Functional magnetic resonance imaging (fMRI)-aided therapeutics of Chinese speech area-related lesions: Screening of fMRI-stimulating mode and its clinical applications
Correspondence Address: Source of Support: National Natural Science Foundation of China
(NSFC, 30872660), Conflict of Interest: None DOI: 10.4103/0028-3886.73741
Source of Support: National Natural Science Foundation of China (NSFC, 30872660), Conflict of Interest: None
Background: Language area-related lesion is a serious issue in neurosurgery. Removing the lesion in the language area and at the same time preserving language functions is a great challenge. Aim: In this study, we aimed to screen functional magnetic resonance imaging (fMRI) based task types suitable for activation of Broca and Wernicke areas in Chinese population, characterize lesion properties of functional area of Chinese language in brain, and assess the potential of fMRI-guided neuronavigation in clinical applications. Materials and Methods: Blood oxygen level-dependent fMRI has been used to localize language area prior to operation. We carried out extensive fMRI analyses and conducted operation on patients with lesions in speech area. Results: fMRI tests revealed that the reciting task in Chinese can steadily activate the Broca area, and paragraph comprehension task in Chinese can effectively activate the Wernicke area. Cortical stimulation of patients when being awake during operation validated the sensitivity and accuracy of fMRI. The safe distance between language activation area and removal of the lesion in language area was determined to be about 10 mm. Further investigation suggested that navigation of fMRI combined with diffuse tensor imaging can decrease the incidence of postoperative dysfunction and increase the success rate for complete removal of lesion. Conclusion: Taken together, these findings may be helpful to clinical therapy for language area-related lesions.
Keywords: Broca area, Chinese speech area, functional magnetic resonance imaging (fMRI), neuronavigation, Wernicke area
The goal of minimally invasive neurosurgery (MIN) involves removal of the lesion at the same time preserving neurological function. Of the neurological functions, preserving of language function is recognized as the most important and complicated. ,, The classical language-processing area of the dominant hemisphere involves Broca area of the frontal lobe and Wernicke area of the temporal lobe. ,, Broca area is involved in linguistic production, and Wernicke area in speech comprehension. ,, Lesions of these areas can result in serious language impairments and the associated disability may cause serious burden to both the family and the society. , It is a challenging task for the neurosurgeons to completely excise the lesion in the linguistic areas while preserving the ability of speech.. Preoperative mapping of linguistic area is one of most effective approches to resolve the conflict between preservation of functional speech area and lesion removal to as large an extent as possible. Language function is specific to human beings and there is lack of reliable clinical method to monitor and localize the language function. Use of anatomic landmarks solely is not enough to dissect speech areas, because there can be shift of the functional area as a result of the lesion, plasticity of neurological functions and complexity of language area. Intraoperative cortical electrical stimulation is often the method used for localizing the speech areas. However, the obvious disadvantages of this method include: serious invasiveness, complex operation, extended operation time, and failure to predict the risk of post-operative functional disorder.
Blood oxygenation level-dependent (BOLD) functional magnetic resonance imaging (fMRI) is one of the powerful approaches to indirectly measure neuron activities in healthy and diseased brains. ,,,, BOLD fMRI detects local changes in relative blood oxygenation that are most probably a direct consequence of neurotransmitter action, thereby reflecting the local neuronal signaling. ,,,,, In the recent years, fMRI has become a leading research tool for mapping brain behaviors, and also has dominated the neuroimaging field because of its low invasiveness, lack of radiation exposure and relatively wide availability. , In general, there are two major experimental designs used in fMRI-based studies: block designs and event-related designs.  Block-designed fMRI is supposed to be the steady state of regional cerebral blood flow and has been applied to examinations of brain activation caused by tasks requiring sustained or repetitive movements. In contrast, event-related fMRI with time resolution of a few seconds allows the mapping of brain activation associated with a single movement according to the transient aspects of the hemodynamic response.  This useful technique is feasible to address at least 3 clinical issues: 1) to characterize anatomically normal patterns of brain function, 2) to distinguish pathological traits and 3) to monitor treatment responses.  Since the first application of BOLD fMRI in preoperational localization of motor cortex in 1994, clinical data have demonstrated that it shows spatial identity to that of electrical stimulation of cortex. Recent accumulated evidence suggests that it can also be extended to language study. ,,, Compared with motor cortex, Chinese language area is much more complicated and exhibits significant individual differences. Far different from the writing system of spelling letters, Chinese characters such as pictographs are an entity of phenotype, pronunciation and meaning. Thus, Chinese recognition is associated with not only the analyses of visual characters but also the cognizing function, indicating specific imaging rules existing in Chinese language area. Elucidating the functional imaging of Chinese characteristics contributes significantly to the preservation of Chinese speech ability. Ideally, fMRI-based mapping of Chinese language area requires the following three key questions resolved: first, screening an appropriate mode of language tasks; second, evaluating the exactness and sensitivity of funtional imaging of Chinese language area; and third, elucidating the encephalic lesion features of Chinese language area.
