Limited diagnostic accuracy of smartphone-based digital biomarkers for Parkinson's disease in a remotely-administered setting (Preprint)

2021 ◽  
Author(s):  
Maria Goni ◽  
Simon B. Eickhoff ◽  
Mehran Sahandi Far ◽  
Kaustubh R. Patil ◽  
Juergen Dukart
Keyword(s):  
Parkinson’S Disease ◽  
Parkinson's Disease ◽  
Relevance Vector Machine ◽  
Machine Learning Algorithms ◽  
Life Tasks ◽  
Moderate Power ◽  
Series Of Experiments ◽  
Voice Data ◽  
Digital Biomarkers ◽  
Age And Sex

BACKGROUND Smartphone-based digital biomarker (DB) assessments provide objective measures of daily-life tasks and thus hold the promise to improve diagnosis and monitoring of Parkinson's disease (PD). To date, little is known about which tasks perform best for these purposes and how different confounds including comorbidities, age and sex affect their accuracy. OBJECTIVE Here we systematically assess the ability of common self-administered smartphone-based tasks to differentiate PD patients and healthy controls (HC) with and without accounting for the above confounds. METHODS Using a large cohort of PD patients and healthy volunteers acquired in the mPower study, we extracted about 700 features commonly reported in previous PD studies for gait, balance, voice and tapping tasks. We perform a series of experiments systematically assessing the effects of age, sex and comorbidities on the accuracy of the above tasks for differentiation of PD patients and HC using several machine learning algorithms. RESULTS When accounting for age, sex and comorbidities, the highest balanced accuracy on hold-out data (67%) was achieved using relevance vector machine on tapping and when combining all tasks. Only moderate accuracies were achieved for other tasks (60% for balance, 56% for gait and 55% for voice data). Not accounting for the confounders consistently yielded higher accuracies of up to 73% (for tapping) for all tasks. CONCLUSIONS Our results demonstrate the importance of controlling DB data for age and comorbidities. They further point to a moderate power of commonly applied DB tasks to differentiate between PD and HC when conducted in poorly controlled self-administered settings.

scholarly journals Limited diagnostic accuracy of smartphone-based digital biomarkers for Parkinson’s disease in a remotely-administered setting

2021 ◽  
Author(s):  
María Goñi ◽  
Simon Eickhoff ◽  
Mehran Sahandi Far ◽  
Kaustubh Patil ◽  
Juergen Dukart
Keyword(s):  
Parkinson’S Disease ◽  
Parkinson's Disease ◽  
Relevance Vector Machine ◽  
Machine Learning Algorithms ◽  
Life Tasks ◽  
Moderate Power ◽  
Series Of Experiments ◽  
Voice Data ◽  
Digital Biomarkers ◽  
Age And Sex

AbstractBackgroundSmartphone-based digital biomarker (DB) assessments provide objective measures of daily-life tasks and thus hold the promise to improve diagnosis and monitoring of Parkinson’s disease (PD). To date, little is known about which tasks perform best for these purposes and how different confounds including comorbidities, age and sex affect their accuracy. Here we systematically assess the ability of common self-administered smartphone-based tasks to differentiate PD patients and healthy controls (HC) with and without accounting for the above confounds.MethodsUsing a large cohort of PD patients and healthy volunteers acquired in the mPower study, we extracted about 700 features commonly reported in previous PD studies for gait, balance, voice and tapping tasks. We perform a series of experiments systematically assessing the effects of age, sex and comorbidities on the accuracy of the above tasks for differentiation of PD patients and HC using several machine learning algorithms.ResultsWhen accounting for age, sex and comorbidities, the highest balanced accuracy on hold-out data (67%) was achieved using relevance vector machine on tapping and when combining all tasks. Only moderate accuracies were achieved for other tasks (60% for balance, 56% for gait and 55% for voice data). Not accounting for the confounders consistently yielded higher accuracies of up to 73% (for tapping) for all tasks.DiscussionOur results demonstrate the importance of controlling DB data for age and comorbidities. They further point to a moderate power of commonly applied DB tasks to differentiate between PD and HC when conducted in poorly controlled self-administered settings.

