Model-based and Model-free Machine Learning Techniques for Diagnostic Prediction and Classification of Clinical Outcomes in Parkinson's Disease

Gait Alterations and Association With Worsening Knee Pain and Physical Function: A Machine Learning Approach With Wearable Sensors in the Multicenter Osteoarthritis Study

OBJECTIVE

The objective of this study was to identify gait alterations related to worsening knee pain and worsening physical function, using machine learning approaches applied to wearable sensor-derived data from a large observational cohort.

METHODS

Participants in the Multicenter Osteoarthritis Study (MOST) completed a 20-m walk test wearing inertial sensors on their lower back and ankles. Parameters describing spatiotemporal features of gait were extracted from these data. We used an ensemble machine learning technique ("super learning") to optimally discriminate between those with and without worsening physical function and, separately, those with and without worsening pain over two years. We then used log-binomial regression to evaluate associations of the top 10 influential variables selected with super learning with each outcome. We also assessed whether the relation of altered gait with worsening function was mediated by changes in pain.

RESULTS

Of 2,324 participants, 29% and 24% had worsening knee pain and function over two years, respectively. From the super learner, several gait parameters were found to be influential for worsening pain and for worsening function. After adjusting for confounders, greater gait asymmetry, longer average step length, and lower dominant frequency were associated with worsening pain, and lower cadence was associated with worsening function. Worsening pain partially mediated the association of cadence with function.

CONCLUSION

We identified gait alterations associated with worsening knee pain and those associated with worsening physical function. These alterations could be assessed with wearable sensors in clinical settings. Further research should determine whether they might be therapeutic targets to prevent worsening pain and worsening function.

Modeling Epidemiology Data with Machine Learning Technique to Detect Risk Factors for Gastric Cancer

PURPOSE

Gastric cancer (GC) ranks as the 7th most common cancer worldwide and a leading cause of cancer mortality. In Iran, stomach malignancies are the most common fatal cancers with higher than world average incidence. In recent years, methods like machine learning that provide the opportunity of merging health issues with computational power and learning capacity have caught considerable attention for prediction and diagnosis of diseases. In this study, we aimed to model GC data to find risk factors and identify GC cases in Golestan Cohort Study (GCS), using gradient boosting as a machine learning technique.

METHODS

Since the GC class (280) was smaller than not-GC (49,467), "Synthetic Minority Oversampling Technique" was used to balance the dataset. Seventy percent of the data was used to train the gradient boosting algorithm and find effective factors on gastric cancer, and the remaining 30% was used for accuracy assessment.

RESULTS

Our results indicated that out of 19 factors, age, social economical status, tea temperature, body mass index, gender, and education were the top six effective factors with impact rates of 0.24, 0.16, 0.13, 0.13, and 0.07, respectively. The trained model classified 70 out of 72 GC patients in the test set, correctly.

CONCLUSION

The results indicate that this model can effectively detect gastric cancer (GC) by utilizing important risk factors, thus avoiding the need for invasive procedures. The model's performance is reliable when provided with an adequate amount of input data, and as the dataset expands, its accuracy and generalization improve significantly. Overall, the trained system's success stems from its ability to identify risk factors and identify cancer patients.

A machine learning model for prediction of sarcopenia in patients with Parkinson's Disease

OBJECTIVE

Patients with Parkinson's disease (PD) have an increased risk of sarcopenia which is expected to negatively affect gait, leading to poor clinical outcomes including falls. In this study, we investigated the gait patterns of patients with PD with and without sarcopenia (sarcopenia and non-sarcopenia groups, respectively) using an app-derived program and explored if gait parameters could be utilized to predict sarcopenia based on machine learning.

METHODS

Clinical and sarcopenia profiles were collected from patients with PD at Hoehn and Yahr (HY) stage ≤ 2. Sarcopenia was defined based on the updated criteria of the Asian Working Group for Sarcopenia. The gait patterns of the patients with and without sarcopenia were recorded and analyzed using a smartphone application. The random forest model was applied to predict sarcopenia in patients with PD.

RESULTS

Data from 38 patients with PD were obtained, among which 9 (23.7%) were with sarcopenia. Clinical parameters were comparable between the sarcopenia and non-sarcopenia groups. Among various clinical and gait parameters, the average range of motion of the hip joint showed the highest association with sarcopenia. Based on the random forest algorithm, the combined difference in knee and ankle angles from standing still before walking to the maximum angle during walking (Kneeankle_diff), the difference between the angle when standing still before walking and the maximum angle during walking for the ankle (Ankle_dif), and the min angle of the hip joint (Hip_min) were the top three features that best predict sarcopenia. The accuracy of this model was 0.949.

CONCLUSIONS

Using smartphone app and machine learning technique, our study revealed gait parameters that are associated with sarcopenia and that help predict sarcopenia in PD. Our study showed potential application of advanced technology in clinical research.

Development and validation of machine learning models for nonalcoholic fatty liver disease

BACKGROUND

Nonalcoholic fatty liver disease (NAFLD) had become the most prevalent liver disease worldwide. Early diagnosis could effectively reduce NAFLD-related morbidity and mortality. This study aimed to combine the risk factors to develop and validate a novel model for predicting NAFLD.

METHODS

We enrolled 578 participants completing abdominal ultrasound into the training set. The least absolute shrinkage and selection operator (LASSO) regression combined with random forest (RF) was conducted to screen significant predictors for NAFLD risk. Five machine learning models including logistic regression (LR), RF, extreme gradient boosting (XGBoost), gradient boosting machine (GBM), and support vector machine (SVM) were developed. To further improve model performance, we conducted hyperparameter tuning with train function in Python package 'sklearn'. We included 131 participants completing magnetic resonance imaging into the testing set for external validation.

RESULTS

There were 329 participants with NAFLD and 249 without in the training set, while 96 with NAFLD and 35 without were in the testing set. Visceral adiposity index, abdominal circumference, body mass index, alanine aminotransferase (ALT), ALT/AST (aspartate aminotransferase), age, high-density lipoprotein cholesterol (HDL-C) and elevated triglyceride (TG) were important predictors for NAFLD risk. The area under curve (AUC) of LR, RF, XGBoost, GBM, SVM were 0.915 [95% confidence interval (CI): 0.886-0.937], 0.907 (95% CI: 0.856-0.938), 0.928 (95% CI: 0.873-0.944), 0.924 (95% CI: 0.875-0.939), and 0.900 (95% CI: 0.883-0.913), respectively. XGBoost model presented the best predictive performance, and its AUC was enhanced to 0.938 (95% CI: 0.870-0.950) with further parameter tuning.

CONCLUSIONS

This study developed and validated five novel machine learning models for NAFLD prediction, among which XGBoost presented the best performance and was considered a reliable reference for early identification of high-risk patients with NAFLD in clinical practice.

Screening for obstructive sleep apnea in patients with cancer - a machine learning approach

Background

Obstructive sleep apnea (OSA) is a highly prevalent sleep disorder associated with daytime sleepiness, fatigue, and increased all-cause mortality risk in patients with cancer. Existing screening tools for OSA do not account for the interaction of cancer-related features that may increase OSA risk.

Study Design and Methods

This is a retrospective study of patients with cancer at a single tertiary cancer institution who underwent a home sleep apnea test (HSAT) to evaluate for OSA. Unsupervised machine learning (ML) was used to reduce the dimensions and extract significant features associated with OSA. ML classifiers were applied to principal components and model hyperparameters were optimized using k-fold cross-validation. Training models for OSA were subsequently tested and compared with the STOP-Bang questionnaire on a prospective unseen test set of patients who underwent an HSAT.

