Distributed AM-FM demodulation of pathological tremor based on electrode network

This paper presents an effective framework of nonstationary time-frequency estimation with high spatial selectivity for the distributed estimation of pathological tremor based on electrode networks that allow to define a spatial EMG representation based on a distributed modelling technique. In a subsequent step, this spatial modelling of surface EMG is exploited to characterize the tremor thanks to the demodulation capabilities of the multicomponent AM-FM representations of spatial models of surface EMG. Methodologically, the generality and the flexibility of our framework allow the use of different demodulation techniques. In such a way, the proposed framework is able to define a large class of decentralized techniques for the multicomponent estimation of the cortical tremor that are irrespective of the medical disease generating the tremor, its physiological modelling, and the realistic assumption of the muscle movements. Experimentations on real cases have been performed and the results show the effectiveness of the proposed methodology, thereby making the proposed approach a favorable choice for a large number of applications in a broad range of emerging technologies of neural engineering.