Enhancing early Parkinson's disease detection through multimodal deep learning and explainable AI: insights from the PPMI database.

Parkinson's is the second most common neurodegenerative disease, affecting nearly 8.5M people and steadily increasing. In this research, Multimodal Deep Learning is investigated for the Prodromal stage detection of Parkinson's Disease (PD), combining different 3D architectures with the novel Excitation Network (EN) and supported by Explainable Artificial Intelligence (XAI) techniques. Utilizing data from the Parkinson's Progression Markers Initiative, this study introduces a joint co-learning approach for multimodal fusion, enabling end-to-end training of deep neural networks and facilitating the learning of complementary information from both imaging and clinical modalities. DenseNet with EN outperformed other models, showing a substantial increase in accuracy when supplemented with clinical data. XAI methods, such as Integrated Gradients for ResNet and DenseNet, and Attention Heatmaps for Vision Transformer (ViT), revealed that DenseNet focused on brain regions believed to be critical to prodromal pathophysiology, including the right temporal and left pre-frontal areas. Similarly, ViT highlighted the lateral ventricles associated with cognitive decline, indicating their potential in the Prodromal stage. These findings underscore the potential of these regions as early-stage PD biomarkers and showcase the proposed framework's efficacy in predicting subtypes of PD and aiding in early diagnosis, paving the way for innovative diagnostic tools and precision medicine.

ADVANCEMENT IN ALPHA-SYNUCLEIN PROTEOMICS: EXPLORING ANALYTICAL TECHNIQUES AND THEIR CLINICAL IMPLICATIONS IN PARKINSON'S DISEASE.

Parkinson's disease (PD) is a prevalent neurodegenerative disorder, affecting around 500,000 to 1 million Americans, with a significant portion diagnosed before age 50. Despite advances in treatments such as dopamine replacement therapy and deep brain stimulation, no therapies currently exist to halt or slow disease progression in advanced stages. Research is increasingly focused on identifying early biomarkers for PD to enable earlier intervention. Alpha-synuclein (α-Syn), a key protein implicated in PD pathology, is studied using various proteomics techniques like mass spectrometry, gel electrophoresis, and chromatography, to understand its role and alterations in PD. These techniques help in extracting, analyzing, and characterizing α-Syn from brain samples, providing insights into disease mechanisms and potential diagnostic and therapeutic applications.

Circulating microRNAs as potential biomarkers for the diagnosis of Parkinson's disease: A meta-analysis.

BACKGROUND AND OBJECTIVE: Parkinson's disease (PD) is the one of the most common neurodegenerative diseases. Many investigators have confirmed the possibility of using circulating miRNAs to diagnose PD. However, the results were inconsistent. Therefore, the aim of this meta-analysis was to systematically evaluate the diagnostic accuracy of circulating miRNAs in the diagnosis of PD. METHODS: We carefully searched PubMed, Embase, Web of Science, Cochrane Library, Wanfang database and China National Knowledge Infrastructure for relevant studies (up to January 1, 2022) based on PRISMA statement. The pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), the diagnostic odds ratio (DOR), and area under the curve (AUC) were calculated to test the diagnostic accuracy. Furthermore, subgroup analyses were performed to identify the potential sources of heterogeneity, and the Deeks' funnel plot asymmetry test was used to evaluate the potential publication bias. RESULTS: Forty-four eligible studies from 16 articles (3298 PD patients and 2529 healthy controls) were included in the current meta-analysis. The pooled sensitivity was 0.79 (95% CI: 0.76-0.81), specificity was 0.82 (95% CI: 0.78-0.84), PLR was 4.3 (95% CI: 3.6-5.0), NLR was 0.26 (95% CI: 0.23-0.30), DOR was 16 (95% CI: 13-21), and AUC was 0.87 (95% CI: 0.84-0.90). Subgroup analysis suggested that miRNA cluster showed a better diagnostic accuracy than miRNA simple. Moreover, there was no significant publication bias. CONCLUSIONS: Circulating miRNAs have great potential as novel non-invasive biomarkers for PD diagnosis.

Tear α-synuclein as a biomarker for Parkinson's disease: A systematic review and meta-analysis.

