Artificial neural networks in neurorehabilitation: A scoping review

Towards Transforming Neurorehabilitation: The Impact of Artificial Intelligence on Diagnosis and Treatment of Neurological Disorders

Background and Objectives: Neurological disorders like stroke, spinal cord injury (SCI), and Parkinson's disease (PD) significantly affect global health, requiring accurate diagnosis and long-term neurorehabilitation. Artificial intelligence (AI), such as machine learning (ML), may enhance early diagnosis, personalize treatment, and optimize rehabilitation through predictive analytics, robotic systems, and brain-computer interfaces, improving outcomes for patients. This systematic review examines how AI and ML systems influence diagnosis and treatment in neurorehabilitation among neurological disorders. Materials and Methods: Studies were identified from an online search of PubMed, Web of Science, and Scopus databases with a search time range from 2014 to 2024. This review has been registered on Open OSF (n) EH9PT. Results: Recent advancements in AI and ML are revolutionizing motor rehabilitation and diagnosis for conditions like stroke, SCI, and PD, offering new opportunities for personalized care and improved outcomes. These technologies enhance clinical assessments, therapy personalization, and remote monitoring, providing more precise interventions and better long-term management. Conclusions: AI is revolutionizing neurorehabilitation, offering personalized, data-driven treatments that enhance recovery in neurological disorders. Future efforts should focus on large-scale validation, ethical considerations, and expanding access to advanced, home-based care.

Innovations in Multiple Sclerosis Care: The Impact of Artificial Intelligence via Machine Learning on Clinical Research and Decision-Making

Artificial intelligence (AI) and its specialized subcomponent machine learning are becoming increasingly popular analytic techniques. With this growth, clinicians and health care professionals should soon expect to see an increase in diagnostic, therapeutic, and rehabilitative technologies and processes that use elements of AI. The purpose of this review is twofold. First, we provide foundational knowledge that will help health care professionals understand these modern algorithmic techniques and their implementation for classification and clustering tasks. The phrases artificial intelligence and machine learning are defined and distinguished, as are the metrics by which they are assessed and delineated. Subsequently, 7 broad categories of algorithms are discussed, and their uses explained. Second, this review highlights several key studies that exemplify advances in diagnosis, treatment, and rehabilitation for individuals with multiple sclerosis using a variety of data sources-from wearable sensors to questionnaires and serology-and elements of AI. This review will help health care professionals and clinicians better understand AI-dependent diagnostic, therapeutic, and rehabilitative techniques, thereby facilitating a greater quality of care.

Improvement of predictive accuracies of functional outcomes after subacute stroke inpatient rehabilitation by machine learning models

OBJECTIVES

Stepwise linear regression (SLR) is the most common approach to predicting activities of daily living at discharge with the Functional Independence Measure (FIM) in stroke patients, but noisy nonlinear clinical data decrease the predictive accuracies of SLR. Machine learning is gaining attention in the medical field for such nonlinear data. Previous studies reported that machine learning models, regression tree (RT), ensemble learning (EL), artificial neural networks (ANNs), support vector regression (SVR), and Gaussian process regression (GPR), are robust to such data and increase predictive accuracies. This study aimed to compare the predictive accuracies of SLR and these machine learning models for FIM scores in stroke patients.

METHODS

Subacute stroke patients (N = 1,046) who underwent inpatient rehabilitation participated in this study. Only patients' background characteristics and FIM scores at admission were used to build each predictive model of SLR, RT, EL, ANN, SVR, and GPR with 10-fold cross-validation. The coefficient of determination (R2) and root mean square error (RMSE) values were compared between the actual and predicted discharge FIM scores and FIM gain.

RESULTS

Machine learning models (R2 of RT = 0.75, EL = 0.78, ANN = 0.81, SVR = 0.80, GPR = 0.81) outperformed SLR (0.70) to predict discharge FIM motor scores. The predictive accuracies of machine learning methods for FIM total gain (R2 of RT = 0.48, EL = 0.51, ANN = 0.50, SVR = 0.51, GPR = 0.54) were also better than of SLR (0.22).

CONCLUSIONS

This study suggested that the machine learning models outperformed SLR for predicting FIM prognosis. The machine learning models used only patients' background characteristics and FIM scores at admission and more accurately predicted FIM gain than previous studies. ANN, SVR, and GPR outperformed RT and EL. GPR could have the best predictive accuracy for FIM prognosis.

Use of artificial neural networks in the prognosis of musculoskeletal diseases-a scoping review

To determine the current evidence on artificial neural network (ANN) in prognostic studies of musculoskeletal diseases (MSD) and to assess the accuracy of ANN in predicting the prognosis of patients with MSD. The scoping review was reported under the Preferred Items for Systematic Reviews and the Meta-Analyses extension for Scope Reviews (PRISMA-ScR). Cochrane Library, Embase, Pubmed, and Web of science core collection were searched from inception to January 2023. Studies were eligible if they used ANN to make predictions about MSD prognosis. Variables, model prediction accuracy, and disease type used in the ANN model were extracted and charted, then presented as a table along with narrative synthesis. Eighteen Studies were included in this scoping review, with 16 different types of musculoskeletal diseases. The accuracy of the ANN model predictions ranged from 0.542 to 0.947. ANN models were more accurate compared to traditional logistic regression models. This scoping review suggests that ANN can predict the prognosis of musculoskeletal diseases, which has the potential to be applied to different types of MSD.

