Parkinson's Disease Classification Research Articles (Page 1)

Introduction Parkinson's disease (PD) is a common neurodegenerative disorder. Traditional diagnostic methods, relying on clinical assessment and imaging, are often invasive, costly, and require specialized personnel, posing barriers to early detection. As approximately 90% of PD patients develop vocal impairments, vocal analysis emerges as a promising non-invasive diagnostic tool. However, individual deep learning models are often limited by overfitting and poor generalizability. Methods This study proposes a PD classification method using spectrogram feature fusion with pre-trained convolutional neural networks (CNNs). Voice recordings were obtained from 61 PD patients and 70 healthy controls (HC) at the First Affiliated Hospital of Henan University of Science and Technology. Preprocessing the raw speech signals yielded 2,476 spectrograms. Three pre-trained models, DenseNet121, MobileNetV3-Large, and ShuffleNetV2, were used for feature extraction. The output of MobileNetV3-Large was adjusted using a 1 × 1 convolutional layer to ensure dimensional alignment before features were fused via summation. Results Evaluation using 5-fold cross-validation demonstrated that models employing feature fusion consistently outperformed individual models across all metrics. Specifically, the fusion of MobileNetV3-Large and ShuffleNetV2 achieved the highest accuracy of 95.56% and an AUC of 0.99. Comparative experiments with existing state-of-the-art methods confirmed the competitive performance of the proposed approach. Discussion The fusion of multi-model features more effectively captures subtle pathological signatures in PD speech, overcoming the limitations of single models. This method provides a reliable, low-cost, and non-invasive tool for auxiliary PD diagnosis, with significant potential for clinical application. The code is available at https://github.com/lvrongfu/pjs .

An optimized framework for Parkinson's disease classification using multimodal neuroimaging data with ensemble-based and data fusion networks.

Integrative approach for early detection of Parkinson's disease and atypical Parkinsonian syndromes leveraging hemodynamic parameters, motion data & advanced AI models.

Multimodal approach for early diagnosis of Parkinson's disease using PET imaging, tremor detection, and machine learning.

Parkinson's disease (PD) is a chronic neurodegenerative disorder characterized by progressive loss of dopaminergic neurons in substantia nigra, resulting in both motor impairments and cognitive decline. Traditional PD classification methods are expert-dependent and time-intensive, while existing deep learning (DL) models often suffer from inconsistent accuracy, limited interpretability, and inability to fully capture PD's clinical heterogeneity. This study proposes a novel framework Enhanced EfficientNet-Extended Multimodal PD Classification with Hybrid Particle Swarm and Grey Wolf Optimizer (EEFN-XM-PDC-HybPS-GWO) to overcome these challenges. The model integrates T1-weighted MRI, DaTscan images, and gait scores from NTUA and PhysioNet repository respectively. Denoising is achieved via Multiscale Attention Variational Autoencoders (MSA-VAE), and critical regions are segmented using Semantic Invariant Multi-View Clustering (SIMVC). The Enhanced EfficientNet-Extended Multimodal (EEFN-XM) model extracts and fuses image and gait features, while HybPS-GWO optimizes classification weights. The system classifies subjects into early-stage PD, advanced-stage PD, and healthy controls (HCs). Ablation analysis confirms the hybrid optimizer's contribution to performance gains. The proposed model achieved 99.2% accuracy with stratified 5-fold cross-validation, outperforming DMFEN-PDC, MMT-CA-PDC, and LSTM-PDD-GS by 7.3%, 15.97%, and 10.43%, respectively, and reduced execution time by 33.33%. EEFN-XM-PDC-HybPS-GWO demonstrates superior accuracy, computational efficiency, and clinical relevance, particularly in early-stage diagnosis and PD classification.

