Post-stroke Brain Functional Network Study Based On Rhythms Of Eeg Signals

Post-stroke survivors often suffer motor function disorders,which are usually associated with anatomical and functional alterations of brain network.However,the mechanism of such changes and the relationship between dysfunctional neural activity and behavioral deficit after stroke is unclear.And during the post-stroke rehabilitation training,the neuroplasticity and compensation mechanism of brain network also remain to be explored.In terms of post-stroke rehabilitation methods,compared with traditional mechanical passive treatment,brain computer interface(BCI)can be combined with some rehabilitation devices to provide patients with a multi-mode,multi-feedback and initiative training.Such training methods could effectively improve the efficiency of post-stroke rehabilitation.And yet,due to the complexity of BCI technology and the difficulty of clinical experiments,these methods haven't been promoted and applied in clinical rehabilitation.In this paper,the changes of brain functional network in stroke patients with hemiplegic side movement were firstly investigated.EEG signals were collected from six stroke patients with right hemiplegia and six healthy subjects when they performing hand grasping tasks,and corresponding brain function network is constructed by partial correlation coefficient(PCC).According to the rhythms of EEG signals during movement,we compared the brain functional network of stroke patients with healthy controls in Alpha and Beta band,respectively.The result suggested that the brain functional network of stroke patients remained “small world” attribute,but the integration and segregation of the whole network were reduced.In particular,the statistical result indicated that Alpha band played an important role in the brain network analysis related to upper limb movement.Next,the BCI-FES upper limb rehabilitation training system based on motor imagination was constructed by combining non-invasive EEG acquisition equipment,functional electrical stimulation(FES)and Unity virtual hand.The system utilized “common space model(CSP)” for feature extraction and “support vector machine(SVM)” for distinguishing the moving and non-moving states during motor imagination,so that patients could control the system through their subjective consciousness and get real-time feedback from virtual hands and FES.Based on this system,we conducted a clinical controlled experiment.Ten stroke patients were randomly divided into the BCI experimental group and the control group to study the impact of BCI-FES rehabilitation training on the brain functional network of stroke patients from the perspective of intergroup and individual.Fugl-Meyer upper limb assessment showed that the scale score of BCI group was significantly higher than that of control group and it exceed minimum clinically significant difference.Brain functional network analysis also showed that the trend of brain network change in BCI group was better than that in control group.Furthermore,directed transfer function(DTF)was used to construct the effective connectivity network during motor imagination of four stroke patients in BCI experimental group.And summarizing the change rules of brain functional network connection mode and network parameters of stroke patients during BCI-FES rehabilitation training,it is concluded that the system could affect the brain functional network of stroke patients during motor imagination,as well as promote the recovery and remodeling of cerebral neural network.