BACKGROUND: Parkinson's disease symptoms mostly manifest after significant and irreversible neuropathology. Hence, there is a need to identify biomarkers that can provide indications of disease before significant neuronal degeneration occurs. OBJECTIVE: To estimate the difference in the concentration of α-synuclein protein in tears between individuals with Parkinson's disease and healthy controls. DATA SOURCES: PubMed, Scopus, and Web of Science. The last database search was on December 20, 2023. STUDY ELIGIBILITY CRITERIA: Primary prospective studies in humans measuring the level of α-synuclein in tears and clinical outcomes reported using mean or median. PARTICIPANTS AND INTERVENTIONS: Individuals with Parkinson's disease and healthy controls. STUDY APPRAISAL AND SYNTHESIS METHODS: The risk of bias was assessed using the Newcastle-Ottawa Scale. The I2 statistic was used to estimate heterogeneity. The outcome measure was the difference in tear total and oligomeric α-synuclein. Mean difference (MD) was used to assess the outcome. The certainty of evidence was rated following the Grading of Recommendations Assessment and Development and Evaluation (GRADE) system. RESULTS: Three hundred twenty-seven Parkinson's disease and 312 healthy control subjects from five studies and 177 Parkinson's disease and 166 healthy control subjects from two studies were included in total α-synuclein levels and oligomeric α-synuclein levels analysis, respectively. Total α-synuclein level was not different between Parkinson's disease and healthy controls (MD = 0.02 ng/mL [95% confidence interval {CI}: 0.00 to 0.05 ng/mL; I2 = 90%; Z = 1.79; p=0.07; number of studies = 5; GRADE rating = very low]). Stratifying the data based on disease duration, total α-synuclein was higher in subjects with Parkinson's disease duration ≥7 years compared with healthy controls (MD = 0.04 ng/mL [95% CI: 0.03 to 0.05 ng/mL; I2 = 0%; Z = 8.24, p<0.00001; number of studies = 2; GRADE rating = low]) but not different between the two groups (MD = -0.12 ng/mL (95% CI: -0.38 to 0.15 ng/mL; I2 = 93%; Z = 0.84, p=0.40; number of studies = 3; GRADE rating = very low]). Oligomeric α-synuclein level was higher in Parkinson's disease compared with controls (MD = 6.50 ng/mL [95% CI: 2.79 to 10.20 ng/mL; I2 = 94%; Z = 3.44; p=0.0006; number of studies = 2; GRADE rating = very low]). LIMITATIONS: High heterogeneity between studies. Potential sources of heterogeneity could not be explored due to the limited number of studies. CONCLUSIONS AND IMPLICATIONS OF KEY FINDINGS: Tear α-synuclein has the potential to be a noninvasive biomarker for Parkinson's disease. Studies are, however, needed to increase certainty in the biomarker and establish how the protein's changes in tears correlate with Parkinson's disease progression and severity.

Teleassessment using five times sit-to-stand test in older adults with and without Parkinson's disease: Guidelines and barriers.

BACKGROUND AND PURPOSE: Assessing lower limb strength, balance, and fall risk are crucial components of rehabilitation, especially for the older adult population. With the growing interest in telehealth, teleassessment has been investigated as an alternative when in-person assessments are not possible. The Five Times Sit-to-Stand test (5TSTS) provides a quick measure of balance during chair transfers, muscle power, endurance, and the hability to change and maintain body position, and is highly recommended by guidelines. However, the literature is unclear about the viability and safety of teleassessment using the 5TSTS in older adults with and without Parkinson's disease (PD). This study aimed to evaluate the reliability of teleassessment using the 5TSTS and to determine its feasibility and safety for older adults with and without PD. METHODS: This cross-sectional study included older adults with and without PD who were evaluated remotely through a videoconference platform. To ensure effective and comprehensive instructions for the test, we developed a guideline called OMPEPE (an acronym for: Objective; Materials; Position-Start; Execution; Position-End; Environment). We assessed the 5TSTS intra- and inter-rater reliability by comparing scores obtained from the same examiner and from different examiners, respectively. Participants and examiners completed online surveys to provide information about feasibility and safety. RESULTS: Twelve older adults with PD and 17 older adults without PD were included in this study (mean ages 69.0 and 67.6 years, respectively). Based on the participants' perspectives and the absence of adverse effects, teleassessment using the 5TSTS is feasible and safe for older adults with and without PD. Excellent intra- and inter-rater reliability (intraclass correlation coefficient >0.90) was found for all measurements of the 5TSTS. DISCUSSION: This study demonstrated the feasibility, safety, and reliability of teleassessment using the 5TSTS. The guidelines developed may help health professionals minimize barriers and safely conduct an online assessment that includes a physical test such as the 5TSTS in older adults with or without PD. In addition to addressing technological barriers, the OMPEPE guideline might ensure the optimal execution of evaluations. CONCLUSION: Teleassessment using the 5TSTS for older adults with and without PD is feasible and safe. Both synchronous (i.e., live) and asynchronous (i.e., recorded) online 5TSTS tests demonstrate excellent intra- and inter-rate reliability.

