A Novel Deep Learning Method Based on an Overlapping Time Window Strategy for Brain-Computer Interface-Based Stroke Rehabilitation

Advances in brain-computer interface controlled functional electrical stimulation for upper limb recovery after stroke

Stroke often results in varying degrees of functional impairment, significantly affecting patients' quality of daily life. In recent years, brain-computer interface-controlled functional electrical stimulation has offered new therapeutic approaches for post-stroke rehabilitation. This paper reviews the application of BCI-FES in the recovery of upper limb function after stroke and explores its underlying mechanisms. By analyzing relevant studies, the aim is to provide a theoretical basis for rehabilitating upper limb function post-stroke, promote BCI-FES, and offer guidance for future clinical practice.

Knowledge mapping and research trends of brain-computer interface technology in rehabilitation: a bibliometric analysis

Background

Although the application of brain-computer interface (BCI) technology in rehabilitation has been extensively studied, a systematic and comprehensive bibliometric analysis of this area remains lacking. Thus, this study aims to analyze the research progress of BCI technology in rehabilitation through bibliometric methods.

Methods

The study retrieved relevant publications on BCI technology in rehabilitation from the Web of Science Core Collection (WoSCC) between January 1, 2004, and June 30, 2024. The search was conducted using thematic queries, and the document types included "original articles" and "review articles." Bibliometric analysis and knowledge mapping were performed using the Bibliometrix package in R software and CiteSpace software.

Results

During the study period, a total of 1,431 publications on BCI technology in rehabilitation were published by 4,932 authors from 1,281 institutions across 79 countries in 386 academic journals. The volume of research literature in this field has shown a steady upward trend. The United States of America (USA) and China are the primary contributors, with Eberhard Karls University of Tübingen being the most active research institution. The journal Frontiers in Neuroscience published the most articles, while the Journal of Neural Engineering was the most cited. Niels Birbaumer not only authored the most articles but also received the highest number of citations. The main research areas include neurology, sports medicine, and ophthalmology. The diverse applications of BCI technology in stroke and spinal cord injury rehabilitation, as well as the evaluation of BCI performance, are current research hotspots. Moreover, deep learning has demonstrated significant potential in BCI technology rehabilitation applications.

Conclusion

This bibliometric study provides an overview of the research landscape and developmental trends of BCI technology in rehabilitation, offering valuable reference points for researchers in formulating future research strategies.

Brain-Computer Interfaces for Upper Limb Motor Recovery after Stroke: Current Status and Development Prospects (Review)

Brain-computer interfaces (BCIs) are a group of technologies that allow mental training with feedback for post-stroke motor recovery. Varieties of these technologies have been studied in numerous clinical trials for more than 10 years, and their construct and software are constantly being improved. Despite the positive treatment results and the availability of registered medical devices, there are currently a number of problems for the wide clinical application of BCI technologies. This review provides information on the most studied types of BCIs and its training protocols and describes the evidence base for the effectiveness of BCIs for upper limb motor recovery after stroke. The main problems of scaling this technology and ways to solve them are also described.

Research hotspots and trends of brain-computer interface technology in stroke: a bibliometric study and visualization analysis

Background

The incidence and mortality rates of stroke are escalating due to the growing aging population, which presents a significant hazard to human health. In the realm of stroke, brain-computer interface (BCI) technology has gained considerable attention as a means to enhance treatment efficacy and improve quality of life. Consequently, a bibliometric visualization analysis was performed to investigate the research hotspots and trends of BCI technology in stroke, with the objective of furnishing reference and guidance for future research.

Methods

This study utilized the Science Citation Index Expanded (SCI-Expanded) within the Web of Science Core Collection (WoSCC) database as the data source, selecting relevant literature published between 2013 and 2022 as research sample. Through the application of VOSviewer 1.6.19 and CiteSpace 6.2.R2 visualization analysis software, as well as the bibliometric online analysis platform, the scientific knowledge maps were constructed and subjected to visualization display, and statistical analysis.

Results

This study encompasses a total of 693 relevant literature, which were published by 2,556 scholars from 975 institutions across 53 countries/regions and have been collected by 185 journals. In the past decade, BCI technology in stroke research has exhibited an upward trend in both annual publications and citations. China and the United States are high productivity countries, while the University of Tubingen stands out as the most contributing institution. Birbaumer N and Pfurtscheller G are the authors with the highest publication and citation frequency in this field, respectively. Frontiers in Neuroscience has published the most literature, while Journal of Neural Engineering has the highest citation frequency. The research hotspots in this field cover keywords such as stroke, BCI, rehabilitation, motor imagery (MI), motor recovery, electroencephalogram (EEG), neurorehabilitation, neural plasticity, task analysis, functional electrical stimulation (FES), motor impairment, feature extraction, and induced movement therapy, which to a certain extent reflect the development trend and frontier research direction of this field.

Conclusion

This study comprehensively and visually presents the extensive and in-depth literature resources of BCI technology in stroke research in the form of knowledge maps, which facilitates scholars to gain a more convenient understanding of the development and prospects in this field, thereby promoting further research work.

Input Shape Effect on Classification Performance of Raw EEG Motor Imagery Signals with Convolutional Neural Networks for Use in Brain-Computer Interfaces

EEG signals are interpreted, analyzed and classified by many researchers for use in brain-computer interfaces. Although there are many different EEG signal acquisition methods, one of the most interesting is motor imagery signals. Many different signal processing methods, machine learning and deep learning models have been developed for the classification of motor imagery signals. Among these, Convolutional Neural Network models generally achieve better results than other models. Because the size and shape of the data is important for training Convolutional Neural Network models and discovering the right relationships, researchers have designed and experimented with many different input shape structures. However, no study has been found in the literature evaluating the effect of different input shapes on model performance and accuracy. In this study, the effects of different input shapes on model performance and accuracy in the classification of EEG motor imagery signals were investigated, which had not been specifically studied before. In addition, signal preprocessing methods, which take a long time before classification, were not used; rather, two CNN models were developed for training and classification using raw data. Two different datasets, BCI Competition IV 2A and 2B, were used in classification processes. For different input shapes, 53.03-89.29% classification accuracy and 2-23 s epoch time were obtained for 2A dataset, 64.84-84.94% classification accuracy and 4-10 s epoch time were obtained for 2B dataset. This study showed that the input shape has a significant effect on the classification performance, and when the correct input shape is selected and the correct CNN architecture is developed, feature extraction and classification can be done well by the CNN architecture without any signal preprocessing.