This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. This subproject is concerned with providing an automatic robust method for registering histological sections with MR images. The specific area for which this method will be optimized is the Medial Temporal Lobe (MTL). The goal in doing this is to use high-resolution MR and the registered histology to build detailed models of hippocampal anatomy and then to use these models to generate probabilistic labellings in vivo imaging data at standard structural resolution. Registering histology directly with MR is a difficult problem, because mounting and staining the tissue slice inevitably results in distortions of the underlying tissue. This means that the mapping from the stained sections to the high-resolution MR is both cross-contrast and non-linear. This is why, instead of attempting a direct approach to this problem, we have chosen to consider an intermediate step in the registration process, namely pre-registering the histology with the uncut block face images. This has several advantages, the most important being that the use of the block face images reduces the complexity of the problem, since one only needs to perform non-linear registration on 2D images.
Showing the most recent 10 out of 1099 publications
