Detecting causality between different frequencies

摘要: Biological systems are usually non-linear and, as a result, the driving signal frequency (say, MHz) is in general not identical with output N Hz). Coherence and causality analysis have been well-developed to measure (directional) correlation between input signals frequencies (N=M), but they applicable cases different (N equal M). In this paper, we propose novel method called frequency-modified (coherence) resolve issue. The or first modulated by up-sampling down-sampling, coherence then applied filtered signals. An optimal found, revealing true input-output relationship successfully tested on data generated from toy model, van der Pol oscillator employed analyze recorded Parkinson's disease (PD) patients.