Optimal combination of multiple cortical surface parcellations

A variety of methodologies have been developed for the parcellation of human cortical surface into sulcal or gyral regions due to its importance in structural and functional mapping of the human brain. However, characterizing the performance of surface parcellation methods and the estimation of ground truth of segmentation are still open problems. In this paper, we present an algorithm for simultaneous truth and performance estimation of various approaches for human cortical surface parcellation. The probabilistic true segmentation is estimated as a weighted combination of the segmentations resulted from multiple methods. Afterward, an Expectation-Maximization (EM) algorithm is used to optimize the weighting depending on the estimated performance level of each method. Furthermore, a spatial homogeneity constraint modeled by the Hidden Markov Random Field (HMRF) theory is incorporated to refine the estimated true segmentation into a spatially homogenous decision. The proposed method has been evaluated using both synthetic and real data. The experimental results demonstrate the validity of the method proposed in this paper. Also, it has been used to generate reference sulci regions to perform a comparison study of three methods for cortical surface parcellation.