superior frontal language (SFL; 100% in verb generation vs 78% in pseudowords) areas. In each case, correlation between fMRI and surgical mapping was possible for each area when it was exposed within the craniotomy. In the frontal lobe, the relevant HCP language areas are SFL, 55b, and areas related to Broca’s complex (i.e., areas 44 and 45 and areas in the inferior frontal sulcus [IFSja and IFSsp]). We analyzed Broca’s complex areas as a group because of their intimate anatomical and functional relation and the difficulty of testing intrasulcal parcellations separately. Similarly, in the temporoparietal area, we grouped closely related parcellations for the estimation of sensitivity and specificity (PSL, STV, and temporo-parieto-occipital junction 1 [TPOJ1]). Area parietalis basalis (PHT) and the superior temporal sulcus dorsal posterior (STSdp) were studied separately, as previous functional studies showed they have different connectivities and functions.20

FIG. 3.
FIG. 3.

Percentage of activation of HCP parcellations in different language protocols. IN = image naming protocol; PW = word versus pseudoword protocol; VG = verb generation protocol.

The overall sensitivities and specificities for language for all studied HCP areas are depicted in Table 2. Considering all language areas, sensitivity was worse in specific language fMRI protocols than in motor tasks and was better when activation of any protocol was considered (88%), but with lower specificity (62%). When evaluating specific areas with each protocol, sensitivity was variable (50%–100%), as was specificity (20%–100%). Notably, areas usually considered eloquent (PSL/STV) showed high specificity with the pseudoword and verb generation protocols (in 10 cases with negative fMRI mapping, no surgical positive mapping was detected), and the negative predictive value was 100% in all protocols. Because of the low number of activations in certain areas, sensitivity could not be determined for all areas and protocols.

Discussion

Efficacy of fMRI in Brain Mapping Gliomas

Various authors have indicated that pathological modifications induced by gliomas decrease fMRI spatial discrimination.8,2124 Gliomas usually present with invasion of eloquent areas but without the patient necessarily showing clinical deficits, on many occasions due to persistence of functional areas within the tumor25,26 or due to neuronal remodeling in peritumoral areas.27,28 fMRI limitations in detecting functional areas within or in the immediate periphery of tumor tissue may be increased due to this model of reorganization.31 Clinically, fMRI can be used as an adjunct to surgical mapping, because preoperatively identifying positive sites helps guide surgical mapping. Functional MRI has also been proposed as a tool to replace awake craniotomy; in this case, a high negative predictive value can be considered the most important factor to obtain in an evaluation of which patient area can be safely resected.

Our study shows variable correlations between fMRI and surgical mapping, with different sensitivities and specificities according to the area tested and the protocol used. High concordance of areas detected using fMRI with specific HCP parcels, such as SFL and 55b (which are not always detected by DES), leads us to rethink the role of cortical mapping in relation to fMRI. While fMRI appears to provide a global vision of the language network in each patient, surgical mapping detects which components of this network behave as hub areas (critical noncompensable nodes), which can be more variable.

Mapping of the Central Lobe

In the literature, the concordance of fMRI with DES is generally good for motor mapping, with sensitivities of 71%–100% and specificities of 68%–100%.10,2931 We obtained similar values in patients with glioma, with 100% sensitivity in the precentral gyrus. The fact that more than half of the observed patients presented with activation of the postcentral gyrus is noteworthy. In the HCP, postcentral areas (3a, 3b, and 2) are classified as sensitive areas. However, these same studies demonstrate connectivity of these areas toward the pyramidal tract,32 and other studies demonstrate the existence of pyramidal motor neurons and motor function in the postcentral gyrus.33 Figure 2A and B shows an example of a resected precentral glioma without worsening paresis, with significant fMRI activation of the postcentral gyrus. It is possible that in certain cases, the neuronal reserve of pyramidal neurons in postcentral areas can functionally compensate for motor function, allowing resection of gliomas in area 4 without generating severe motor deficit. Our hypothesis is that in certain cases, resection of the precentral gyrus is possible if the postcentral gyrus is preserved; activation of the postcentral gyrus in motor fMRI would support this pattern of compensation. The fact that in all cases of precentral gyrus gliomas this plasticity has been expressed in the postcentral gyrus supports the idea that compensation patterns are predictable and identifiable by fMRI.

