Real-time functional mapping: potential tool for improving language outcome in pediatric epilepsy surgery

Case report

Full access

Accurate language localization expands surgical treatment options for epilepsy patients and reduces the risk of postsurgery language deficits. Electrical cortical stimulation mapping (ESM) is considered to be the clinical gold standard for language localization. While ESM affords clinically valuable results, it can be poorly tolerated by children, requires active participation and compliance, carries a risk of inducing seizures, is highly time consuming, and is labor intensive. Given these limitations, alternative and/or complementary functional localization methods such as analysis of electrocorticographic (ECoG) activity in high gamma frequency band in real time are needed to precisely identify eloquent cortex in children.

In this case report, the authors examined 1) the use of real-time functional mapping (RTFM) for language localization in a high gamma frequency band derived from ECoG to guide surgery in an epileptic pediatric patient and 2) the relationship of RTFM mapping results to postsurgical language outcomes. The authors found that RTFM demonstrated relatively high sensitivity (75%) and high specificity (90%) when compared with ESM in a “next-neighbor” analysis. While overlapping with ESM in the superior temporal region, RTFM showed a few other areas of activation related to expressive language function, areas that were eventually resected during the surgery. The authors speculate that this resection may be associated with observed postsurgical expressive language deficits. With additional validation in more subjects, this finding would suggest that surgical planning and associated assessment of the risk/benefit ratio would benefit from information provided by RTFM mapping.

Abbreviations used in this paper:ECoG = electrocorticography; EEG = electroencephalography; ESM = electrical cortical stimulation mapping; RTFM = real-time functional mapping; SIGFRIED = SIGnal modeling For Real-time Identification and Event Detection.

Abstract

Accurate language localization expands surgical treatment options for epilepsy patients and reduces the risk of postsurgery language deficits. Electrical cortical stimulation mapping (ESM) is considered to be the clinical gold standard for language localization. While ESM affords clinically valuable results, it can be poorly tolerated by children, requires active participation and compliance, carries a risk of inducing seizures, is highly time consuming, and is labor intensive. Given these limitations, alternative and/or complementary functional localization methods such as analysis of electrocorticographic (ECoG) activity in high gamma frequency band in real time are needed to precisely identify eloquent cortex in children.

In this case report, the authors examined 1) the use of real-time functional mapping (RTFM) for language localization in a high gamma frequency band derived from ECoG to guide surgery in an epileptic pediatric patient and 2) the relationship of RTFM mapping results to postsurgical language outcomes. The authors found that RTFM demonstrated relatively high sensitivity (75%) and high specificity (90%) when compared with ESM in a “next-neighbor” analysis. While overlapping with ESM in the superior temporal region, RTFM showed a few other areas of activation related to expressive language function, areas that were eventually resected during the surgery. The authors speculate that this resection may be associated with observed postsurgical expressive language deficits. With additional validation in more subjects, this finding would suggest that surgical planning and associated assessment of the risk/benefit ratio would benefit from information provided by RTFM mapping.

In the US, epilepsy affects nearly 326,000 children younger than 16 years; 10%–30% of these cases are resistant to antiepileptic medications.3 Epilepsy surgery is a potentially curative option for these patients. However, it is often not given full consideration due to concern about the impact of surgery on language and cognitive function. Deficits may result when there is an overlap between seizure focus and language-specific regions, similar to risks in adults. However, in children there is added concern about the risk of interrupting developing networks that are connected to, but do not necessarily directly overlap, seizure onset zones. Accurate language localization in children may provide much needed surgical guidance and expand surgical epilepsy treatment options, thereby reducing the risk of postsurgical language deficit and benefiting children's quality of life, educational capacity, and long-term employment potential.

Electrical cortical stimulation mapping (ESM) is considered to be the clinical gold standard for language localization.18 Although many studies have shown the value of ESM mapping, it can be poorly tolerated by children, requires active participation and compliance, carries a risk of inducing seizures, and is highly time and effort consuming. Electrical cortical stimulation mapping utilizes a “reversible lesion” approach by inhibiting cortical function through electrical stimulation of subdural electrode pairs. This necessarily involves stimulating one pair of electrodes at a time, adding to the challenges noted above. Also, ESM has some restrictions related to the timing of stimulation and subjectivity of interpretation when there is language interruption, especially when there is subtle language interruption (for example, hesitation vs speech arrest). Hence, an activation (rather than suppression) task that would allow for instant visualization of activated networks during language tasks would offer considerable advantages. Functional imaging techniques such as functional MRI share some of the limitations of ESM, with the added concern about specificity and spatial resolution of areas that are engaged during activation tasks.

Many studies over the past 15 years have shown that electrical signals recorded directly from the surface of the brain (electrocorticography [ECoG]) in patients with subdural grids implanted for clinical purposes readily detect and localize signal changes due to different auditory, motor, or language tasks.10,24 Also, many studies have suggested that ECoG amplitudes in the high gamma band (> 70 Hz) may be most valuable for assessing the task-related activity of cortical neurons directly underneath the electrodes.1,8,9,19,27 Therefore, task-related test batteries based on the analysis of ECoG high gamma activity can be successfully applied for localization of eloquent cortex. In this way, ECoG-based localization utilizing offline processing approaches has been successfully demonstrated for sensory-motor,1,9,19,27 auditory,7 and language function.12,28,29

Schalk et al.20,21 went beyond the offline processing of task-related ECoG changes. They implemented a real-time ECoG signal analysis approach (real-time functional mapping [RTFM]). This created the potential for instant visualization of task-related changes in ECoG signals, which may be used in clinical settings, either directly at the patient's bedside or intraoperatively.