Results

From a training dataset of 249 patients, kernel principal component analysis (PCA) extracted eight components through dimension reduction to explain the maximum variance with OSA at 98%. Predictors of OSA were smoking, asthma, chronic kidney disease, STOP-Bang score, race, diabetes, radiation to head/neck/thorax (RT-HNT), type of cancer, and cancer metastases. Of the ML models, PCA + RF had the highest sensitivity (96.8%), specificity (92.3%), negative predictive value (92%), F1 score (0.93), and ROC-AUC score (0.88). The PCA + RF screening algorithm also performed better than the STOP-Bang questionnaire alone when tested on a prospective unseen test set.

Conclusions

The PCA + RF ML model had the highest accuracy in screening for OSA in patients with cancer. History of RT-HNT, cancer metastases, and type of cancer were identified as cancer-related risk factors for OSA.

A-GSTCN: An Augmented Graph Structural-Temporal Convolution Network for Medication Recommendation Based on Electronic Health Records

Medication recommendation based on electronic health records (EHRs) is a significant research direction in the biomedical field, which aims to provide a reasonable prescription for patients according to their historical and current health conditions. However, the existing recommended methods have many limitations in dealing with the structural and temporal characteristics of EHRs. These methods either only consider the current state while ignoring the historical situation, or fail to adequately assess the structural correlations among various medical events. These factors result in poor recommendation quality. To solve this problem, we propose an augmented graph structural-temporal convolutional network (A-GSTCN). Firstly, an augmented graph attention network is used to model the structural features among medical events of patients' EHRs. Next, the dilated convolution combined with residual connection is applied in the proposed model, which can improve the temporal prediction capability and further reduce the complexity. Moreover, the cache memory module further enhances the model's learning of the history of EHRs. Finally, the A-GSTCN model is compared with the baselines through experiments, and the efficiency of the A-GSTCN model is verified by Jaccard, F1 and PRAUC. Not only that, the proposed model also reduces the training parameters by an order of magnitude.

Classifying Tremor Dominant and Postural Instability and Gait Difficulty Subtypes of Parkinson's Disease from Full-Body Kinematics

Characterizing motor subtypes of Parkinson's disease (PD) is an important aspect of clinical care that is useful for prognosis and medical management. Although all PD cases involve the loss of dopaminergic neurons in the brain, individual cases may present with different combinations of motor signs, which may indicate differences in underlying pathology and potential response to treatment. However, the conventional method for distinguishing PD motor subtypes involves resource-intensive physical examination by a movement disorders specialist. Moreover, the standardized rating scales for PD rely on subjective observation, which requires specialized training and unavoidable inter-rater variability. In this work, we propose a system that uses machine learning models to automatically and objectively identify some PD motor subtypes, specifically Tremor-Dominant (TD) and Postural Instability and Gait Difficulty (PIGD), from 3D kinematic data recorded during walking tasks for patients with PD (MDS-UPDRS-III Score, 34.7 ± 10.5, average disease duration 7.5 ± 4.5 years). This study demonstrates a machine learning model utilizing kinematic data that identifies PD motor subtypes with a 79.6% F1 score (N = 55 patients with parkinsonism). This significantly outperformed a comparison model using classification based on gait features (19.8% F1 score). Variants of our model trained to individual patients achieved a 95.4% F1 score. This analysis revealed that both temporal, spectral, and statistical features from lower body movements are helpful in distinguishing motor subtypes. Automatically assessing PD motor subtypes simply from walking may reduce the time and resources required from specialists, thereby improving patient care for PD treatments. Furthermore, this system can provide objective assessments to track the changes in PD motor subtypes over time to implement and modify appropriate treatment plans for individual patients as needed.

Gait and turning characteristics from daily life increase ability to predict future falls in people with Parkinson's disease

Objectives

To investigate if digital measures of gait (walking and turning) collected passively over a week of daily activities in people with Parkinson's disease (PD) increases the discriminative ability to predict future falls compared to fall history alone.

Methods

We recruited 34 individuals with PD (17 with history of falls and 17 non-fallers), age: 68 ± 6 years, MDS-UPDRS III ON: 31 ± 9. Participants were classified as fallers (at least one fall) or non-fallers based on self-reported falls in past 6 months. Eighty digital measures of gait were derived from 3 inertial sensors (Opal® V2 System) placed on the feet and lower back for a week of passive gait monitoring. Logistic regression employing a "best subsets selection strategy" was used to find combinations of measures that discriminated future fallers from non-fallers, and the Area Under Curve (AUC). Participants were followed via email every 2 weeks over the year after the study for self-reported falls.

Results

Twenty-five subjects reported falls in the follow-up year. Quantity of gait and turning measures (e.g., number of gait bouts and turns per hour) were similar in future fallers and non-fallers. The AUC to discriminate future fallers from non-fallers using fall history alone was 0.77 (95% CI: [0.50-1.00]). In contrast, the highest AUC for gait and turning digital measures with 4 combinations was 0.94 [0.84-1.00]. From the top 10 models (all AUCs>0.90) via the best subsets strategy, the most consistently selected measures were variability of toe-out angle of the foot (9 out of 10), pitch angle of the foot during mid-swing (8 out of 10), and peak turn velocity (7 out of 10).

Conclusions

These findings highlight the importance of considering precise digital measures, captured via sensors strategically placed on the feet and low back, to quantify several different aspects of gait (walking and turning) during daily life to improve the classification of future fallers in PD.

Bayesian Optimization with Support Vector Machine Model for Parkinson Disease Classification

Parkinson's disease (PD) has become widespread these days all over the world. PD affects the nervous system of the human and also affects a lot of human body parts that are connected via nerves. In order to make a classification for people who suffer from PD and who do not suffer from the disease, an advanced model called Bayesian Optimization-Support Vector Machine (BO-SVM) is presented in this paper for making the classification process. Bayesian Optimization (BO) is a hyperparameter tuning technique for optimizing the hyperparameters of machine learning models in order to obtain better accuracy. In this paper, BO is used to optimize the hyperparameters for six machine learning models, namely, Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), Naive Bayes (NB), Ridge Classifier (RC), and Decision Tree (DT). The dataset used in this study consists of 23 features and 195 instances. The class label of the target feature is 1 and 0, where 1 refers to the person suffering from PD and 0 refers to the person who does not suffer from PD. Four evaluation metrics, namely, accuracy, F1-score, recall, and precision were computed to evaluate the performance of the classification models used in this paper. The performance of the six machine learning models was tested on the dataset before and after the process of hyperparameter tuning. The experimental results demonstrated that the SVM model achieved the best results when compared with other machine learning models before and after the process of hyperparameter tuning, with an accuracy of 92.3% obtained using BO.

Prediction model for missed abortion of patients treated with IVF-ET based on XGBoost: a retrospective study

Aim

In this study, we established a model based on XGBoost to predict the risk of missed abortion in patients treated with in vitro fertilization-embryo transfer (IVF-ET), evaluated its prediction ability, and compared the model with the traditional logical regression model.

Methods

We retrospectively collected the clinical data of 1,017 infertile women treated with IVF-ET. The independent risk factors were screened by performing a univariate analysis and binary logistic regression analysis, and then, all cases were randomly divided into the training set and the test set in a 7:3 ratio for constructing and validating the model. We then constructed the prediction models by the traditional logical regression method and the XGBoost method and tested the prediction performance of the two models by resampling.

