In Vivo Measurement、Model Validation And Optimization Of Transcranial Electrical Stimulation In Human Brain

Transcranial electrical stimulation(t ES),as a non-invasive brain modulation technique,has great significance for brain mechanism and brain health in both clinical and research applications.Studies have shown that the t ES-induced spatial electric field(EF)within the corresponding brain regions are likely a significant factor for the biological effects of t ES.However,it is extremely difficult to obtain the EF distributions directly from live human brains during the stimulation process.It is remarkable that the actual spatial EF distribution characteristics remain largely unclear under different types of t ES,the biophysical properties need to be studied.As an alternative way,simulation models have been developed to understand the spatial electric field distributions throughout the whole brain during t ES.However,there is still a lack of in vivo evidence to assess and validate model prediction performance.How to improve and optimize the model performance need to be investigated further.In this thesis,owing to the advantage of the wider area coverage and minimally invasive for stereotactic-electroencephalographic(s EEG),we focus on the following studies: 1)To measure the t ES-induced EF distributions in vivo,and further study the biophysical characteristics of t ES in human brain.2)To assess and validate the prediction accuracy of simulation models by combining the in vivo measured data.3)To improve the simulation model based on tissue conductivity and segmentation optimization.Firstly,the simulation model of t ES was constructed,and the feasibility and reliability of the measurement method was validated in vitro by combining the s EEG electrode with Agar phantom;Secondly,by using the implanted s EEG electrodes in epilepsy participants,the EF distributions in human brain were measured in vivo during t ES.The effects of stimulation parameters,such as stimulation intensity,frequency(AC/ DC),waveform and electrode montage,on the EF distribution were studied.Our in vivo results validated the rationality of the quasi-static hypothesis of t ES modeling,and proved the stability and repeatability of the EF generated by t ES;Thirdly,we found that the simulation model of t ES can accurately provide predicted values in different age and gender groups,For three kinds of structural disorders in the brain,including malacia,brain parenchyma resection and skull defect,the simulation model has good generalization performance,and the validity of the electrode optimization method was also proved by in vivo measurements;Lastly,two model optimization methods,individual conductivity optimization and accurate tissue segmentation which combined MRI and CT data,are proposed respectively.The prediction performance of the simulation model is improved.The main innovations of this study are:1)We conducted systematically in vivo measurements for intracranial EF generated by t ES.By using the implanted s EEG electrodes in human brain,the t ESinduced EFs were measured under different stimulation intensity,frequency,waveform and electrode montage.We developed the largest and most comprehensive dataset of human intracranial recording during t ES.The in vivo measurements in the live human brain indicated that the t ES-induced electric field accords with the quasi-static assumption,and found the induced EF magnitudes were equal for t DCS and t ACS at the same stimulation intensity.Our results provide confirmatory in vivo evidence for linear superposition of t ES.2)The accuracy and generalization of the simulation model are systematically evaluated.Combined with these in vivo data,the prediction accuracy of the simulation model for different gender and age groups is validated.It is also proved that the simulation model has good generalization performance for special individual cases such as malacia,brain parenchyma resection and skull defect.3)Two optimization methods(conductivity and segmentation optimization)were proposed to improve the model prediction performance.Based on these measured data,the simulation model was improved by individual conductivity optimization and accurate tissue segmentation which combined MRI and CT data.