In our previous study14 we demonstrated the feasibility and usefulness of a new RTFM stimulation paradigm implementation during presurgical evaluation. We showed that block design paradigms used in functional MRI, such as story processing or verb generation task, may be optimal for this purpose. However, it is unclear how RTFM activation is related to postsurgical language outcomes.

Case Report

The institutional review board of the Florida Hospital for Children, Orlando, Florida, approved the study protocol.

History and Examination

This 13-year-old girl, diagnosed with Turner syndrome shortly after birth, presented with nearly daily seizures despite being on oxcarbazepine monotherapy. Several other antiseizure drugs were also ineffective at managing the patient's complex focal seizures with occasional generalization that began at the age of 2 years. She had experienced cognitive decline for the past few years. Additional patient characteristics are presented in Table 1. The patient underwent the standard preoperative assessment to determine the epileptogenic area and eloquent cortex.

TABLE 1:

Characteristics of the patient, a 13-year-old right-handed girl*

ParameterPreopTime of Testing
Postop
2 Mos1 Yr
diagnosis
 epilepsy focustemporalNANA
 sideleftNANA
 epilepsy onset (yrs)2NANA
 epilepsy duration (yrs)11NANA
WISC-IV & IQ scores
 VCI67NA63
 PRI73NA67
 WMI80NA59
 PSI59NA53
 FSIQ62NA63
Likert scale
 overall severity0
 overall bothersome0
 treatment change+7

FSIQ = Full Scale IQ; NA = not applicable; PRI = Perceptual Reasoning Index, PSI = Processing Speed Index, VCI = Verbal Comprehension Index, WISC-IV = Wechsler Intelligence Scale for Children, Fourth Edition; WMI = Working Memory Index.

General test interpretation = 1 standard deviation or 15 standard score reduction.

The Likert scale ranges from 0 to 7 for overall severity and overall bothersome and −7 to +7 for treatment change.

Presurgical Noninvasive Evaluation

The evaluation included a neuropsychological examination, long-term video-electroencephalography (video-EEG) to review seizure onset and clinical semiology, MRI with epilepsy surgery protocol, and FDG-PET. Ictal EEG showed diffuse rhythmic theta activity evolving in the left fronto-temporal region, MRI demonstrated diffuse left temporal lobe atrophy without clear evidence of mesial temporal sclerosis, and FDG-PET revealed extensive hypometabolism involving both mesial and lateral structures in the left temporal lobe.

Presurgical Neuropsychological Evaluation

A neuropsychologist (C.M.S.) performed the neuropsychological evaluation, which included the Wechsler Intelligence Scale for Children, Fourth Edition. The relevant index scores are provided in Table 1. Of note are the patient's extremely low verbal reasoning and concept formation skills, as measured by the Verbal Comprehension Index.

Consistent with her left temporal lobe epilepsy, the patient also showed extremely low performance on visual confrontation naming (with latent responses and frequent paraphasic errors), very low phonological fluency, borderline semantic fluency, and low average verbal memory. Her nonverbal reasoning skills, as measured by the Perceptual Reasoning Index was borderline. Her processing speed, as measured by Processing Speed Index, was extremely low, and her working memory, as measured by Working Memory Index, was low average.

Invasive EEG Monitoring

After a comprehensive review of preoperative data by the epileptologist and neurosurgeon, subdural grid placement for seizure localization and language mapping were recommended since the ictal semiology, postictal characteristics, and review of neuropsychological testing implicated seizure onset zone close to language cortex. For invasive monitoring, the intracranial electrode placement included a 16-contact grid covering the inferior frontal area and a 32-contact grid covering the temporal lobe. A single 6-contact electrode strip was placed in each of the following regions: temporal pole, anterior basal temporal, and middle basal temporal. An 8-contact electrode strip covered the posterior basal temporal region.

MRI, CT, and ECoG Coregistration

Three-dimensional volumetric MRI, performed using a 3-T GE scanner (T1 FSPGR, 512 × 512 matrix, FOV 250, 1 mm slice thickness), and a high-resolution CT scanning (512 × 512 matrix, FOV 250, 1-mm slice thickness) were carried out to precisely visualize the location of the subdural electrodes on the brain surface. The MRI and CT were coregistered and 3D segmentation of the brain surface and grid positions was performed using Analyze 11 (Analyze Direct).

Functional Mapping

Two types of functional mapping were used to determine localization of eloquent language cortex: ESM and RTFM.

ESM. Electrical cortical stimulation mapping was performed for identification of eloquent language cortex. Stimulations consisted of 5-second trains of 200-msec square-wave pulses of alternating polarity at 50 Hz applied to the adjacent electrode pairs using an Ojemann Cortical Stimulator (Integra LifeSciences Corp.).

Functional mapping was conducted by stimulating adjacent pairs of electrodes on the left temporal grid. Stimulation was initiated at low amperage (4–6 mA) and increased in 2-mA increments until an after-discharge was elicited, a functional response was produced, or a maximum of 10 mA was attained. Sites where stimulation induced speech arrest or other language disruptions were noted and recorded.