Tremor Identification Using Machine Learning in Parkinson's Disease

2019 ◽  
pp. 128-151
Author(s):  
Angana Saikia ◽  
Vinayak Majhi ◽  
Masaraf Hussain ◽  
Sudip Paul ◽  
Amitava Datta
Keyword(s):  
Machine Learning ◽  
Parkinson’S Disease ◽  
Support Vector Machine ◽  
Parkinson's Disease ◽  
Discriminant Analysis ◽  
Learning Algorithms ◽  
The Body ◽  
Machine Learning Algorithms ◽  
Support Vector

Tremor is an involuntary quivering movement or shake. Characteristically occurring at rest, the classic slow, rhythmic tremor of Parkinson's disease (PD) typically starts in one hand, foot, or leg and can eventually affect both sides of the body. The resting tremor of PD can also occur in the jaw, chin, mouth, or tongue. Loss of dopamine leads to the symptoms of Parkinson's disease and may include a tremor. For some people, a tremor might be the first symptom of PD. Various studies have proposed measurable technologies and the analysis of the characteristics of Parkinsonian tremors using different techniques. Various machine-learning algorithms such as a support vector machine (SVM) with three kernels, a discriminant analysis, a random forest, and a kNN algorithm are also used to classify and identify various kinds of tremors. This chapter focuses on an in-depth review on identification and classification of various Parkinsonian tremors using machine learning algorithms.

scholarly journals Prevalence and type of cervical deformities among adults with Parkinson's disease: a cross-sectional study

2016 ◽  
Vol 24 (4) ◽  
pp. 527-534 ◽  
Author(s):  
Bong Ju Moon ◽  
Justin S. Smith ◽  
Christopher P. Ames ◽  
Christopher I. Shaffrey ◽  
Virginie Lafage ◽  
...  
Keyword(s):  
Parkinson’S Disease ◽  
Parkinson's Disease ◽  
Logistic Regression ◽  
Regression Analysis ◽  
Cervical Spine ◽  
Logistic Regression Analysis ◽  
Control Group ◽  
Spine Deformities ◽  
Sagittal Malalignment ◽  
Age And Sex

OBJECT To identify the characteristics of cervical deformities in Parkinson's disease (PD) and the role of severity of PD in the development of cervical spine deformities, the authors investigated the prevalence of the cervical deformities, cervical kyphosis (CK), and cervical positive sagittal malalignment (CPSM) in patients with PD. They also analyzed the association of severity of cervical deformities with the stage of PD in the context of global sagittal spinopelvic alignment. METHODS This study was a prospective assessment of consecutively treated patients (n = 89) with PD. A control group of the age- and sex-matched patients was selected from patients with degenerative cervical spine disease but without PD. Clinical and demographic parameters including age, sex, duration of PD, and Hoehn and Yahr (H&Y) stage were collected. Full-length standing radiographs were used to assess spinopelvic parameters. CK was defined as a C2–7 Cobb angle < 0°. CPSM was defined as C2–7 sagittal vertical axis (SVA) > 4 cm. RESULTS A significantly higher prevalence of CPSM (28% vs 1.1%, p < 0.001), but not CK (12% vs 10.1%, p = 0.635), was found in PD patients compared with control patients. Among patients with PD, those with CK were younger (62.1 vs 69.0 years, p = 0.013) and had longer duration of PD (56.4 vs 36.2 months, p = 0.034), but the severity of PD was not significantly different. Logistic regression analysis revealed that the presence of CK was associated with younger age, higher mismatch between pelvic incidence and lumbar lordosis, and lower C7–S1 SVA. The patients with CPSM had significantly greater thoracic kyphosis (TK) (p < 0.001) and a trend toward more advanced H&Y stage (p = 0.05). Logistic regression analysis revealed that CPSM was associated with male sex, greater TK, and more advanced H&Y stage. CONCLUSIONS Patients with PD have a significantly higher prevalence of CPSM compared with age- and sex-matched control patients with cervical degenerative disease but without PD. Among patients with PD, CK is not associated with the severity of PD but is associated with overall global sagittal malalignment. In contrast, the presence of CPSM is associated more with the severity of PD than it is with the presence of global sagittal malalignment. Collectively, these data suggest that the neuromuscular pathogenesis of PD may affect the development of CPSM more than of CK.

scholarly journals Can Gut Microbiota Be a Good Predictor for Parkinson’s Disease? A Machine Learning Approach

2020 ◽  
Vol 10 (4) ◽  
pp. 242 ◽  
Author(s):  
Daniele Pietrucci ◽  
Adelaide Teofani ◽  
Valeria Unida ◽  
Rocco Cerroni ◽  
Silvia Biocca ◽  
...  
Keyword(s):  
Machine Learning ◽  
Parkinson’S Disease ◽  
Parkinson's Disease ◽  
Random Forest ◽  
Gut Microbiota ◽  
Biological Data ◽  
Machine Learning Algorithms ◽  
Support Vector ◽  
Published Data ◽  
Promising Tool

