Background: Data collected by an actigraphy device worn on the wrist or waist can provide objective measurements for studies related to physical activity; however, some data may contain intervals where values are missing. In previous studies, statistical methods have been applied to impute missing values on the basis of statistical assumptions. Deep learning algorithms, however, can learn features from the data without any such assumptions and may outperform previous approaches in imputation tasks. Objective: The aim of this study was to impute missing values in data using a deep learning approach. Methods: To develop an imputation model for missing values in accelerometer-based actigraphy data, a denoising convolutional autoencoder was adopted. We trained and tested our deep learning–based imputation model with the National Health and Nutrition Examination Survey data set and validated it with the external Korea National Health and Nutrition Examination Survey and the Korean Chronic Cerebrovascular Disease Oriented Biobank data sets which consist of daily records measuring activity counts. The partial root mean square error and partial mean absolute error of the imputed intervals (partial RMSE and partial MAE, respectively) were calculated using our deep learning–based imputation model (zero-inflated denoising convolutional autoencoder) as well as using other approaches (mean imputation, zero-inflated Poisson regression, and Bayesian regression). Results: The zero-inflated denoising convolutional autoencoder exhibited a partial RMSE of 839.3 counts and partial MAE of 431.1 counts, whereas mean imputation achieved a partial RMSE of 1053.2 counts and partial MAE of 545.4 counts, the zero-inflated Poisson regression model achieved a partial RMSE of 1255.6 counts and partial MAE of 508.6 counts, and Bayesian regression achieved a partial RMSE of 924.5 counts and partial MAE of 605.8 counts. Conclusions: Our deep learning–based imputation model performed better than the other methods when imputing missing values in actigraphy data.

Objective We aimed to present the study design and baseline cross-sectional participant characteristics of biobank innovations for chronic cerebrovascular disease with Alzheimer’s disease study (BICWALZS) participants.Methods A total of 1,013 participants were enrolled in BICWALZS from October 2016 to December 2020. All participants underwent clinical assessments, basic blood tests, and standardized neuropsychological tests (n=1,013). We performed brain magnetic resonance imaging (MRI, n=817), brain amyloid positron emission tomography (PET, n=713), single nucleotide polymorphism microarray chip (K-Chip, n=949), locomotor activity assessment (actigraphy, n=200), and patient-derived dermal fibroblast sampling (n=175) on a subset of participants.Results The mean age was 72.8 years, and 658 (65.0%) were females. Based on clinical assessments, total of 168, 534, 211, 80, and 20 had subjective cognitive decline, mild cognitive impairment (MCI), Alzheimer’s dementia, vascular dementia, and other types of dementia or not otherwise specified, respectively. Based on neuroimaging biomarkers and cognition, 199, 159, 78, and 204 were cognitively normal (CN), Alzheimer’s disease (AD)-related cognitive impairment, vascular cognitive impairment, and not otherwise specified due to mixed pathology (NOS). Each group exhibited many differences in various clinical, neuropsychological, and neuroimaging results at baseline. Baseline characteristics of BICWALZS participants in the MCI, AD, and vascular dementia groups were generally acceptable and consistent with 26 worldwide dementia cohorts and another independent AD cohort in Korea.Conclusion The BICWALZS is a prospective and longitudinal study assessing various clinical and biomarker characteristics in older adults with cognitive complaints. Details of the recruitment process, methodology, and baseline assessment results are described in this paper.

Lee DY, et al. J Korean Med Sci. 2019 Sep;34(41):e263. https://doi.org/10.3346/jkms.2019.34.e263

This cross-sectional, observational study aimed to integrate the analyses of relationships of physical activity, depression, and sleep with cognitive function in community-dwelling older adults using a single model. To this end, physical activity, sleep, depression, and cognitive function in 864 community-dwelling older adults from the Suwon Geriatric Mental Health Center were assessed using the International Physical Activity Questionnaire, Montgomery-Asberg Depression Rating Scale, Pittsburgh Sleep Quality Index, and Mini-Mental State Examination for Dementia Screening, respectively. Their sociodemographic characteristics were also recorded. After adjusting for confounders, multiple linear regression analysis was performed to investigate the effects of physical activity, sleep, and depression on cognitive function. Models 4, 5, 7, and 14 of PROCESS were applied to verify the mediating and moderating effects of all variables. Physical activity had a direct effect on cognitive function (effect = 0.97, p < 0.01) and indirect effect (effect = 0.36; confidence interval: 0.18, 0.57) through depression. Moreover, mediated moderation effects of sleep were confirmed in the pathways where physical activity affects cognitive function through depression (F-coeff = 13.37, p < 0.001). Furthermore, these relationships differed with age. Thus, the associations among physical activity, depression, and sleep are important in interventions for the cognitive function of community-dwelling older adults. Such interventions should focus on different factors depending on age.