Language mapping was conducted using a variety of tasks. Prior to conducting the mapping, baseline assessment of visual confrontation naming and reading was conducted with the mapping stimuli. It was done to identify stimuli that the patient could readily and consistently respond to with a high degree of accuracy. The patient was able to maintain optimal responding to naming stimuli, and most of the mapping was conducted with repetition of the visual naming items (that is, name pictures of objects presented on a computer screen), or responding to pictures with a brief sentence identifying an object (for example, “This is a bird”). These tasks were alternated with repetition of over-learned sequences (for example, counting and reciting the alphabet), and singing along with videos. Areas in which a change in language was noted during stimulation were restimulated to verify the responses before being recorded. One hundred percent interrater reliability between 2 neuropsychologists was achieved.

Baseline assessment prior to ESM also included testing for receptive language. However, baseline testing indicated that the patient was not able to consistently respond in a reliable manner adequate for identifying receptive language due to the complexity of the task and the relatively brief stimulation train (5 seconds).

RTFM. Electrocorticography signals from subdural electrodes implanted for presurgical evaluation were separated into a clinical and a research signal path through a splitter box. This ensured that the ECoG recording for seizures remained uninterrupted during the procedure. The clinical signal path went to the Nihon Kohden EEG 1200 system for clinical diagnosis, while the research signal path went to an array of g.USBamp amplification/digitization devices (g.tec). The g.USBamp system was controlled by a dedicated research personal computer that records ECoG signals and delivers auditory and visual stimuli through speakers and a monitor.

BCI2000 software22 acquired signals from the g.USBamps, provided visual/auditory stimulation, and performed real-time analyses using the SIGnal modeling For Real-time Identification and Event Detection (SIGFRIED) module (Schalk et al.21 and Brunner et al.5). Subdural electrodes away from the seizure zone and eloquent areas served as the ground and reference electrodes. The patient sat upright in a raised bed in front of a computer screen and was instructed to relax and remain as still as possible.

Baseline cortical activity was first recorded for about 6 minutes followed by administration of several different paradigms, including those that assess receptive and expressive language function. More detailed paradigm descriptions can be found in Korostenskaja et al.14 For determining the areas associated with speech comprehension (Wernicke's areas), stories were presented to the subject through the speakers.25 In contrast to ESM testing in which the patient was unable to respond in a reliable manner adequate for identifying receptive language, we were able to successfully conduct RTFM testing for receptive function localization. For determining expressive language function, we used a picture-naming paradigm, where the subject was asked to name pictures of different objects presented on the computer screen.15

The results from the real-time signal analyses (described in detail by Schalk et al.21 and Brunner et al.5) were visualized at the bedside using a topographic interface. For each task (for example, story processing) the interface included a display that contained 1 circle at each electrode's location. For each circle, the size of the circle and its tint were proportional to the coefficient of determination (r2), calculated between brain signals at rest and brain signals during the particular task. Thus, a large red circle represented a large statistical difference between the corresponding task and rest, whereas a small black dot indicated a small statistical difference (Fig. 1).

Fig. 1.
Fig. 1.

Cortical RTFM activation map indicating language activation elicited during picture naming (left) and the story processing task (right). Locations with significant levels of activation are presented as large red circles; grid placement locations are indicated as black dots.

Establishing Concordance Between Functional Mapping Modalities

We referred to electrode sites at which we observed RTFM activation (red circle on RTFM map) as RTFM positive, and those at which we did not as RTFM negative. We also referred to electrode sites at which ESM affected task performance as ESM positive and those at which they did not as ESM negative, (for the list of main findings, see Table 3). The sensitivity and specificity of RTFM were calculated only for sites where both RTFM and ESM were tested (see Tables 4 and 5). The RTFM sensitivity was calculated as the percentage of sites that were both RTFM positive and ESM positive among all ESM positive sites, and RTFM specificity was calculated as the percentage of sites that were both RTFM negative and ESM negative among all ESM negative sites.

TABLE 3:

Calculation of sensitivity and specificity of language mapping with RTFM and ESM*

ParameterFormulaActual Data
sensitivity(ESM pos + RTFM pos)/all ESM pos × 100%1/4 × 100% = 25%
specificity(ESM neg + RTFM neg)/all ESM neg × 100%10/11 × 100% = 90.0%

neg = negative; pos = positive.

TABLE 4:

Calculation of sensitivity and specificity of language mapping with RTFM and ESM by using “next-neighbors” comparison method

ParameterFormulaActual Data
sensitivity(ESM pos + RTFM pos)/all ESM pos × 100%3/4 × 100% = 75%
specificity(ESM neg + RTFM neg)/all ESM neg × 100%10/11 × 100% = 90.0%
TABLE 5:

Results of ESM and RTFM mapping*

ElectrodeESM ResultsRTFM Result
LT1not tested
LT2not tested
LT3not tested
LT4not tested
LT5not tested
LT6not tested
LT7not tested
LT8not tested
LT9not tested
LT10not tested
LT11not tested
LT12not tested
LT13not tested
LT14
LT15
LT16
LT17not tested+
LT18not tested
LT19
LT20
LT21
LT22
LT23
LT24+
LT25not tested
LT26not tested
LT27
LT28
LT29+
LT30+
LT31+
LT32++
LTP1not tested+
LTP2not tested
LTP3not tested
LTP4not tested
LTP5not tested
LTP6not tested

+ = positive response; − = negative response.