Development of a Tremor Detection Algorithm for Use in an Academic Movement Disorders Center.

Tremor, defined as an "involuntary, rhythmic, oscillatory movement of a body part", is a key feature of many neurological conditions including Parkinson's disease and essential tremor. Clinical assessment continues to be performed by visual observation with quantification on clinical scales. Methodologies for objectively quantifying tremor are promising but remain non-standardized across centers. Our center performs full-body behavioral testing with 3D motion capture for clinical and research purposes in patients with Parkinson's disease, essential tremor, and other conditions. The objective of this study was to assess the ability of several candidate processing pipelines to identify the presence or absence of tremor in kinematic data from patients with confirmed movement disorders and compare them to expert ratings from movement disorders specialists. We curated a database of 2272 separate kinematic data recordings from our center, each of which was contemporaneously annotated as tremor present or absent by a movement physician. We compared the ability of six separate processing pipelines to recreate clinician ratings based on F1 score, in addition to accuracy, precision, and recall. The performance across algorithms was generally comparable. The average F1 score was 0.84±0.02 (mean ± SD; range 0.81-0.87). The second highest performing algorithm (cross-validated F1=0.87) was a hybrid that used engineered features adapted from an algorithm in longstanding clinical use with a modern Support Vector Machine classifier. Taken together, our results suggest the potential to update legacy clinical decision support systems to incorporate modern machine learning classifiers to create better-performing tools.

Computerized assessment of handwriting in de novo Parkinson's disease: A kinematic study.

INTRODUCTION: Dysgraphia, a recognized PD motor symptom, lacks effective clinical assessment. Current evaluation relies on motor assessment scales. Computational methods introduced over the past decade offer an objective dysgraphia assessment, considering size, duration, speed, and handwriting fluency. Objective evaluation of dysgraphia may be of help for early diagnosis of PD. OBJECTIVE: Computerized assessment of dysgraphia in de novo PD patients and its correlation with clinical scales. METHODS: We evaluated 38 recently diagnosed, premedication PD patients and age-matched controls without neurological disorders. Participants wrote "La casa de Pamplona es bonita" three times on paper and once on a Wacom tablet under the paper, totaling four phrases. Writing segments of 5-10 s were analyzed. The Wacom tablet captured kinematic data, including mean velocity, mean acceleration, and pen pressure. Data were saved in.svc format and analyzed using specialized software developed by Tecnocampus Mataró. Standard clinical practice data, Hoehn & Yahr staging, and UPDRS scales were used for evaluation. RESULTS: Significant kinematic differences existed; patients had lower mean speed (27 ± 12 vs. 48 ± 18, p < 0.0001) and mean acceleration (7.2 ± 3.9 vs. 15.01 ± 7, p < 0.0001) than controls. Mean speed and mean acceleration correlated significantly with UPDRS III scores (speed: r = -0.52, p < 0.0007; acceleration: r = 0.60, p < 0.0001), indicating kinematic parameters' potential in PD evaluation. CONCLUSIONS: Dysgraphia is identifiable in PD patients, even de novo, indicating an early symptom and correlates with clinical scales, offering potential for objective PD patient evaluation.

Metabolomics Unveils Disrupted Pathways in Parkinson's Disease: Toward Biomarker-Based Diagnosis.