The involvement of the gut microbiota in Parkinson’s disease (PD), investigated in several studies, identified some common alterations of the microbial community, such as a decrease in Lachnospiraceae and an increase in Verrucomicrobiaceae families in PD patients. However, the results of other bacterial families are often contradictory. Machine learning is a promising tool for building predictive models for the classification of biological data, such as those produced in metagenomic studies. We tested three different machine learning algorithms (random forest, neural networks and support vector machines), analyzing 846 metagenomic samples (472 from PD patients and 374 from healthy controls), including our published data and those downloaded from public databases. Prediction performance was evaluated by the area under curve, accuracy, precision, recall and F-score metrics. The random forest algorithm provided the best results. Bacterial families were sorted according to their importance in the classification, and a subset of 22 families has been identified for the prediction of patient status. Although the results are promising, it is necessary to train the algorithm with a larger number of samples in order to increase the accuracy of the procedure.

scholarly journals Cancerous Inhibitor of Protein Phosphatase 2A (CIP2A): Could It Be a Promising Biomarker and Therapeutic Target in Parkinson’s Disease?

2022 ◽  
Author(s):  
Sijia Yin ◽  
Chao Han ◽  
Yun Xia ◽  
Fang Wan ◽  
Junjie Hu ◽  
...  
Keyword(s):  
Parkinson’S Disease ◽  
Parkinson's Disease ◽  
Protein Phosphatase ◽  
Essential Role ◽  
Protein Phosphatase 2A ◽  
Area Under The Curve ◽  
Alpha Synuclein ◽  
Phosphatase 2A ◽  
Induced Aggregation ◽  
Age And Sex

AbstractParkinson’s disease (PD) is an incurable neurodegenerative disease characterized by aggregation of pathological alpha-synuclein (α-syn) and loss of dopaminergic neuron in the substantia nigra. Inhibition of phosphorylation of the α-syn has been shown to mediate alleviation of PD-related pathology. Protein phosphatase 2A (PP2A), an important serine/threonine phosphatase, plays an essential role in catalyzing dephosphorylation of the α-syn. Here, we identified and validated cancerous inhibitor of PP2A (CIP2A), as a potential diagnostic biomarker for PD. Our data showed that plasma CIP2A concentrations in PD patients were significantly lower compared to age- and sex-matched controls, 1.721 (1.435–2.428) ng/ml vs 3.051(2.36–5.475) ng/ml, p < 0.0001. The area under the curve of the plasma CIP2A in distinguishing PD from the age- and sex-matched controls was 0.776. In addition, we evaluated the role of CIP2A in PD-related pathogenesis in PD cellular and MPTP-induced mouse model. The results demonstrated that CIP2A is upregulated in PD cellular and MPTP-induced mouse models. Besides, suppression of the CIP2A expression alleviates rotenone induced aggregation of the α-syn as well as phosphorylation of the α-syn in SH-SY5Y cells, which is associated with increased PP2A activity. Taken together, our data demonstrated that CIP2A plays an essential role in the mechanisms related to PD development and might be a novel PD biomarker.

scholarly journals BODY COMPOSITION AND POSTURAL INSTABILITY IN PEOPLE WITH IDIOPATHIC PARKINSON’S DISEASE

2016 ◽  
pp. 1-6
Author(s):  
A.S. Diab ◽  
L.A. Hale ◽  
M.A. Skinner ◽  
G. Hammond-Tooke ◽  
A.L. Ward ◽  
...  
Keyword(s):  
Parkinson’S Disease ◽  
Parkinson's Disease ◽  
Body Composition ◽  
Postural Stability ◽  
Postural Instability ◽  
Composite Score ◽  
Step Test ◽  
Timed Up And Go ◽  
Composite Scores ◽  
Age And Sex