Operation

After the presurgical evaluation, the patient underwent a tailored temporal lobectomy according to the intracranial ictal EEG onset and the ESM language map. Based on the surgical report and postsurgical MRI results (Fig. 2), the resection included removal of the left middle, inferior, and anterior part of the superior temporal gyrus. The neurosurgeon worked posteriorly up until approximately 1.5 cm anterior to the area identified as the “language cortex” by ESM in the superior temporal gyrus. Therefore, there was partial sparing of the posterior and superior temporal lobes to avoid possible postsurgical expressive language deficits predicted by ESM mapping. The resection also included medial temporal lobe structures together with the hippocampus and amygdala. Importantly, the anterior temporal pole was removed, which was associated with expressive language function demonstrated by RTFM (Fig. 3).

Fig. 2.
Fig. 2.

Magnetic resonance images obtained before (A and B) and after (C and D) resection of epileptic foci. The imaging results are presented in axial (A and C) and coronal (B and D) planes. Radiological side labeling is used. The presurgical images show that there is abnormal T2 and FLAIR T2 signal involving the entire left mesial temporal lobe, including the amygdala, hippocampus, and parahippocampal gyrus. There is mild thinning throughout, most notable at the amygdala. Findings suggest mesial temporal lobe sclerosis. Focal cortical dysplasia could also be considered. The postsurgical images show that there has been interval resection of the left mesial temporal lobe and anterior temporal pole as well as the lateral portion of the inferior, middle, and partial superior temporal lobe. There is partial sparing of the posterior and superior temporal lobe.

Fig. 3.
Fig. 3.

Results of the left temporal lobe's eloquent language cortex mapping with ESM and RTFM. Both ESM and RTFM results are overlaid offline on the 3D-rendered cortical surface of the patient's brain. Electrode locations are presented as blue dots.

Postoperative Course

Seizure Outcome

The patient had no seizures after surgery. Antiseizure medication (oxcarbazepine) was discontinued at the 12-month postoperative follow-up, and the patient remained seizure free (at the 18-months postoperative follow-up). The pathology specimen revealed mesial temporal sclerosis without evidence of focal cortical dysplasia.

Postsurgical Neuropsychological Evaluation

A 2-month acute postsurgical neuropsychological testing follow-up showed expressive language decline (Table 2). The patient demonstrated very low performance in visual confrontation naming compared with same-aged peers. The patient exhibited a mild decline in phonemic fluency and a noticeable decline in semantic fluency. Furthermore, the patient's mother's report during clinical interviews suggested a decline in language functioning within the patient's everyday environment.

TABLE 2:

Summary of neuropsychological examination results for patient's language functioning*

Time of TestingBNT/SSDKEFS Verbal FluencyCELF-IV
Letter Fluency Total Correct Raw Score/SSCategory Fluency Total Correct Raw Score/SS Category Switching Total Correct Raw Score/SSCategory Switching Total Switching Accuracy Raw Score/SSExpressive Language Composite/SSReceptive Language Content Composite/SSLanguage ContentLanguage Memory
preop38/6310/6521/759/858/90
postop
 2 mos39/666/559/557/706/809/5723/8622/8415/69
 1 yr44/743/558/557/706/853/4511/6212/643/45

BNT = Boston Naming Test; CELF-IV = Clinical Evaluation of Language Fundamentals, 4th Edition; DKEFS = Delis Kaplan Executive Function System; SS = standard score; — = not administered at this phase.

General test interpretation = 1 standard deviation or 15 standard score reduction.

Twelve-month postsurgical neuropsychological testing follow-up revealed a moderate (or 24 standard score point) decline in receptive language skills compared with postacute assessment. The patient demonstrated more prominent difficulty with sentence repetition and comprehension of contextually based information. She also showed a severe (or 45 standard score point) decline in her ability to make semantic associations. Notably, she showed a mild to moderate decline in complex attention or working memory (21 standard point reduction), which may have been a contributing factor to her reduced performance on complex verbal comprehension tasks. Consistent with her longstanding history of epilepsy, the patient exhibited ongoing confrontation naming as well as phonological fluency deficits that have remained stable despite surgical intervention (that is, no significant improvement over longer follow-up duration). In summary, although the patient exhibited low baseline performance, she exhibited further deterioration in category fluency skills, which are typically mediated by the temporal lobe. At the 1-year follow-up the patient also showed a more dramatic decline in receptive language as well as complex attention. These findings reflect a true decline due to surgery as well as a developmental lag compared with her same-aged peers.

Discussion

This study for the first time demonstrates the relationship between RTFM and postoperative language outcome as measured by detailed neuropsychological testing. It also provides important details on how RTFM can contribute to presurgical functional language mapping. Understandably, postsurgical deficits cannot always be predicted with ESM.11,13 This may make RTFM a valuable technique with the potential to reduce postsurgical language morbidity. We took advantage of the detailed ESM mapping conducted in our facilities and demonstrated that when the “next-neighbor” approach is used, RTFM shows high sensitivity and specificity during mapping of expressive language function when compared with ESM.

