Research On The System Of Intelligent Recognition And Intervention Of FOG In PD

Parkinson’s Disease(PD)is the second-largest neurodegenerative disease.Its typical symptoms include tremor,bradykinesia,muscular rigidity,postural instability and gait disorder.Freezing of gait(FOG)is the most common motor symptom in patients with advanced PD.FOG seriously affects the mobility of patients and can cause falls,which is disabling.FOG can be alleviated or relieved by some intervention methods,such as visual,auditory,and somatosensory intervention.There are already several FOG intervention devices,but their intervention method is single and requires manual control to start and stop the intervention device.They are only suitable for use in hospitals or laboratories and cannot meet patients’ needs for rehabilitation training and fall prevention in daily life.In this paper,the algorithm of FOG recognition is studied.By analyzing the signals of FOG in time domain and frequency domain based on Matlab,a total of 7 features are extracted,and the accurate recognition of FOG is realized by feature fusion.Compared with freezing index(FI)alone,fusion features significantly improve the recognition performance of FOG,with accuracy,sensitivity,and specificity of 96.62%,90.03% and 88.18% on public data sets,respectively.To verify the performance of the algorithm in clinical data,an induction experiment is designed based on the pathogenesis of FOG.Total 16 patients are collected in the cooperative hospital,and 105 times of FOG events occurred during the experiment.The accuracy,sensitivity,and specificity of the algorithm on clinical data sets are94.28%,92.07% and 90.82%,respectively.An automatic recognition and intelligent intervention system for FOG is designed and implemented in this paper.The system has experimental mode and independent working mode.It realizes visual and auditory combined intervention through laser emitter and buzzer.Based on C programming language,it realizes real-time recognition of FOG,intelligent control of intervention device and event marking of FOG in experimental mode.The system is used for clinical experiments of FOG intervention.Compared with single intervention,the visual and auditory combined intervention shows more significant relief of FOG and improvement of gait quality in almost all patients.In particular,the best effect is obtained when the laser line spacing is about one step length and the frequency of auditory cue is about 1.1 times of normal walking.Then,the system is evaluated comprehensively in clinical experiments.The result shows that the system could recognize most FOG events.Besides,the intelligent intervention significantly shortens the duration of FOG,and improves the turn speed,stride frequency and stride length of patients.The score of the system increases with the severity of the FOG of patients.