Simultaneous analysis of noisy signals obtained from multiple experiments, with application to deriving brain functional images

We proposed a new approach based on the linear least square method for the simultaneous analysis of noisy signals obtained from multiple experiments. To demonstrate the effectiveness of this approach, we applied it to noisy signals obtained from multiple studies with positron emission tomography for deriving brain functional images (cerebral metabolic rate of oxygen). The results show a significant improvement in the estimation accuracy and reliability of the determined functional parameters. Furthermore, this new approach reduces the likelihood of errors being propagated due to multiple experiments and is able to fully utilize the information content provided by the measured signals