Deep brain stimulation (DBS) can effectively ameliorate motor symptoms in Parkinson's disease (PD), but patient outcomes remain variable. Clinical predictors lack reliability and only explain a small proportion of outcome variance, outlining a need for biomarkers that can enhance prediction accuracy. Functional magnetic resonance imaging (fMRI) could address this, offering insight into the relative impact of functional brain health on DBS outcomes. Preoperative resting-state fMRI was retrospectively collected for 120 patients with PD and DBS targeting the subthalamic nucleus or pallidum. Motor network connectivity was computed, and clinical predictors extracted including age, sex, target, hemisphere treated, and preoperative medication responsiveness. Pre-to-post changes in antiparkinson medications defined outcomes. Regression analysis selected important connectivity features and evaluated additional outcome variance explained by fMRI when combined with clinical predictors. Last, classification analysis assessed fMRI feature specificity to PD compared to Huntington's disease and healthy controls. Alongside clinical predictors, fMRI features explained more outcome variance (R2 adjusted = 0.36) than clinical predictors alone (R2 adjusted = 0.13). Five connectivity pairs bridging the cortico-basal ganglia-cerebellar network were predictive of outcomes and distinguished PD from controls and Huntington's disease with reasonable accuracy (67% and 73%). Exploratory analysis revealed that intracerebellar connectivity, although a less stable predictor of DBS outcomes, dramatically increased PD classification accuracy (≥98%). Patient-specific motor connectivity enhances DBS outcome prediction and could aid detection and monitoring of basal ganglia disorders. Future work will validate and extend our models for multi-parameter MR prediction of standardized outcomes, toward the development of decision support tools for DBS. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Parkinson's disease (PD) is increasingly recognized as a heterogeneous neurodegenerative entity with diverse clinical presentations, genetic contributors, and neuropathological features. Central to its pathogenesis is misfolded and aggregated α-synuclein, which collectively form Lewy pathology. Recent advances in biomarker and genetic research have enabled biologically grounded models of PD classification, diagnosis and staging. This review summarizes key principles, differences, and ongoing challenges of two emerging research frameworks: the SynNeurGe criteria and the Neuronal α-Synuclein Disease Integrated Staging System (NSD-ISS)-the former proposed a biologically based classification, while the latter proposed a more restrictive biological definition and staging schema. SynNeurGe incorporates synucleinopathy (S), neurodegeneration (N), genetic risk (G) and clinical status (C) to classify etiologic subtypes across the disease spectrum, emphasizing clinical heterogeneity and multifaceted underlying biological processes. In contrast, the NSD-ISS defines "neuronal α-synuclein disease" (NSD) based on specific molecular (S) and dopaminergic dysfunction (D) markers and a single genetic anchor (SNCA) (G), and maps disease progression across seven clinical stages. While both aim to improve early detection and to advance PD research, they differ in scope, operational definitions, implementation principles, and intended applications. Prevailing challenges include current limitations in mechanistic insights, biomarker standardization and accessibility, underrepresentation of genetic diversity, and ethical considerations around disease labeling and risk disclosure, particularly in asymptomatic cases. These frameworks represent a pivotal shift toward biologically based concepts of PD and related disorders, with future success contingent on continued refinement, validation, and equitable implementation.

Parkinson’s Disease (PD) is the second most common neurodegenerative disorder, whose symptoms worsen over time, making early diagnosis a challenging task. Changes in speech have been identified as an early symptom of PD identification. However, medical datasets often have a small sample size, while speech signal analysis generates high-dimensional data. Therefore, rigorous feature selection is essential for obtaining the best set of PD characteristics. This paper proposes a hybrid filter-wrapper feature selection approach for PD classification using a publicly available speech dataset (188 PD, 64 healthy subjects). Maximum Relevancy Minimum Redundancy (mRMR) and Relief algorithms are used to select top-ranked features, followed by the Modified Whale Optimization Algorithm (mWOA) to refine the selection for obtaining an optimized feature subset. The class imbalance issue is addressed using SMOTE. A stacked ensemble model is developed, integrating base learners, Decision Tree, Support Vector Machine, Naïve Bayes, k-Nearest Neighbour, and deep networks like shallow and deep with hyperparameters tuned via a grid search mechanism. The proposed approach is evaluated against state-of-the-art methods based on accuracy, precision, recall, and F1-score. Results demonstrate that hybrid feature selection and hyperparameter tuning reduce computational burden while improving classification accuracy, making it a promising framework for PD detection from speech data.