ESM Versus RTFM

Findings from published studies that attempt to delineate the relationship between ECoG-related changes and ESM mapping are inconsistent. For example, Brunner et al.5 showed highly congruent results between these 2 mapping approaches with a motor task—a “next-neighbor” evaluation showed no false negatives, and only 0.46% and 1.10% false positives for hand and tongue maps, respectively. These authors speculated about possible higher sensitivity and lower specificity of the RTFM procedure when compared with ESM. In their real-time language mapping study, Miller et al.17 argued that RTFM might have lower sensitivity and higher specificity to detect language-specific regions, which is the opposite from what was observed in the motor modality by Brunner et al.5

However, offline analysis of the ECoG signal in the frequency domain “electrocorticographic frequency alteration mapping” (see the study by Leuthardt et al.16) did not match ESM results accurately on a language task with an 83.9% sensitivity and a 40.4% specificity found in a study with 7 patients.29 In addition, Brown et al.4 evaluated the concordance between ECoG and ESM findings in 4 children. They reported larger activation in ECoG data when compared with ESM at similar electrode sites, but they did not evaluate sensitivity and specificity of ECoG-based activation relative to ESM. Their offline-derived study results are in line with our observed real-time language activation pattern (based on SIGFRIED methodology), which is more broadly represented in RTFM than in ESM.

Several more studies have been published that demonstrated discrepancies between ECoG mapping and ESM findings. Sinai et al.23 showed 38% sensitivity and 78% specificity for a picture-naming task for ECoG-based mapping against ESM. Similar results with low sensitivity and high specificity of ECoG-based mapping when compared with ESM mapping were demonstrated in a study by Bauer et al.2 in a picture-naming task. These authors suggested that the different nature of ESM and ECoG-based mapping (“all-or-none” ESM principle vs variable and changing in time ECoG activation) may be responsible for the discrepancies. Moreover, different from ECoG-based mapping, ESM does not allow testing language function on the electrode sites, associated with mouth-related motor function. This, in turn, could have affected the comparison of ECoG-based mapping and ESM results. A study by Cervenka et al.,6 who used both visual object naming and auditory descriptive naming tasks, revealed discrepancies between sensitivity and specificity of ECoG-based mapping to predict postsurgical language deficits. The authors suggested that taking into account the results from both mapping approaches would lead to improved postsurgical language outcomes.

Possibility to Test Receptive Language and Additional Electrode Sites

The language network distribution was more broadly represented with RTFM than with ESM. The eloquent language cortex determined by RTFM was located more anterior and closer to the resection margin than the language cortex determined by ESM, although similar tasks were implemented (for example, picture naming). This may reflect essential differences between RTFM and ESM approaches (“activation” vs “inhibition”). The wider distribution of RTFM activation is in line with the ability of RTFM to visualize activation of a language network related to a particular task. Additional advantages of RTFM include the ability to more extensively test language function. This is clinically relevant since children may have variable abilities to cooperate and respond to language tasks during mapping, and the parameters of ESM are limiting. In addition, medications administered to reduce seizures during stimulation produce fatigue. Stimulation duration is limited due to increased risk of after-discharges with increasing length of stimulation, and children with processing-speed deficits (as was the case for our patient) may not be able to respond with adequate reliability or accuracy within the limited time frame. In this patient, we were unable to perform receptive language mapping with ESM, but RTFM identified areas of activation related to language processing. In addition, the length of time needed for sustained cooperation may present a challenge for many children, resulting in the need to select regions of interest rather than mapping the entire electrode array. In contrast, all implanted electrodes can be monitored simultaneously with RTFM. Moreover, there is no need to limit or artificially constrain the response parameters (for example, requiring responses within a particular time frame). The advantages of flexible brain mapping approaches that are free of time constraints were demonstrated by Vansteensel et al.26 These authors were able to analyze ECoG activity in frequency domains by choosing periods of ECoG data related to different movements performed by patients during real-time events, such as grabbing eating utensils and other goal-directed movements. Finally, RTFM may improve prediction of postsurgical outcomes because more ecologically oriented language responses can be recorded and mapped. While ESM relies mainly on eliciting specific language responses to regulated presentations of stimuli within a limited time frame, RTFM is capable of recording activity during conversations and other more natural acts.

Relation to Postsurgical Language Outcome

The main purpose of presurgical mapping is the reduction of functional morbidity postsurgery. The RTFM approach is still in its early development stages, and studies determining its usefulness for localization of eloquent cortex have yet to be performed. Taking advantage of the comprehensive approach used by our neuropsychologists for assessing language function pre- and postsurgery, we have made it our priority to evaluate the impact of RTFM on postsurgical language outcomes. We examined the relationship between RTFM and ESM findings in conjunction with cortical resection data and comprehensive postsurgical neuropsychological testing. Real-time functional mapping demonstrated relatively high sensitivity and high specificity when compared with ESM by using a “next-neighbors” approach. Moreover, it provided valuable results not available from ESM mapping. For example, in addition to areas stimulated with ESM showing language in the posterior temporal lobe, RTFM showed several other areas of activation related to expressive language function, which were eventually resected during the surgery. This resection may have been associated with observed postsurgical expressive language deficits. It is well known that patients who undergo dominant anterior temporal lobectomy, including medial temporal lobe structures, may exhibit language and verbal memory deficits after surgery. In our patient, immediate postoperative changes in language were found in areas related to function that was activated during RTFM. Real-time functional mapping was also able to test language comprehension with narrative information. This language skill had a significant decline at 1-year follow-up. Detection of this area with RTFM could have had direct implications on this patient's outcome. The mapping results allow us to speculate that surgical planning and assessment of the risk/benefit ratio may benefit from RTFM information provided in conjunction with ESM results.

