A review of machine learning and deep learning algorithms for Parkinson's disease detection using handwriting and voice datasets

Heliyon, Journal Year: 2024, Volume and Issue: 10(3), P. e25469 - e25469

Published: Feb. 1, 2024

Parkinson's Disease (PD) is a prevalent neurodegenerative disorder with significant clinical implications. Early and accurate diagnosis of PD crucial for timely intervention personalized treatment. In recent years, Machine Learning (ML) Deep (DL) techniques have emerged as promis-ing tools improving diagnosis. This review paper presents detailed analysis the current state ML DL-based diagnosis, focusing on voice, handwriting, wave spiral datasets. The study also evaluates effectiveness various DL algorithms, including classifiers, these datasets highlights their potential in enhancing diagnostic accuracy aiding decision-making. Additionally, explores identifi-cation biomarkers using techniques, offering insights into process. discussion encompasses different data formats commonly employed methods providing comprehensive overview field. serves roadmap future research, guiding development detection. It expected to benefit both scientific community medical practitioners by advancing our understanding ultimately patient outcomes.

Language: Английский

---

PIDGN: An explainable multimodal deep learning framework for early prediction of Parkinson's disease

Journal of Neuroscience Methods, Journal Year: 2025, Volume and Issue: unknown, P. 110363 - 110363

Published: Jan. 1, 2025

Language: Английский

---

A practical Alzheimer’s disease classifier via brain imaging-based deep learning on 85,721 samples

Journal Of Big Data, Journal Year: 2022, Volume and Issue: 9(1)

Published: Oct. 13, 2022

Abstract Beyond detecting brain lesions or tumors, comparatively little success has been attained in identifying disorders such as Alzheimer’s disease (AD), based on magnetic resonance imaging (MRI). Many machine learning algorithms to detect AD have trained using limited training data, meaning they often generalize poorly when applied scans from previously unseen scanners/populations. Therefore, we built a practical MRI-based diagnostic classifier deep learning/transfer dataset of unprecedented size and diversity. A retrospective MRI pooled more than 217 sites/scanners constituted one the largest samples date (85,721 50,876 participants) between January 2017 August 2021. Next, state-of-the-art convolutional neural network, Inception-ResNet-V2, was sex with high generalization capability. The achieved 94.9% accuracy served base model transfer for objective diagnosis AD. After learning, fine-tuned classification 90.9% leave-sites-out cross-validation Disease Neuroimaging Initiative (ADNI, 6,857 samples) 94.5%/93.6%/91.1% direct tests three independent datasets (AIBL, 669 / MIRIAD, 644 OASIS, 1,123 samples). When this tested images mild cognitive impairment (MCI) patients, MCI patients who converted were 3 times likely be predicted did not convert (65.2% vs. 20.6%). Predicted scores showed significant correlations illness severity. In sum, proposed offers medical-grade marker that potential integrated into practice.

Language: Английский

---

Choice of Voxel-based Morphometry processing pipeline drives variability in the location of neuroanatomical brain markers

Communications Biology, Journal Year: 2022, Volume and Issue: 5(1)

Published: Sept. 6, 2022

Abstract Fundamental and clinical neuroscience has benefited tremendously from the development of automated computational analyses. In excess 600 human neuroimaging papers using Voxel-based Morphometry (VBM) are now published every year a number different processing pipelines used, although it remains to be systematically assessed whether they come up with same answers. Here we examined variability between four commonly used VBM in two large brain structural datasets. Spatial similarity between-pipeline reproducibility processed gray matter maps were generally low pipelines. Examination sex-differences age-related changes revealed considerable differences terms specific regions identified. Machine learning-based multivariate analyses allowed accurate predictions sex age, however accuracy differed Our findings suggest that choice pipeline alone leads markers which poses serious challenge for interpretation.

Language: Английский

---

Novel nested patch-based feature extraction model for automated Parkinson's Disease symptom classification using MRI images

Computer Methods and Programs in Biomedicine, Journal Year: 2022, Volume and Issue: 224, P. 107030 - 107030

Published: July 16, 2022

Language: Английский

---

Gray matter, white matter and cerebrospinal fluid abnormalities in Parkinson’s disease: A voxel-based morphometry study

Frontiers in Psychiatry, Journal Year: 2022, Volume and Issue: 13

Published: Oct. 17, 2022

Parkinson's disease (PD) is a chronic neurodegenerative disorder characterized by bradykinesia, tremor, and rigidity among other symptoms. With 70% cumulative prevalence of dementia in PD, cognitive impairment neuropsychiatric symptoms are frequent.In this study, we looked at anatomical brain differences between groups patients controls. A total 138 people with PD were compared to 64 age-matched healthy using voxel-based morphometry (VBM). VBM fully automated technique that allows for the identification regional gray matter (GM), white (WM), cerebrospinal fluid (CSF) allowing an objective comparison brains different people. We used statistical parametric mapping image processing analysis.In controls, had lower GM volumes left middle cingulate, lingual gyrus, right calcarine fusiform also indicated WM calcarine, inferior occipital gyrus. Moreover, group demonstrated higher CSF caudate controls.Physical fragility impairments may be detected more easily if abnormalities cingulate lobe level identified. Thus, our findings shed light on role aid better understanding events occur patients.