The Role of Neural Network for the Detection of Parkinson's Disease: A Scoping Review

Background: Parkinson's Disease (PD) is a chronic neurodegenerative disorder that has been ranked second after Alzheimer's disease worldwide. Early diagnosis of PD is crucial to combat against PD to allow patients to deal with it properly. However, there is no medical test(s) available to diagnose PD conclusively. Therefore, computer-aided diagnosis (CAD) systems offered a better solution to make the necessary data-driven decisions and assist the physician. Numerous studies were conducted to propose CAD to diagnose PD in the early stages. No comprehensive reviews have been conducted to summarize the role of AI tools to combat PD. Objective: The study aimed to explore and summarize the applications of neural networks to diagnose PD. Methods: PRISMA Extension for Scoping Reviews (PRISMA-ScR) was followed to conduct this scoping review. To identify the relevant studies, both medical databases (e.g., PubMed) and technical databases (IEEE) were searched. Three reviewers carried out the study selection and extracted the data from the included studies independently. Then, the narrative approach was adopted to synthesis the extracted data. Results: Out of 1061 studies, 91 studies satisfied the eligibility criteria in this review. About half of the included studies have implemented artificial neural networks to diagnose PD. Numerous studies included focused on the freezing of gait (FoG). Biomedical voice and signal datasets were the most commonly used data types to develop and validate these models. However, MRI- and CT-scan images were also utilized in the included studies. Conclusion: Neural networks play an integral and substantial role in combating PD. Many possible applications of neural networks were identified in this review, however, most of them are limited up to research purposes.

Estimation of Axle Torque for an Agricultural Tractor Using an Artificial Neural Network

The objective of this study was to develop a model to estimate the axle torque (AT) of a tractor using an artificial neural network (ANN) based on a relatively low-cost sensor. ANN has proven to be useful in the case of nonlinear analysis, and it can be applied to consider nonlinear variables such as soil characteristics, unlike studies that only consider tractor major parameters, thus model performance and its implementation can be extended to a wider range. In this study, ANN-based models were compared with multiple linear regression (MLR)-based models for performance verification. The main input data were tractor engine parameters, major tractor parameters, and soil physical properties. Data of soil physical properties (i.e., soil moisture content and cone index) and major tractor parameters (i.e., engine torque, engine speed, specific fuel consumption, travel speed, tillage depth, and slip ratio) were collected during a tractor field experiment in four Korean paddy fields. The collected soil physical properties and major tractor parameter data were used to estimate the AT of the tractor by the MLR- and ANN-based models: 250 data points were used for developing and training the model were used, the 50 remaining data points were used to test the model estimation. The AT estimated with the developed MLR- and ANN-based models showed agreement with actual measured AT, with the R² value ranging from 0.825 to 0.851 and from 0.857 to 0.904, respectively. These results suggest that the developed models are reliable in estimating tractor AT, while the ANN-based model showed better performance than the MLR-based model. This study can provide useful results as a simple method using ANNs based on relatively inexpensive sensors that can replace the existing complex tractor AT measurement method is emphasized.

Classification of Parkinson's disease and essential tremor based on balance and gait characteristics from wearable motion sensors via machine learning techniques: a data-driven approach

BACKGROUND

Parkinson's disease (PD) and essential tremor (ET) are movement disorders that can have similar clinical characteristics including tremor and gait difficulty. These disorders can be misdiagnosed leading to delay in appropriate treatment. The aim of the study was to determine whether balance and gait variables obtained with wearable inertial motion sensors can be utilized to differentiate between PD and ET using machine learning. Additionally, we compared classification performances of several machine learning models.

METHODS

This retrospective study included balance and gait variables collected during the instrumented stand and walk test from people with PD (n = 524) and with ET (n = 43). Performance of several machine learning techniques including neural networks, support vector machine, k-nearest neighbor, decision tree, random forest, and gradient boosting, were compared with a dummy model or logistic regression using F1-scores.

RESULTS

Machine learning models classified PD and ET based on balance and gait characteristics better than the dummy model (F1-score = 0.48) or logistic regression (F1-score = 0.53). The highest F1-score was 0.61 of neural network, followed by 0.59 of gradient boosting, 0.56 of random forest, 0.55 of support vector machine, 0.53 of decision tree, and 0.49 of k-nearest neighbor.

CONCLUSIONS

This study demonstrated the utility of machine learning models to classify different movement disorders based on balance and gait characteristics collected from wearable sensors. Future studies using a well-balanced data set are needed to confirm the potential clinical utility of machine learning models to discern between PD and ET.