Background/Objectives: Parkinson's disease (PD) is a progressive neurodegenerative disorder that requires early and accurate diagnosis. This study aimed to classify individuals with and without PD using volumetric brain MRI data and to improve model interpretability using explainable artificial intelligence (XAI) techniques. Methods: This retrospective study included 79 participants (39 PD patients, 40 controls) recruited at Inonu University Turgut Ozal Medical Center between 2013 and 2025. A deep neural network (DNN) was developed using a multilayer perceptron architecture with six hidden layers and ReLU activation functions. Seventeen volumetric brain features were used as the input. To ensure robust evaluation and prevent overfitting, a stratified five-fold cross-validation was applied, maintaining class balance in each fold. Model transparency was explored using two complementary XAI techniques: the Contrastive Explanation Method (CEM) and Local Interpretable Model-Agnostic Explanations (LIME). CEM highlights features that support or could alter the current classification, while LIME provides instance-based feature attributions. Results: The DNN model achieved high diagnostic performance with 94.1% accuracy, 98.3% specificity, 90.2% sensitivity, and an AUC of 0.97. The CEM analysis suggested that reduced hippocampal volume was a key contributor to PD classification (-0.156 PP), whereas higher volumes in the brainstem and hippocampus were associated with the control class (+0.035 and +0.150 PP, respectively). The LIME results aligned with these findings, revealing consistent feature importance (mean = 0.1945) and faithfulness (0.0269). Comparative analyses showed different volumetric patterns between groups and confirmed the DNN's superiority over conventional machine learning models such as SVM, logistic regression, KNN, and AdaBoost. Conclusions: This study demonstrates that a deep learning model, enhanced with CEM and LIME, can provide both high diagnostic accuracy and interpretable insights for PD classification, supporting the integration of explainable AI in clinical neuroimaging.

ABSTRACT Parkinson’s Disease (PD) is a progressive neurodegenerative disorder and the early diagnosis is crucial for managing symptoms and slowing disease progression. This paper proposes a framework named Federated Learning Enabled Waterwheel Shuffled Shepherd Optimization-based Efficient-Fuzzy Deep Maxout Network (FedL_WSSO based Eff-FDMNet) for PD detection and classification. In local training model, the input image from the database “Image and Data Archive (IDA)” is given for preprocessing that is performed using Gaussian filter. Consequently, image augmentation takes place and feature extraction is conducted. These processes are executed for every input image. Therefore, the collected outputs of images are used for PD detection using Shepard Convolutional Neural Network Fuzzy Zeiler and Fergus Net (ShCNN-Fuzzy-ZFNet). Then, PD classification is accomplished using Eff-FDMNet, which is trained using WSSO. At last, based on CAViaR, local updation and aggregation are changed in server. The developed method obtained highest accuracy as 0.927, mean average precision as 0.905, lowest false positive rate (FPR) as 0.082, loss as 0.073, Mean Squared Error (MSE) as 0.213, and Root Mean Squared Error (RMSE) as 0.461. The high accuracy and low error rates indicate that the potent framework can enhance patient outcomes by enabling more reliable and personalized diagnosis.

Parkinson's disease (PD) is a complex neurodegenerative disorder characterized by a high rate of misdiagnosis, underscoring the critical importance of early and accurate diagnosis. Although existing computer-aided diagnostic systems integrate clinical assessment scales with neuroimaging data, they typically rely on superficial feature concatenation, which fails to capture the deep inter-modal dependencies essential for effective multimodal fusion. To address this limitation, we propose ModFus-PD, Contrastive learning effectively aligns heterogeneous modalities such as imaging and clinical text, while the cross-modal attention mechanism further exploits semantic interactions between them to enhance feature fusion. The framework comprises three key components: (1) a contrastive learning-based feature alignment module that projects MRI data and clinical text prompts into a unified embedding space via pretrained image and text encoders; (2) a bidirectional cross-modal attention module in which textual semantics guide MRI feature refinement for improved sensitivity to PD-related brain regions, while MRI features simultaneously enhance the contextual understanding of clinical text; (3) a hierarchical classification module that integrates the fused representations through two fully connected layers to produce final PD classification probabilities. Experiments on the PPMI dataset demonstrate the superior performance of ModFus-PD, achieving an accuracy of 0.903, AUC of 0.892, and F1 score of 0.840, surpassing several state-of-the-art baselines. These results validate the effectiveness of our cross-modal fusion strategy, which enables interpretable and reliable diagnostic support, holding promise for future clinical translation.

Parkinson's disease (PD) is a neurodegenerative disorder of the brain that primarily affects motor function. Clinical challenges associated with this condition include accurately diagnosing patients in the early stages of the disease and predicting how the condition will progress. This project aims to enhance PD detection by integrating feature selection and classification using supervised learning techniques. Two publicly available datasets—the speech and PD classification datasets—are utilized to evaluate model performance across diverse features. The proposed work employs class balancing through the Synthetic Minority Oversampling Technique (SMOTE) to address the issue of class imbalance in this highly unbalanced dataset. Subsequently, the Relief algorithm is used for feature selection to identify the most relevant predictors. An ensemble of models is applied using the RF-XGBoost-KNN classifiers due to their superior accuracy compared to other classifier combinations. The RF-XGBoost-KNN model stack achieved classification accuracies of 94.56% and 93.53% for the PD speech dataset and Parkinson's Disease Classification Dataset, respectively, demonstrating its potential as a robust tool for early and accurate PD diagnosis.