Language: Английский

---

Explainable classification of Parkinson’s disease using deep learning trained on a large multi-center database of T1-weighted MRI datasets

NeuroImage Clinical, Journal Year: 2023, Volume and Issue: 38, P. 103405 - 103405

Published: Jan. 1, 2023

Parkinson's disease (PD) is a severe neurodegenerative that affects millions of people. Early diagnosis important to facilitate prompt interventions slow down progression. However, accurate PD can be challenging, especially in the early stages. The aim this work was develop and evaluate robust explainable deep learning model for classification trained from one largest collections T1-weighted magnetic resonance imaging datasets.A total 2,041 MRI datasets 13 different studies were collected, including 1,024 patients 1,017 age- sex-matched healthy controls (HC). skull stripped, resampled isotropic resolution, bias field corrected, non-linearly registered MNI PD25 atlas. Jacobian maps derived deformation fields together with basic clinical parameters used train state-of-the-art convolutional neural network (CNN) classify HC subjects. Saliency generated display brain regions contributing most task as means artificial intelligence.The CNN using an 85%/5%/10% train/validation/test split stratified by diagnosis, sex, study. achieved accuracy 79.3%, precision 80.2%, specificity 81.3%, sensitivity 77.7%, AUC-ROC 0.87 on test set while performing similarly independent set. computed data highlighted frontotemporal regions, orbital-frontal cortex, multiple gray matter structures important.The developed model, large heterogenous database, able differentiate subjects high clinically feasible explanations. Future research should investigate combination modalities validating these results prospective trial decision support system.

Language: Английский

---

AI in Rehabilitation Medicine: Opportunities and Challenges

Annals of Rehabilitation Medicine, Journal Year: 2023, Volume and Issue: 47(6), P. 444 - 458

Published: Dec. 14, 2023

Artificial intelligence (AI) tools are increasingly able to learn from larger and more complex data, thus allowing clinicians scientists gain new insights the information they collect about their patients every day. In rehabilitation medicine, AI can be used find patterns in huge amounts of healthcare data. These then leveraged at individual level, design personalized care strategies interventions optimize each patient’s outcomes. However, building effective requires many careful considerations how we handle train models, interpret results. this perspective, discuss some current opportunities challenges for rehabilitation. We first review recent trends screening, diagnosis, treatment, continuous monitoring disease or injury, with a special focus on different types data these applications. examine potential barriers designing integrating into clinical workflow, propose an end-to-end framework address guide development Finally, present ideas future work pave way implementation real-world practices.

Language: Английский

---

An improved method for diagnosis of Parkinson’s disease using deep learning models enhanced with metaheuristic algorithm

BMC Medical Imaging, Journal Year: 2024, Volume and Issue: 24(1)

Published: June 24, 2024

Abstract Parkinson's disease (PD) is challenging for clinicians to accurately diagnose in the early stages. Quantitative measures of brain health can be obtained safely and non-invasively using medical imaging techniques like magnetic resonance (MRI) single photon emission computed tomography (SPECT). For accurate diagnosis PD, powerful machine learning deep models as well effectiveness tools assessing neurological are required. This study proposes four with a hybrid model detection PD. simulation study, two standard datasets chosen. Further improve performance models, grey wolf optimization (GWO) used automatically fine-tune hyperparameters models. The GWO-VGG16, GWO-DenseNet, GWO-DenseNet + LSTM, GWO-InceptionV3 GWO-VGG16 InceptionV3 applied T1,T2-weighted SPECT DaTscan datasets. All performed near or above 99% accuracy. highest accuracy 99.94% AUC 99.99% achieved by (GWO-VGG16 InceptionV3) dataset 100% 99.92% recorded dataset.

Language: Английский

---

Review on computational methods for the detection and classification of Parkinson's Disease

Computers in Biology and Medicine, Journal Year: 2025, Volume and Issue: 187, P. 109767 - 109767

Published: Feb. 11, 2025

Language: Английский

---