Objectives: Idopathic Parkinson’s disease (PD) is the second most common neurodegenerative disorder. Our objective was to investigate the relationship between body composition and postural instability in people with PD, and age- and sex-matched controls. Design: Cross-sectional study among PD sufferers and age- and sex-matched controls. Setting: University of Otago’s Balance Clinic, School of Physiotherapy. Participants: Forty-seven people with PD and 58 age- and sex-matched controls. Measurements: Postural stability was assessed with the Sensory Organization Test, Motor Control Test, Timed Up and Go Test, and Step Test. Body composition was measured by dual energy x-ray absorptiometry (DXA). Movement Disorders Society-Unified Parkinson’s Disease Rating Scale was applied to assess PD severity. Results: Mean group differences between PD and controls for the equilibrium composite score, Timed Up and Go Tests, and Step Test were statistically significant (p<0.05); strategy and latency composite scores and body composition variables were not (p>0.05). Three PD participants were sarcopenic; 15 PD and 24 controls were obese. In PD participants, total body lean mass and age predicted latency composite scores. Disease, age, and leg fat mass predicted the Timed Up and Go Test results (p<0.05). Sex and disease predicted the equilibrium composite score (p<0.01). Conclusion: The prevalence of obesity was high and sarcopenia low in the PD group, which is a novel finding. Not surprisingly, participants with PD had reduced postural stability compared to controls. Disease status, age and sex were influential factors in the weak relationships found between postural stability and body composition. These findings may have clinical relevance for the treatment of the physical symptoms of those suffering from PD.

scholarly journals Machine Learning Prediction of Parkinson's Disease Onset and Subtype Using Germline Variants

2021 ◽  
Author(s):  
Saya R Dennis ◽  
Tanya Simuni ◽  
Yuan Luo
Keyword(s):  
Machine Learning ◽  
Parkinson’S Disease ◽  
Parkinson's Disease ◽  
Neurodegenerative Disorder ◽  
Disease Onset ◽  
The United States ◽  
Machine Learning Algorithms ◽  
Progression Rate ◽  
High Importance ◽  
Germline Variants

Parkinson's Disease is the second most common neurodegenerative disorder in the United States, and is characterized by a largely irreversible worsening of motor and non-motor symptoms as the disease progresses. A prominent characteristic of the disease is its high heterogeneity in manifestation as well as the progression rate. For sporadic Parkinson's Disease, which comprises ~90% of all diagnoses, the relationship between the patient genome and disease onset or progression subtype remains largely elusive. Machine learning algorithms are increasingly adopted to study the genomics of diseases due to their ability to capture patterns within the vast feature space of the human genome that might be contributing to the phenotype of interest. In our study, we develop two machine learning models that predict the onset as well as the progression subtype of Parkinson's Disease based on subjects' germline mutations. Our best models achieved an ROC of 0.77 and 0.61 for disease onset and subtype prediction, respectively. To the best of our knowledge, our models present state-of-the-art prediction performances of PD onset and subtype solely based on the subjects' germline variants. The genes with high importance in our best-performing models were enriched for several canonical pathways related to signaling, immune system, and protein modifications, all of which have been previously associated with PD symptoms or pathogenesis. These high-importance gene sets provide us with promising candidate genes for future biomedical and clinical research.

Levels of oligomeric α-Synuclein in reflex tears distinguish Parkinson's disease patients from healthy controls

2019 ◽  
Vol 13 (17) ◽  
pp. 1447-1457 ◽  
Author(s):  
Sarah F Hamm-Alvarez ◽  
Srikanth R Janga ◽  
Maria C Edman ◽  
Danielle Feigenbaum ◽  
Daniel Freire ◽  
...  
Keyword(s):  
Parkinson’S Disease ◽  
Parkinson's Disease ◽  
Nerve Fibers ◽  
Afferent Nerve ◽  
Healthy Controls ◽  
Active Engagement ◽  
Potential Source ◽  
Afferent Nerve Fibers ◽  
Tear Flow ◽  
Age And Sex

Aim: Due to active engagement of sensory and afferent nerve fibers in reflex tearing which could be affected in Parkinson's disease (PD), we tested reflex tears as a source of potential PD biomarkers. Patients & methods: Reflex tears collected from 84 PD and 84 age- and sex-equivalent healthy controls (HC) were used to measure levels of oligomeric α-Syn (α-SynOligo), total α-Syn (α-SynTotal), CCL2, DJ-1, lactoferrin and MMP9. Results: α-synOligo (p < 0.0001), CCL2 (p = 0.003) and lactoferrin (p = 0.002) were significantly elevated in PD patient tears relative to HC tears. Tear flow was significantly lower in PD relative to HC (p = 0.001). Conclusion: Reflex tears are a potential source for detection of characteristic changes in PD patients.

Sign in / Sign up

Export Citation Format

Share Document