MIT
MGH
HMS
LCLN

LABORATORY FOR COMPUTATIONAL LONGITUDINAL NEUROIMAGING (LCLN)

Rotate to view desktop version ...

Computational Neuroimaging, Computer Vision and Machine Learning
Brain

Neuroimaging focusses on imaging structure and function of the brain. Computer vision is the process of automatically processing images and shapes in order to extract information. Machine learning focuses on the construction of algorithms and models that can learn from and make predictions on data and is related to computational statistics.

Our research spans the MR acquisition (treatment of head motion, longitudinal acquisition), image processing (registration, segmentation, surface reconstruction), image analysis and statistical modeling/machine learning for computer-aided diagnosis and prognosis. See for example the image at the left for a cortical segmentation of a brain, left hemisphere. We have performed statistical morphometric population studies of subcortical brain structures in schizotypal personality disorder, Alzheimer's disease, Huntington's disease and a variety of other pathologies, including brain tumor (Glioblastoma).

We have contributed highly accurate robust registration procedures with applications in tumor treatment assessment, as well as reliable and sensitive methods for longitudinal image analysis of large brain MRI datasets (part of the widely used FreeSurfer package), to identify novel biomarkers and evaluate drug trials. Among other large cohort studies, for instance, the Alzheimer's Disease Neuroimaging Initiative (ADNI) employs our methods for the sensitive detection of early anatomical gray and white matter changes in the human brain. Our final morphometric estimates are made available by ADNI online. More recently we combined our early work in shape analysis with neuroimaging to study geometry and symmetry of neuroanatomical structures for the early and sensitive detection of presymptomatic disease. For this we have developed the BrainPrint (a holistic and compact shape descriptor) and incorporated it into a predictive model which has achieved the second place in the MICCAI'14 Computer-aided diagnosis of dementia challenge (CADDEMENTIA). Ongoing work focuses on disease lateralization in subfields, prediction of disease onset, shape heritability of neuroanatomical structures, bio age classification, patient stratification and disease subgrouping among other exciting topics.

Links:
Selected Publications (full list):
clustrmaps.com
Martin Reuter - MIT - Cambridge, MA, USA - EMail: reu...@mit.edu
Last modified 17.04.2023 -- [xhtml] -- [css] -- [welcome]