Although RTFM identified areas relevant to predicting outcome, not all of the postoperative changes could be predicted from the ESM and RTFM alone. This patient showed a significant decline in working memory, which would also affect performance on many language tasks that have a working memory load (for example, recalling sentences, concepts, and following directions). Additionally, several language tasks that have a verbal memory load also showed decline, a cognitive ability that frequently shows decline following dominant temporal lobectomy. While this report focuses on predicting language outcomes, RTFM procedures have the potential for developing mapping for other cognitive functions that cannot be mapped with ESM.

Conclusions

Our case study illustrated that RTFM may be a valuable tool for guiding and assisting ESM or other functional mapping methodologies, thus decreasing duration and improving the accuracy of eloquent cortex localization, resulting in reduction of postsurgical language morbidity. Whether RTFM truly provides more accurate language mapping results than ESM, and thus better neuropsychological outcomes, needs to be validated through a large-scale prospective study.

Acknowledgments

We would like to thank Ashley Webster for her help in technical matters pertaining to ECoG recordings. We also would like to express our gratitude to Drs. Scott Holland and Jennifer Vannest for sharing their protocol for mapping receptive language function (“story processing” task). Lastly, we would like to thank Tere Richard for her help with manuscript editing.

Disclosure

The authors have no financial or personal relationships to disclose. Dr. Schalk states that his institution licensed the underlying technology to g.tec Medical Engineering. This study was funded in part by the US Army Research Office (W911NF-08-1-0216 [G.S.]) and the NIH/NIBIB (EB00856 [G.S.]). Funding organizations had no role in the study design, data collection, analysis, interpretation, or writing of results.

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 to the study and manuscript preparation include the following. Conception and design: Korostenskaja, Brunner, Schalk, Baumgartner, Lee. Acquisition of data: Korostenskaja, Chen. Analysis and interpretation of data: Korostenskaja, Chen, Salinas, Westerveld, Cook, Lee. Drafting the article: Korostenskaja, Salinas. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Korostenskaja. Administrative/technical/material support: Korostenskaja, Brunner, Schalk. Study supervision: Korostenskaja, Lee.

This article contains some figures that are displayed in color online but in black-and-white in the print edition.

References

  • 1

    Anderson NRBlakely TSchalk GLeuthardt ECMoran DW: Electrocorticographic (ECoG) correlates of human arm movements. Exp Brain Res 223:1102012

  • 2

    Bauer PRVansteensel MJBleichner MGHermes DFerrier CHAarnoutse EJ: Mismatch between electrocortical stimulation and electrocorticography frequency mapping of language. Brain Stimulat 6:5245312013

  • 3

    Berg ATShinnar SLevy SRTesta FMSmith-Rapaport SBeckerman B: Early development of intractable epilepsy in children: a prospective study. Neurology 56:144514522001

  • 4

    Brown ECRothermel RNishida MJuhász CMuzik OHoechstetter K: In vivo animation of auditory-language-induced gamma-oscillations in children with intractable focal epilepsy. Neuroimage 41:112011312008

  • 5

    Brunner PRitaccio ALLynch TMEmrich JFWilson JAWilliams JC: A practical procedure for real-time functional mapping of eloquent cortex using electrocorticographic signals in humans. Epilepsy Behav 15:2782862009

  • 6

    Cervenka MCCorines JBoatman-Reich DFEloyan ASheng XFranaszczuk PJ: Electrocorticographic functional mapping identifies human cortex critical for auditory and visual naming. Neuroimage 69:2672762013

  • 7

    Crone NEBoatman DGordon BHao L: Brazier Award-winning article, 2001. Induced electrocorticographic gamma activity during auditory perception. Clin Neurophysiol 112:5655822001

  • 8

    Crone NEMiglioretti DLGordon BLesser RP: Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. II Event-related synchronization in the gamma band. Brain 121:230123151998

  • 9

    Crone NEMiglioretti DLGordon BSieracki JMWilson MTUematsu S: Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. I Alpha and beta event-related desynchronization. Brain 121:227122991998

  • 10

    Crone NESinai AKorzeniewska A: High-frequency gamma oscillations and human brain mapping with electrocorticography. Prog Brain Res 159:2752952006

  • 11

    Davies KGRisse GLGates JR: Naming ability after tailored left temporal resection with extraoperative language mapping: increased risk of decline with later epilepsy onset age. Epilepsy Behav 7:2732782005

  • 12

    Edwards ENagarajan SSDalal SSCanolty RTKirsch HEBarbaro NM: Spatiotemporal imaging of cortical activation during verb generation and picture naming. Neuroimage 50:2913012010

  • 13

    Hermann BPPerrine KChelune GJBarr WLoring DWStrauss E: Visual confrontation naming following left anterior temporal lobectomy: a comparison of surgical approaches. Neuropsychology 13:391999

  • 14

    Korostenskaja MWilson AJRose DFBrunner PSchalk GLeach J: Real-time functional mapping with electrocorticography in pediatric epilepsy: comparison with fMRI and ESM findings. Clin EEG Neurosci 2013