Another clinical implication of these findings is language mapping in the precentral gyrus. Our study supports a semantic role for area 55b because it presented with a high percentage of activation in the pseudoword protocol (Fig. 3; example shown in Fig. 4). Three of these cases were positive using DES (100% sensitivity with image naming, 67% with pseudowords). This finding suggests that resection of this area could cause aphasia, as has been previously suggested in the literature,34 which would imply the need to perform language mapping in this location.

FIG. 4.
FIG. 4.

A and B: Examples of SFL mapping. Images obtained in a patient with a glioma of the left SFG with activation of the SFL (yellow). The area was tested in surgical mapping with no elicited deficit. The patient showed no postoperative language worsening (but the area was not resected). This was labeled as false positive. C: Examples of mapping of area 55b. Image naming protocol in a patient with a left parietal glioma, showing activation in the anterior part of the precentral gyrus (yellow). D: Intraoperative photograph obtained in the same patient, with speech arrest in the corresponding area (blue circle). It was labeled as true positive. E and F: Examples of mapping in Broca’s complex. Images obtained in a patient with a frontotemporal insular glioma with positive fMRI mapping during image naming in area 45 and within the inferior frontal sulcus (IFSsp). Surgical mapping was negative in area 45, and it was resected to create a corridor for tumor excision. The patient presented with no postoperative deficits. It was labeled as false positive for area 45. G and H: Images obtained in a patient with glioma of the IFG and pars opercularis, and no activation on fMRI in Broca’s complex. The pars opercularis (area 44) was resected without language worsening. It was labeled as a true negative.

Mapping the Dorsal Frontal Cortex

A previous study indicated that in patients with glioma excision in the superior frontal gyrus (SFG), approximately 40% have long-term language problems.35 In our series, the fact that the SFL was universally detected by the verb generation protocol is noteworthy. The connectivity of this area through the frontal aslant tract with Broca’s complex suggests an important role in the expression of language. In our series, 4 patients had exposure of this area in the craniotomy, and 1 presented both positive mapping on fMRI and language deficit elicited by DES (Fig. 1). The direct excision of this area can explain more severe symptoms of mutism, which can evolve into permanent language problems.

Mapping in the Lateral Frontal Cortex

Broca’s area, canonically located in BAs 44 and 45 of the left hemisphere, is usually considered an eloquent site. The HCP states that it probably does not correspond to one area, but to several separate areas, with specific connectivity and functions. In these areas, fMRI appears to have better sensitivity when using the verb generation protocol, with a sensitivity of 80% obtained when the positive findings of any of the 3 protocols are used as a criterion. Allowing the information to be provided by fMRI mapping may permit excision of the left hemisphere pars triangularis or opercularis without worsening of language in some cases (Fig. 4), but maintain surgical mapping in this area as the gold standard.

Mapping in the Temporoparietal Cortex

Traditional nomenclature mentions Wernicke’s area at the level of the posterior part of the upper left temporal gyrus, within BA 22.36 The HCP parcel subdivides this area into several discrete parcellations, including intrasulcal areas such as STSdp. The detection of PSL stands out in this region, which presents a much more marked activation percentage with the pseudoword discrimination protocol (Fig. 3). Despite our series having a good proportion of temporoparietal gliomas, this very precise area was infrequently tested. This PSL detection confirms what has been stated in other series,5 in which small craniotomies guided by neuronavigation usually provide negative mapping (small eloquent areas are frequently not exposed). Here, fMRI appears to have good specificity, without false negatives (including cases of tumor invasion of this area), but with two false-positive cases. This finding probably indicates areas involved in language, but ones that are compensable after resection.