Parkinson's disease (PD) is a neurodegenerative disorder characterized by diverse symptoms, where accurate diagnosis remains challenging. Traditional clinical observation methods often result in misdiagnosis, highlighting the need for biomarker-based diagnostic approaches. This study utilizes ultraperformance liquid chromatography coupled to an electrospray ionization source and quadrupole time-of-flight untargeted metabolomics combined with biochemometrics to identify novel serum biomarkers for PD. Analyzing a Brazilian cohort of serum samples from 39 PD patients and 15 healthy controls, we identified 15 metabolites significantly associated with PD, with 11 reported as potential biomarkers for the first time. Key disrupted metabolic pathways include caffeine metabolism, arachidonic acid metabolism, and primary bile acid biosynthesis. Our machine learning model demonstrated high accuracy, with the Rotation Forest boosting model achieving 94.1% accuracy in distinguishing PD patients from controls. It is based on three new PD biomarkers (downregulated: 1-lyso-2-arachidonoyl-phosphatidate and hypoxanthine and upregulated: ferulic acid) and surpasses the general 80% diagnostic accuracy obtained from initial clinical evaluations conducted by specialists. Besides, this machine learning model based on a decision tree allowed for visual and easy interpretability of affected metabolites in PD patients. These findings could improve the detection and monitoring of PD, paving the way for more precise diagnostics and therapeutic interventions. Our research emphasizes the critical role of metabolomics and machine learning in advancing our understanding of the chemical profile of neurodegenerative diseases.

Simultaneous time-frequency analysis of gait signals of both legs in classifying neurodegenerative diseases.

BACKGROUND: Neurodegenerative diseases (NDDs) pose significant challenges due to their debilitating nature and limited therapeutic options. Accurate and timely diagnosis is crucial for optimizing patient care and treatment strategies. Gait analysis, utilizing wearable sensors, has shown promise in assessing motor abnormalities associated with NDDs. RESEARCH QUESTION: Research Question 1 To what extent can analyzing the interaction of both limbs in the time-frequency domain serve as a suitable methodology for accurately classifying NDDs? Research Question 2 How effective is the utilization of color-coded images, in conjunction with deep transfer learning models, for the classification of NDDs? METHODS: GaitNDD database was used, comprising recordings from patients with Huntington's disease, amyotrophic lateral sclerosis, Parkinson's disease, and healthy controls. The gait signals underwent signal preparation, wavelet coherence analysis, and principal component analysis for feature enhancement. Deep transfer learning models (AlexNet, GoogLeNet, SqueezeNet) were employed for classification. Performance metrics, including accuracy, sensitivity, specificity, precision, and F1 score, were evaluated using 5-fold cross-validation. RESULTS: The classification performance of the models varied depending on the time window used. For 5-second gait signal segments, AlexNet achieved an accuracy of 95.91 %, while GoogLeNet and SqueezeNet achieved accuracies of 96.49 % and 92.73 %, respectively. For 10-second segments, AlexNet outperformed other models with an accuracy of 99.20 %, while GoogLeNet and SqueezeNet achieved accuracies of 96.75 % and 95.00 %, respectively. Statistical tests confirmed the significance of the extracted features, indicating their discriminative power for classification. SIGNIFICANCE: The proposed method demonstrated superior performance compared to previous studies, offering a non-invasive and cost-effective approach for the automated diagnosis of NDDs. By analyzing the interaction between both legs during walking using wavelet coherence, and utilizing deep transfer learning models, accurate classification of NDDs was achieved.

Association between plasma Netrin-1 levels and motor and nonmotor symptoms in Parkinson's disease.

BACKGROUND: Parkinson's disease (PD) is a prevalent neurodegenerative disorder characterized by dopaminergic neuron degeneration and diverse motor and nonmotor symptoms. Early diagnosis and intervention are crucial but challenging due to reliance on clinical presentation. Recent research suggests potential biomarkers for early detection, including plasma netrin-1 (NTN-1), a protein implicated in neuronal survival. METHODS: This cross-sectional study recruited 105 PD patients and 65 healthy controls, assessing plasma NTN-1 levels and correlating them with clinical characteristics. Statistical analyses explored associations between NTN-1 levels and PD symptoms, considering demographic factors. RESULTS: PD patients exhibited significantly lower plasma NTN-1 levels compared to controls. NTN-1 demonstrated moderate potential as a PD biomarker. Positive correlations were found between NTN-1 levels and motor, depression, and cognitive symptoms. Multiple regression analysis revealed disease duration and NTN-1 levels as key factors influencing symptom severity. Gender also impacted symptom scores. CONCLUSION: Reduced plasma NTN-1 levels correlate with PD severity, suggesting its potential as a biomarker. However, further research is needed to elucidate the roles of NTN-1 in PD pathophysiology and validate its diagnostic and therapeutic implications. Understanding the involvement of NTN-1 may lead to personalized management strategies for PD.

Limitations of the α-Synuclein Seed Amplification Assay in Clinical Practice: Understanding the Pathological Diversity of Parkinson Syndrome.

This Viewpoint cautions against premature adoption of the α-synuclein seed amplification assay as a biomarker test for Parkinson disease in general neurology practice.