Parkinson's disease is a neurodegenerative disorder affecting motor abilities, with a prevalence of 329 cases per 100,000 individuals. Early diagnosis is crucial to prevent complications. This study classifies Parkinson's disease using the Extreme Gradient Boosting (XGBoost) algorithm with hyperparameter tuning via Grid Search and Random Search. The dataset from Kaggle consists of 2105 records from 2024 and includes 32 clinical and demographic features such as age, gender, BMI, medical history, and Parkinson's symptoms. The XGBoost method effectively manages large and complex data and reduces. Tuning was performed with 5-fold cross-validation for result validity. After tuning with Grid Search, the model achieved 93.35% accuracy in 44 minutes 51 seconds, with optimal parameters gamma=5, max depth=3, learning rate=0.3, n estimators=100, and subsample=0.7. Meanwhile, Random Search with 50 iterations achieved 93.97% accuracy in 3 minutes 4 seconds with optimal parameters gamma=5, max depth=3, learning rate=0.262, n estimators=58, and subsample=0.631. Random Search also shows better time efficiency than Grid Search, although with relatively similar accuracy. The results of this study confirm that hyperparameter tuning using Random Search not only produces competitive accuracy performance but also minimizes computation time, making it a more optimal choice for Parkinson's disease classification.

Robust computation of subcortical functional connectivity guided by quantitative susceptibility mapping: An application in Parkinson's disease diagnosis.

Cerebellum has a stronger individual specificity of functional signals than the brain and is associated with a variety of neuropsychiatric disorders, and increasing attention is being paid to neuropsychiatric symptoms caused by cerebellar dysfunction. However, there is a lack of a suitable cerebellar partition utilizing researchers to fully understand the functional and structural organization of the cerebellum, reduce data dimensionality, and improve the applicability of various types of models to cerebellar functional imaging data, impeding progress in cerebellum-related research. In this study, we use order-preserving variations with spatial constraints to optimize functional connectivity matrices and employ a spectral clustering algorithm combined with a clustering ensemble technique to construct a cerebellar partitioning algorithm with a variable number of partitions. Our method was initially validated by using two separate sets of functional magnetic resonance data (fMRI), demonstrating high reproducibility across individuals. Comparative analysis revealed that our partitions exhibited enhanced signal coherence and greater spatial congruence with established cerebellar structural templates compared to four publicly available cerebellar atlases. Furthermore, preliminarily applying these partitions to Parkinson's disease (PD) data, we extracted cerebellar connectivity network features and constructed a classification model using a logistic regression model with L2 regularization. The connectivity features derived from our newly constructed cerebellar partitions substantially improved the usability of the Parkinson's classification model, with the classification of PD optimized at a number of partitions equal to 185, suggesting that the optimal number of cerebellar partitions may also vary based on the problem under study. Notably, cerebellar regions implicated in motor execution were identified to exhibit higher feature importance in the Parkinson's classification model, offering an important direction for feature selection in the multimodal classification models of PD.

Parkinson's disease (PD) is a progressive neurodegenerative disorder affecting millions worldwide, characterized by a wide range of motor and non-motor symptoms. Among these symptoms, alterations in speech and voice quality stand out as early and prominent indicators of the disease. Recently, the emergence of speech foundation models has revolutionized the field by providing powerful tools for speech processing and feature extraction. In this article, we investigate the capabilities of three state-of the art speech foundation models, wav2vec2.0, Whisper and SeamlessM4T, to develop robust and accurate methods for PD detection from voice recordings. We experiment with both direct feature extraction and finetuning of the foundation models for the PD classification task, and validate the results against clinical and neuroimaging data. We achieve promising results using both pretrained features and models' finetuning, with finetuning providing stronger performance, up to 91.35% for AUC, which is the new state of the art on the ICEBERG dataset. The predictions of our models also show good correlation with clinical as well as DaTSCAN scores, proving the feasibility to apply speech foundation models for detection of early PD.

