Characteristic Neural Signal Analysis Of Sensorimotor–related Brain Areas In Rhesus Macaque During Grasping

Reaching out to grasp the objects in the environment is a very important life ability in human daily life.Due to different reasons,such as trauma and disease,the loss of some people's movement and/or sensory functions seriously affects their daily life.For these patients,mild cases can be improved by surgery,drugs and rehabilitation.But for patients with severe disease,such as patients with complete loss of motor and sensory function in the upper limbs,the current clinical practice is mainly to perform an alternative or supplement to limb movement and sensory function through brain-computer interface technology.Some research teams implanted electrode arrays in multiple motion-related brain areas.Using spike signals collected by the electrodes,they continuously controlled the prosthesis/mechanical arm for three-dimensional stretching through an optimized decoding algorithm and allowed the subject to generate sensory feedback by electrically stimulating the primary somatosensory cortex,improving the patients' motor or sensory function and their quality of life.Although clinical progress has been made in clinical trials through implantable brain-computer interface technology,there are still lack of researches to figure out how to achieve precise and stable control of the prosthetic/mechanical arm under more complex environmental variables.This study intends to preliminarily explore the changes of neural signals in the primary motor cortex,primary somatosensory cortex and posterior parietal cortex under the condition of complex environmental variables,thus providing some theory for the design and optimization of decoding algorithms.In this study,two macaques were used as research objects,and a multivariable controllable reach-to-grasp research platform was built.The implanted brain-computer interface technique was used to implant multielectrodes into the primary motor cortex,primary somatosensory cortex and posterior parietal cortex of macaque monkeys.The neural signals in the process of interacting with variables such as time,motion state of the object,and motion disturbance are collected in real time when the macaque performs the reach-to-grasp task.On this basis,this research explored the patterns of signal changes under the conditions of complex environmental changes during the process by using spike and local field potential signal to analyze the firing rates,the power spectrum of the local field potential,as well as the coherence between the brain areas.It is found that:(1)In the process of interacting with the time variable,when the time interference condition is set in the preparation stage,some neurons shows a difference in firing rates between the two groups of short-term waiting and long-term waiting during reaction,reach and grasp stages;and the mean beta power of the first 300 ms as well as the maximum speed of beta power at the beginning of the preparation phase showed higher in short-term waiting group than the long-term waiting group;the short-term waiting group has a latent period of peak beta power in the preparation stage that is significantly shorter than the long-term waiting group;while for the random group,the average beta power of the first 300 ms and the maximum speed of beta power change showed no statistical difference between groups.When the time variable intervention was in reach stage,some neurons in the primary motor cortex,the primary somatosensory cortex and the posterior parietal cortex exhibited tunable changes in the firing rates at different stages of the movement;(2)in the process of interacting with the motion state variables of the object,some neurons in the primary motor cortex,primary somatosensory cortex and posterior parietal cortex showed a tuning change in motion directions at different stages of the movement;the results of neural signal and trajectory analysis based on Kalman filter showed that individual neurons exhibit the preference for decoding different motion parameters and the variability between different trials;SVM-based directional classification results showed that the direction decoding accuracy is gradually improved from the reaction stage to the grasp stage;based on coherence analysis,the results showed that the local field potentials in the delta and theta bands showed a decrease in the coherence coefficient in the reaction period in all three brain regions;(3)in the process of interacting with the motion disturbance variables,some neurons in the primary motor cortex,the primary sensory cortex and the posterior parietal cortex showed tunning changes for movement directions in different stages of the movement;some of the primary motor cortex and primary sensory cortex neurons showed two peaks in the reaction and reach periods;the analysis of neural signal and trajectory based on Kalman filter showed that individual neurons exhibit preference for decoding of different motion parameters and the variability between different trials;SVM-based directional classification results showed that the direction decoding accuracy reached morte than 79% during the grasp phase and more than 89% in the later stage of the grasp period;based on coherence analysis,the local field potentials of the delta and theta bands showed a decrease in the coherence coefficient of the reaction period in all three brain areas.From the above results,it can be concluded that:(1)some neurons in the primary motor cortex,the primary somatosensory cortex and the posterior parietal cortex can exhibit plasticity changes in firing rates during temporal perception of different lengths;the power spectrum of the beta band during the preparation period may reflect the working memory of the task sequence;(2)some neurons in the primary motor cortex,the primary somatosensory cortex and the posterior parietal cortex showed tunning variation in firing rates for different movement directions,and as the object motion direction information increases,it exhibits a tuning change with increased classification accuracy;at the same time,the low-frequency local field potential(delta and theta)show obvious tunable change in the reaction period for information interaction between brain areas;(3)some neurons in the primary motor cortex,primary somatosensory cortex and posterior parietal cortex show varying degrees of directional tuning changes in their firing rates during the process of grasping moving object under disturbance;and some neurons in the primary motor cortex,primary somatosensory cortex with "bimodal" characteristics that are significantly tuned to changes in motion during the experimental process;in addition,the low-frequency deta and theta bands also exhibit significant tuning changes in the reaction period between brain areas.This study preliminarily explores the changes of neural signals in the sensorimotor-related brain areas to time variables,object motion state variables and motion disturbance variables,and provides part of theoretical basis for the design and optimization of decoding algorithms based on neural signals obtained from implantable electrodes.