  • 15

    Kunii NKamada KOta TKawai KSaito N: A detailed analysis of functional magnetic resonance imaging in the frontal language area: a comparative study with extraoperative electrocortical stimulation. Neurosurgery 69:5905972011

  • 16

    Leuthardt ECMiller KAnderson NRSchalk GDowling JMiller J: Electrocorticographic frequency alteration mapping: a clinical technique for mapping the motor cortex. Neurosurgery 60:4 Suppl 22602712007

  • 17

    Miller KJAbel TJHebb AOOjemann JG: Rapid online language mapping with electrocorticography. Clinical article. J Neurosurg Pediatr 7:4824902011

  • 18

    Ojemann GOjemann JLettich EBerger M: Cortical language localization in left, dominant hemisphere. An electrical stimulation mapping investigation in 117 patients. J Neurosurg 71:3163261989

  • 19

    Qian TZhou WLing ZGao SLiu HHong B: Fast presurgical functional mapping using task-related intracranial high gamma activity. Laboratory investigation. J Neurosurg 119:26362013

  • 20

    Schalk GBrunner PGerhardt LABischof HWolpaw JR: Brain-computer interfaces (BCIs): detection instead of classification. J Neurosci Methods 167:51622008

  • 21

    Schalk GLeuthardt ECBrunner POjemann JGGerhardt LAWolpaw JR: Real-time detection of event-related brain activity. Neuroimage 43:2452492008

  • 22

    Schalk GMcFarland DJHinterberger TBirbaumer NWolpaw JR: BCI2000: a general-purpose brain-computer interface (BCI) system. IEEE Trans Biomed Eng 51:103410432004

  • 23

    Sinai ABowers CWCrainiceanu CMBoatman DGordon BLesser RP: Electrocorticographic high gamma activity versus electrical cortical stimulation mapping of naming. Brain 128:155615702005

  • 24

    Sinai ACrone NEWied HMFranaszczuk PJMiglioretti DBoatman-Reich D: Intracranial mapping of auditory perception: event-related responses and electrocortical stimulation. Clin Neurophysiol 120:1401492009

  • 25

    Vannest JJKarunanayaka PRAltaye MSchmithorst VJPlante EMEaton KJ: Comparison of fMRI data from passive listening and active-response story processing tasks in children. J Magn Reson Imaging 29:9719762009

  • 26

    Vansteensel MJBleichner MGDintzner LTAarnoutse EJLeijten FSHermes D: Task-free electrocorticography frequency mapping of the motor cortex. Clin Neurophysiol 124:116911742013

  • 27

    Wray CDBlakely TMPoliachik SLPoliakov AMcDaniel SSNovotny EJ: Multimodality localization of the sensorimotor cortex in pediatric patients undergoing epilepsy surgery. Clinical article. J Neurosurg Pediatr 10:162012

  • 28

    Wu HCNagasawa TBrown ECJuhasz CRothermel RHoechstetter K: γ-oscillations modulated by picture naming and word reading: intracranial recording in epileptic patients. Clin Neurophysiol 122:192919422011

  • 29

    Wu MWisneski KSchalk GSharma MRoland JBreshears J: Electrocorticographic frequency alteration mapping for extraoperative localization of speech cortex. Neurosurgery 66:E407E4092010

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

Address correspondence to: Milena Korostenskaja, Ph.D., Functional Brain Mapping and Brain-Computer Interface Lab, Florida Hospital for Children, P.O. Box 207, 601 Rollins St., Orlando, FL 32803. email: milena.korostenskaja@gmail.com.

Please include this information when citing this paper: published online July 4, 2014; DOI: 10.3171/2014.6.PEDS13477.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Cortical RTFM activation map indicating language activation elicited during picture naming (left) and the story processing task (right). Locations with significant levels of activation are presented as large red circles; grid placement locations are indicated as black dots.

  • View in gallery

    Magnetic resonance images obtained before (A and B) and after (C and D) resection of epileptic foci. The imaging results are presented in axial (A and C) and coronal (B and D) planes. Radiological side labeling is used. The presurgical images show that there is abnormal T2 and FLAIR T2 signal involving the entire left mesial temporal lobe, including the amygdala, hippocampus, and parahippocampal gyrus. There is mild thinning throughout, most notable at the amygdala. Findings suggest mesial temporal lobe sclerosis. Focal cortical dysplasia could also be considered. The postsurgical images show that there has been interval resection of the left mesial temporal lobe and anterior temporal pole as well as the lateral portion of the inferior, middle, and partial superior temporal lobe. There is partial sparing of the posterior and superior temporal lobe.

  • View in gallery

    Results of the left temporal lobe's eloquent language cortex mapping with ESM and RTFM. Both ESM and RTFM results are overlaid offline on the 3D-rendered cortical surface of the patient's brain. Electrode locations are presented as blue dots.

