Publications by authors named "K Sakatani"
Article Synopsis
- * A deep learning model was developed using data from 2,897 cases to predict cognitive function and brain atrophy based on age and basic blood tests, with key predictors being age, nutrition, and organ function.
- * The study supports using routine blood tests for assessing dementia risk and suggests personalized dietary interventions, aligning with traditional Chinese medicine's holistic approach to brain health.
Article Synopsis
- - The study explores a deep learning method designed to predict dementia risk using blood test data, particularly examining the role of nutritional status indicators, like serum albumin, in older adults.
- - Researchers analyzed data from 1,287 patients, separating them by age (65 and above vs. below 65) to assess how including serum albumin influenced the model's accuracy in predicting cognitive function measured by the Mini-Mental State Examination.
- - Results showed that including serum albumin improved prediction accuracy slightly for both age groups, with a more significant enhancement in those under 65, indicating the potential importance of nutritional status in dementia risk assessment.
Article Synopsis
- - This study examines how data augmentation using generative adversarial networks (GANs) can enhance the prediction of dementia risk through deep neural networks (DNNs), specifically focusing on blood test and periodontal examination data.
- - Challenges in creating accurate models include high costs, limited sample sizes, and missing data from various tests, which can hinder effective dementia risk predictions.
- - The results showed that DNNs using GAN-synthesised data outperformed those using real data, with improved accuracy and robustness against missing information, indicating a promising avenue for better dementia risk prediction methods.
The purpose of this study was to examine the effects on prefrontal cortex (PFC) activity of listening to pleasant sounds (PS) while walking, gum chewing (GCh), or performing the dual task of walking and gum chewing at the same time (walking + GCh). A total of 11 healthy adult male volunteers participated in the study (mean age: 29.54 ± 3.
View Article and Find Full Text PDFArticle Synopsis
- The study examines the use of data augmentation techniques to enhance the prediction of dementia risk through machine learning models, specifically focusing on easily obtainable blood test results and periodontal examination data.
- It employs advanced techniques like SMOGN, GAN, and CTGAN on datasets comprising cognitive assessments, demographics, and test results, utilizing various machine learning models for evaluation.
- Findings indicate that while normal GANs improve data diversity, CTGANs better preserve data structure, significantly enhancing the performance of linear regression models when applied to synthesized data.