Analysis And Inversion Research On The Thalamic-Cortial Coupled Neuron Mass Model

Different neural mass models can simulate the generation of EEG signals with different rhythms,and can simulate the relationships between different neural cell mass in the brain.The parameters in the model have physiological significance,so exploring the impact of parameters on the model output and spectrum,as well as identifying biomarkers related to brain function through inversion analysis,is the focus of this study.By combining the forward analysis of the single channel Janson&Rit neural group model,Wendling neural group model,improved alpha rhythm model（i ARM）,and Thalamic cortical coupled neuron mass model（TCC-NMM）,as well as the inversion analysis of the TCC-NMM model and real EEG data,the impact of parameters on mild cognitive impairment disease is studied,to provide new methods for delaying the transition from mild cognitive impairment to Alzheimer’s disease.Firstly,a forward analysis is conducted on the impact of parameter changes on signal rhythm in the single channel Janson&Rit neural mass model,Wendling neural mass model,and i ARM model.The simulation results showed that a decrease in parameters in the excitatory loop and fast inhibitory loop,or an increase in parameters in the slow inhibitory loop,in the model would result inαdecreased rhythm.Exploring the influence of coupling on the main frequency and high-frequency power spectrum of TCC-NMM output,simulation results show that the coupling parameters K_t-c,excitatory and inhibitory synaptic gain parameters H_e,H_i,A,B,and time constantsτ_i.The changes in a,b,and synaptic connectivity coefficients C_t1,C_t2,C₁,and C₂ may lead to a decrease in the frequency spectrum.Secondly,in inversion analysis,the unscented Kalman filtering algorithm and the genetic algorithm fused with vectors are used to identify and track parameters in TCC-NMM.The simulation results of single parameter identification show that the algorithm has a certain degree of accuracy.Combined with the conclusion of forward analysis,it was found that parameters with less impact on the high-frequency power spectrum of the model output are more difficult to identify during parameter identification.The results of identifying multiple parameters in the coupled model by module indicate that the recognition accuracy of the cortical module is higher than that of the thalamic module,and there is a certain balance between the multiple parameters.From the forward analysis of the impact of multiple parameter changes on the model’s output main frequency,it was found that when conducting joint analysis with{τ_e,τ_i},{A,B,G}and{a,b,g},τ_i,G and g have almost no effect on the main frequency of the model’s output,while parameter identification for the same group is accurate.Finally,the EEG data of patients with mild cognitive impairment were inverted and analyzed using the unscented Kalman filtering algorithm and the fusion vector based genetic algorithm.The recognition results using the unscented Kalman filter algorithm show that the recognition value of parameter C₁ in the mild cognitive impairment group is significantly lower than that in the normal control group,especially in the frontal and occipital areas.The recognition value of parameter C₁ is significantly lower than that in the normal control group,which is significantly correlated with the results of Mo CA and MMSE scales,indicating that the reduction of the number of synaptic connections between pyramidal somatic cell cells and excitatory interneuron may be one of the neural mechanisms leading to the reduction of MCI spectrum.The recognition results of the genetic algorithm based on fusion vectors indicate that the time constantτ_i significantly increased.Indicating that there may be damage or degradation between the thalamic reticular nucleus cells and the thalamic relay nucleus neurons,resulting in slow or ineffective information transmission,which may be one of the neural mechanisms in patients with mild cognitive impairment.