Features Extraction And Model Analysis Of Acupuncture Encoding Based On Neural Electrical Signals

Neural system characterizes information in external stimulations by spa-tiotemporal encoding. Complex firing pattern can be observed in spinal dorsalroot when acupuncture is taken on ‘Zusanli’ point. Experiments are designed thatmanual acupuncture (MA) manipulations with diferent types and frequencies aretaken at ‘Zusanli’ points of experiment rats. Acupuncture features are extractedfrom the perspective of encoding, decoding and complex network. Acupunctureis further understood by applying network models.The firings of neuronal clusters are distinguished by extracting wavelet fea-tures of each spike shapes. Spatiotemporal coding features are investigated. Thenmean firing rate and coefcient of heterogeneity of encoding, which is a new co-efcient for characterizing neuronal encoding selectivity, are applied to quantifythese features in all experiment sessions. The diferences between features of MA‘nb’‘nx’ and those of MA ‘tb’‘tx’ are obvious. Neuronal adaptivity and sat-uration phenomenon are observed when acupuncture with diferent frequenciesare taken. However, neuronal selectivity of encoding is not obvious in diferentacupuncture frequencies.Types of acupuncture manipulations taken on the rats are inferred with ahigh probability by Bayesian decoding algorithm based on each single trial. Datain the first200ms from acupuncture onset are recognized to play a crucial role inincreasing the decoding performance in all sessions. These results are proved to besignificant by statistical analysis. Furthermore, mutual information is applied toquantify the decoding process. These studies may help to construct the interfacebetween neural systems and machines and improve the clinical study.To further understand the experimental results, spiking rate and regular-ity are studied in feedforward network of FitzHugh-Nagumo(FHN) neurons withdiferent excitatory and inhibitory connections, which is analogous to the trans-mission path of acupuncture signals. The phenomenon observed in experimentscan be explained by variation of parameters in network model to some extent. Finally, we study efects of network topology and external stimuli with diferentfrequencies on synchronization of neuronal network.