Results

The results of the binary logistic regression analysis showed that several factors, including the age of men and women, abnormal ovarian structure, prolactin (PRL), anti-Müllerian hormone (AMH), activated partial thromboplastin time (APTT), anticardiolipin antibody (ACA), and thyroid peroxidase antibody (TPO-Ab), independently influenced missed abortion significantly (P < 0.05). The area under the receiver operating characteristic curve (AUC) score and the F1 score with the training set of the XGBoost model (0.877 ± 0.014 and 0.730 ± 0.019, respectively) were significantly higher than those of the logistic model (0.713 ± 0.013 and 0.568 ± 0.026, respectively). In the test set, the AUC and F1 scores of the XGBoost model (0.759 ± 0.023 and 0.566 ± 0.042, respectively) were also higher than those of the logistic model (0.695 ± 0.030 and 0.550 ± 049, respectively).

Conclusions

We established a prediction model based on the XGBoost algorithm, which can accurately predict the risk of missed abortion in patients with IVF-ET. This model performed better than the traditional logical regression model.

Application of novel hybrid machine learning systems and radiomics features for non-motor outcome prediction in Parkinson's disease

Objectives.Parkinson's disease (PD) is a complex neurodegenerative disorder, affecting 2%-3% of the elderly population. Montreal Cognitive Assessment (MoCA), a rapid nonmotor screening test, assesses different cognitive dysfunctionality aspects. Early MoCA prediction may facilitate better temporal therapy and disease control. Radiomics features (RF), in addition to clinical features (CF), are indicated to increase clinical diagnoses, etc, bridging between medical imaging procedures and personalized medicine. We investigate the effect of RFs, CFs, and conventional imaging features (CIF) to enhance prediction performance using hybrid machine learning systems (HMLS).Methods.We selected 210 patients with 981 features (CFs, CIFs, and RFs) from the Parkinson's Progression-Markers-Initiative database. We generated 4 datasets, namely using (i), (ii) year-0 (D1) or year-1 (D2) features, (iii) longitudinal data (D3, putting datasets in years 0 and 1 longitudinally next to each other), and (iv) timeless data (D4, effectively doubling dataset size by listing both datasets from years 0 and 1 separately). First, we directly applied 23 predictor algorithms (PA) to the datasets to predict year-4 MoCA, which PD patients this year have a higher dementia risk. Subsequently, HMLSs, including 14 attribute extraction and 10 feature selection algorithms followed by PAs were employed to enhance prediction performances. 80% of all datapoints were utilized to select the best model based on minimum mean absolute error (MAE) resulting from 5-fold cross-validation. Subsequently, the remaining 20% was used for hold-out testing of the selected models.Results.When applying PAs without ASAs/FEAs to datasets (MoCA outcome range: [11,30]), Adaboost achieved an MAE of 1.74 ± 0.29 on D4 with a hold-out testing performance of 1.71. When employing HMLSs, D4 + Minimum_Redundancy_Maximum_Relevance (MRMR)+K_Nearest_Neighbor Regressor achieved the highest performance of 1.05 ± 0.25 with a hold-out testing performance of 0.57.Conclusion.Our study shows the importance of using larger datasets (timeless), and utilizing optimized HMLSs, for significantly improved prediction of MoCA in PD patients.

Development of a 32-gene signature using machine learning for accurate prediction of inflammatory bowel disease

Inflammatory bowel disease (IBD) is a chronic inflammatory condition caused by multiple genetic and environmental factors. Numerous genes are implicated in the etiology of IBD, but the diagnosis of IBD is challenging. Here, XGBoost, a machine learning prediction model, has been used to distinguish IBD from healthy cases following elaborative feature selection. Using combined unsupervised clustering analysis and the XGBoost feature selection method, we successfully identified a 32-gene signature that can predict IBD occurrence in new cohorts with 0.8651 accuracy. The signature shows enrichment in neutrophil extracellular trap formation and cytokine signaling in the immune system. The probability threshold of the XGBoost-based classification model can be adjusted to fit personalized lifestyle and health status. Therefore, this study reveals potential IBD-related biomarkers that facilitate an effective personalized diagnosis of IBD.

Guidance for good practice in the application of machine learning in development of toxicological quantitative structure-activity relationships (QSARs)

Recent years have seen a substantial growth in the adoption of machine learning approaches for the purposes of quantitative structure-activity relationship (QSAR) development. Such a trend has coincided with desire to see a shifting in the focus of methodology employed within chemical safety assessment: away from traditional reliance upon animal-intensive in vivo protocols, and towards increased application of in silico (or computational) predictive toxicology. With QSAR central amongst techniques applied in this area, the emergence of algorithms trained through machine learning with the objective of toxicity estimation has, quite naturally, arisen. On account of the pattern-recognition capabilities of the underlying methods, the statistical power of the ensuing models is potentially considerable-appropriate for the handling even of vast, heterogeneous datasets. However, such potency comes at a price: this manifesting as the general practical deficits observed with respect to the reproducibility, interpretability and generalisability of the resulting tools. Unsurprisingly, these elements have served to hinder broader uptake (most notably within a regulatory setting). Areas of uncertainty liable to accompany (and hence detract from applicability of) toxicological QSAR have previously been highlighted, accompanied by the forwarding of suggestions for "best practice" aimed at mitigation of their influence. However, the scope of such exercises has remained limited to "classical" QSAR-that conducted through use of linear regression and related techniques, with the adoption of comparatively few features or descriptors. Accordingly, the intention of this study has been to extend the remit of best practice guidance, so as to address concerns specific to employment of machine learning within the field. In doing so, the impact of strategies aimed at enhancing the transparency (feature importance, feature reduction), generalisability (cross-validation) and predictive power (hyperparameter optimisation) of algorithms, trained upon real toxicity data through six common learning approaches, is evaluated.

Relation of gait measures with mild unilateral knee pain during walking using machine learning

Gait alterations in those with mild unilateral knee pain during walking may provide clues to modifiable alterations that affect progression of knee pain and osteoarthritis (OA). To examine this, we applied machine learning (ML) approaches to gait data from wearable sensors in a large observational knee OA cohort, the Multicenter Osteoarthritis (MOST) study. Participants completed a 20-m walk test wearing sensors on their trunk and ankles. Parameters describing spatiotemporal features of gait and symmetry, variability and complexity were extracted. We used an ensemble ML technique ("super learning") to identify gait variables in our cross-sectional data associated with the presence/absence of unilateral knee pain. We then used logistic regression to determine the association of selected gait variables with odds of mild knee pain. Of 2066 participants (mean age 63.6 [SD: 10.4] years, 56% female), 21.3% had mild unilateral pain while walking. Gait parameters selected in the ML process as influential included step regularity, sample entropy, gait speed, and amplitude dominant frequency, among others. In adjusted cross-sectional analyses, lower levels of step regularity (i.e., greater gait variability) and lower sample entropy(i.e., lower gait complexity) were associated with increased likelihood of unilateral mild pain while walking [aOR 0.80 (0.64-1.00) and aOR 0.79 (0.66-0.95), respectively].

