Advancing Brain-Computer Interface Closed-Loop Systems for Neurorehabilitation: A Systematic Review of AI and Machine Learning Innovations in Biomedical Engineering (Preprint)

Published: Feb. 5, 2025

Background: Brain-Computer Interface (BCI) closed-loop systems have emerged as a promising tool in healthcare and wellness monitoring, particularly neurorehabilitation cognitive assessment. With the increasing burden of neurological disorders, including Alzheimer’s Disease Related Dementias (AD/ADRD), there is critical need for real-time, non-invasive monitoring technologies. BCIs enable direct communication between brain external devices, leveraging artificial intelligence (AI) machine learning (ML) to interpret neural signals. However, challenges such signal noise, data processing limitations, privacy concerns hinder widespread implementation. This review explores integration ML AI BCI systems, evaluating their effectiveness improving assessments interventions. Objective: The primary objective this study investigate role enhancing applications. Specifically, we aim analyze methods parameters used these assess different techniques, identify key development implementation, propose framework utilizing longitudinal AD/ADRD patients. By addressing aspects, seeks provide comprehensive overview potential limitations AI-driven healthcare. Methods: A systematic literature was conducted following PRISMA guidelines, focusing on studies published 2019 2024. Research articles were sourced from PubMed, IEEE, ACM, Scopus using predefined keywords related BCIs, AI, AD/ADRD. total 220 papers initially identified, with 18 meeting final inclusion criteria. Data extraction followed structured matrix approach, categorizing based methods, algorithms, proposed solutions. comparative analysis performed synthesize findings trends AI-enhanced monitoring. Results: identified several Transfer Learning, Support Vector Machines, Convolutional Neural Networks, that enhance performance. These improve classification, feature extraction, real-time adaptability, enabling accurate states. long calibration sessions, computational costs, security risks, variability signals also highlighted. To address issues, emerging solutions improved sensor technology, efficient protocols, advanced decoding models are being explored. Additionally, show alert support caregivers managing Conclusions: when integrated ML, offer significant advancements healthcare, neurorehabilitation. Despite potential, accuracy, security, scalability must be addressed clinical adoption. Future research should focus refining models, processing, user accessibility. continued advancements, AI-powered can revolutionize personalized by providing continuous, adaptive intervention patients disorders.

Language: Английский

Correction: Investigating the Best Practices for Engagement in Remote Participatory Design: Mixed Methods Analysis of 4 Remote Studies With Family Caregivers (Preprint)

Published: Dec. 6, 2024

Digital health interventions are a promising method for delivering timely support to underresourced family caregivers. The uptake of digital among caregivers may be improved by engaging in participatory design (PD). In recent years, there has been shift toward conducting PD remotely, which enable participation previously hard-to-reach groups. However, little is known regarding how best facilitate engagement remote This study aims (1) understand the context, quality, and outcomes caregivers’ experiences (2) learn aspects observed approach facilitated or need improved. We analyzed qualitative quantitative data from evaluation reflection surveys interviews completed research community partners (family caregivers) across 4 studies. Studies focused on building For each study, met with 5 sessions 6 months. After session, an survey. 1 studies, survey interview. Descriptive statistics were used summarize data, while reflexive thematic analysis was data. 62.9% (83/132) evaluations projects 1-3, participants described session as “very effective.” 74% (28/38) project 4, feeling “extremely satisfied” session. Qualitative relating context identified that identities partners, technological PD, partners’ understanding their role all influenced engagement. Within domain relationship-building co-learning; satisfaction prework, activities, time allotted, final prototype; inclusivity distribution influence contributed experience Outcomes included ongoing interest after its conclusion, gratitude participation, sense meaning self-esteem. These results indicate high processes few losses specific PD. also demonstrate ways can changed improve partner outcomes. Community should involved inception defining problem solved, used, roles within project. Throughout process, online tools check perceptions power-sharing. Emphasis placed increasing psychosocial benefits (eg, purpose) opportunities participate disseminating findings future

Language: Английский