Ica-based Segmentation of the Brain on Perfusion Data

3. Comon P. Signal Processing. 36:11-20 (1994) 4. Hyvarinen A, IEEE Trans. Neural Networks 10:626-634 (1999) 2. Rempp et al. Radiology 193:637-641 (1994). • Perfusion indices such as Cerebral Blood Volume (CBV), Cerebral Blood Flow (CBF), and Mean Transit Time (MTT) are calculated based on the temporal response of the MR signal to the administered contrast agent. Since the temporal response is altered in the case of brain pathology. CBV, CBF, and MTT are good indicators of compromised regions. Segmentation of perfusion images allows CBV, CBF, and MTT to be evaluated per tissue type. • Independent Comopenent Analysis (ICA) applied to perfusion data, differentiates tissue types based on the temporal response. GM, WM, and CSF responses are distinct. Hence ICA can serve as a segmentation technique as well as identify pathology since the changes would be reflected in the perfusion profiles.