In this study, we attempted to screen the task mode appropriate for Chinese population through extensive fMRI-based analyses of characteristics of Chinese language area when performing different language tasks. The exactness and sensitivity were assessed with the cerebral electrical stimulation of patients with dysfunction of language area. We also determined the relatively safe distance between the functional language area and the complete removal of lesion, providing important basis for clinical application.
The volunteers for this study were ten healthy Chinese undergraduates (5 males and 5 females; with an average age of 22.8 years) with no normal language function and hearing from the Third Military Medical University and were determined to be of dextromanuality using Edinburgh Handedness Questionnaire. All the experiments were approved by the ethics committee of our hospital, and the study volunteers signed written agreement of consent.
Eighteen patients (8 males and 10 females, mean age 42 years) with left hemispherical lesion on MRI were included in the study. Further, the patients were divided into three subgroups based on the location of the lesion: 5 with frontal lobe, 6 with temporal lobe and 7 with frontal-temporal lobe. The lesion diameter ranged from 1.1 to 8.1 cm. The pathology of the lesions was: glioma in Grade I (2); glioma Grade II-III, (7); glioblastoma (1); oligodentroglioma (4); angiocavernoma (2); vascular abnormality (1); and arachnoid cyst, (1). Prior to hospitalization, 11 patients exhibited normal language function, while the remaining 7 patients showed some language functional disorder in either linguistic production or language comprehension.
Functional magnetic resonance imaging
fMRI examinations were conducted on the SONATA1.5T magnetic resonance apparatus (Siemens, Germany). T1WI structural image data was collected using the turbo spin echo (TSE) (Time Repetition (TR): 500 ms, Time Echo (TE): 11 ms and data matrix: 192Χ144). Acquisition of BOLD data was based on echo planar imaging (EPI) (TR: 3,000 ms, TE: 39 ms, data matrix: 64Χ64). For the stimulus scheme, there were 5 cycles of the module, which consisted of the resting period and the stimulating period. To relieve the anxiousness of the volunteers, the entire test process was explained to them in detail prior to formal MRI. To address the activation of Wernicke area in different tasks of Chinese listening, three kinds of listening tasks (voice strength, 65 dB) were employed, viz., a group of music enjoyment tasks, a group of phrase comprehension tasks and a group of paragraph comprehension tasks, respectively. In the music group, uncommon symphony was played with simple voice as negative control during the stimulatory period of 30 seconds. In the group of phrase comprehension, volunteers were required to pick up a unique one from 7 terms in 30 seconds. In the group of paragraph comprehension, the volunteers were required to understand exactly the content of a brief introduction, a paragraph of common science, in 30 seconds.
For further probing the Broca area response to the varied tasks of Chinese language, four groups of language tasks were designed, including silent-reading group, reciting group, picture-naming group and verbal reading group. In the silent-reading group, a paragraph comprising of Chinese words was required to be gone through repeatedly. In the reciting group, the same content as that of the silent-reading group was to be recited time and again. In the picture-naming group, E-prime was used to process photographs. Photograph collection was maintained in synchronization with signal stimulation. Projection of images was done by LCD (liquid crystal display) every 2 seconds during the stimulatory period. The volunteers were required to speak what they saw. In the last group (i.e., verbal reading group), volunteers were required to read at normal speed the document comprising of Chinese letters displayed by LCD.
The images acquired were processed with the fMRI Automatic Data Analysis (FADA) software (Texas University Medical Center) and analyzed using the software AFNI (analysis of functional neuro-images).
Diffusion tensor imaging
The image was generated using single activation of echo planar imagining technology. Diffusion tensor signals from 6 directions were collected (b, 1000; n, 10; thickness, 3 mm; distance, 0.8 mm; TR, 6,100 ms; TE, 106 ms). The images were processed with the software DTI-Studio (version 2.02) for reconstruction of plots of average apparent coefficient (ADC mean), anisotropic coefficient (FA, Fractional Anisotropy) relative FA (RA), volume ratio (VR), tracing white matter fibers.
