Freezing of gait (FoG) is a prevalent symptom among individuals with Parkinson’s disease and related disorders. FoG detection from videos has been developed recently; however, the process requires using videos filmed within a controlled environment. We attempted to establish an automatic FoG detection method from videos taken in uncontrolled environments such as in daily clinical practices. Motion features of 16 patients were extracted from timed-up-and-go test in 109 video data points, through object tracking and three-dimension pose estimation. These motion features were utilized to form the FoG detection model, which combined rule-based and machine learning-based models. The rule-based model distinguished the frames in which the patient was walking from those when the patient has stopped, using the pelvic position coordinates; the machine learning-based model distinguished between FoG and stop using a combined one-dimensional convolutional neural network and long short-term memory (1dCNN-LSTM). The model achieved a high intraclass correlation coefficient of 0.75–0.94 with a manually-annotated duration of FoG and %FoG. This method is novel as it combines object tracking, 3D pose estimation, and expert-guided feature selection in the preprocessing and modeling phases, enabling FoG detection even from videos captured in uncontrolled environments.

The risk of pneumonia and death is higher in acute stroke patients with signs of pulmonary infection on chest computed tomography (CT) at admission. However, few reports have examined the incidence of pneumonia and its predictors in subacute stroke patients. The aim of this study was to examine factors related to post-stroke pneumonia in subacute stroke patients. A total of 340 subacute stroke patients were included. Univariable logistic regression analysis was performed using variables that may contribute to pneumonia, with the development of pneumonia as the dependent variable. Multivariable logistic regression analysis using the three independent variables with the lowest p-values on the univariable logistic regression analysis was also performed to calculate adjusted odds ratios. Twenty-two patients developed pneumonia during hospitalization. The univariable logistic regression analysis showed that the top three items were serum albumin (Alb), functional Oral Intake Scale (FOIS) score, and signs of pulmonary infection on chest CT at admission. Multivariable logistic regression analysis adjusted for these three items showed that the presence of signs of pulmonary infection on chest CT at admission was the independent variable (OR: 4.45; 95% CI: 1.54–12.9). When signs of pulmonary infection are seen on admission chest CT, careful follow-up is necessary because pneumonia is significantly more likely to occur during hospitalization.

Objectives: Stepwise linear regression (SLR) is the most common approach to predicting activities of daily living at discharge with the Functional Independence Measure (FIM) in stroke patients, but noisy nonlinear clinical data decrease the predictive accuracies of SLR. Machine learning is gaining attention in the medical field for such nonlinear data. Previous studies reported that machine learning models, regression tree (RT), ensemble learning (EL), artificial neural networks (ANNs), support vector regression (SVR), and Gaussian process regression (GPR), are robust to such data and increase predictive accuracies. This study aimed to compare the predictive accuracies of SLR and these machine learning models for FIM scores in stroke patients. Methods: Subacute stroke patients (N = 1,046) who underwent inpatient rehabilitation participated in this study. Only patients’ background characteristics and FIM scores at admission were used to build each predictive model of SLR, RT, EL, ANN, SVR, and GPR with 10-fold cross-validation. The coefficient of determination (R²) and root mean square error (RMSE) values were compared between the actual and predicted discharge FIM scores and FIM gain. Results: Machine learning models (R² of RT = 0.75, EL = 0.78, ANN = 0.81, SVR = 0.80, GPR = 0.81) outperformed SLR (0.70) to predict discharge FIM motor scores. The predictive accuracies of machine learning methods for FIM total gain (R² of RT = 0.48, EL = 0.51, ANN = 0.50, SVR = 0.51, GPR = 0.54) were also better than of SLR (0.22). Conclusions: This study suggested that the machine learning models outperformed SLR for predicting FIM prognosis. The machine learning models used only patients’ background characteristics and FIM scores at admission and more accurately predicted FIM gain than previous studies. ANN, SVR, and GPR outperformed RT and EL. GPR could have the best predictive accuracy for FIM prognosis.

In 2020, COVID-19 spread throughout the world, and international measures such as travel bans, quarantines, and increased social distancing were implemented. In Japan, the number of infected people increased, and a state of emergency was declared from 16 April to 25 May 2020. Such a change in physical activity could lead to a decline in physical function in people with disabilities. A retrospective study was conducted to determine the impact of the pandemic on the physical function of disabled persons living in the community. Data were collected at four points in time: two points before the declaration of the state of emergency was issued and two points after the declaration period had ended. Time series data of physical function at four points in time were compared for 241 people with disabilities. The mean age was 72.39 years; 157 had stroke, 59 musculoskeletal disease, and 26 other diseases. Overall, there was a long-term decrease in walking speed (p < 0.001) and a worsening of the Timed Up-and-Go (TUG) score (p < 0.001) after the period of the state of emergency. The TUG score worsened only in the group with a walking speed of 1.0 m/s or less before the state of emergency (p = 0.064), suggesting that this group was more susceptible.