General Considerations About Language Mapping Using DES

The DES image naming protocol with Ojemann parameters is generally assumed to be the gold standard for cortical mapping. However, there are conceptual considerations regarding its use. The new cortical anatomofunctional parcellations show an increasingly complex network underlying language, with many areas dedicated to very specific functions of semantic understanding, syntax, naming, etc. The literature supports the existence of nodes within the network in which relatively small lesions can cause significant deficits; these are the areas that cortical mapping aims to detect. HCP parcellations allow us to propose that certain areas are more likely hub areas, based on their connectivity and belonging to multiple functional clusters. These areas may include PSL or areas 44 and 45. Our study suggests that these areas can occasionally be resected; i.e., they do not always behave as critical nodes, and therefore other more recently described areas, such as SFL or 55b, can behave like hubs within the network in particular patients. The HCP studies and the high consistency in the detection of these areas by fMRI in a clinical context suggest that these areas are relatively fixed; what seems to vary in each patient is the ability of the network to adapt to the injury of one of them.

Limitations of the Study

One limitation of this study comes from the fact that both low- and high-grade gliomas are included, while the way the BOLD response underlying the fMRI mapping is affected might be different in different types of gliomas. Further studies are still needed in that direction to consolidate understanding of functional mapping in glioma. The analysis method proposed here is based on expert reading of the activation maps superimposed onto the HCP parcellation. Even though this method is human dependent—and not systematized such as calculation of the Euclidean distance between activated clusters, for instance—it is based on expert knowledge, with various years of experience, and applied in a double-blinded manner to minimize bias in the interpretation. This allows the observation to be conducted closer to brain anatomy.

Conclusions

Cortical areas described by HCP studies can be detected by fMRI in a clinical setting, including recent areas such as SFL and 55b. Different fMRI protocols have variable detection capacities depending on the type of area, with the verb generation protocol being the most efficient in areas of phonological production. The sensitivity and specificity of fMRI in detecting critical areas when compared with DES are also variable, being higher for motor function than for language. Although there is relative agreement on the detection of HCP areas by fMRI in patients with glioma, there is individual variability in which of these areas behave as critical nodes when evaluated using surgical mapping. Further work is needed to study the individual connectome to predict compensation patterns and determine the noncompensable areas with noninvasive mapping.

Acknowledgments

The authors acknowledge funding from the following grants: Agencia Nacional de Investigación y Desarrollo, Fondo de Investigacion y Desarrollo en Salud FONIS SA17I0124 (C.B., S.C., M.G., G.T., R.R., N.L., A.V.), and FONDECYT Iniciación 11201046 (A.V.).

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Chabert, Bennett. Acquisition of data: Chabert, Bennett, González, Tapia, Riveros, Torres, Loyola. Analysis and interpretation of data: Chabert, Bennett, González, Tapia, Torres, Loyola, Veloz. Drafting the article: Chabert, Bennett, González. Critically revising the article: all authors. Reviewed submitted version of manuscript: Chabert, Bennett. Approved the final version of the manuscript on behalf of all authors: Chabert. Statistical analysis: Chabert, Veloz. Administrative/technical/material support: Chabert, Bennett, González, Tapia, Riveros, Torres. Study supervision: Chabert, Bennett.

References

  • 1

    Rueda E, Sierra M, Infante J, et al. Controversial aspects in WHO grade II gliomas management: review of recent literature.Article in Spanish. Rev Neurol. 2011;53(12):747757.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 2

    Morshed RA, Young JS, Hervey-Jumper SL, Berger MS. The management of low-grade gliomas in adults. J Neurosurg Sci. 2019;63(4):450457.

  • 3

    Choi J, Kim SH, Ahn SS, et al. Extent of resection and molecular pathologic subtype are potent prognostic factors of adult WHO grade II glioma. Sci Rep. 2020;10(1):2086.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4

    Duffau H. Long-term outcomes after supratotal resection of diffuse low-grade gliomas: a consecutive series with 11-year follow-up. Acta Neurochir (Wien). 2016;158(1):5158.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 5

    Sanai N, Mirzadeh Z, Berger MS. Functional outcome after language mapping for glioma resection. N Engl J Med. 2008;358(1):1827.

