Network-based predictive models for artificial intelligence: an interpretable application of machine learning techniques in the assessment of depression in stroke patients DOI Creative Commons

Wenwei Zuo,

Xuelian Yang

BMC Geriatrics, Journal Year: 2025, Volume and Issue: 25(1)

Published: March 22, 2025

Depression is a common complication after stroke that may lead to increased disability and decreased quality of life. The objective this study was develop validate an interpretable predictive model assess the risk depression in patients using machine learning (ML) methods. This included 1143 from NHANES database between 2005 2020. First, factors for were determined by univariate multivariate logistic regression analysis. Next, five algorithms used construct models, several evaluation metrics (including area under curve (AUC)) compare performance models. In addition, SHAP (Shapley Additive Explanations) method rank importance features interpret final model. We screened seven Among 5 XGBoost (extreme gradient boosting) showed best discriminative ability, with AUC ROC (receiver operating characteristic curve) test set 0.746 accuracy 0.834. prediction results interpreted detail algorithm. also developed web-based calculator provides convenient tool predicting at following link: https://prediction-model-for-depression.streamlit.app . Our serves as auxiliary clinical judgment, aimed early effective identification patients.

Language: Английский

Multiscale brain modeling: bridging microscopic and macroscopic brain dynamics for clinical and technological applications DOI Creative Commons
Ondřej Krejcar, Hamidreza Namazi

Frontiers in Cellular Neuroscience, Journal Year: 2025, Volume and Issue: 19

Published: Feb. 19, 2025

The brain's complex organization spans from molecular-level processes within neurons to large-scale networks, making it essential understand this multiscale structure uncover brain functions and address neurological disorders. Multiscale modeling has emerged as a transformative approach, integrating computational models, advanced imaging, big data bridge these levels of organization. This review explores the challenges opportunities in linking microscopic phenomena macroscopic functions, emphasizing methodologies driving progress field. It also highlights clinical potential including their role advancing artificial intelligence (AI) applications improving healthcare technologies. By examining current research proposing future directions for interdisciplinary collaboration, work demonstrates how can revolutionize both scientific understanding practice.

Language: Английский

Citations

2
Quantum deep learning in neuroinformatics: a systematic review DOI Creative Commons
Nabil Anan Orka, Md. Abdul Awal, Píetro Lió

et al.

Artificial Intelligence Review, Journal Year: 2025, Volume and Issue: 58(5)

Published: Feb. 14, 2025

Language: Английский

Citations

0
Combined deep and reinforcement learning with gaming to promote healthcare in neurodevelopmental disorders: a new hypothesis DOI Creative Commons
Fabrizio Stasolla, Anna Passaro,

Enza Curcio

et al.

