Innovative Speech-Based Deep Learning Approaches for Parkinson’s Disease Classification: A Systematic Review

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(17), P. 7873 - 7873

Published: Sept. 4, 2024

Parkinson’s disease (PD), the second most prevalent neurodegenerative disorder worldwide, frequently presents with early-stage speech impairments. Recent advancements in Artificial Intelligence (AI), particularly deep learning (DL), have significantly enhanced PD diagnosis through analysis of data. Nevertheless, progress research is restricted by limited availability publicly accessible speech-based datasets, primarily due to privacy concerns. The goal this systematic review explore current landscape DL approaches for classification, based on 33 scientific works published between January 2020 and March 2024. We discuss their available resources, capabilities, potential limitations, issues related bias, explainability, privacy. Furthermore, provides an overview datasets open-source material PD. identified are categorized into end-to-end (E2E) learning, transfer (TL), acoustic feature extraction (DAFE). Among E2E approaches, Convolutional Neural Networks (CNNs) prevalent, though Transformers increasingly popular. face challenges such as data computational especially Transformers. TL addresses these providing more robust better generalizability across languages. DAFE aims improve explainability interpretability results examining specific effects features both other traditional machine (ML) methods. However, it often underperforms compared approaches.

Language: Английский

---

Digital and Computational Pathology Applications in Bladder Cancer: Novel Tools Addressing Clinically Pressing Needs

Modern Pathology, Journal Year: 2024, Volume and Issue: 38(1), P. 100631 - 100631

Published: Oct. 12, 2024

Language: Английский

---

Machine Learning Algorithms for Iron Deficiency Anemia Detection in Children Using Palm Images

International Journal of Education and Management Engineering, Journal Year: 2024, Volume and Issue: 14(1), P. 1 - 15

Published: Feb. 2, 2024

Anemia is a common condition among adults, particularly in children and pregnant women. defined as lack of healthy red blood cells or hemoglobin. Early identification anemia critical for excellent health well-being, which contributes to the sustainable development goals (SDGs), notably SDG 3. The intrusive way detecting has several hurdles, including anxiety cost, impedes development. With advent technology, it create non-invasive techniques diagnose that can minimize costs while also improving detection efficacy. A distinct technique developed this study employing machine learning (ML) models. This study's dataset contains 4260 observations non-anemic (0) anemic (1) children. To train dataset, six (6) different ML models were employed: k-Nearest Neighbor (KNN), decision tree (DT), logistic regression (LR), nave bayes (NB), random forest (RF), kernel-support vector (KSVM). DT RF obtained highest accuracy 99.92%, followed by KNN at 98.98%. used produced substantial results. received high marks on performance evaluation metrics such accuracy, recall, F1-score, Area Under Curve-Receiver Operating Characteristics (AUC-ROC). When compared other models, had best precision (1.000), recall (0.9987), F1-score (0.9994), AUC-ROC (0.9994) ratings. According findings, are crucial using technique, facilities boost efficiency quality healthcare. Various detect palm images. Finally, findings confirm earlier studies effectiveness algorithms means iron deficiency anemia.

Language: Английский

---

Artificial Intelligence can Facilitate Application of Risk Stratification Algorithms to Bladder Cancer Patient Case Scenarios

Clinical Medicine Insights Oncology, Journal Year: 2024, Volume and Issue: 18

Published: Jan. 1, 2024

Background: Chat Generative Pre-Trained Transformer (ChatGPT) has previously been shown to accurately predict colon cancer screening intervals when provided with clinical data and context in the form of guidelines. The National Comprehensive Cancer Network ® (NCCN ) guideline on non-muscle invasive bladder (NMIBC) includes criteria for risk stratification into low-, intermediate-, high-risk groups based patient disease characteristics. aim this study is evaluate ability ChatGPT apply NCCN Guidelines stratify theoretical scenarios related NMIBC. Methods: Thirty-six hypothetical NMIBC were created submitted GPT-3.5 GPT-4 at two separate time points. First, both models prompted patients without any additional provided. Custom instructions then as textual using written versions Guidelines, followed by repeat stratification. Finally, was an image table, again performed. Results: correctly stratified 68% (24.5 36) context, slightly increasing 74% (26.5 context. Using GPT-4, model had accuracy 83% (30 reaching 100% (36 ( P = .025). maintained similar 81% (29 36). generally performed poorly stratifying intermediate (33%-63%). When incorrect, most responses overestimations risk. Conclusions: can respect containing Overestimation more common than underestimation, likely be incorrectly stratified. With further validation, become a tool practice.

Language: Английский

---

Fluorescence Confocal Microscopy in Urological Malignancies: Current Applications and Future Perspectives

Life, Journal Year: 2023, Volume and Issue: 13(12), P. 2301 - 2301

Published: Dec. 5, 2023

Fluorescence confocal microscopy (FCM) represents a novel diagnostic technique able to provide real-time histological images from non-fixed specimens. As consequence of its recent developments, FCM is gaining growing popularity in urological practice. Nevertheless, evidence still sparse, and, at the moment, applications are heterogeneous. We performed narrative review current literature on this topic. Papers were selected Pubmed, Embase, and Medline archives. focused prostate cancer (PCa), urothelial carcinoma (UC), renal cell (RCC). Articles investigating both office intraoperative settings included. The showed that displays promising accuracy as compared conventional histopathology. These results represent significant steps along path FCM's formal validation an innovative ready-to-use support Instant access reliable evaluation may indeed significantly influence physicians' decision-making process. In regard, addresses unmet clinical need introduces intriguing perspectives into future pathways. Further studies required thoroughly assess whole potential technique.

Language: Английский

---