Automatic Brain Image Segmentation in Alzheimer's Disease
Author Information
Author(s): Heckemann Rolf A., Keihaninejad Shiva, Aljabar Paul, Gray Katherine R., Nielsen Casper, Rueckert Daniel, Hajnal Joseph V., Hammers Alexander
Primary Institution: The Neurodis Foundation (Fondation Neurodis), Lyon, France
Hypothesis
Can automatic segmentation of brain images improve the understanding of Alzheimer's disease and mild cognitive impairment?
Conclusion
The study provides a valuable repository of brain image segmentations that can help researchers analyze Alzheimer's disease and mild cognitive impairment.
Supporting Evidence
- The repository includes segmentations of 996 screening and baseline images.
- Segmentations were validated against manual methods and showed high accuracy.
- The study found significant differences in brain regions between diagnostic groups.
- The method demonstrated robustness across different MRI field strengths.
Takeaway
This study created a free collection of brain images to help scientists understand Alzheimer's disease better.
Methodology
The study used multi-atlas propagation with enhanced registration to segment brain images into 83 regions.
Potential Biases
Potential bias due to the reliance on a specific atlas for segmentation.
Limitations
The automatic segmentation may misclassify tissues due to white-matter disease.
Participant Demographics
Participants included healthy elderly subjects and patients with mild cognitive impairment or Alzheimer's disease.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website