Scoring the Sit-to-Stand Performance of Parkinson's Patients with a Single Wearable Sensor

Monitoring disease progression in Parkinson's disease is challenging. Postural transfers by sit-to-stand motions are adapted to trace the motor performance of subjects. Wearable sensors such as inertial measurement units allow for monitoring motion performance. We propose quantifying the sit-to-stand performance based on two scores compiling kinematics, dynamics, and energy-related variables. Three groups participated in this research: asymptomatic young participants (n = 33), senior asymptomatic participants (n = 17), and Parkinson's patients (n = 20). An unsupervised classification was performed of the two scores to differentiate the three populations. We found a sensitivity of 0.4 and a specificity of 0.96 to distinguish Parkinson's patients from asymptomatic subjects. In addition, seven Parkinson's patients performed the sit-to-stand task "ON" and "OFF" medication, and we noted the scores improved with the patients' medication states (MDS-UPDRS III scores). Our investigation revealed that Parkinson's patients demonstrate a wide spectrum of mobility variations, and while one inertial measurement unit can quantify the sit-to-stand performance, differentiating between PD patients and healthy adults and distinguishing between "ON" and "OFF" periods in PD patients is still challenging.

Imperative Role of Machine Learning Algorithm for Detection of Parkinson’s Disease: Review, Challenges and Recommendations

Parkinson’s disease (PD) is a neurodegenerative disease that affects the neural, behavioral, and physiological systems of the brain. This disease is also known as tremor. The common symptoms of this disease are a slowness of movement known as ‘bradykinesia’, loss of automatic movements, speech/writing changes, and difficulty with walking at early stages. To solve these issues and to enhance the diagnostic process of PD, machine learning (ML) algorithms have been implemented for the categorization of subjective disease and healthy controls (HC) with comparable medical appearances. To provide a far-reaching outline of data modalities and artificial intelligence techniques that have been utilized in the analysis and diagnosis of PD, we conducted a literature analysis of research papers published up until 2022. A total of 112 research papers were included in this study, with an examination of their targets, data sources and different types of datasets, ML algorithms, and associated outcomes. The results showed that ML approaches and new biomarkers have a lot of promise for being used in clinical decision-making, resulting in a more systematic and informed diagnosis of PD. In this study, some major challenges were addressed along with a future recommendation.

Characterization of Early Stage Parkinson's Disease From Resting-State fMRI Data Using a Long Short-Term Memory Network

Parkinson's disease (PD) is a common and complex neurodegenerative disorder with five stages on the Hoehn and Yahr scaling. Characterizing brain function alterations with progression of early stage disease would support accurate disease staging, development of new therapies, and objective monitoring of disease progression or treatment response. Functional magnetic resonance imaging (fMRI) is a promising tool in revealing functional connectivity (FC) differences and developing biomarkers in PD. While fMRI and FC data have been utilized for diagnosis of PD through application of machine learning approaches such as support vector machine and logistic regression, the characterization of FC changes in early-stage PD has not been investigated. Given the complexity and non-linearity of fMRI data, we propose the use of a long short-term memory (LSTM) network to distinguish the early stages of PD and understand related functional brain changes. The study included 84 subjects (56 in stage 2 and 28 in stage 1) from the Parkinson's Progression Markers Initiative (PPMI), the largest-available public PD dataset. Under a repeated 10-fold stratified cross-validation, the LSTM model reached an accuracy of 71.63%, 13.52% higher than the best traditional machine learning method and 11.56% higher than a CNN model, indicating significantly better robustness and accuracy compared with other machine learning classifiers. Finally, we used the learned LSTM model weights to select the top brain regions that contributed to model prediction and performed FC analyses to characterize functional changes with disease stage and motor impairment to gain better insight into the brain mechanisms of PD.

An Efficient Rotation Forest-Based Ensemble Approach for Predicting Severity of Parkinson’s Disease

Parkinson's disease (PD) is a neurodegenerative nervous system disorder that mainly affects body movement, and it is one of the most common diseases, particularly in elderly individuals. This paper proposes a new machine learning approach to predict Parkinson's disease severity using UCI's Parkinson's telemonitoring voice dataset. The proposed method analyses the patient's voice data and classifies them into “severe” and “nonsevere” classes. At first, a subset of features was selected, then a novel approach with a combination of Rotation Forest and Random Forest was applied on selected features to determine each patient's disease severity. Analysis of the experimental results shows that the proposed approach can detect the severity of PD patients in the early stages. Moreover, the proposed model is compared with several algorithms, and the results indicate that the model is highly successful in classifying records and outperformed the other methods concerning classification accuracy and F1-measure rate.

Artificial intelligence for distinguishment of hammering sound in total hip arthroplasty

Recent studies have focused on hammering sound analysis during insertion of the cementless stem to decrease complications in total hip arthroplasty. However, the nature of the hammering sound is complex to analyse and varies widely owing to numerous possible variables. Therefore, we performed a preliminary feasibility study that aimed to clarify the accuracy of a prediction model using a machine learning algorithm to identify the final rasping hammering sound recorded during surgery. The hammering sound data of 29 primary THA without complication were assessed. The following definitions were adopted. Undersized rasping: all undersized stem rasping before the rasping of the final stem size, Final size rasping: rasping of the final stem size, Positive example: hammering sound during final size rasping, Negative example A: hammering sound during minimum size stem rasping, Negative example B: hammering sound during all undersized rasping. Three datasets for binary classification were set. Finally, binary classification was analysed in six models for the three datasets. The median values of the ROC-AUC in models A–F among each dataset were dataset a: 0.79, 0.76, 0.83, 0.90, 0.91, and 0.90, dataset B: 0.61, 0.53, 0.67, 0.69, 0.71, and 0.72, dataset C: 0.60, 0.48, 0.57, 0.63, 0.67, and 0.63, respectively. Our study demonstrated that artificial intelligence using machine learning was able to distinguish the final rasping hammering sound from the previous hammering sound with a relatively high degree of accuracy. Future studies are warranted to establish a prediction model using hammering sound analysis with machine learning to prevent complications in THA.

Co-evolution of machine learning and digital technologies to improve monitoring of Parkinson's disease motor symptoms

Parkinson’s disease (PD) is a neurodegenerative disorder characterized by motor impairments such as tremor, bradykinesia, dyskinesia, and gait abnormalities. Current protocols assess PD symptoms during clinic visits and can be subjective. Patient diaries can help clinicians evaluate at-home symptoms, but can be incomplete or inaccurate. Therefore, researchers have developed in-home automated methods to monitor PD symptoms to enable data-driven PD diagnosis and management. We queried the US National Library of Medicine PubMed database to analyze the progression of the technologies and computational/machine learning methods used to monitor common motor PD symptoms. A sub-set of roughly 12,000 papers was reviewed that best characterized the machine learning and technology timelines that manifested from reviewing the literature. The technology used to monitor PD motor symptoms has advanced significantly in the past five decades. Early monitoring began with in-lab devices such as needle-based EMG, transitioned to in-lab accelerometers/gyroscopes, then to wearable accelerometers/gyroscopes, and finally to phone and mobile & web application-based in-home monitoring. Significant progress has also been made with respect to the use of machine learning algorithms to classify PD patients. Using data from different devices (e.g., video cameras, phone-based accelerometers), researchers have designed neural network and non-neural network-based machine learning algorithms to categorize PD patients across tremor, gait, bradykinesia, and dyskinesia. The five-decade co-evolution of technology and computational techniques used to monitor PD motor symptoms has driven significant progress that is enabling the shift from in-lab/clinic to in-home monitoring of PD symptoms.