Frontiers in Human Neuroscience, Journal Year: 2025, Volume and Issue: 19

Published: March 14, 2025

Children and adolescents with neurodevelopmental disorders (NDD) may experience significant problems dealing daily activities and/or everyday life environmental requests. Besides intellectual disabilities, communication challenging behaviors occur. Commonly, isolation passivity are acknowledged. Accordingly, social interactions be relevantly compromised. NDD usually has an early onset, a variable clinical manifestation, wide range of severity, recognized comorbidity (Howner et al., 2018;Malik 2023). Their conditions have negative outcomes on their quality-of-lifefamilies' caregivers' burden meaningfully increased consequences overall management healthcare (Kanniappan 2024;Lefton-Greif 2024;Materula To tackle this issue assessment is crucial (Chorna 2024;Henry 2016). Thus, either standard tests or technology-based solutions available (Ceruti 2024;Niu 2022;Woodcock Blackwell, 2020). Traditional relies neuropsychological evaluation (Haddad 2024;Hamadelseed Technology-aided options represent functional bridge between personal skills requests by enhancing self-determination positive occupation accordingly (Passaro 2024). Recently, artificial intelligence-based programs (AI) emerged (Boubakri Nafil, Both rehabilitative goals targeted (Anbarasi 2024;Climent-Pérez 2024).Deep learning (DL), as part machine (ML), been growingly used to evaluate the normal brain functioning differentiate individuals who development at risk developmental (Kucewicz 2023;Li 2024;Swinckels For example, DL algorithms convolutional neural networks (CNN) progressed allowing future amount data patterns which enables subjectivity in extraction procedure. Successful implementations CNN documented. positively investigated through magnetic resonance (fMRI) main domains (Hu 2023).Reinforcement Learning (RL) further ML adopted for purposes.That is, intelligent agent continuously interacting participant cognitive task reinforced such interaction capable it. Based interaction, it will provide optimal task. Consequently, ensures participants highly customized tailored along all working sessions ideal process ensured (Zini 2022). RL-based principles emotional regulation neurodegenerative diseases (Stasolla 2024;Stasolla Di Gioia, 2023).Gamification considered advanced technological cornerstone both purposes. Educational recovery targeted. Education, healthcare, rehabilitation objectives pursued. Significant improvements reported disabilities. Self-determination, independence, fulfillment fostered embedding features challenges, competitions, rewards. gamification can help persons active role constructive engagement 2024).A literature overview was performed Scopus. Neurodevelopmental disorders, quality life, DL, RL, gamification, assessment, were merged keywords. Although detailed widely (Alves 2020a;Bakır 2023;Brzosko 2019;Nahar 2024;Ouyang 2024;Pandya 2024;Rahman 2024;Rodulfo-Cárdenas 2023;Wyatt 2024;Zhao 2024), no records found integration. In line above, aims current opinion paper (a) reader concise framework use gaming strategy 63 including five reviews published 2020 2024 2024;Soybilgic Avcin, 2020;Swinckels 2024;Wang Li al. (2024) conducted comprehensive review electroencephalography (EEG) method that changes activity marker identification autism spectrum (ASD). The included methods. Future perspectives challenges highlighted automatically diagnose ASD EEG signals emphasize automated identification. Kucewicz RL also represented 60 documents Scopus last two decades (Brzosko 2019;Meyer 2005;Rodulfo-Cárdenas 2023;Swan 2016;Wyatt By inspection, contribution Meyer (2005) animal model detection schizophrenia, irrelevant work, not accordingly. Conversely, exploitation decision-making associated default network regions. Data interpreted context architecture useful support flexible switching externally internally directed processes, mandatory adaptive purposeful behaviors. Moreover, they surveyed studies involving neurodevelopmental, neuropsychological, neuropsychiatric well lifespan diseases. differences exploring-exploiting observed across populations corroborating modes supported independent circuits. Comprehensive circuity mapping behavioral correlates exploration-exploitation humans warranted. A new trans-diagnostic approach surveillance, intervention decline dysfunction mental health population putted forward.Gamification 3 2020b;Bakır 2023;Boubakri Alves (2020) Boubakri Nafil explored potential impact accessibility issues According PRISMA guidelines, seven databases. Fifty-three selected. revealed suitable blindness, visual impairments, improving processes Nevertheless, gaps remained filled need more accurate integration emerging technologies like AI-based customize proven effective targeting balanced required.Specifically, focused identify its benefits address existing claim generate inclusive experiences.An illustrative example might combined immersive system AR, VR, gamification. One argue funny promote executive functions, communicative skills, interactions. Different rigorously designed. principles, one envisage different tasks properly participant's capacities adapted his/her performance. responses recorded, monitored, tracked. Stasolla (2025) proposed scoping specific topic.Considering hypothesis integrated solution based proposed. Matched recently outlined three-step hierarchical recommended purposes (see figure 1). design first step exploring during tasks. Once differentiated, (i.e., second step), plan funny, educational, promoting redirecting into adaptive, occupational (Chiapparino 2011).In third step, validity assessed expert external raters validation procedures 2019). depending functioning.For instance, severe profound multiple suppose basic discrimination emotions situations eliciting emotions. moderate level virtual reality (VR), opportunities enhance high estimated borderline mild implement access literacy leisure 2011;Lancioni 2007). Finally, supporting academic needs, fully environment (Bennewith 2024;Panzeri Individuals levels disabilities difficulties contexts settings. Because commonly present intellectual, motor, communicative, sensorial constantly rely families' assistance. This condition deleterious image, status, desirability. fact, seriously hamper life. overcome issue, technology-aided helpful previous findings demonstrated (Kinsella 2017;Paul 2023;Pham valid avenue children (Mengi Malhotra, Here following considerations suggested.First, self-determination, engagement, inclusion fostered. role, occupation, reduction passivity. Assessment critically (Harris, 2020;Kwan 2021;Lancioni 2008Lancioni , 2004a;;Stasolla 2014bStasolla 2014a;;Zimmer Dunn, 2021).Second, individuals' outlined.For simple very low limited repertoire. Otherwise, Furthermore, complex managed (Michalski 2021).Third, include unique program. embedded. Through systems people disorders. Gamification play dual educational (Gao 2024;Liu 2024;Shariat 2024).Fourth, reduced. assessed. being constructively engaged, occupied, solutions, easily involved settings (Song 2022).Despite promising postulated outcomes, some relevant should First, empirical available. Systematic reviews, metanalysis, single-subject comparisons matched longitudinal carried out. Second, sustainability carefully considered. Human, financial, resources investigate affordability suitability combination. Ethical additionally Third, differentiation currently lacking large body ASD. Other rare genetic studies. research deal topics experimental collected, (b) systematic groups investigations single subjects sought, (c) targeted, (d) focus prioritized.

Language: Английский

Citations

0
Network-based predictive models for artificial intelligence: an interpretable application of machine learning techniques in the assessment of depression in stroke patients DOI Creative Commons

Wenwei Zuo,

Xuelian Yang

BMC Geriatrics, Journal Year: 2025, Volume and Issue: 25(1)

Published: March 22, 2025

Depression is a common complication after stroke that may lead to increased disability and decreased quality of life. The objective this study was develop validate an interpretable predictive model assess the risk depression in patients using machine learning (ML) methods. This included 1143 from NHANES database between 2005 2020. First, factors for were determined by univariate multivariate logistic regression analysis. Next, five algorithms used construct models, several evaluation metrics (including area under curve (AUC)) compare performance models. In addition, SHAP (Shapley Additive Explanations) method rank importance features interpret final model. We screened seven Among 5 XGBoost (extreme gradient boosting) showed best discriminative ability, with AUC ROC (receiver operating characteristic curve) test set 0.746 accuracy 0.834. prediction results interpreted detail algorithm. also developed web-based calculator provides convenient tool predicting at following link: https://prediction-model-for-depression.streamlit.app . Our serves as auxiliary clinical judgment, aimed early effective identification patients.

Language: Английский

Citations

0