FreeSurfer

FreeSurfer. It is a freely available, open-source software suite for processing, analyzing, and visualizing structural MRI data of the human brain. Developed primarily at the Martinos Center for Biomedical Imaging, part of Massachusetts General Hospital and Harvard Medical School, it is a cornerstone tool in the fields of neuroscience and neurology. The software provides a fully automated pipeline for constructing models of the cortical surface and segmenting subcortical white matter and deep gray matter volumetric structures, enabling detailed quantitative analysis of brain anatomy.

Overview

The primary purpose of the software is to provide a comprehensive set of tools for the analysis of neuroanatomy from MRI scans. Its automated processing stream reconstructs the brain's cortical surface, a task that involves delineating the boundary between gray matter and cerebrospinal fluid as well as between gray and white matter. This allows researchers to measure cortical thickness, surface area, and curvature, metrics that are crucial for studying brain development, aging, and various neurological and psychiatric disorders. The suite is widely used in large-scale studies, including those affiliated with the Alzheimer's Disease Neuroimaging Initiative and the Human Connectome Project.

Technical details

The core algorithms rely on a combination of image segmentation, surface reconstruction, and spherical registration techniques. A key component is the use of an atlas, such as the Desikan-Killiany atlas, for labeling cortical gyri and sulci based on a subject's unique anatomy. The pipeline involves several stages, including motion correction, intensity normalization, removal of non-brain tissue via a process akin to skull stripping, and Talairach transformation for approximate spatial alignment. Advanced methods like spherical inflation and registration to a common spherical coordinate system, such as that defined by the Caret PALS atlas, facilitate cross-subject comparisons and group analyses.

Applications

It is extensively applied in both clinical research and basic neuroscience to investigate the structural correlates of brain function and disease. In clinical contexts, it aids in the study of Alzheimer's disease, schizophrenia, epilepsy, and multiple sclerosis, providing biomarkers like hippocampal volume or cortical thinning patterns. Cognitive neuroscience studies utilize it to explore relationships between brain structure and abilities measured by instruments like the Wechsler Adult Intelligence Scale. Its outputs are also frequently integrated with data from other modalities, such as diffusion tensor imaging from FSL or functional data from Statistical Parametric Mapping, for multimodal brain analysis.

Development and history

The project was initiated in the late 1990s by a team led by Bruce Fischl at the Martinos Center for Biomedical Imaging. Its development has been sustained by continuous funding from institutions like the National Institutes of Health, particularly the National Institute of Biomedical Imaging and Bioengineering and the National Institute of Neurological Disorders and Stroke. The software has evolved through numerous versions, with major algorithmic advancements in surface-based registration and segmentation accuracy. Its development philosophy emphasizes rigorous validation, often benchmarking against manual tracings by neuroanatomists, and it has been a critical tool in the Brain Imaging, Structure, Phenotype and Neuroinformatics research community.

Related software

It exists within a broader ecosystem of neuroimaging tools. For volumetric analysis and segmentation, alternatives include FSL from the University of Oxford and Statistical Parametric Mapping from the Wellcome Trust Centre for Neuroimaging. Surface-based analysis is also offered by Caret and BrainVoyager. For multimodal integration and visualization, platforms like 3D Slicer and MRtrix are commonly used alongside it. Many researchers use it in conjunction with Python libraries such as NiBabel and scikit-learn for further statistical analysis and machine learning applications. Category:Neuroimaging software Category:Harvard Medical School Category:Free science software Category:Computational neuroscience