The input of the combined structural images, fMRI and FA of diffusion tensor imaging (DTI) into the neuronavigation system (GE Insta Trak 3500) showed the relationship among lesion, language cortex, white matter fibers, with the aid of image fusion. Following neuronavigation-aided craniotomy, lesion was differentiated from the language cortex. On the surface of cerebral cortex, sterilized slips (5 Χ 5 mm) were used to label the boundary of tumor projection and exact location of Chinese language region.
During the state of being awake, electrical stimulation of cerebral cortex was done to map the Chinese language area (stimulation parameters: square wave pulse of di-electrode, 200 ms; pulse interval, 500 ms; impulse frequency, 50 Hz; current intensity, 3-15 mA). Patients were requested to execute three different kinds of tasks, and the positive regions identified were marked using the sterilized slips with numbers. These tasks were as follows: 1) counting Arabic numbers (patients were required to count the continuous Arabic numbers clearly and loudly (1, 2, 3, …); 2) picture-naming (patients were required to speak out the names of the objects on photo slides displayed at a speed of 1/3 seconds); 3) language comprehension (an easy quiz was given to evaluate systemically patients' abilities, especially those of comprehension/ expression/ counting). After the exact location of language cortex, we avoided the language region and white matter fibers to form an appropriate entry and then remove the lesion region under microscopy.
Screening for fMRI-specific Chinese activation mode in Wernicke area
Three groups of different listening tasks that included music group (negative control), phrase comprehension group and paragraph comprehension group were employed to determine which one could activate Wernicke area steadily in fMRI [Figure 1]. As expected, activations were recorded in both temporal lobes, majority of which were right temporal lobe activations in the music group as negative control [[Figure 1]a and b], while the tendency of time-signal intensity varied steadily [Figure 1]c. For the phrase comprehension group, activation signals could be observed in both temporal lobes but seemed to be located mainly in the left temporal lobe, i.e., Wernicke area [[Figure 1]d and e]. Moreover, the dynamic curve of time-signal intensity was relatively unstable [Figure 1]f. In the paragraph comprehension group, the activation signals in the brain of all the tested volunteers were found to be much clearer and stronger, all of which were concentrated in the Wernicke area [[Figure 1]g and h]. Moreover, the dynamic curve of time-signal intensity was extremely steady [Figure 1]i. Considering the above observations, we proposed that the task of Chinese paragraph comprehension could be the most suitable fMRI-specific mode for steadily activating the Wernicke area in healthy individuals.
Determination of fMRI-specific Chinese stimulating mode in Broca area
To gain further understanding of Broca area in normal Chinese individuals, a series of BOLD fMRI tests were performed. For this, four different groups of language tasks were designed: silent-reading group, reciting group, picture-naming group and reading group. In the silent-reading group, there was no obvious signal in the Broca area, and only slight activation could be observed in the cerebellar area [Figure 2]a. Subsequently, we noted that the result in the picture-naming group was very similar to that in the reading group [Figure 2]b. In brief, first, intensive activation of the visual cortex accompanied with Broca area response, ruling out the feasibility of its being extended to clinical application; second, the stability of the dynamic curve of time-signal intensity failed to satisfy the basic requirement. Intriguingly, the activation signal of Broca area was clearly found in nearly all the tested volunteers belonging to the reciting group [Figure 2]c. Moreover, the dynamic curve of time-signal intensity was stable, which indicated that the mode of language task was very consistent with the activation of Broca area [Figure 2]d. Therefore, reciting task in Chinese was considered to be a good mode that could activate the Broca area steadily in fMRI, having the potential of being applied in the clinical operation of language area's lesions.
Evaluation of fMRI in Chinese language area using cortical stimulation
As we know that BOLD-fMRI data indirectly reflects the activity of neuron, and may not be in complete agreement with the genuine activation area of neuron, it is essential to assess both the accuracy and sensitivity of fMRI in probing Chinese language area. Here we utilized cortical stimulation to address this issue. Twenty-nine language tasks were given to 18 patients. Among them, 15 patients completed the requested tasks in an awake state. The system error was less than 2 mm, and the average time of patients being awake was 19 minute (range, 10-50 minutes), indicating appreciable. Due to either the restriction in operating the window of skull or poor state of being awake, there were totally 16 patients joining in the comparison of core area of brain activation with cortical stimulation [Table 1]. In 9 patients the deviation in distance was less than 1 cm and could be overlapped well and in the remaining 7 patients the deviation in distance of no more than 2 cm [Table 1].
