Cerebral networks in sensorimotor disturbances

Increasing evidence suggests that the human brain employs multiple, interconnected brain areas for information processing and control of behavior, including the performance of laboratory tasks. Brain diseases are expected to affect these networks directly by interference and indirectly as a consequence of deficit compensation. Covariance analyses applied to functional brain imaging data open the opportunity to study neural networks and their disease-related changes in the human brain. Here, we review our analytic approach based on principal component analysis (PCA) to address such questions. We will discuss its methodological foundations and applications in patients with sensorimotor disorders. We will show that PCA in combination with, both, hypothesis-driven testing and correlation statistics provides a powerful tool for elucidating disease-related abnormalities and postlesional reorganization of neural networks in the human brain.