Abstract: Traditional analysis of brain signals is often realized under assumptions such as stationarity, linearity, predetermined frequency bands and basis functions. These strong assumptions imposed on brain signals might cause distortions leading to improper results. This study adapts a data-driven approach called ‘empirical mode decomposition’ in order to avoid unrealistic assumptions and minimize the parameter space. In this respect, we confront with the issues of band range determination and coherent source localization. Magnetoencephalographic...
(read more)
Topics: 
Artificial intelligence
Neuroscience