References

1

Anderson NRBlakely TSchalk GLeuthardt ECMoran DW: Electrocorticographic (ECoG) correlates of human arm movements. Exp Brain Res 223:1102012

2

Bauer PRVansteensel MJBleichner MGHermes DFerrier CHAarnoutse EJ: Mismatch between electrocortical stimulation and electrocorticography frequency mapping of language. Brain Stimulat 6:5245312013

3

Berg ATShinnar SLevy SRTesta FMSmith-Rapaport SBeckerman B: Early development of intractable epilepsy in children: a prospective study. Neurology 56:144514522001

4

Brown ECRothermel RNishida MJuhász CMuzik OHoechstetter K: In vivo animation of auditory-language-induced gamma-oscillations in children with intractable focal epilepsy. Neuroimage 41:112011312008

5

Brunner PRitaccio ALLynch TMEmrich JFWilson JAWilliams JC: A practical procedure for real-time functional mapping of eloquent cortex using electrocorticographic signals in humans. Epilepsy Behav 15:2782862009

6

Cervenka MCCorines JBoatman-Reich DFEloyan ASheng XFranaszczuk PJ: Electrocorticographic functional mapping identifies human cortex critical for auditory and visual naming. Neuroimage 69:2672762013

7

Crone NEBoatman DGordon BHao L: Brazier Award-winning article, 2001. Induced electrocorticographic gamma activity during auditory perception. Clin Neurophysiol 112:5655822001

8

Crone NEMiglioretti DLGordon BLesser RP: Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. II Event-related synchronization in the gamma band. Brain 121:230123151998

9

Crone NEMiglioretti DLGordon BSieracki JMWilson MTUematsu S: Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. I Alpha and beta event-related desynchronization. Brain 121:227122991998

10

Crone NESinai AKorzeniewska A: High-frequency gamma oscillations and human brain mapping with electrocorticography. Prog Brain Res 159:2752952006

11

Davies KGRisse GLGates JR: Naming ability after tailored left temporal resection with extraoperative language mapping: increased risk of decline with later epilepsy onset age. Epilepsy Behav 7:2732782005

12

Edwards ENagarajan SSDalal SSCanolty RTKirsch HEBarbaro NM: Spatiotemporal imaging of cortical activation during verb generation and picture naming. Neuroimage 50:2913012010

13

Hermann BPPerrine KChelune GJBarr WLoring DWStrauss E: Visual confrontation naming following left anterior temporal lobectomy: a comparison of surgical approaches. Neuropsychology 13:391999

14

Korostenskaja MWilson AJRose DFBrunner PSchalk GLeach J: Real-time functional mapping with electrocorticography in pediatric epilepsy: comparison with fMRI and ESM findings. Clin EEG Neurosci 2013

15

Kunii NKamada KOta TKawai KSaito N: A detailed analysis of functional magnetic resonance imaging in the frontal language area: a comparative study with extraoperative electrocortical stimulation. Neurosurgery 69:5905972011

16

Leuthardt ECMiller KAnderson NRSchalk GDowling JMiller J: Electrocorticographic frequency alteration mapping: a clinical technique for mapping the motor cortex. Neurosurgery 60:4 Suppl 22602712007

17

Miller KJAbel TJHebb AOOjemann JG: Rapid online language mapping with electrocorticography. Clinical article. J Neurosurg Pediatr 7:4824902011

18

Ojemann GOjemann JLettich EBerger M: Cortical language localization in left, dominant hemisphere. An electrical stimulation mapping investigation in 117 patients. J Neurosurg 71:3163261989

19

Qian TZhou WLing ZGao SLiu HHong B: Fast presurgical functional mapping using task-related intracranial high gamma activity. Laboratory investigation. J Neurosurg 119:26362013

20

Schalk GBrunner PGerhardt LABischof HWolpaw JR: Brain-computer interfaces (BCIs): detection instead of classification. J Neurosci Methods 167:51622008

21

Schalk GLeuthardt ECBrunner POjemann JGGerhardt LAWolpaw JR: Real-time detection of event-related brain activity. Neuroimage 43:2452492008

22

Schalk GMcFarland DJHinterberger TBirbaumer NWolpaw JR: BCI2000: a general-purpose brain-computer interface (BCI) system. IEEE Trans Biomed Eng 51:103410432004

23

Sinai ABowers CWCrainiceanu CMBoatman DGordon BLesser RP: Electrocorticographic high gamma activity versus electrical cortical stimulation mapping of naming. Brain 128:155615702005

24

Sinai ACrone NEWied HMFranaszczuk PJMiglioretti DBoatman-Reich D: Intracranial mapping of auditory perception: event-related responses and electrocortical stimulation. Clin Neurophysiol 120:1401492009

25

Vannest JJKarunanayaka PRAltaye MSchmithorst VJPlante EMEaton KJ: Comparison of fMRI data from passive listening and active-response story processing tasks in children. J Magn Reson Imaging 29:9719762009

26

Vansteensel MJBleichner MGDintzner LTAarnoutse EJLeijten FSHermes D: Task-free electrocorticography frequency mapping of the motor cortex. Clin Neurophysiol 124:116911742013

27

Wray CDBlakely TMPoliachik SLPoliakov AMcDaniel SSNovotny EJ: Multimodality localization of the sensorimotor cortex in pediatric patients undergoing epilepsy surgery. Clinical article. J Neurosurg Pediatr 10:162012

28

Wu HCNagasawa TBrown ECJuhasz CRothermel RHoechstetter K: γ-oscillations modulated by picture naming and word reading: intracranial recording in epileptic patients. Clin Neurophysiol 122:192919422011

29

Wu MWisneski KSchalk GSharma MRoland JBreshears J: Electrocorticographic frequency alteration mapping for extraoperative localization of speech cortex. Neurosurgery 66:E407E4092010

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 132 132 31
PDF Downloads 130 130 18
EPUB Downloads 0 0 0

PubMed

Google Scholar