Novel Prenylated Indole Alkaloids with Neuroprotection on SH-SY5Y Cells against Oxidative Stress Targeting Keap1–Nrf2

Oxidative stress has been implicated in the etiology of Parkinson’s disease (PD). Molecules non-covalently binding to the Keap1–Nrf2 complex could be a promising therapeutic approach for PD. Herein, two novel prenylated indole alkaloids asperpenazine (1), and asperpendoline (2) with a scarce skeleton of pyrimido[1,6-a]indole were discovered from the co-cultivated fungi of Aspergillus ochraceus MCCC 3A00521 and Penicillium sp. HUBU 0120. Compound 2 exhibited potential neuroprotective activity on SH-SY5Y cells against oxidative stress. Molecular mechanism research demonstrated that 2 inhibited Keap1 expression, resulting in the translocation of Nrf2 from the cytoplasm to the nucleus, activating the downstream genes expression of HO-1 and NQO1, leading to the reduction in reactive oxygen species (ROS) and the augment of glutathione. Molecular docking and dynamic simulation analyses manifested that 2 interacted with Keap1 (PDB ID: 1X2R) via forming typical hydrogen and hydrophobic bonds with residues and presented less fluctuation of RMSD and RMSF during a natural physiological condition.

Longitudinal clustering analysis and prediction of Parkinson's disease progression using radiomics and hybrid machine learning

BACKGROUND

We employed machine learning approaches to (I) determine distinct progression trajectories in Parkinson's disease (PD) (unsupervised clustering task), and (II) predict progression trajectories (supervised prediction task), from early (years 0 and 1) data, making use of clinical and imaging features.

METHODS

We studied PD-subjects derived from longitudinal datasets (years 0, 1, 2 & 4; Parkinson's Progressive Marker Initiative). We extracted and analyzed 981 features, including motor, non-motor, and radiomics features extracted for each region-of-interest (ROIs: left/right caudate and putamen) using our standardized standardized environment for radiomics analysis (SERA) radiomics software. Segmentation of ROIs on dopamine transposer - single photon emission computed tomography (DAT SPECT) images were performed via magnetic resonance images (MRI). After performing cross-sectional clustering on 885 subjects (original dataset) to identify disease subtypes, we identified optimal longitudinal trajectories using hybrid machine learning systems (HMLS), including principal component analysis (PCA) + K-Means algorithms (KMA) followed by Bayesian information criterion (BIC), Calinski-Harabatz criterion (CHC), and elbow criterion (EC). Subsequently, prediction of the identified trajectories from early year data was performed using multiple HMLSs including 16 Dimension Reduction Algorithms (DRA) and 10 classification algorithms.

RESULTS

We identified 3 distinct progression trajectories. Hotelling's t squared test (HTST) showed that the identified trajectories were distinct. The trajectories included those with (I, II) disease escalation (2 trajectories, 27% and 38% of patients) and (III) stable disease (1 trajectory, 35% of patients). For trajectory prediction from early year data, HMLSs including the stochastic neighbor embedding algorithm (SNEA, as a DRA) as well as locally linear embedding algorithm (LLEA, as a DRA), linked with the new probabilistic neural network classifier (NPNNC, as a classifier), resulted in accuracies of 78.4% and 79.2% respectively, while other HMLSs such as SNEA + Lib_SVM (library for support vector machines) and t_SNE (t-distributed stochastic neighbor embedding) + NPNNC resulted in 76.5% and 76.1% respectively.

CONCLUSIONS

This study moves beyond cross-sectional PD subtyping to clustering of longitudinal disease trajectories. We conclude that combining medical information with SPECT-based radiomics features, and optimal utilization of HMLSs, can identify distinct disease trajectories in PD patients, and enable effective prediction of disease trajectories from early year data.

LncRNA RMST Regulates Neuronal Apoptosis and Inflammatory Response via Sponging miR-150-5p in Parkinson's Disease

INTRODUCTION

LncRNA rhabdomyosarcoma 2-associated transcript (RMST) serves as a key regulator in neural stem cell fate and is involved in the progression of different neurological diseases. In this research, the serum level and clinical value of RMST in Parkinson's disease (PD) patients were detected, and the underlying mechanism was explored.

METHODS

Ninety-nine PD patients and 93 healthy individuals were collected for clinical experiments. SH-SY5Y cells were treated with the dopaminergic neurotoxin 1-methyl-4-phenylpyridinium (MPP+) to establish PD cell models. qRT-PCR was used for the detection of mRNA levels. CCK-8 and flow cytometry were used to detect neuronal viability and apoptosis. The target relationship of RMST with miR-15a-5p was confirmed applying luciferase reporter assay.

RESULTS

RMST was present at high levels in both serum of PD patients and PD cell models. Serum RMST had a certain clinical value for the diagnosis of PD with the AUC of 0.892 at a cutoff value of 1.225. Serum RMST was positively associated with the levels of TNF-α (r = 0.421, p < 0.001) and IL-1β (r = 0.567, p < 0.001) in PD patients. Knockdown of RMST alleviated the apoptosis and inflammatory response of SH-SY5Y cells induced by MPP+. miR-150-5p was the target gene of RMST and less expressed in the clinical serum samples and PD cell models.

CONCLUSION

Serum RMST serves as a promising biomarker for the diagnosis of PD. RMST downregulation may regulate the occurrence and development of PD through inhibiting neuron cell apoptosis and the release of inflammatory cytokines via targeting miR-150-5p.

Classification of Parkinson's disease with freezing of gait based on 360° turning analysis using 36 kinematic features

BACKGROUND

Freezing of gait (FOG) is a sensitive problem, which is caused by motor control deficits and requires greater attention during postural transitions such as turning in people with Parkinson's disease (PD). However, the turning characteristics have not yet been extensively investigated to distinguish between people with PD with and without FOG (freezers and non-freezers) based on full-body kinematic analysis during the turning task. The objectives of this study were to identify the machine learning model that best classifies people with PD and freezers and reveal the associations between clinical characteristics and turning features based on feature selection through stepwise regression.

METHODS

The study recruited 77 people with PD (31 freezers and 46 non-freezers) and 34 age-matched older adults. The 360° turning task was performed at the preferred speed for the inner step of the more affected limb. All experiments on the people with PD were performed in the "Off" state of medication. The full-body kinematic features during the turning task were extracted using the three-dimensional motion capture system. These features were selected via stepwise regression.

RESULTS

In feature selection through stepwise regression, five and six features were identified to distinguish between people with PD and controls and between freezers and non-freezers (PD and FOG classification problem), respectively. The machine learning model accuracies revealed that the random forest (RF) model had 98.1% accuracy when using all turning features and 98.0% accuracy when using the five features selected for PD classification. In addition, RF and logistic regression showed accuracies of 79.4% when using all turning features and 72.9% when using the six selected features for FOG classification.

CONCLUSION

We suggest that our study leads to understanding of the turning characteristics of people with PD and freezers during the 360° turning task for the inner step of the more affected limb and may help improve the objective classification and clinical assessment by disease progression using turning features.

Machine learning trained with quantitative susceptibility mapping to detect mild cognitive impairment in Parkinson's disease

BACKGROUND

Cognitive decline is commonly observed in Parkinson's disease (PD). Identifying PD with mild cognitive impairment (PD-MCI) is crucial for early initiation of therapeutic interventions and preventing cognitive decline.

OBJECTIVE

We aimed to develop a machine learning model trained with magnetic susceptibility values based on the multi-atlas label-fusion method to classify PD without dementia into PD-MCI and normal cognition (PD-CN).

