A machine learning method to process voice samples for identification of Parkinson's disease

Voice biomarkers as prognostic indicators for Parkinson's disease using machine learning techniques

Many people suffer from Parkinson's disease globally, a complicated neurological condition caused by the deficiency of dopamine, an organic chemical responsible for regulating movement in individuals. Patients with Parkinson face muscle stiffness or rigidity, tremors, vocal impairment, slow movement, loss of facial expressions, and problems with balance and coordination. As there is no cure for Parkinson, early diagnosis can help prevent the progression of this disease. The study explores the potential of vocal measures as significant indicators for early prediction of Parkinson. Different machine learning models such as Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), and Decision Tree (DT) are used to detect Parkinson using voice measures and differentiate between the healthy and Parkinson patients. The dataset contains 195 vocal recordings from 31 patients. The Synthetic Minority Over-Sampling Technique (SMOTE) is used for handling class imbalance to improve the performance of the models. The Principal Component Analysis (PCA) method was used for feature selection. The study uses different parameters to evaluate the model's classification results. The results highlight RF as the most effective model with an accuracy of 94% and a precision of 94%. In addition, SVM achieves an accuracy score of 92%, and precision of 91%. However, with the PCA method, SVM achieves an accuracy of 89%, 92%, and 87% for RF and DT respectively. This study highlights the significance of using vocal features along with advanced machine learning methods to reliably diagnose Parkinson's disease, considering the challenges associated with early detection.

Explainable artificial intelligence to diagnose early Parkinson's disease via voice analysis

Parkinson's disease (PD) is a neurodegenerative disorder affecting motor control, leading to symptoms such as tremors and stiffness. Early diagnosis is essential for effective treatment, but traditional methods are often time-consuming and expensive. This study leverages Artificial Intelligence (AI) and Machine Learning (ML) techniques, using voice analysis to detect early signs of PD. We applied a hybrid model combining Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Multiple Kernel Learning (MKL), and Multilayer Perceptron (MLP) to a dataset of 81 voice recordings. Acoustic features such as Mel-Frequency Cepstral Coefficients (MFCCs), jitter, and shimmer were analyzed. The model achieved 91.11% accuracy, 92.50% recall, 89.84% precision, 91.13% F1 score, and an area-under-the-curve (AUC) of 0.9125. SHapley Additive exPlanations (SHAP) provided data explainability, identifying key features driving the PD diagnosis, thus enhancing AI interpretability and trustability. Furthermore, a probability-based scoring system was developed to enable PD patients and clinicians to track disease progression. This AI-driven approach offers a non-invasive, cost-effective, and rapid tool for early PD detection, facilitating personalized treatment through vocal biomarkers.

Pre-trained convolutional neural networks identify Parkinson's disease from spectrogram images of voice samples

Machine learning approaches including deep learning models have shown promising performance in the automatic detection of Parkinson's disease. These approaches rely on different types of data with voice recordings being the most used due to the convenient and non-invasive nature of data acquisition. Our group has successfully developed a novel approach that uses convolutional neural network with transfer learning to analyze spectrogram images of the sustained vowel /a/ to identify people with Parkinson's disease. We tested this approach by collecting a dataset of voice recordings via analog telephone lines, which support limited bandwidth. The convolutional neural network with transfer learning approach showed superior performance against conventional machine learning methods that collapse measurements across time to generate feature vectors. This study builds upon our prior results and presents two novel contributions: First, we tested the performance of our approach on a larger voice dataset recorded using smartphones with wide bandwidth. Our results show comparable performance between two datasets generated using different recording platforms despite the differences in most important features resulting from the limited bandwidth of analog telephonic lines. Second, we compared the classification performance achieved using linear-scale and mel-scale spectrogram images and showed a small but statistically significant gain using mel-scale spectrograms.

Development of Neurodegenerative Disease Diagnosis and Monitoring from Traditional to Digital Biomarkers

Monitoring and assessing the progression of symptoms in neurodegenerative diseases, including Alzheimer's and Parkinson's disease, are critical for improving patient outcomes. Traditional biomarkers, such as cerebrospinal fluid analysis and brain imaging, are widely used to investigate the underlying mechanisms of disease and enable early diagnosis. In contrast, digital biomarkers derived from phenotypic changes-such as EEG, eye movement, gait, and speech analysis-offer a noninvasive and accessible alternative. Leveraging portable and widely available devices, such as smartphones and wearable sensors, digital biomarkers are emerging as a promising tool for ND diagnosis and monitoring. This review highlights the comprehensive developments in digital biomarkers, emphasizing their unique advantages and integration potential alongside traditional biomarkers.