A novel voice classification based on Gower distance for Parkinson disease detection.

BACKGROUND: Traditional classifier for the classification of diseases, such as K-Nearest Neighbors (KNN), Linear Discriminant Analysis (LDA), Random Forest (RF), Logistic Regression (LR), and Support Vector Machine (SVM), often struggle with high-dimensional medical datasets. OBJECTIVE: This study presents a novel classifier to overcome the limitations of traditional classifiers in Parkinson's disease (PD) detection based on Gower distance. METHODS: We present the Gower distance metric to handle diverse feature sets in voice recordings, which acts as a dissimilarity measure for all feature types, making the model adept at identifying subtle patterns indicative of PD. Additionally, the Cuckoo Search algorithm is employed for feature selection, reducing dimensionality by focusing on key features, thereby lessening the computational load associated with high-dimensional datasets. RESULTS: The proposed classifier based on Gower distance resulted in an accuracy rate of 98.3% with feature selection and achieved an accuracy of 94.92% without the feature selection method. It outperforms traditional classifiers and recent studies in PD detection from voice recordings. CONCLUSIONS: This accuracy shows the capability of the approach in the correct classification of instances and points out the potential of the approach as a reliable diagnostic tool for the medical practitioner. The findings state that the proposed approach holds promise for improving the diagnosis and monitoring of PD, both within medical institutions and at homes for the elderly.

Combining features on vertical ground reaction force signal analysis for multiclass diagnosing neurodegenerative diseases.

Neurodegenerative diseases (NDDs), which are caused by the degeneration of neurons and their functions, affect a significant part of the world's population. Although gait disorders are one of the critical and common markers to determine the presence of NDDs, diagnosing which NDD the patients have among a group of NDDs using gait data is still a significant challenge to be addressed. In this study, we addressed the multi-class classification of NDDs and aim to diagnose Parkinson's disease (PD), Amyotrophic lateral sclerosis disease (AD), and Huntington's disease (HD) from a group containing NDDs and healthy control subjects. We also examined the impact of disease-specific identified features derived from VGRF signals. Detrended Fluctuation Analysis (DFA), Dynamic Time Warping (DTW) and Autocorrelation (AC) were used for feature extraction on Vertical Ground Reaction Force (VGRF) signals. To compare the performance of the features, we employed Support Vector Machines, K-Nearest Neighbors, and Neural Networks as classifiers. In three-class problem addressing the classification of AD, PD and HD 93.3% accuracy rate was achieved, while in the four classes case, in which NDDs and HC groups were considered together, 93.5% accuracy rate was yielded. Considering the disease-specific impact of features, it is revealed that while DFA based features diagnose patients with AD with the highest accuracy, DTW has been shown to be more successful in diagnosing PD. AC based features provided the highest accuracy in diagnosing HD. Although gait disorder is common for NDDs, each disease may have its own distinctive gait rhythms; therefore, it is important to identify disease-specific patterns and parameters for the diagnosis of each disease. To increase the diagnostic accuracy, it is necessary to use a combination of features, which were effective for each disease diagnosis. Determining a limited number of disease-specific features would provide NDD diagnostic systems suitable to be deployed in edge-computing environments.

α-Synuclein in speculative neuronal extracellular vesicles: A marker for Parkinson's disease risk?

This opinion piece describes major limitations of using α-synuclein in speculative neuronally enriched for diagnosing or predicting Parkinson's disease risk from prodromal conditions such as REM behaviour disorder. It concludes that such an approach is unreliable and recommends that future researchers divert away to more widely accepted approaches such as seed amplification assays.

Digital biomarkers in Parkinson's disease.

Digital biomarker (DB) assessments provide objective measures of daily life tasks and thus hold promise to improve diagnosis and monitoring of Parkinson's disease (PD) patients especially those with advanced stages. Data from DB studies can be used in advanced analytics such as Artificial Intelligence and Machine Learning to improve monitoring, treatment and outcomes. Although early development of inertial sensors as accelerometers and gyroscopes in smartphones provided encouraging results, the use of DB remains limited due to lack of standards, harmonization and consensus for analytical as well as clinical validation. Accordingly, a number of clinical trials have been developed to evaluate the performance of DB vs traditional assessment tools with the goal of monitoring disease progression, improving quality of life and outcomes. Herein, we update current evidence on the use of DB in PD and highlight potential benefits and limitations and provide suggestions for future research study.