METHODS

This multicenter observational cohort study retrospectively reviewed 61 PD-MCI and 59 PD-CN cases for the internal validation cohort and 22 PD-MCI and 21 PD-CN cases for the external validation cohort. The multi-atlas method parcellated the quantitative susceptibility mapping (QSM) images into 20 regions of interest and extracted QSM-based magnetic susceptibility values. Random forest, extreme gradient boosting, and light gradient boosting were selected as machine learning algorithms.

RESULTS

All classifiers demonstrated substantial performances in the classification task, particularly the random forest model. The accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve for this model were 79.1%, 77.3%, 81.0%, and 0.78, respectively. The QSM values in the caudate nucleus, which were important features, were inversely correlated with the Montreal Cognitive Assessment scores (right caudate nucleus: r = -0.573, 95% CI: -0.801 to -0.298, p = 0.003; left caudate nucleus: r = -0.659, 95% CI: -0.894 to -0.392, p < 0.001).

CONCLUSIONS

Machine learning models trained with QSM values successfully classified PD without dementia into PD-MCI and PD-CN groups, suggesting the potential of QSM values as an auxiliary biomarker for early evaluation of cognitive decline in patients with PD.

Modeling and prediction of pressure injury in hospitalized patients using artificial intelligence

BACKGROUND

Hospital-acquired pressure injuries (PIs) induce significant patient suffering, inflate healthcare costs, and increase clinical co-morbidities. PIs are mostly due to bed-immobility, sensory impairment, bed positioning, and length of hospital stay. In this study, we use electronic health records and administrative data to examine the contributing factors to PI development using artificial intelligence (AI).

METHODS

We used advanced data science techniques to first preprocess the data and then train machine learning classifiers to predict the probability of developing PIs. The AI training was based on large, incongruent, incomplete, heterogeneous, and time-varying data of hospitalized patients. Both model-based statistical methods and model-free AI strategies were used to forecast PI outcomes and determine the salient features that are highly predictive of the outcomes.

RESULTS

Our findings reveal that PI prediction by model-free techniques outperform model-based forecasts. The performance of all AI methods is improved by rebalancing the training data and by including the Braden in the model learning phase. Compared to neural networks and linear modeling, with and without rebalancing or using Braden scores, Random forest consistently generated the optimal PI forecasts.

CONCLUSIONS

AI techniques show promise to automatically identify patients at risk for hospital acquired PIs in different surgical services. Our PI prediction model provide a first generation of AI guidance to prescreen patients at risk for developing PIs.

CLINICAL IMPACT

This study provides a foundation for designing, implementing, and assessing novel interventions addressing specific healthcare needs. Specifically, this approach allows examining the impact of various dynamic, personalized, and clinical-environment effects on PI prevention for hospital patients receiving care from various surgical services.

XGBoost based machine learning approach to predict the risk of fall in older adults using gait outcomes

This study aimed to identify the optimal features of gait parameters to predict the fall risk level in older adults. The study included 746 older adults (age: 63–89 years). Gait tests (20 m walkway) included speed modification (slower, preferred, and faster-walking) while wearing the inertial measurement unit sensors embedded in the shoe-type data loggers on both outsoles. A metric was defined to classify the fall risks, determined based on a set of questions determining the history of falls and fear of falls. The extreme gradient boosting (XGBoost) model was built from gait features to predict the factor affecting the risk of falls. Moreover, the definition of the fall levels was classified into high- and low-risk groups. At all speeds, three gait features were identified with the XGBoost (stride length, walking speed, and stance phase) that accurately classified the fall risk levels. The model accuracy in classifying fall risk levels ranged between 67–70% with 43–53% sensitivity and 77–84% specificity. Thus, we identified the optimal gait features for accurate fall risk level classification in older adults. The XGBoost model could inspire future works on fall prevention and the fall-risk assessment potential through the gait analysis of older adults.

Decision Support Systems in Temporomandibular Joint Osteoarthritis: A review of Data Science and Artificial Intelligence Applications

With the exponential growth of computational systems and increased patient data acquisition, dental research faces new challenges to manage a large quantity of information. For this reason, data science approaches are needed for the integrative diagnosis of multifactorial diseases, such as Temporomandibular joint (TMJ) Osteoarthritis (OA). The Data science spectrum includes data capture/acquisition, data processing with optimized web-based storage and management, data analytics involving in-depth statistical analysis, machine learning (ML) approaches, and data communication. Artificial intelligence (AI) plays a crucial role in this process. It consists of developing computational systems that can perform human intelligence tasks, such as disease diagnosis, using many features to help in the decision-making support. Patient's clinical parameters, imaging exams, and molecular data are used as the input in cross-validation tasks, and human annotation/diagnosis is also used as the gold standard to train computational learning models and automatic disease classifiers. This paper aims to review and describe AI and ML techniques to diagnose TMJ OA and data science approaches for imaging processing. We used a web-based system for multi-center data communication, algorithms integration, statistics deployment, and process the computational machine learning models. We successfully show AI and data-science applications using patients' data to improve the TMJ OA diagnosis decision-making towards personalized medicine.

Moving towards accurate and early prediction of language delay with network science and machine learning approaches

Due to wide variability of typical language development, it has been historically difficult to distinguish typical and delayed trajectories of early language growth. Improving our understanding of factors that signal language disorder and delay has the potential to improve the lives of the millions with developmental language disorder (DLD). We develop predictive models of low language (LL) outcomes by analyzing parental report measures of early language skill using machine learning and network science approaches. We harmonized two longitudinal datasets including demographic and standardized measures of early language skills (the MacArthur-Bates Communicative Developmental Inventories; MBCDI) as well as a later measure of LL. MBCDI data was used to calculate several graph-theoretic measures of lexico-semantic structure in toddlers' expressive vocabularies. We use machine-learning techniques to construct predictive models with these datasets to identify toddlers who will have later LL outcomes at preschool and school-age. This approach yielded robust and reliable predictions of later LL outcome with classification accuracies in single datasets exceeding 90%. Generalization performance between different datasets was modest due to differences in outcome ages and diagnostic measures. Grammatical and lexico-semantic measures ranked highly in predictive classification, highlighting promising avenues for early screening and delineating the roots of language disorders.

Data-driven biomarker analysis using computational omics approaches to assess neurodegenerative disease progression

The complexity of biological systems suggests that current definitions of molecular dysfunctions are essential distinctions of a complex phenotype. This is well seen in neurodegenerative diseases (ND), such as Alzheimer's disease (AD) and Parkinson's disease (PD), multi-factorial pathologies characterized by high heterogeneity. These challenges make it necessary to understand the effectiveness of candidate biomarkers for early diagnosis, as well as to obtain a comprehensive mapping of how selective treatment alters the progression of the disorder. A large number of computational methods have been developed to explain network-based approaches by integrating individual components for modeling a complex system. In this review, high-throughput omics methodologies are presented for the identification of potent biomarkers associated with AD and PD pathogenesis as well as for monitoring the response of dysfunctional molecular pathways incorporating multilevel clinical information. In addition, principles for efficient data analysis pipelines are being discussed that can help address current limitations during the experimental process by increasing the reproducibility of benchmarking studies.

An artificial neural network approach to detect presence and severity of Parkinson's disease via gait parameters

Introduction

Gait deficits are debilitating in people with Parkinson’s disease (PwPD), which inevitably deteriorate over time. Gait analysis is a valuable method to assess disease-specific gait patterns and their relationship with the clinical features and progression of the disease.

Objectives

Our study aimed to i) develop an automated diagnostic algorithm based on machine-learning techniques (artificial neural networks [ANNs]) to classify the gait deficits of PwPD according to disease progression in the Hoehn and Yahr (H-Y) staging system, and ii) identify a minimum set of gait classifiers.