  • 6

    De Witt Hamer PC, Robles SG, Zwinderman AH, Duffau H, Berger MS. Impact of intraoperative stimulation brain mapping on glioma surgery outcome: a meta-analysis. J Clin Oncol. 2012;30(20):25592565.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Fernández Coello A, Moritz-Gasser S, Martino J, Martinoni M, Matsuda R, Duffau H. Selection of intraoperative tasks for awake mapping based on relationships between tumor location and functional networks. J Neurosurg. 2013;119(6):13801394.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8

    Roux FE, Boulanouar K, Lotterie JA, Mejdoubi M, LeSage JP, Berry I. Language functional magnetic resonance imaging in preoperative assessment of language areas: correlation with direct cortical stimulation. Neurosurgery. 2003;52(6):13351347.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9

    Bizzi A, Blasi V, Falini A, et al. Presurgical functional MR imaging of language and motor functions: validation with intraoperative electrocortical mapping. Radiology. 2008;248(2):579589.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 10

    Meier MP, Ilmberger J, Fesl G, Ruge MI. Validation of functional motor and language MRI with direct cortical stimulation. Acta Neurochir (Wien). 2013;155(4):675683.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 11

    Trinh VT, Fahim DK, Maldaun MV, et al. Impact of preoperative functional magnetic resonance imaging during awake craniotomy procedures for intraoperative guidance and complication avoidance. Stereotact Funct Neurosurg. 2014;92(5):315322.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12

    Kuchcinski G, Mellerio C, Pallud J, et al. Three-tesla functional MR language mapping: comparison with direct cortical stimulation in gliomas. Neurology. 2015;84(6):560568.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13

    Petrovich N, Holodny AI, Tabar V, et al. Discordance between functional magnetic resonance imaging during silent speech tasks and intraoperative speech arrest. J Neurosurg. 2005;103(2):267274.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14

    Qiu TM, Gong FY, Gong X, et al. Real-time motor cortex mapping for the safe resection of glioma: an intraoperative resting-state fMRI study. AJNR Am J Neuroradiol. 2017;38(11):21462152.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15

    Kapsalakis IZ, Kapsalaki EZ, Gotsis ED, et al. Preoperative evaluation with FMRI of patients with intracranial gliomas. Radiol Res Pract. 2012;2012:727810.

    • Search Google Scholar
    • Export Citation
  • 16

    Glasser MF, Coalson TS, Robinson EC, et al. A multi-modal parcellation of human cerebral cortex. Nature. 2016;536(7615):171178.

  • 17

    Kaplan E, Goodglass H, Weintraub S. Boston Naming Test. Lea & Febiger, 1983.

  • 18

    Mathôt S, Schreij D, Theeuwes J. OpenSesame: an open-source, graphical experiment builder for the social sciences. Behav Res Methods. 2012;44(2):314324.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19

    Klein A, Tourville J. 101 labeled brain images and a consistent human cortical labeling protocol. Front Neurosci. 2012;6:171.

  • 20

    Baker CM, Burks JD, Briggs RG, et al. A connectomic atlas of the human cerebrum—Chapter 6: The temporal lobe. Oper Neurosurg (Hagerstown). 2018;15(suppl 1):S245-S294.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21

    Håberg A, Kvistad KA, Unsgård G, Haraldseth O. Preoperative blood oxygen level-dependent functional magnetic resonance imaging in patients with primary brain tumors: clinical application and outcome. Neurosurgery. 2004;54(4):902915.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 22

    Holodny AI, Schulder M, Ybasco A, Liu WC. Translocation of Broca’s area to the contralateral hemisphere as the result of the growth of a left inferior frontal glioma. J Comput Assist Tomogr. 2002;26(6):941943.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 23

    Hossmann KA, Linn F, Okada Y. Bioluminescence and fluoroscopic imaging of tissue pH and metabolites in experimental brain tumors of cat. NMR Biomed. 1992;5(5):259264.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 24

    Segebarth C, Belle V, Delon C, et al. Functional MRI of the human brain: predominance of signals from extracerebral veins. Neuroreport. 1994;5(7):813816.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 25

    Ojemann JG, Miller JW, Silbergeld DL. Preserved function in brain invaded by tumor. Neurosurgery. 1996;39(2):253259.

  • 26

    Skirboll SS, Ojemann GA, Berger MS, Lettich E, Winn HR. Functional cortex and subcortical white matter located within gliomas. Neurosurgery. 1996;38(4):678685.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 27

    Thiel A, Herholz K, Koyuncu A, et al. Plasticity of language networks in patients with brain tumors: a positron emission tomography activation study. Ann Neurol. 2001;50(5):620629.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 28

    Duffau H. Brain plasticity: from pathophysiological mechanisms to therapeutic applications. J Clin Neurosci. 2006;13(9):885897.

