Neural Computational Modeling For Sleep:Theory,Method And Application

With the acceleration of the pace of modern society and the increase of pressure on working and living,more and more people suffer from sleep disorders(SD).SD can cause low body immunity,deterioration of the memory,even trigger neurological diseases and mental diseases.There is a critical unmet need for exploring how to modulate and control sleep disorders effectively in clinics.To address this need,it is necessary to have a good understanding of the underlying mechanisms of SD.Several lines of evidence indicate that sleep disorders are closely related to abnormal discharge activities of neurons in cortex,thalamus,and hypothalamus etc.Therefore,it is helpful to explore the details of neural discharge activities for understanding SD.Neural computational models(NCMs),which are formulated mathematically,can be applied to investigate and gain the mechanistic insight into the dynamics of electrophysiological activities.Electroencephalogram(EEG)provides a tool to record abnormal electrical activities during SD,which contains a large amount of pathophysiological information.Thus,it is a potential way to improve the rationality of modeling and efficiency of its application in clinics by combining the sleep EEGs and NCMs.Guided by such novel modeling strategy,this paper focuses on developing a systematical method for sleep neural computational modeling and its applications in assisting the modulation and controlling of SD.Details are presented as follows:(1)The study of sleep NCM with cholinergic neuromodulation and its applications in modulation of sleep rhythms.In this part,we first develop a new sleep neural computational model TC-ACH by incorporating a cholinergic neuron population(CH)into the classical thalamocortical circuitry,and mathematically modeling the projection mechanisms of neurotransmitter acetylcholine(ACH)released by CH on thalamocortical circuitry.Next,we applied the developed model TC-ACH to study the effects of ACH modulation on thalamocortical rhythms during NREM sleep.Simulation results reveal that1)transitions between different patterns of thalamocortical oscillations are dramatically modulated through diverse projection pathways;2)more spindles appear,and the duration of oscillation patterns during NREM sleep including K-complexes,spindles,and slow oscillations is longer when cholinergic input from CH to thalamocortical neurons becomes stronger.(2)The study of sleep NCM with cholinergic neuromodulation,noradrenergic neuromodulation,and GABAergic neuromodulation and its applications in modulation of sleep rhythms.In this part,we first develop a new sleep neural computational model TC-ANG by: 1)incorporating the cholinergic neuron population,noradrenergic neuron population,and GABAergic neuron population into the classical thalamocortical circuitry;2)mathematically modelling the projection mechanisms of neurotransmitter acetylcholine(ACH),norepinephrine(NE),and GABA on thalamocortical circuitry,which are released by cholinergic neuron,noradrenergic neuron and GABAergic neuron,respectively.Next,we applied the developed model TC-ANG to study the effects of ACH,NE,and GABA modulation on thalamocortical rhythms during NREM sleep.Simulation results demonstrate that 1)transitions between wakefulness,NREM sleep and REM sleep,as well as the corresponding patterns of thalamocortical oscillations are dramatically modulated through the concentration changes of ACH,NE and GABA;2)more spindles and slow waves appear when the concentration of NE decreases.(3)The study of neural modulation mechanisms underlying sleep-related epilepsy(i.e.,Rolandic epilepsy,RE)by combining model and data together.A new neural computational model for RE(RE-NCM)is first proposed by formulating the pathological mechanisms of RE in the neural loop level and cell molecular level separately.Secondly,a model parameter estimation approach was designed,which is based on the Markov Chain Monte Carlo method.Thirdly,performances of the proposed method were verified by numerical experiments on 5 RE children at Massachusetts General Hospital(MGH).Numerical results show that 1)epileptiform spikes can be triggered and promoted by either a reduced NMDA current or h-type current;2)the changes of inhibitory transmission in thalamus lead to a negative relationship between epileptiform spikes and spindles;3)the increase of inhibitory transmission in thalamus promotes spindles,while reduces epileptiform spikes.(4)The study of sleep status trend analysis by combining model and data together.A computational model for sleeping cortex(SC-NCM)is first introduced by incorporating the effects of neurotransmitter acetylcholine and norepinephrine on the discharge activity of excitatory neurons.Secondly,the automatic parameter estimation method was designed by particle swarm optimization algorithm.Thirdly,performances of the proposed method were verified by numerical experiments on Sleep-edf and Dreams databases.Numerical results show that 1)the cortical input from thalamus,inhibitory postsynaptic potential and inhibitory synaptic activity,play a central role in the generation and modulation of the EEG rhythms during stage N1 & N3,N2 and wakefulness,respectively;2)for patients with SD,the cortical input from thalamus increases during sleep stage N1,and the inhibitory synaptic activity decreases during wakefulness,as well as the E-I ratio decreases from light sleep to deep sleep.