Methods

We evaluated 76 PwPD (H-Y stage 1–4) and 67 healthy controls (HCs) by computerized gait analysis. We computed the time-distance parameters and the ranges of angular motion (RoMs) of the hip, knee, ankle, trunk, and pelvis. Principal component analysis was used to define a subset of features including all gait variables. An ANN approach was used to identify gait deficits according to the H-Y stage.

Results

We identified a combination of a small number of features that distinguished PwPDs from HCs (one combination of two features: knee and trunk rotation RoMs) and identified the gait patterns between different H-Y stages (two combinations of four features: walking speed and hip, knee, and ankle RoMs; walking speed and hip, knee, and trunk rotation RoMs).

Conclusion

The ANN approach enabled automated diagnosis of gait deficits in several symptomatic stages of Parkinson’s disease. These results will inspire future studies to test the utility of gait classifiers for the evaluation of treatments that could modify disease progression.

Lupus or not? SLE Risk Probability Index (SLERPI): a simple, clinician-friendly machine learning-based model to assist the diagnosis of systemic lupus erythematosus

OBJECTIVES

Diagnostic reasoning in systemic lupus erythematosus (SLE) is a complex process reflecting the probability of disease at a given timepoint against competing diagnoses. We applied machine learning in well-characterised patient data sets to develop an algorithm that can aid SLE diagnosis.

METHODS

From a discovery cohort of randomly selected 802 adults with SLE or control rheumatologic diseases, clinically selected panels of deconvoluted classification criteria and non-criteria features were analysed. Feature selection and model construction were done with Random Forests and Least Absolute Shrinkage and Selection Operator-logistic regression (LASSO-LR). The best model in 10-fold cross-validation was tested in a validation cohort (512 SLE, 143 disease controls).

RESULTS

A novel LASSO-LR model had the best performance and included 14 variably weighed features with thrombocytopenia/haemolytic anaemia, malar/maculopapular rash, proteinuria, low C3 and C4, antinuclear antibodies (ANA) and immunologic disorder being the strongest SLE predictors. Our model produced SLE risk probabilities (depending on the combination of features) correlating positively with disease severity and organ damage, and allowing the unbiased classification of a validation cohort into diagnostic certainty levels (unlikely, possible, likely, definitive SLE) based on the likelihood of SLE against other diagnoses. Operating the model as binary (lupus/not-lupus), we noted excellent accuracy (94.8%) for identifying SLE, and high sensitivity for early disease (93.8%), nephritis (97.9%), neuropsychiatric (91.8%) and severe lupus requiring immunosuppressives/biologics (96.4%). This was converted into a scoring system, whereby a score >7 has 94.2% accuracy.

CONCLUSIONS

We have developed and validated an accurate, clinician-friendly algorithm based on classical disease features for early SLE diagnosis and treatment to improve patient outcomes.

Machine-learning-based knowledge discovery in rheumatoid arthritis-related registry data to identify predictors of persistent pain

Early detection of patients with chronic diseases at risk of developing persistent pain is clinically desirable for timely initiation of multimodal therapies. Quality follow-up registries may provide the necessary clinical data; however, their design is not focused on a specific research aim, which poses challenges on the data analysis strategy. Here, machine-learning was used to identify early parameters that provide information about a future development of persistent pain in rheumatoid arthritis (RA). Data of 288 patients were queried from a registry based on the Swedish Epidemiological Investigation of RA. Unsupervised data analyses identified the following 3 distinct patient subgroups: low-, median-, and high-persistent pain intensity. Next, supervised machine-learning, implemented as random forests followed by computed ABC analysis-based item categorization, was used to select predictive parameters among 21 different demographic, patient-rated, and objective clinical factors. The selected parameters were used to train machine-learned algorithms to assign patients pain-related subgroups (1000 random resamplings, 2/3 training, and 1/3 test data). Algorithms trained with 3-month data of the patient global assessment and health assessment questionnaire provided pain group assignment at a balanced accuracy of 70%. When restricting the predictors to objective clinical parameters of disease severity, swollen joint count and tender joint count acquired at 3 months provided a balanced accuracy of RA of 59%. Results indicate that machine-learning is suited to extract knowledge from data queried from pain- and disease-related registries. Early functional parameters of RA are informative for the development and degree of persistent pain.

Gait Analysis with Wearables Can Accurately Classify Fallers from Non-Fallers: A Step toward Better Management of Neurological Disorders

Falls are the leading cause of mortality, morbidity and poor quality of life in older adults with or without neurological conditions. Applying machine learning (ML) models to gait analysis outcomes offers the opportunity to identify individuals at risk of future falls. The aim of this study was to determine the effect of different data pre-processing methods on the performance of ML models to classify neurological patients who have fallen from those who have not for future fall risk assessment. Gait was assessed using wearables in clinic while walking 20 m at a self-selected comfortable pace in 349 (159 fallers, 190 non-fallers) neurological patients. Six different ML models were trained on data pre-processed with three techniques such as standardisation, principal component analysis (PCA) and path signature method. Fallers walked more slowly, with shorter strides and longer stride duration compared to non-fallers. Overall, model accuracy ranged between 48% and 98% with 43-99% sensitivity and 48-98% specificity. A random forest (RF) classifier trained on data pre-processed with the path signature method gave optimal classification accuracy of 98% with 99% sensitivity and 98% specificity. Data pre-processing directly influences the accuracy of ML models for the accurate classification of fallers. Using gait analysis with trained ML models can act as a tool for the proactive assessment of fall risk and support clinical decision-making.

Using machine learning-based analytics of daily activities to identify modifiable risk factors for falling in Parkinson's disease

BACKGROUND

Although risk factors that lead to falling in Parkinson's disease (PD) have been previously studied, the established predictors are mostly non-modifiable. A novel method for fall risk assessment may provide more insight into preventable high-risk activities to reduce future falls.

OBJECTIVES

To explore the prediction of falling in PD patients using a machine learning-based approach.

METHOD

305 PD patients, with or without a history of falls within the past month, were recruited. Data including clinical demographics, medications, and balance confidence, scaled by the 16-item Activities-Specific Balance Confidence Scale (ABC-16), were entered into the supervised machine learning models using XGBoost to explore the prediction of fallers/recurrent fallers in two separate models.

RESULTS

99 (32%) patients were fallers and 58 (19%) were recurrent fallers. The accuracy of the model to predict falls was 72% (p = 0.001). The most important factors were item 7 (sweeping the floor), item 5 (reaching on tiptoes), and item 12 (walking in a crowded mall) in the ABC-16 scale, followed by disease stage and duration. When recurrent falls were analysed, the models had higher accuracy (81%, p = 0.02). The strongest predictors of recurrent falls were item 12, 5, and 10 (walking across parking lot), followed by disease stage and current age.

CONCLUSION

Our machine learning-based study demonstrated that predictors of falling combined demographics of PD with environmental factors, including high-risk activities that require cognitive attention and changes in vertical and lateral orientations. This enables physicians to focus on modifiable factors and appropriately implement fall prevention strategies for individual patients.