Parkinson’s Disease (PD) is a deteriorating condition that mostly affects older people. The lack of conclusive treatment for PD makes diagnosis very challenging. However, using patterns like tremors for early diagnosis, handwriting analysis has become a useful diagnostic technique. This work aims to improve early PD diagnosis by proposing a hybrid deep fusion model that blends ResNet-50 and GoogLeNet (RGG-Net). We demonstrated the RGG-Net model in a series of steps such as preprocessing images, ResNet-50 and GoogLeNet models for feature extraction, combining the features using the Adaptive Feature Fusion technique and selecting the relevant features using the attention process, making the models stronger through Hierarchical Ensemble Learning. The grad-CAM technique is used for decision-making in PD prediction. The proposed model is a reliable way to analyze handwritten images using advanced techniques like adaptive feature fusion, hierarchical ensemble learning, and eXplainable Artificial Intelligence. Here, we analyzed ten pre-trained models to determine which model best captures the relevant features for PD classification using handwritten images. The models included are AlexNet, DenseNet-201, SqueezeNet1.1, VGG-16, VGG-19, ResNet-50, ResNet-101, GoogleNet, MobileNetV1, and MobileNetV2. The proposed deep transfer learning model showed an accuracy of 99.12%, outperforming the other state-of-the-art methods, indicating the model’s excellence and vigor. The proposed model performs better than all pre-trained models with and without freezing convolutional layers. These results underscore the efficacy of the proposed approach in enhancing accuracy and transparency in Parkinson’s disease prediction and the potential of deep learning in promoting early diagnosis.

This study presents a hybrid ensemble learning framework for the joint detection and motor severity prediction of Parkinson's disease (PD) using biomedical voice features. The proposed architecture integrates a deep multimodal fusion model with dense expert pathways, multi-head self-attention, and multitask output branches to simultaneously perform binary classification and regression. To ensure data quality and improve model generalization, preprocessing steps included outlier removal via Isolation Forest, two-stage feature scaling (RobustScaler followed by MinMaxScaler), and augmentation through polynomial and interaction terms. Borderline-SMOTE was employed to address class imbalance in the classification task. To enhance prediction performance, ensemble learning strategies were applied by stacking outputs from the fusion model with tree-based regressors (Random Forest, Gradient Boosting, and XGBoost), using diverse meta-learners including XGBoost, Ridge Regression, and a deep neural network. Among these, the Stacking Ensemble with XGBoost (SE-XGB) achieved the best results, with an R2 of 99.78% and RMSE of 0.3802 for UPDRS regression and 99.37% accuracy for PD classification. Comparative analysis with recent literature highlights the superior performance of our framework, particularly in regression settings. These findings demonstrate the effectiveness of combining advanced feature engineering, deep learning, and ensemble meta-modeling for building accurate and generalizable models in voice-based PD monitoring. This work provides a scalable foundation for future clinical decision support systems.

Background: Parkinson's disease (PD) is a progressive neurodegenerative condition that impairs motor and non-motor functions. Early and accurate diagnosis is critical for effective management and care. Leveraging machine learning (ML) techniques, this study aimed to develop a robust prediction system for PD using a stacked ensemble learning approach, addressing challenges such as imbalanced datasets and feature optimization. Methods: An open-access PD dataset comprising 22 vocal attributes and 195 instances from 31 subjects was utilized. To prevent data leakage, subjects were divided into training (22 subjects) and testing (9 subjects) groups, ensuring no subject appeared in both sets. Preprocessing included data cleaning and normalization via min-max scaling. The synthetic minority oversampling technique (SMOTE) was applied exclusively to the training set to address class imbalance. Feature selection techniques-forward search, gain ratio, and Kruskal-Wallis test-were employed using subject-wise cross-validation to identify significant attributes. The developed system combined support vector machine (SVM), random forest (RF), K-nearest neighbor (KNN), and decision tree (DT) as base classifiers, with logistic regression (LR) as the meta-classifier in a stacked ensemble learning framework. Performance was evaluated using both recording-wise and subject-wise metrics to ensure clinical relevance. Results: The stacked ensemble learning model achieved realistic performance with a recording-wise accuracy of 84.7% and subject-wise accuracy of 77.8% on completely unseen subjects, outperforming individual classifiers including KNN (81.4%), RF (79.7%), and SVM (76.3%). Cross-validation within the training set showed 89.2% accuracy, with the performance difference highlighting the importance of proper validation methodology. Feature selection results showed that using the top 10 features ranked by gain ratio provided optimal balance between performance and clinical interpretability. The system's methodological robustness was validated through rigorous subject-wise evaluation, demonstrating the critical impact of validation methodology on reported performance. Conclusions: By implementing subject-wise validation and preventing data leakage, this study demonstrates that proper validation yields substantially different (and more realistic) results compared to flawed recording-wise approaches. The findings underscore the critical importance of validation methodology in healthcare ML applications and provide a template for methodologically sound PD classification research. Future research should focus on validating the model with larger, multi-center datasets and implementing standardized validation protocols to enhance clinical applicability.

Parkinson disease classification: a comparison of quantum and RBF kernels using support vector machine