Applications of resting-state fMRI and machine learning in cognitive impairment in type 2 diabetes mellitus: A scoping review

This is the authors' abstract. We don't add key points for this paper.

Type 2 Diabetes Mellitus (T2DM) is a common metabolic disorder that adversely affects cognitive function and heightens the risk of neurodegenerative diseases. This review examines cutting-edge developments in utilizing machine learning techniques to assess brain function changes in T2DM patients, with a focus on cognitive impairment (CI). Through a comprehensive search across major medical databases, we identified and evaluated six studies that used resting-state functional MRI (rs-fMRI) and machine learning classifiers to analyze brain connectivity patterns in T2DM patients. Our analysis indicates that machine learning methods can effectively distinguish between T2DM patients with and without CI, revealing abnormal functional connectivity patterns linked to cognitive decline. These findings suggest that machine learning combined with neuroimaging holds promising initial findings for guiding early interventions and treatment strategies, with the goal of mitigating CI in T2DM patients and improving clinical outcomes. • Integration of Machine Learning and rs-fMRI: This review highlights the notable progress made in utilizing machine learning techniques to analyze resting-state functional MRI (rs-fMRI) data in patients with Type 2 Diabetes Mellitus (T2DM). By combining these methodologies, researchers can detect subtle changes in brain connectivity that are linked to cognitive impairment. • Cognitive Impairment Detection: The studies examined in this review illustrate that machine learning algorithms, especially support vector machines, are effective in classifying T2DM patients according to their cognitive status. This method provides a more objective alternative to conventional cognitive assessments, thereby minimizing the potential for subjective bias in diagnosis. • Call for Future Research: The review underscores the necessity for future studies to involve larger and more diverse participant groups, adopt multimodal research approaches, and implement advanced feature selection techniques. Such enhancements could deepen our understanding of cognitive decline in T2DM and inform early intervention and treatment strategies.