Algorithmic Development Of Acute Pain Detection Based On Cortical Neural Ensemble Spikes And LFP Signals In Rats

Brain is a senior center in the nervous system,responsible for all cognitive functions of the body,while studying the structure and function of the brain has become the hottest field of science at present.Pain is a complex sensory experience and the most serious problem afflicting human health today.Decoding pain neural signals have always been an important research topic in the field of neuroscience,which can not only help people understand the mechanism of pain processing in brain,and then promote the emergence of new treatment strategies,but also will have important guiding significance and application value for clinical and closed-loop brain-machine interface.Firstly,according to the neural signals of different modality,this paper makes an in-depth study on the pain problems in different states of rats from many angles.The putative spontaneous pain episodes were identified by combining animal behavior with neurophysiology records,and different multimodal neural responses were found between evoked pain and spontaneous pain:1)We find stronger phase-amplitude coupling in the S1 than the ACC in naive and chronic pain rats;2)The pre-behavior S1 gamma-ERS/ERD(event-related desynchronization/synchronization)correlates with the post-behavior ACC beta-ERS/ERD during spontaneous pain;3)Pain-modulated ACC and S1 neuronal firing correlates with the amplitude of stimulus-evoked event-related poten-tials(ERPs);4)The integration of ACC and S1 ensemble spikes and local field potential(LFP)fea-tures provides imformative cues for detecting pain signals.Together,these results suggest distinct neural mechanisms between evoked and spontaneous pain at both LFP and cellular levels as well as different coding roles between S1 and ACC in pain processing.Secondly,based on the ensemble neural spikes,on the basis of Poisson Linear Dynamic Sys-tem(PLDS)model,this paper proposes a decoding algorithm called " Ensembles of Change-point Detectors(ECPDs)from the point of view of improving the accuracy of acute pain detection,which utilizes the idea of ensemble learning to integrate a series of independent"Weak" detectors and develop a majority voting mechanism to achieve the goal of improving detection accuracy.Through the test of multiple computer simulations and experimental recordings,it is shown that the detection performance of ECPDs integrated decoding algorithm proposed in this paper is ob-viously superior than that of single detector.Finally,based on the LFP signal,this paper proves the feasibility of using LFP signal to detect acute pain by decoding and analyzing the acute pain intensity of rats in different physiological states.According to the role played by the power characteristics of theta frequency band and high-gamma frequency band in the study of distinguishing pain intensity in the LFP signal,this paper also proposes a method of detecting acute pain events based on Steady-state Kalman Filter,which is verified by multiple experimental records,and the average true positive rate reaches more than 80%,with a false positive rate of less than 20%.