  • 29

    Schulder M, Maldjian JA, Liu WC, et al. Functional image-guided surgery of intracranial tumors located in or near the sensorimotor cortex. J Neurosurg. 1998;89(3):412418.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30

    Lehéricy S, Duffau H, Cornu P, et al. Correspondence between functional magnetic resonance imaging somatotopy and individual brain anatomy of the central region: comparison with intraoperative stimulation in patients with brain tumors. J Neurosurg. 2000;92(4):589598.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 31

    Spena G, Nava A, Cassini F, et al. Preoperative and intraoperative brain mapping for the resection of eloquent-area tumors. A prospective analysis of methodology, correlation, and usefulness based on clinical outcomes. Acta Neurochir (Wien). 2010;152(11):18351846.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 32

    Baker CM, Burks JD, Briggs RG, et al. A connectomic atlas of the human cerebrum—Chapter 2: The lateral frontal lobe. Oper Neurosurg (Hagerstown). 2018;15(suppl 1):S10-S74.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 33

    Fogassi L, Luppino G. Motor functions of the parietal lobe. Curr Opin Neurobiol. 2005;15(6):626631.

  • 34

    Hazem SR, Awan M, Lavrador JP, et al. Middle frontal gyrus and area 55b: perioperative mapping and language outcomes. Front Neurol. 2021;12:646075.

  • 35

    Liu W, Lai JJ, Qu YM. Surgical treatment of gliomas involving the supplementary motor area in the superior frontal gyrus.Article in Chinese. Zhonghua Wai Ke Za Zhi. 2004;42(13):781783.

    • Search Google Scholar
    • Export Citation
  • 36

    Binder JR. The Wernicke area: modern evidence and a reinterpretation. Neurology. 2015;85(24):21702175.

  • Collapse
  • Expand
  • View in gallery
    FIG. 1.

    Images showing the methodology used in the study. A: Motor and language fMRI is performed (hand movement depicted in the picture). B: Activation zones are superimposed on a normalized AAL map, which allows an allocation according to BAs, which in turn can be correlated with HCP areas. C: To ensure that displacement caused by the tumor has been considered, sulcal and gyral anatomy is carefully examined by two neurosurgeons. In this case, the IFS (blue), precentral sulcus (red), postcentral sulcus (black), and intraparietal sulcus (green) are marked. D: Of the studied areas, the patient is shown to have activations in motor fMRI in HCP areas 4, 3a, 3b, and 2. E: The patient is operated on using awake craniotomy; an intraoperative photograph is taken to correlate DES findings with fMRI findings. F: Three-dimensional fMRI with superimposition of cortical vessels is compared with the intraoperative photograph (E). The black outline indicates the area exposed during surgery, and the arrow indicates identical localization of activation, labeled as true positive.

  • View in gallery
    FIG. 2.

    Examples of motor mapping. A: Functional MRI showing activation (yellow) in the postcentral gyrus (and absence of activation in the precentral gyrus) in a patient with a precentral glioma. Surgical mapping of the precentral gyrus was negative. B: Postoperative MRI obtained in the same patient, showing resection of the precentral gyrus without motor worsening. It was labeled as a true negative in area 4. C: Three-dimensional fMRI of a patient with activation of both the pre- and postcentral gyrus in a patient with a postcentral glioma. The black outline indicates the area exposed during surgery. D: Intraoperative photograph obtained in the same patient in panel C shows the absence of elicited deficit in the postcentral gyrus in the localization (arrow). This area was labeled as a false positive in area 2.

  • View in gallery
    FIG. 3.

    Percentage of activation of HCP parcellations in different language protocols. IN = image naming protocol; PW = word versus pseudoword protocol; VG = verb generation protocol.

  • View in gallery
    FIG. 4.

