Artificial neural networks in neurorehabilitation: A scoping review

Abstract

BACKGROUND: Advances in medical technology produce highly complex datasets in neurorehabilitation clinics and research laboratories. Artificial neural networks (ANNs) have been utilized to analyze big and complex datasets in various fields, but the use of ANNs in neurorehabilitation is limited. OBJECTIVE: To explore the current use of ANNs in neurorehabilitation. METHODS: PubMed, CINAHL, and Web of Science were used for literature search. Studies in the scoping review (1) utilized ANNs, (2) examined populations with neurological conditions, and (3) focused rehabilitation outcomes. The initial search identified 1,136 articles. A total of 19 articles were included. RESULTS: ANNs were used for prediction of functional outcomes and mortality (n = 11) and classification of motor symptoms and cognitive status (n = 8). Most ANN-based models outperformed regression or other machine learning models (n = 11) and showed accurate performance (n = 6; no comparison with other models) in predicting clinical outcomes and accurately classifying different neurological impairments. CONCLUSIONS: This scoping review provides encouraging evidence to use ANNs for clinical decision-making of complex datasets in neurorehabilitation. However, more research is needed to establish the clinical utility of ANNs in diagnosing, monitoring, and rehabilitation of individuals with neurological conditions.

Accuracy of machine learning algorithms for the assessment of upper-limb motor impairments in patients with post-stroke hemiparesis: A systematic review and meta-analysis

“…40 Moon et al conducted a scoping review to explore the use of artificial neural networks in neurorehabilitation in various pathologies, including stroke, particularly in the prediction of variables such as functional recovery and rehospitalization. 41 In the same vein, Sirsat et al performed a narrative review about the use of machine learning in stroke patients, grouping them according to their use for the identification of associated risk factors, diagnosis, treatment, and prognosis. 42 In summary, current reviews studying the application of machine learning in stroke patients focus on its use as a plausible tool for prediction and classification of neurological and motor impairments, as well as the assessment of rehabilitation progress.…”

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

“…It will help researchers to understand how neural network is used efficiently for detecting PD in early stages. Compared with other reviews [ 106 , 107 , 108 ] that do not focus on PD disease, this review is unique in its field because it describes and summarizes features of the identified neural network models, datasets, available repository, type of PD evaluation, validation, and research implication. Moreover, this review is different from the previously mentioned reviews by following the latest version of PRISMA-ScR [ 16 ].…”

Rehabilitación neuropsicológica en pacientes pre y posquirúrgicos de lobectomía temporal por epilepsia fármaco resistente

“…El uso de la inteligencia artificial (IA) en escenarios clínicos de relevancia resulta ser muy eficaz debido a su alta capacidad predictiva y analítica en procesos de neurorrehabilitación e investigación neurocientífica, porque permite analizar grandes grupos de datos en dichos procesos para tomar las mejores decisiones, en función de mitigar los daños colaterales de todo tipo de neuropatologías. De hecho, se ha demostrado por medio de revisiones sistemáticas de la literatura que el uso de redes neuronales artificiales permite una notable mejora en los procesos de recuperación de pacientes con lesiones cerebrales tanto adquiridas como congénitas, debido a que se toman decisiones más precisas por parte del equipo de profesionales gracias a los patrones que la IA encuentra en sus observaciones [48]. Sin embargo, algunos autores están de acuerdo con que el uso de la IA en los procesos clínicos aún es muy limitado y poco explorado, por lo que actualmente su uso no va más allá de funcionar como una herramienta predictiva, en función de hacerse una idea anticipada de los posibles resultados al final del proceso, y que puede interpretarse como una segunda opinión para el paciente y los profesionales de parte del algoritmo [48][49].…”