Functional MRI In Language Mapping And Its Clinical Application

Objective:To explore the normal language regions of Chinese healthy volunteers using three different language tasks, and select the most suitable task for clinical application. Materials and Methods:In our study,20 right-handed normal healthy volunteers, including 10 male and 10 female with mean age at 41.3 years, underwent blood oxygen level dependent-functional magnetic resonance imaging (BOLD-fMRI) using Siemens Verio 3.0T MR scanner. Three language tasks, including picture naming, semantic imaging and verb generation, were given during the scanning of BOLD-fMRI. The imaging data were processed using the soft ware SPM8. A independent student's t-test was used to compare the number of fuctionally activations in Broca area between any pairs of the three language tasks. The lateralization indexs (LIs) of three language tasks were computed using the formula LI=(L-R)/(L+R), in which L represented the number of fuctionally activations in Broca area, while R represented right hemisphere Broca homologue area. Results:The main activated areas by picture naming task included lots of brain regions as follows:bilateral inferior frontal gyrus(IFG), bilateral middle frontal gyrus(MFG), left frontal lobe(PFL), bilateral superior temporal gyrus(STG), bilateral insula lobe(IL), bilateral premotor area, and supplementary motor area(SMA). The main activated areas by semantic imaging task were as follows:left IL, bilateral regions of IFG(left predominant), STG, premotor areas(left predominant), SMA. The main activated areas by verb generation task included brain regions were as follows:bilateral MFG, left regions of IFG, PFL, STG, IL, inferior parietal lobe(IPL), premotor areas, SMA. The activated volumes by semantic imaging task were significantly more than both picture naming and verb generation (P<0.05). LIs of picture naming, semantic imaging and verb generation task were 0.05,0.79 and 1, respectively. Conclusions:All three language tasks could effectively localize language regions of normal people; semantic imaging task was more suitable for clinical application with more fuctionally activations in Broca area. LIs of both semantic imaging and verb generation language tasks obviously demonstrated left hemisphere dominance. Objective:To explore the feasibility of resting sate fMRI functional connectivity in localizing language regions of normal healthy volunteers and its accuracy comparing with task-evoked fMRI. Materials and Methods:In our study,20 right-handed normal healthy volunteers (the same people as selected in the first part) underwent both task-free fMRI(resting state fMRI) and task-evoked fMRI(semantic imaging the same as used in part one) using Siemens Verio 3.0 T MR scanner. Resting state fMRI imaging data were processed with soft wares REST v1.3 and GIFT v1.3, using both Seeds and independent component analysis(ICA) methods. Four seeds were selected, including two seeds in Broca area (left BA44 and BA45), and another two seeds in right hemisphere Broca homologue area(right BA44 and BA45). ICA could directly retrieve data information from the raw data using spatial ICA (sICA). Using Broca area activated by task-evoked fMRI as a standard, x2 test was used to compare the accuracy between the two methods of resting state fMRI. The statistical analyses were carried out using SPSS 16.0. A P-value< 0.05 was considered to be statistically significant. Results:Functional connectivities between Broca and right hemisphere Broca homologue area could be shown using anyone of the four seeds. When selecting left BA45 as the seed, Broca area showed more volumes than right hemisphere Broca homologue area; while the other three seeds showed approximately equivalent volumes between bilateral Broca areas. ICA method could also demonstrate bilateral Broca areas effectively and clearly. Comparing with Broca area activated by task-evoked fMRI, the accuracy of the Seeds method(right BA44 and right BA45, respectively) and ICA method were 70%,85%and 90%,respectively. No statistically significant difference was observed between any pairs of them (P>0.05). Conclusions:It was feasible to localize Broca area in normal people using resting sate fMRI functional connectivity. Both the seed method and ICA method could effectively and accurately show these areas. ICA method without the limitation of model-dependent as the seed method could be used more conveniently. Objective:To explore the feasibility of resting sate fMRI in presurgical language mapping and verify its efficiency by comparing with task-evoked fMRI and electrocortical stimulation mapping (ESM). Materials and Methods:There were 34 right-handed patients with brain tumors localized within or adjacent to Broca's area included in our study. All of them underwent both resting state fMRI and task-evoked fMRI(semantic imaging) using Siemens Verio 3.0 T MR scanner. Resting state fMRI imaging data were processed through both Seeds method(right Broca area as the ROI) and independent component analysis(ICA) method. Using Broca's area activated by task-evoked fMRI as a standard, x2 test was used to compare the accuracy between the two methods of resting state fMRI. Comparing with ESM, the efficiency of resting state fMRI in localizing language regions around the tumor was evaluated. Several soft wares, including SPM, REST v1.3, GIFT v1.3, were applied in the imaging data post-processing. The statistical analyses were carried out using the SPSS 16.0. A P-value< 0.05 was considered to be statistically significant. Results:Among the 34 patients, good activated language mappings had been obtained from 28 patients. Of them,17 cases showed activation around the tumor as well as right Broca area, six cases around the tumor only, two cases in right Broca area only, three cases within the tumor as well as around the tumor or in right Broca area. Using Broca area activated by task-evoked fMRI as a standard, among 28 patients, language regions could be successfully shown in 24 patients (85.71%) and 26 patients (92.86%) by Seeds method and ICA method, respectively. No statistically significant difference was observed between the two methods (P>0.05). Effective activated language mappings failed to obtain from six patients. However, we got their language mappings from resting state functional connectivity in the use of Seed and ICA methods. Combining with the two methods, resting state fMRI could localize the language regions in all patients effectively. The language regions obtained from resting state fMRI were quitely similar to the regions defined by task-evoked fMRI and ESM. Conclusions:It was possible to map the language cortex in presurgical patients using resting sate fMRI functional connectivity. This task-free paradigm may provide a powerful approach to map functional anatomy in patients without task compliance.