Enhancing identification performance of cognitive impairment high-risk based on a semi-supervised learning method.

BACKGROUND: Cognitive assessment plays a pivotal role in the early detection of cognitive impairment, particularly in the prevention and management of cognitive diseases such as Alzheimer's and Lewy body dementia. Large-scale screening relies heavily on cognitive assessment scales as primary tools, with some low sensitivity and others expensive. Despite significant progress in machine learning for cognitive function assessment, its application in this particular screening domain remains underexplored, often requiring labor-intensive expert annotations. AIMS: This paper introduces a semi-supervised learning algorithm based on pseudo-label with putback (SS-PP), aiming to enhance model efficiency in predicting the high risk of cognitive impairment (HR-CI) by utilizing the distribution of unlabeled samples. DATA: The study involved 189 labeled samples and 215,078 unlabeled samples from real world. A semi-supervised classification algorithm was designed and evaluated by comparison with supervised methods composed by 14 traditional machine-learning methods and other advanced semi-supervised algorithms. RESULTS: The optimal SS-PP model, based on GBDT, achieved an AUC of 0.947. Comparative analysis with supervised learning models and semi-supervised methods demonstrated an average AUC improvement of 8% and state-of-art performance, repectively. CONCLUSION: This study pioneers the exploration of utilizing limited labeled data for HR-CI predictions and evaluates the benefits of incorporating physical examination data, holding significant implications for the development of cost-effective strategies in relevant healthcare domains.

Validity and reliability of a computerized cognitive function evaluation battery (CogEvo) as a screening tool.

Aim: The aim of this study was to determine the validity and reliability of cognitive function evaluation battery, CogEvo, a recently developed computerized cognitive function evaluation battery, as a screening tool for decreased cognitive function. Methods: The study sample comprised 123 (age: 57-97 years) community-dwelling elderly people. They were required to perform five CogEvo tasks and complete two questions-based neuropsychological tests, including the Mini-Mental State Examination, so that the correlations could be analyzed. The validity and reliability of CogEvo were examined using factor analysis, MacDonald's omega reliability coefficient, logistic regression analysis, and receiver operating characteristic curve analysis. Results: Exploratory factor analysis revealed the orientation/spatial cognitive function (orientation and spatial cognition) and attention/executive function (attention, memory, and execution) factors. Structural validity was supported by confirmatory factor analysis. All two-factor-based subtasks showed adequate internal consistency (MacDonald's omega ≥0.6). The total CogEvo score and two-factor scores were significantly correlated with neuropsychological test results. Based on the total CogEvo score, the cognitively normal and cognitive decline groups were identified by receiver operating characteristic curve analysis with a moderate predictive performance. The cognitive decline group was well identified using the orientation/spatial cognitive function factor. Conclusions: CogEvo is a valid and reliable screening tool for cognitive function evaluation. It proved useful in the early identification of cognitive decline in our study sample.