Cognitive Impairment of Patient With Neurological Cerebrovascular Disease Using the Artificial Intelligence Technology Guided by MRI

Abstract

This study was to explore the application of MRI based on artificial intelligence technology combined with neuropsychological assessment to the cognitive impairment of patients with neurological cerebrovascular diseases. A total of 176 patients were divided into a control group, a vascular cognitive impairment non-dementia (VCIND) group, a vascular dementia (VD) group, and an Alzheimer's disease (AD) group. All patients underwent MRI and neuropsychological evaluation and examination, and an improved fuzzy C-means (FCM) clustering algorithm was proposed for MRI processing. It was found that the segmentation accuracy (SA) and similarity (KI) data of the improved FCM algorithm used in this study were higher than those of the standard FCM algorithm, bias-corrected FCM (BCFCM) algorithm, and rough FCM (RFCM) algorithm (p < 0.05). In the activities of daily living (ADL), the values in the VCIND group (23.55 ± 6.12) and the VD group (28.56 ± 3.1) were higher than that in the control group (19.17 ± 3.67), so the hippocampal volume was negatively correlated with the ADL (r = −0.872, p < 0.01). In the VCIND group (52.4%), VD group (31%), and AD group (26.1%), the proportion of patients with the lacunar infarction distributed on both sides of the brain and the number of multiple cerebral infarction lesions (76.2, 71.4, and 71.7%, respectively) were significantly higher than those in the control group (23.9 and 50%). In short, the improved FCM algorithm showed a higher segmentation effect and SA for MRI of neurological cerebrovascular disease. In addition, the distribution, number, white matter lesions, and hippocampal volume of lacunar cerebral infarction were related to the cognitive impairment of patients with cerebrovascular diseases.