Feasibility of differentiating gait in Parkinson's disease and spinocerebellar degeneration using a pose estimation algorithm in two-dimensional video.

BACKGROUND: Although pose estimation algorithms have been used to analyze videos of patients with Parkinson's disease (PD) to assess symptoms, their feasibility for differentiating PD from other neurological disorders that cause gait disturbances has not been evaluated yet. We aimed to determine whether it was possible to differentiate between PD and spinocerebellar degeneration (SCD) by analyzing video recordings of patient gait using a pose estimation algorithm. METHODS: We videotaped 82 patients with PD and 61 patients with SCD performing the timed up-and-go test. A pose estimation algorithm was used to extract the coordinates of 25 key points of the participants from these videos. A transformer-based deep neural network (DNN) model was trained to predict PD or SCD using the extracted coordinate data. We employed a leave-one-participant-out cross-validation method to evaluate the predictive performance of the trained model using accuracy, sensitivity, and specificity. As there were significant differences in age, weight, and body mass index between the PD and SCD groups, propensity score matching was used to perform the same experiment in a population that did not differ in these clinical characteristics. RESULTS: The accuracy, sensitivity, and specificity of the trained model were 0.86, 0.94, and 0.75 for all participants and 0.83, 0.88, and 0.78 for the participants extracted by propensity score matching. CONCLUSION: The differentiation of PD and SCD using key point coordinates extracted from gait videos and the DNN model was feasible and could be used as a collaborative tool in clinical practice and telemedicine.

Mixed effects models but not t-tests or linear regression detect progression of apathy in Parkinson's disease over seven years in a cohort: a comparative analysis.

INTRODUCTION: While there is an interest in defining longitudinal change in people with chronic illness like Parkinson's disease (PD), statistical analysis of longitudinal data is not straightforward for clinical researchers. Here, we aim to demonstrate how the choice of statistical method may influence research outcomes, (e.g., progression in apathy), specifically the size of longitudinal effect estimates, in a cohort. METHODS: In this retrospective longitudinal analysis of 802 people with typical Parkinson's disease in the Luxembourg Parkinson's study, we compared the mean apathy scores at visit 1 and visit 8 by means of the paired two-sided t-test. Additionally, we analysed the relationship between the visit numbers and the apathy score using linear regression and longitudinal two-level mixed effects models. RESULTS: Mixed effects models were the only method able to detect progression of apathy over time. While the effects estimated for the group comparison and the linear regression were smaller with high p-values (+ 1.016/ 7 years, p = 0.107, -0.056/ 7 years, p = 0.897, respectively), effect estimates for the mixed effects models were positive with a very small p-value, indicating a significant increase in apathy symptoms by + 2.345/ 7 years (p < 0.001). CONCLUSION: The inappropriate use of paired t-tests and linear regression to analyse longitudinal data can lead to underpowered analyses and an underestimation of longitudinal change. While mixed effects models are not without limitations and need to be altered to model the time sequence between the exposure and the outcome, they are worth considering for longitudinal data analyses. In case this is not possible, limitations of the analytical approach need to be discussed and taken into account in the interpretation.

Simplification of Mobility Tests and Data Processing to Increase Applicability of Wearable Sensors as Diagnostic Tools for Parkinson's Disease.

Quantitative mobility analysis using wearable sensors, while promising as a diagnostic tool for Parkinson's disease (PD), is not commonly applied in clinical settings. Major obstacles include uncertainty regarding the best protocol for instrumented mobility testing and subsequent data processing, as well as the added workload and complexity of this multi-step process. To simplify sensor-based mobility testing in diagnosing PD, we analyzed data from 262 PD participants and 50 controls performing several motor tasks wearing a sensor on their lower back containing a triaxial accelerometer and a triaxial gyroscope. Using ensembles of heterogeneous machine learning models incorporating a range of classifiers trained on a set of sensor features, we show that our models effectively differentiate between participants with PD and controls, both for mixed-stage PD (92.6% accuracy) and a group selected for mild PD only (89.4% accuracy). Omitting algorithmic segmentation of complex mobility tasks decreased the diagnostic accuracy of our models, as did the inclusion of kinesiological features. Feature importance analysis revealed that Timed Up and Go (TUG) tasks to contribute the highest-yield predictive features, with only minor decreases in accuracy for models based on cognitive TUG as a single mobility task. Our machine learning approach facilitates major simplification of instrumented mobility testing without compromising predictive performance.