Analysis of Different Modality of Data to Diagnose Parkinson's Disease Using Machine Learning and Deep Learning Approaches: A Review

The dynamic nature of Parkinson's disease (PD) is that it gradually impacts regions of the brain that are responsible for the production of the dopamine hormone. Despite continuous efforts, no effective treatment or preventative approach exists for PD. Nonetheless, the disease can be detected. Our goal is to create a Machine Learning and Deep Learning‐based system that can detect Parkinson's disease from a variety of data sources with high accuracy, sensitivity, specificity and interpretability. However, there have been significant advancements in the field of research, especially the use of artificial intelligence in the Parkinson's disease diagnostic process. We reviewed articles that were released between 2018 and 2024, concentrating on the most current studies that had been published. We chose 70 research articles for our review paper based on a set of criteria from a variety of online databases, including IEEExpress, medical databases like PubMed, Google Scholar, ResearchGate and ScienceDirect, and various publishers, including Elsevier, Taylor & Francis, Springer, MDPI, Plos One and so forth. According to our review, the majority of works make use of voice data. Our review study found that the highest accuracy level of most papers was above 90%, and the most commonly used algorithms were CNN and SVM. The main goal of this review study is to look into and put together information about the different ways that artificial intelligence, especially Machine Learning, can be used to find Parkinson's disease. Using diverse data gathered from multiple public and private datasets, we can infer that the application of artificial intelligence, particularly Machine Learning algorithms, for identifying Parkinson's disease plays a crucial role in the medical field.

Pre-trained Convolutional Neural Networks Identify Parkinson's Disease from Spectrogram Images of Voice Samples

Machine learning approaches including deep learning models have shown promising performance in the automatic detection of Parkinson's disease. These approaches rely on different types of data with voice recordings being the most used due to the convenient and non-invasive nature of data acquisition. Our group has successfully developed a novel approach that uses convolutional neural network with transfer learning to analyze spectrogram images of the sustained vowel /a/ to identify people with Parkinson's disease. We tested this approach by collecting a dataset of voice recordings via telephone lines, which have limited bandwidth. This study builds upon our prior results in two major ways: First, we tested the performance of our approach on a larger voice dataset recorded using smartphones with wide bandwidth. Our results show comparable performance between two datasets generated using different recording platforms where we report differences in most important features resulting from the limited bandwidth of telephonic lines. Second, we compared the classification performance achieved using linear-scale and mel-scale spectrogram images and showed a small but statistically significant gain using mel-scale spectrograms. The convolutional neural network with transfer learning approach showed superior performance against conventional machine learning methods that collapse measurements across time to generate feature vectors.

In Silico Decoding of Parkinson's: Speech & Writing Analysis

Background: Parkinson's disease (PD) has transitioned from a rare condition in 1817 to the fastest-growing neurological disorder globally. The significant increase in cases from 2.5 million in 1990 to 6.1 million in 2016, coupled with predictions of a further doubling by 2040, underscores an impending healthcare challenge. This escalation aligns with global demographic shifts, including rising life expectancy and a growing global population. The economic impact, notably in the U.S., reached $51.9 billion in 2017, with projections suggesting a 46% increase by 2037, emphasizing the substantial socio-economic implications for both patients and caregivers. Coupled with a worldwide demand for health workers that is expected to rise to 80 million by 2030, we have fertile ground for a pandemic. Methods: Our transdisciplinary research focused on early PD detection through running speech and continuous handwriting analysis, incorporating medical, biomedical engineering, AI, and linguistic expertise. The cohort comprised 30 participants, including 20 PD patients at stages 1-4 on the Hoehn and Yahr scale and 10 healthy controls. We employed advanced AI techniques to analyze correlation plots generated from speech and handwriting features, aiming to identify prodromal PD biomarkers. Results: The study revealed distinct speech and handwriting patterns in PD patients compared to controls. Our ParkinsonNet model demonstrated high predictive accuracy, with F1 scores of 95.74% for speech and 96.72% for handwriting analyses. These findings highlight the potential of speech and handwriting as effective early biomarkers for PD. Conclusions: The integration of AI as a decision support system in analyzing speech and handwriting presents a promising approach for early PD detection. This methodology not only offers a novel diagnostic tool but also contributes to the broader understanding of PD's early manifestations. Further research is required to validate these findings in larger, diverse cohorts and to integrate these tools into clinical practice for timely PD pre-diagnosis and management.

