Robust Tissue Segmentation of Human Brain Images Acquired with a Surface Coil at Ultrahigh Field

Surface coil has been commonly employed in human brain MRI and in vivo MRS research since it offers the best attainable signal-to-noise ratio (SNR) in the brain region of interest. However, quantitative tissue segmentation becomes a challenging task due to the inhomogeneous B1 field of the surface coil, which leads to large variation in the MR signal and image contrast. In this study, we implemented and tested an automatic brain tissue segmentation method including bias field correction and partial volume estimation (PVE) to reliably quantify tissue contents and distributions in the brain region covered by a surface coil at 7T. The results indicate that this segmentation method is robust for differentiating various brain tissues; and the CSF volume can be more accurately estimated by PVE with model parameters as compared to one without optimization. Therefore, this advanced segmentation method will provide a robust and valuable tool for many quantitative brain MRI/MRS studied, and it is particularly critical for ultrahigh-field applications using either a surface coil or transmit-receive array coil or even head volume coil. In all these cases, an inhomogeneous B1 distribution is present in the human head owing to the complication of the RF wave behavior at high/ultrahigh fields.

This abstract and the presentation materials are available to members only; a login is required.

Join Here