An Artificial Neural Network Algorithm for the Evaluation of Postoperative Rehabilitation of Patients.

Enas Abdulhay,Kunhao Tang,Ruogu Luo,Sanhua Zhang

doi:10.1155/2021/3959844

Enas Abdulhay, Kunhao Tang + Show 2 more

Open Access

https://doi.org/10.1155/2021/3959844

Copy DOI

Abstract

In order to explore the application of artificial neural network in rehabilitation evaluation, a kind of ANN stable and reliable artificial intelligence algorithm is proposed. By learning the existing clinical gait data, this method extracted the gait characteristic parameters of patients with different ages, disease types and course of disease, and repeated data iteration and finally simulated the corresponding gait parameters of patients. Experiments showed that the trained ANN had the same score as the human for most of the data (82.2%, Cohen's kappa = 0.743). There was a strong correlation between ANN and improved Ashworth scores as assessed by human raters (r = 0.825, P < 0.01). As a stable and reliable artificial intelligence algorithm, ANN can provide new ideas and methods for clinical rehabilitation evaluation.

Highlights

Artificial intelligence (AI) has become a neighborhood with many practical applications and active research topics
Based on the fusion of sensor technology and artificial intelligence theory, the automatic evaluation of upper limb motor function is deeply studied, and the motion measurement system is designed from the motor dysfunction of arm, hand, and upper limb activity function. ree different depth learning methods based on sensor data are proposed to realize the feature extraction of clinical scale [4]
According to the results of existing studies, at least 20 subjects (5 in each group) are required to achieve a statistical power of 0.95 in the test to detect differences in limb mobility among patients with different degrees of impairment. erefore, a total of 36 healthy subjects (20 males, 16 females, mean age 23 ± 1.5 years) were enrolled in this study, and each participant was subjected to a Brunnstrom II–VI simulation of upper limb motor dysfunction in stroke patients

Summary

Introduction

Artificial intelligence (AI) has become a neighborhood with many practical applications and active research topics. Artificial neural network (ANN), as an important branch of modern artificial intelligence technology, has been widely and deeply applied to modern medical activities due to its powerful learning ability and stable feature recognition and prediction functions [1]. Erefore, in recent years, artificial intelligence experts have designed and developed a variety of ANN algorithms, and in the real clinical environment to evaluate the rehabilitation of patients, they have achieved satisfactory results. Stroke has become a common disease in clinical practice due to its high morbidity and disability rate It is often accompanied by a variety of functional disorders, among which the upper limb motor dysfunction is the most widely affecting limb dysfunction. Based on the fusion of sensor technology and artificial intelligence theory, the automatic evaluation of upper limb motor function is deeply studied, and the motion measurement system is designed from the motor dysfunction of arm, hand, and upper limb activity function. ree different depth learning methods based on sensor data are proposed to realize the feature extraction of clinical scale [4]

Methods

Results

Conclusion