Pilot study of fMRI combined with DTI in clinical operation
Diffusion tensor imaging has emerged as a robust technique that can be applied to structural characterization of white matter fibers of live body. More importantly, DTI technique has been demonstrated to enhance the level of removal of pathological lesion in clinical operations. In this pilot clinical investigation, we combined DTI with fMRI. In total, 18 patients with language area-related lesions were admitted for operation-based therapy. Among them, 11 patients behaved normally in their language functions, while the remaining 7 patients were found to have language dysfunction to varied extent. Generally, the following guidelines were followed: 1) Relatively safe distance between the Chinese language activation area and complete removal of pathological lesions is 10 mm. 2) When the distance is less than 10 mm, cortical stimulation is recommended for aiding in the localization of language area during the awake state of the patients. Depending upon the operative findings, 7 patients underwent total removal of the tumor, 5 patients underwent subtotal removal of the tumor, and 6 patients underwent partial removal of the tumor. Prognosis in 4 patients was much better when compared with to the preoperative state. Eleven patients exhibited normal functioning, and 3 patients manifested clinical signs of transient loss of speech function [Table 2]. Permanent language dysfunction was not observed in our investigation [Table 2].
A 31-years-old male prior to operation, was believed to be lagging in response with poor comprehension. The lesion was located in the left temporal lobe [Figure 3]a. EPI-derived results suggested the Wernicke area on left temporal lobe was in retrocession and fMRI-based results also indicated the elbow of tumor growth caused Wernicke area to be in retrocession [[Figure 3]b and c]. As expected, the two Wernicke areas: one mapped by fMRI and the other by cortical stimulation - were observed to match well during the operation [Figure 3]d. During operation the Chinese language area was avoided and most of the tumor lesion (glioma Grade II) was removed [Figure 3]e, Postoperative computerized tomography (CT) examination showed partial removal of the tumor [Figure 3]f. Postoperatively he could speak clearly, and his memory and comprehension have improved to a great extent.
It seems that fMRI-based clinical localization in Chinese language area is fully validated by cortical stimulation, suggesting that it is feasible that fMRI can be applied to operation therapy for Chinese language area-related lesions.
Language system is known to be a complex network in which different language tasks are performed using complicated connections of multiple brain areas. Generally, the language system in brain is considered to easily respond to visual and auditory stimulation. However, the varied language tasks always can be attributed to the so-called core area of brain. As we know, Chinese is one of the complicated languages controlled by a unique language area in the brain. ,,, BOLD fMRI is a powerful approach to characterize indirectly local neuronal signals through detecting local changes in relative blood oxygenation of individuals. ,, In recent years, BOLD fMRI has generated interest not only as a tool for mapping brain activation but also as a means of studying the dynamics of neural networks by tracking fMRI response characteristics across various spatial and temporal scales.  In this study, we aimed to unravel the relationship between the BOLD fMRI signal and the underlying neural activity specific to the functional area of Chinese language. A couple of volunteers were invited to join our pilot study with emphasis on describing BOLD fMRI-based characteristics of both Wernicke area [Figure 1] and Broca area [Figure 2].
We successfully screened the optimal stimulating mode specific to Wernicke area of normal Chinese individuals, which is the Chinese paragraph comprehension task. In contrast, the reciting task in Chinese is suggested to be a nice steadily activating pattern for the Broca area in an fMRI-based survey. The above two types of stimulating tasks obtained could establish solid basis for directing clinical operation. Meanwhile, it is also comprehensive information that documents, for the first time, theoretical mechanism underlying language area-related lesions. To further test the potential of the mapped language area in clinical applications, we combined fMRI with DTI to carry out the clinical operation of language area-related lesions. Consequently, the results showed that the clinical requirements could be greatly fulfilled, indicating that the combination of fMRI with DTI (followed by neuronavigation) is feasible in the performance of the operation. Moreover, it validated that the stimulating modes are useful for decreasing the side effect, viz., injury.
Collectively, we reported a systematic exploration of fMRI-aided treatment for Chinese speech area-related lesions, which is mainly based on the optimal stimulating mode in fMRI and beyond. , It may be developed to become a promising therapeutic mode in the near future for patients with such lesions.
This work was supported by National Natural Science Foundation of China (NSFC, 30872660).
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2]