    A and B: Examples of SFL mapping. Images obtained in a patient with a glioma of the left SFG with activation of the SFL (yellow). The area was tested in surgical mapping with no elicited deficit. The patient showed no postoperative language worsening (but the area was not resected). This was labeled as false positive. C: Examples of mapping of area 55b. Image naming protocol in a patient with a left parietal glioma, showing activation in the anterior part of the precentral gyrus (yellow). D: Intraoperative photograph obtained in the same patient, with speech arrest in the corresponding area (blue circle). It was labeled as true positive. E and F: Examples of mapping in Broca’s complex. Images obtained in a patient with a frontotemporal insular glioma with positive fMRI mapping during image naming in area 45 and within the inferior frontal sulcus (IFSsp). Surgical mapping was negative in area 45, and it was resected to create a corridor for tumor excision. The patient presented with no postoperative deficits. It was labeled as false positive for area 45. G and H: Images obtained in a patient with glioma of the IFG and pars opercularis, and no activation on fMRI in Broca’s complex. The pars opercularis (area 44) was resected without language worsening. It was labeled as a true negative.

  • 1

    Rueda E, Sierra M, Infante J, et al. Controversial aspects in WHO grade II gliomas management: review of recent literature.Article in Spanish. Rev Neurol. 2011;53(12):747757.

    • Search Google Scholar
    • Export Citation
  • 2

    Morshed RA, Young JS, Hervey-Jumper SL, Berger MS. The management of low-grade gliomas in adults. J Neurosurg Sci. 2019;63(4):450457.

  • 3

    Choi J, Kim SH, Ahn SS, et al. Extent of resection and molecular pathologic subtype are potent prognostic factors of adult WHO grade II glioma. Sci Rep. 2020;10(1):2086.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4

    Duffau H. Long-term outcomes after supratotal resection of diffuse low-grade gliomas: a consecutive series with 11-year follow-up. Acta Neurochir (Wien). 2016;158(1):5158.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5

    Sanai N, Mirzadeh Z, Berger MS. Functional outcome after language mapping for glioma resection. N Engl J Med. 2008;358(1):1827.

  • 6

    De Witt Hamer PC, Robles SG, Zwinderman AH, Duffau H, Berger MS. Impact of intraoperative stimulation brain mapping on glioma surgery outcome: a meta-analysis. J Clin Oncol. 2012;30(20):25592565.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7

    Fernández Coello A, Moritz-Gasser S, Martino J, Martinoni M, Matsuda R, Duffau H. Selection of intraoperative tasks for awake mapping based on relationships between tumor location and functional networks. J Neurosurg. 2013;119(6):13801394.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8

    Roux FE, Boulanouar K, Lotterie JA, Mejdoubi M, LeSage JP, Berry I. Language functional magnetic resonance imaging in preoperative assessment of language areas: correlation with direct cortical stimulation. Neurosurgery. 2003;52(6):13351347.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9

    Bizzi A, Blasi V, Falini A, et al. Presurgical functional MR imaging of language and motor functions: validation with intraoperative electrocortical mapping. Radiology. 2008;248(2):579589.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 10

    Meier MP, Ilmberger J, Fesl G, Ruge MI. Validation of functional motor and language MRI with direct cortical stimulation. Acta Neurochir (Wien). 2013;155(4):675683.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 11

    Trinh VT, Fahim DK, Maldaun MV, et al. Impact of preoperative functional magnetic resonance imaging during awake craniotomy procedures for intraoperative guidance and complication avoidance. Stereotact Funct Neurosurg. 2014;92(5):315322.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12

    Kuchcinski G, Mellerio C, Pallud J, et al. Three-tesla functional MR language mapping: comparison with direct cortical stimulation in gliomas. Neurology. 2015;84(6):560568.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13

    Petrovich N, Holodny AI, Tabar V, et al. Discordance between functional magnetic resonance imaging during silent speech tasks and intraoperative speech arrest. J Neurosurg. 2005;103(2):267274.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14

    Qiu TM, Gong FY, Gong X, et al. Real-time motor cortex mapping for the safe resection of glioma: an intraoperative resting-state fMRI study. AJNR Am J Neuroradiol. 2017;38(11):21462152.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15

    Kapsalakis IZ, Kapsalaki EZ, Gotsis ED, et al. Preoperative evaluation with FMRI of patients with intracranial gliomas. Radiol Res Pract. 2012;2012:727810.

    • Search Google Scholar
    • Export Citation
  • 16

    Glasser MF, Coalson TS, Robinson EC, et al. A multi-modal parcellation of human cerebral cortex. Nature. 2016;536(7615):171178.

