Purpose or Case Report: Functional MRI (fMRI) for preoperative language mapping of Wernicke’s and Broca’s areas in clinical pediatric epilepsy presents various challenges resulting in poor patient compliance. We hypothesized that higher fMRI success rates can be achieved using intensive patient preparation in a fMRI practice session using a mock MRI scanner.
Methods & Materials: We performed a retrospective IRB approved review of epilepsy patients that received preoperative language fMRI between 2010 and 2017. We recorded age, FSIQ scores, scan information and the radiologist’s reports. We defined success of the fMRI by the radiologist’s ability to locate Broca’s and Wernicke’s areas in their report.

Under the old protocol, patient preparation consisted of a 10-minute verbal explanation before the MRI consisting of 2 language tasks. Under the new protocol, the practice session lasted 60 minutes using a mock MRI scanner. Tasks were first practiced seated and then supine in the mock scanner. Coaching with real-time feedback was given during the practice and exam. Information from practice was used to select and modify tasks according to patient’s ability. We increased the number of language tasks, and an additional specialist was responsible for task practice and coaching during the exam. On average 6 language tasks were administered during the MRI, increasing total length over the old protocol.
Results: We identified 33 preoperative language-mapping fMRI scans of children with epilepsy. 13 of 33 were done under the old protocol with a mean age of 16 (range 12 – 19 years). FSIQ scores were available for 10 of the 14 patients with a mean score of 84 (range 62 – 108). 19 of 33 scans were done after institution of an new protocol with mean age of 14 (range 8 – 19 years). FSIQ scores were available for 14 of the 19 patients with a mean score of 75 (range 49 – 111).

The radiologist could identify Broca’s area in 9 of 14 scans (64%) under the old protocol, and 18/19 (95%) with the new protocols. The radiologist could identify Wernicke’s area in 9 of 14 scans (64%) under the old protocol, and 19/19 (100%) of scans with the new protocol. Using the z-statistic for two group proportions, the success rate between new and old protocols were significantly different (p<0.05), with p=0.035 for Broca’s and p=0.005 for Wernicke’s.
Conclusions: Higher pediatric fMRI success rates can be achieved with intensive patient preparation in an fMRI practice session using a mock scanner.