Recent progress and technical advances in catheter ablation have dramatically improved the success rate and safety of pulmonary vein isolation (PVI) for atrial fibrillation (AF). AF recurrences occur predominantly because of reconnections in previously isolated pulmonary veins. The aim of this study is to analyze an acute pulmonary vein reconnection (PVR) using graph convolutional networks (GCN). The subjects of the patient population survey were 30 patients (20 males and 10 females), who underwent ablation for AF from May 2019 to March 2020. The average age of patients is 65.3 years old (from 53 to 89 years old). The targeted atrial fibrillation is paroxysmal atrial fibrillation (PAF), persistent atrial fibrillation (PeAF), and chronic atrial fibrillation (CAF). The target samples included 11 cases of PVR (spontaneous PV-LA reconnection or dormant conduction (DC)) and 19 cases of non-PVR. The feature parameters analyzed in this study are as follows: Location X, Y, Z, Duration time (DT), Average force (A-Force), Max temperature (MT), Max power (MP), Base impedance (BI), Impedance drop (ΔImp), Force time integral (FTI), Ablation index (AI), Unipolar (Uni), Bipolar (Bi). We conducted a comparative study of predicting the AF recurrences using GCN and other popular machine learning methods. As a result, we found that GCN has better prediction accuracy than the other methods. Its prediction accuracy still needs to be increased. However, in this field, our study is the first trial of integrating the location of the cauterization points and the various indices with GCN. It is expected that GCN will be very useful in predicting the AF recurrences during radiofrequency catheter ablation (RFCA).

抄録全体を表示

This paper presents a novel human-computer interaction system that has the potential to be used for autonomous or safety assisted driving in future vehicles. Convolutional neural networks were built and used for single shot detection and blink detection. The single shot detection method has been used to accomplish the detection of dynamic targets. The blink detection is performed by feeding multiple images of open and closed eyes into the network for deep learning, and based on the learned data models can detect the open or closed state of the subject’s eyes. In addition, eye tracking technology is used to identify the direction of the driver’s gaze. The human-computer interaction system is empirically validated in a super-compact electric vehicle, and it can accurately detect external dynamic targets, while the driver can control the vehicle by blinking and gaze direction.

抄録全体を表示

The number of elderly people is increasing day by day. Elderly people are physically very weak. Most of them have to consult doctors regularly. Wireless Sensor Network is designed to keep incessant observations on many elderly people at a time. This WSN is comprised of many sensor nodes. For the recognition of elderly activities, accelerometer sensors are broadly used in WSN since they are available on almost every phone. Recognition of elderly activities is being widely researched and it has great importance in the medical field. An elderly is very likely to be a victim of a sudden accident. It is an urgent need to detect that accident so that immediate treatment can be ensured. In this paper, we have mentioned some important features in the time, frequency, and time-frequency domain which can be used as a combination to recognize activities from accelerometer data more accurately. We have extracted those efficient features from the HASC dataset to perform activity recognition by exploiting Support Vector Machine (SVM). We have obtained 97.05% accuracy on HASC dataset. Considering the challenges involved in recognizing human activities from accelerometer data, the result is quite satisfactory.

抄録全体を表示

The purpose of this pilot cross-sectional study was to investigate the relationship between depressed mood, based on the Kihon Checklist (KCL), and higher-level functional capacity in community-dwelling older adults. We analyzed 112 older adults (age 78.2 ± 7.3 years, 91 women, 81.3%) recruited from local senior centers and through local newspaper advertisements in Ibaraki City. Depressed mood was assessed using the depressed domain of the KCL. The Japan Science and Technology Agency Index of Competence (JST-IC) was used to assess higher-level functional capacity. Age, sex, cognitive function, current medical history, number of medications taken, certification of needed support, grip strength, and walking speed were assessed. The two groups were compared based on the presence or absence of depressed mood. Factors associated with depressed mood were analyzed using age- and sex-adjusted logistic regression analyses. Thirty-three (29.5%) participants were found to have depressed mood. The median JST-IC was 10.0. Participants with depressed mood were significantly older and had lower JST-IC (total and daily life management), cognitive function, and physical function and more medication than those without. Logistic regression analysis showed no significant factors associated with depressed mood. Further investigation is needed in the future using simple depressive mood screening from the KCL.

抄録全体を表示

Our aim is to develop high accurate Extracorporeal Shock Wave Lithotripsy (ESWL) outcome prediction by introducing complex machine learning models. To obtain interpretation of these complex models, factor importance evaluation by SHapley Additive exPlanations (SHAP) value was performed. In this retrospective study, the data was collected from 214 subjects, where single session ESWL outcome was defined by the residue of stone fragments smaller than 4 mm. Outcome prediction by three machine learning algorithms was performed, and accuracy verification and factor evaluation by SHAP were executed. As a result, we obtained superior accuracy for the two types of ensemble tree models compared to the single decision tree, and clear model interpretations by SHAP. These results may encourage the use of complex models with interpretability difficulties and allow more accurate ESWL outcome prediction.

抄録全体を表示