  • 17

    Kaplan E, Goodglass H, Weintraub S. Boston Naming Test. Lea & Febiger, 1983.

  • 18

    Mathôt S, Schreij D, Theeuwes J. OpenSesame: an open-source, graphical experiment builder for the social sciences. Behav Res Methods. 2012;44(2):314324.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19

    Klein A, Tourville J. 101 labeled brain images and a consistent human cortical labeling protocol. Front Neurosci. 2012;6:171.

  • 20

    Baker CM, Burks JD, Briggs RG, et al. A connectomic atlas of the human cerebrum—Chapter 6: The temporal lobe. Oper Neurosurg (Hagerstown). 2018;15(suppl 1):S245-S294.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21

    Håberg A, Kvistad KA, Unsgård G, Haraldseth O. Preoperative blood oxygen level-dependent functional magnetic resonance imaging in patients with primary brain tumors: clinical application and outcome. Neurosurgery. 2004;54(4):902915.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 22

    Holodny AI, Schulder M, Ybasco A, Liu WC. Translocation of Broca’s area to the contralateral hemisphere as the result of the growth of a left inferior frontal glioma. J Comput Assist Tomogr. 2002;26(6):941943.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 23

    Hossmann KA, Linn F, Okada Y. Bioluminescence and fluoroscopic imaging of tissue pH and metabolites in experimental brain tumors of cat. NMR Biomed. 1992;5(5):259264.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 24

    Segebarth C, Belle V, Delon C, et al. Functional MRI of the human brain: predominance of signals from extracerebral veins. Neuroreport. 1994;5(7):813816.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 25

    Ojemann JG, Miller JW, Silbergeld DL. Preserved function in brain invaded by tumor. Neurosurgery. 1996;39(2):253259.

  • 26

    Skirboll SS, Ojemann GA, Berger MS, Lettich E, Winn HR. Functional cortex and subcortical white matter located within gliomas. Neurosurgery. 1996;38(4):678685.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 27

    Thiel A, Herholz K, Koyuncu A, et al. Plasticity of language networks in patients with brain tumors: a positron emission tomography activation study. Ann Neurol. 2001;50(5):620629.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 28

    Duffau H. Brain plasticity: from pathophysiological mechanisms to therapeutic applications. J Clin Neurosci. 2006;13(9):885897.

  • 29

    Schulder M, Maldjian JA, Liu WC, et al. Functional image-guided surgery of intracranial tumors located in or near the sensorimotor cortex. J Neurosurg. 1998;89(3):412418.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30

    Lehéricy S, Duffau H, Cornu P, et al. Correspondence between functional magnetic resonance imaging somatotopy and individual brain anatomy of the central region: comparison with intraoperative stimulation in patients with brain tumors. J Neurosurg. 2000;92(4):589598.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 31

    Spena G, Nava A, Cassini F, et al. Preoperative and intraoperative brain mapping for the resection of eloquent-area tumors. A prospective analysis of methodology, correlation, and usefulness based on clinical outcomes. Acta Neurochir (Wien). 2010;152(11):18351846.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 32

    Baker CM, Burks JD, Briggs RG, et al. A connectomic atlas of the human cerebrum—Chapter 2: The lateral frontal lobe. Oper Neurosurg (Hagerstown). 2018;15(suppl 1):S10-S74.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 33

    Fogassi L, Luppino G. Motor functions of the parietal lobe. Curr Opin Neurobiol. 2005;15(6):626631.

  • 34

    Hazem SR, Awan M, Lavrador JP, et al. Middle frontal gyrus and area 55b: perioperative mapping and language outcomes. Front Neurol. 2021;12:646075.

  • 35

    Liu W, Lai JJ, Qu YM. Surgical treatment of gliomas involving the supplementary motor area in the superior frontal gyrus.Article in Chinese. Zhonghua Wai Ke Za Zhi. 2004;42(13):781783.

    • Search Google Scholar
    • Export Citation
  • 36

    Binder JR. The Wernicke area: modern evidence and a reinterpretation. Neurology. 2015;85(24):21702175.

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 649 649 76
PDF Downloads 750 750 95
EPUB Downloads 0 0 0