Introduction: Although several studies have examined the individual relationships among digital literacy, cognitive function, and depressive symptoms, few have integrated all three factors into a single model. This study aimed to address this gap by investigating the mediating effect of depressive symptoms on the relationship between digital literacy and cognition. In doing so, we hoped to contribute to a more comprehensive understanding of the complex interplay among these variables and their implications for mental health and well-being.Methods: Participants were 7,988 older adults (65 years or older) who participated in the Living Profiles of Older People Survey 2020. The main type of exposure was digital literacy (communication, information, media, and online transaction literacy). The main outcomes were depressive symptoms measured using the Short Geriatric Depression Scale of Korean version and cognitive function measured using the Mini-Mental State Examination score. Multiple linear regression and mediation analyses were also performed.Results: After adjusting for covariates, our analysis found a significant association between digital literacy and both depressive symptoms and cognitive function (β of four types of digital literacy and depressive symptoms = −0.123, −0.172, −0.702, and − 0.639, respectively; β of four types of digital literacy and cognitive function = 2.102, 2.217, 1.711, and 1.436, respectively). Moreover, our study showed that depressive symptoms play a mediating role in the relationship between media and online transaction literacy and cognitive function (95% CI of indirect effects = 0.0647–0.1212 and 0.0639–0.1277, respectively), implying an indirect pathway (digital literacy, depressive symptoms, and cognitive function).Discussion: This study sheds light on the relationship between digital literacy, depressive symptoms, and cognitive function in older adults. We found that depressive symptoms mediated the association between specific aspects of digital literacy (online transaction and media literacy) and cognitive function. Our results indicate that community-based digital literacy programs could be effective in reducing depression and preserving or improving cognitive function in older adults.

Background: Large language models (LLMs) have emerged as promising tools for addressing global disparities in mental health care. However, cloud-based proprietary models raise concerns about data privacy and limited adaptability to local health care systems. In contrast, open-source LLMs offer several advantages, including enhanced data security, the ability to operate offline in resource-limited settings, and greater adaptability to non-English clinical environments. Nevertheless, their performance in psychiatric applications involving non-English language inputs remains largely unexplored. Objective: This study aimed to systematically evaluate the clinical reasoning capabilities and diagnostic accuracy of a locally deployable open-source LLM in both Korean and English psychiatric contexts. Methods: The openbuddy-mistral-7b-v13.1 model, fine-tuned from Mistral 7B to enable conversational capabilities in Korean, was selected. A total of 200 deidentified psychiatric interview notes, documented during initial assessments of emergency department patients, were randomly selected from the electronic medical records of a tertiary hospital in South Korea. The dataset included 50 cases each of schizophrenia, bipolar disorder, depressive disorder, and anxiety disorder. The model translated the Korean notes into English and was prompted to extract 5 clinically meaningful diagnostic clues and generate the 2 most likely diagnoses using both the original Korean and translated English inputs. The hallucination rate and clinical relevance of the generated clues were manually evaluated. Top-1 and top-2 diagnostic accuracy were assessed by comparing the model’s prediction with the ground truth labels. Additionally, the model’s performance on a structured diagnostic task was evaluated using the psychiatry section of the Korean Medical Licensing Examination and its English-translated version. Results: The model generated 997 clues from Korean interview notes and 1003 clues from English-translated notes. Hallucinations were more frequent with Korean input (n=301, 30.2%) than with English (n=134, 13.4%). Diagnostic relevance was also higher in English (n=429, 42.8%) compared to Korean (n=341, 34.2%). The model showed significantly higher top-1 diagnostic accuracy with English input (74.5% vs 59%; P<.001), while top-2 accuracy was comparable (89.5% vs 90%; P=.56). Across 115 questions from the medical licensing examination, the model performed better in English (n=53, 46.1%) than in Korean (n=37, 32.2%), with superior results in 7 of 11 diagnostic categories. Conclusions: This study provides an in-depth evaluation of an open-source LLM in multilingual psychiatric settings. The model’s performance varied notably by language, with English input consistently outperforming Korean. These findings highlight the importance of assessing LLMs in diverse linguistic and clinical contexts. To ensure equitable mental health artificial intelligence, further development of high-quality psychiatric datasets in underrepresented languages and culturally adapted training strategies will be essential.