Locomotion in virtual environments predicts cardiovascular responsiveness to subsequent stressful challenges

Individuals differ in their physiological responsiveness to stressful challenges, and stress potentiates the development of many diseases. Heart rate variability (HRV), a measure of cardiac vagal break, is emerging as a strong index of physiological stress vulnerability. Thus, it is important to develop tools that identify predictive markers of individual differences in HRV responsiveness without exposing subjects to high stress. Here, using machine learning approaches, we show the strong predictive power of high-dimensional locomotor responses during novelty exploration to predict HRV responsiveness during stress exposure. Locomotor responses are collected in two ecologically valid virtual reality scenarios inspired by the animal literature and stress is elicited and measured in a third threatening virtual scenario. Our model's predictions generalize to other stressful challenges and outperforms other stress prediction instruments, such as anxiety questionnaires. Our study paves the way for the development of behavioral digital phenotyping tools for early detection of stress-vulnerable individuals.

Data-Driven Models for Objective Grading Improvement of Parkinson's Disease

Parkinson's disease (PD) is a progressive disorder of the central nervous system that causes motor dysfunctions in affected patients. Objective assessment of symptoms can support neurologists in fine evaluations, improving patients' quality of care. Herein, this study aimed to develop data-driven models based on regression algorithms to investigate the potential of kinematic features to predict PD severity levels. Sixty-four patients with PD (PwPD) and 50 healthy subjects of control (HC) were asked to perform 13 motor tasks from the MDS-UPDRS III while wearing wearable inertial sensors. Simultaneously, the clinician provided the evaluation of the tasks based on the MDS-UPDRS scores. One hundred-ninety kinematic features were extracted from the inertial motor data. Data processing and statistical analysis identified a set of parameters able to distinguish between HC and PwPD. Then, multiple feature selection methods allowed selecting the best subset of parameters for obtaining the greatest accuracy when used as input for several predicting regression algorithms. The maximum correlation coefficient, equal to 0.814, was obtained with the adaptive neuro-fuzzy inference system (ANFIS). Therefore, this predictive model could be useful as a decision support system for a reliable objective assessment of PD severity levels based on motion performance, improving patients monitoring over time.

A comparative study of machine learning algorithms for predicting acute kidney injury after liver cancer resection

Objective

Machine learning methods may have better or comparable predictive ability than traditional analysis. We explore machine learning methods to predict the likelihood of acute kidney injury after liver cancer resection.

Methods

This is a secondary analysis cohort study. We reviewed data from patients who had undergone resection of primary hepatocellular carcinoma between January 2008 and October 2015.

Results

The analysis included 1,173 hepatectomy patients, 77 (6.6%) of whom had AKI and 1,096 (93.4%) who did not. The importance matrix for the Gbdt algorithm model shows that age, cholesterol, tumor size, surgery duration and PLT were the five most important parameters. Figure 1 shows that Age, tumor size and surgery duration had weak positive correlations with AKI. Cholesterol and PLT also had weak negative correlations with AKI. The models constructed by the four machine learning algorithms in the training group were compared. Among the four machine learning algorithms, random forest and gbm had the highest accuracy, 0.989 and 0.970 respectively. The precision of four of the five algorithms was 1, random forest being the exception. Among the test group, gbm had the highest accuracy (0.932). Random forest and gbm had the highest precision, both being 0.333. The AUC values for the four algorithms were: Gbdt (0.772), gbm (0.725), forest (0.662) and DecisionTree (0.628).

Conclusions

Machine learning technology can predict acute kidney injury after hepatectomy. Age, cholesterol, tumor size, surgery duration and PLT influence the likelihood and development of postoperative acute kidney injury.

Predicting Hepatitis B Virus Infection Based on Health Examination Data of Community Population

Despite a decline in the prevalence of hepatitis B in China, the disease burden remains high. Large populations unaware of infection risk often fail to meet the ideal treatment window, resulting in poor prognosis. The purpose of this study was to develop and evaluate models identifying high-risk populations who should be tested for hepatitis B surface antigen. Data came from a large community-based health screening, including 97,173 individuals, with an average age of 54.94. A total of 33 indicators were collected as model predictors, including demographic characteristics, routine blood indicators, and liver function. Borderline-Synthetic minority oversampling technique (SMOTE) was conducted to preprocess the data and then four predictive models, namely, the extreme gradient boosting (XGBoost), random forest (RF), decision tree (DT), and logistic regression (LR) algorithms, were developed. The positive rate of hepatitis B surface antigen (HBsAg) was 8.27%. The area under the receiver operating characteristic curves for XGBoost, RF, DT, and LR models were 0.779, 0.752, 0.619, and 0.742, respectively. The Borderline-SMOTE XGBoost combined model outperformed the other models, which correctly predicted 13,637/19,435 cases (sensitivity 70.8%, specificity 70.1%), and the variable importance plot of XGBoost model indicated that age was of high importance. The prediction model can be used to accurately identify populations at high risk of hepatitis B infection that should adopt timely appropriate medical treatment measures.

Selecting Clinically Relevant Gait Characteristics for Classification of Early Parkinson's Disease: A Comprehensive Machine Learning Approach

Parkinson's disease (PD) is the second most common neurodegenerative disease; gait impairments are typical and are associated with increased fall risk and poor quality of life. Gait is potentially a useful biomarker to help discriminate PD at an early stage, however the optimal characteristics and combination are unclear. In this study, we used machine learning (ML) techniques to determine the optimal combination of gait characteristics to discriminate people with PD and healthy controls (HC). 303 participants (119 PD, 184 HC) walked continuously around a circuit for 2-minutes at a self-paced walk. Gait was quantified using an instrumented mat (GAITRite) from which 16 gait characteristics were derived and assessed. Gait characteristics were selected using different ML approaches to determine the optimal method (random forest with information gain and recursive features elimination (RFE) technique with support vector machine (SVM) and logistic regression). Five clinical gait characteristics were identified with RFE-SVM (mean step velocity, mean step length, step length variability, mean step width, and step width variability) that accurately classified PD. Model accuracy for classification of early PD ranged between 73-97% with 63-100% sensitivity and 79-94% specificity. In conclusion, we identified a subset of gait characteristics for accurate early classification of PD. These findings pave the way for a better understanding of the utility of ML techniques to support informed clinical decision-making.

Using gait analysis' parameters to classify Parkinsonism: A data mining approach

INTRODUCTION

Parkinson's disease (PD) is the second most common neurodegenerative disorder in the world, while Progressive Supranuclear Palsy (PSP) is an atypical Parkinsonism resembling PD, especially in early stage. Assumed that gait dysfunctions represent a major motor symptom for both pathologies, gait analysis can provide clinicians with subclinical information reflecting subtle differences between these diseases. In this scenario, data mining can be exploited in order to differentiate PD patients at different stages of the disease course and PSP using all the variables acquired through gait analysis.

METHODS

A cohort of 46 subjects (divided into three groups) affected by PD patients at different stages and PSP patients was acquired through gait analysis and spatial and temporal parameters were analysed. Synthetic Minority Over-sampling Technique was used to balance our imbalanced dataset and cross-validation was applied to provide different training and testing sets. Then, Random Forests and Gradient Boosted Trees were implemented.

RESULTS

Accuracy, error, precision, recall, specificity and sensitivity were computed for each group and for both algorithms, including 16 features. Random Forests obtained the highest accuracy (86.4%) but also specificity and sensitivity were particularly high, overcoming the 90% for PSP group.

CONCLUSION

The novelty of the study is the use of a data mining approach on the spatial and temporal parameters of gait analysis in order to classify patients affected by typical (PD) and atypical Parkinsonism (PSP) based on gait patterns. This application would be helpful for clinicians to distinguish PSP from PD at early stage, when the differential diagnosis is particularly challenging.