Yuki Hachiman RT, Presenter: Nothing to Disclose

Norio Hayashi PhD, Abstract Co-Author: Nothing to Disclose

Mitsuru Kikuchi, Abstract Co-Author: Nothing to Disclose

Shigeru Sanada PhD, Abstract Co-Author: Stockholder, Hologic, Inc

Yoshio Minabe, Abstract Co-Author: Nothing to Disclose

Toru Arishita, Abstract Co-Author: Nothing to Disclose

Yuko Yoshimura, Abstract Co-Author: Nothing to Disclose

Kiyomi Shitamichi, Abstract Co-Author: Nothing to Disclose

Sanae Ueno, Abstract Co-Author: Nothing to Disclose

We have developed a method for creating a personalized brain template for superimposition of MEG without MR imaging. Using this method, the MEG data can be accurately superimposed on the brain surface without the necessity of performing MRI for mapping.

Magnetoencephalography (MEG) is a noninvasive method with a high time resolution that is very useful for predicting the degree of pediatric psychiatric disorders. An anatomical position is obtained by superimposing the MEG data on a magnetic resonance (MR) image; however, MR imaging is difficult in children. In this study, we propose a novel method for creating a personalized brain template for mapping in MEG without an MR image.

RMSE between the subjects’ data and the personalized template data was 8.3 ± 1.8 mm for the conventional method and 6.9 ± 1.2 mm for our proposed method. RMSE of the proposed method was thus significantly smaller than that of the conventional method (P < 0.05). The action potential was shown in the primary auditory cortex when the personalized image and the data showing the MEG results of a hearing stimulus were superimposed.

Institutional review board approval and written informed consent were obtained. Thirty-eight subjects (aged 3–5 years) in whom no abnormalities were observed on a 3.0 Tesla MRI were included in this study. For evaluation, 5 subjects underwent MEG. Our method for creation of the brain template consisted of the following three steps: 1) plotting of five points on the head surface; 2) selection of suitable template data; 3) and registration of the personalized data on the template. Brain template data was personalized by the method of selecting an image in which the error of the selected five points on the head surface was the smallest after performing an affine transformation on MR images of all children. The error between the personalized template data and MR images of the subjects was evaluated using the root mean squared error (RMSE). Finally, we confirmed whether the action potential would be plotted in the correct position on personalized data superimposed on the hearing stimulus data provided by MEG.

Hachiman, Y, Hayashi, N, Kikuchi, M, Sanada, S, Minabe, Y, Arishita, T, Yoshimura, Y, Shitamichi, K, Ueno, S, A Novel Method of Creating a Brain Template for Mapping in MEG Without the Need of an MRI.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12025396.html