Cluster analysis in individual functional brain images: Some new techniques to enhance the sensitivity of activation detection methods
✍ Scribed by Dr. Jean-Baptiste Poline; Bernard Mazoyer
- Publisher
- John Wiley and Sons
- Year
- 1994
- Tongue
- English
- Weight
- 806 KB
- Volume
- 2
- Category
- Article
- ISSN
- 1065-9471
No coin nor oath required. For personal study only.
✦ Synopsis
Low signal-to-noise ratio is the fundamental limit of rurrent voxel-based strategies for detecting activations in functional brain maps. We propose some new techniques to enhance detection sensitivity in the analysis of brain activation maps. These new techniques are: 1) a multi-filtering strategy; and 2) the use of a hierarchical decomposition. Multi-filtering is used to optimize detection sensitivity when multiple signals of various sizes are present, while hierarchical decomposition allows selection of activation foci on the basis of their spatial extent and magnitude. Both techniques are combined within a single testing procedure that controls the overall type I error. This approach shows significantly higher detection sensitivity on both simulated and experimental single-subject datasets.