Exploring Spectrogram-Based Audio Classification for Parkinson's Disease: A Study on Speech Classification and Qualitative Reliability Verification

Patients suffering from Parkinson's disease suffer from voice impairment. In this study, we introduce models to classify normal and Parkinson's patients using their speech. We used an AST (audio spectrogram transformer), a transformer-based speech classification model that has recently outperformed CNN-based models in many fields, and a CNN-based PSLA (pretraining, sampling, labeling, and aggregation), a high-performance model in the existing speech classification field, for the study. This study compares and analyzes the models from both quantitative and qualitative perspectives. First, qualitatively, PSLA outperformed AST by more than 4% in accuracy, and the AUC was also higher, with 94.16% for AST and 97.43% for PSLA. Furthermore, we qualitatively evaluated the ability of the models to capture the acoustic features of Parkinson's through various CAM (class activation map)-based XAI (eXplainable AI) models such as GradCAM and EigenCAM. Based on PSLA, we found that the model focuses well on the muffled frequency band of Parkinson's speech, and the heatmap analysis of false positives and false negatives shows that the speech features are also visually represented when the model actually makes incorrect predictions. The contribution of this paper is that we not only found a suitable model for diagnosing Parkinson's through speech using two different types of models but also validated the predictions of the model in practice.

Harnessing Voice Analysis and Machine Learning for Early Diagnosis of Parkinson's Disease: A Comparative Study Across Three Datasets

OBJECTIVE

This study evaluates the efficacy of voice analysis combined with machine learning (ML) techniques in enabling the diagnosis of Parkinson's disease (PD).

METHODS

Voice data, phonation of the vowel "a," from three distinct datasets (two from the University of California Irvine ML Repository and one from figshare) for 432 participants (278 PD patients) were analyzed. We employed four ML models-Artificial Neural Networks, Random Forest, Gradient Boosting (GB), and Support Vector Machine (SVM)-alongside two ensemble methods (soft voting classifier-Ensemble Voting Classifier and stacking method-Ensemble Stacking Model (ESM)). The models underwent 50 iterations of evaluation, involving various data splits and 10-fold cross-validation. Comparative analysis was done using one-way Analysis of Variance followed by Bonferroni posthoc corrections.

RESULTS

The ESM, SVM, and GB models emerged as the top performers, demonstrating superior performance across metrics, including accuracy, sensitivity, specificity, precision, F1 score, and area under the receiver operating characteristic curve (ROC AUC). Despite data heterogeneity and variable selection limitations, the models showed high values for all metrics.

CONCLUSIONS

ML integration with voice analysis, mainly through ESM, SVM, and GB, is promising for early PD diagnosis. Using multi-source data and a large sample size enhances our findings' validity, reliability, and generalizability.

SIGNIFICANCE

Integrating advanced ML techniques with voice analysis demonstrates substantial potential for improving early PD detection, offering valuable tools for speech-language pathologists (SLPs). These findings provide clinically relevant insights that can be applied within the scope of SLP practice to refine diagnostic processes and facilitate early intervention.

Parkinson's disease diagnosis using deep learning: A bibliometric analysis and literature review

Parkinson's Disease (PD) is a progressive neurodegenerative illness triggered by decreased dopamine secretion. Deep Learning (DL) has gained substantial attention in PD diagnosis research, with an increase in the number of published papers in this discipline. PD detection using DL has presented more promising outcomes as compared with common machine learning approaches. This article aims to conduct a bibliometric analysis and a literature review focusing on the prominent developments taking place in this area. To achieve the target of the study, we retrieved and analyzed the available research papers in the Scopus database. Following that, we conducted a bibliometric analysis to inspect the structure of keywords, authors, and countries in the surveyed studies by providing visual representations of the bibliometric data using VOSviewer software. The study also provides an in-depth review of the literature focusing on different indicators of PD, deployed approaches, and performance metrics. The outcomes indicate the firm development of PD diagnosis using DL approaches over time and a large diversity of studies worldwide. Additionally, the literature review presented a research gap in DL approaches related to incremental learning, particularly in relation to big data analysis.