Analyzing brain structural differences associated with categories of blood pressure in adults using empirical kernel...

Area of Science:

• Neuroimaging
• Machine Learning
• Cardiovascular Health

Background:

• Hypertension poses risks for heart disease and cognitive impairment.
• Accurate blood pressure classification is crucial for patient management but understudied with machine learning.
• Brain structural changes associated with hypertension require further investigation.

Purpose of the Study:

• To develop and evaluate a machine learning model for discriminating blood pressure grades using structural brain MRI.
• To identify specific brain regions and features indicative of different hypertension stages.
• To explore the potential of machine learning in early detection of hypertension-related brain alterations.

Main Methods:

• Proposed an Empirical Kernel Mapping-based Kernel Extreme Learning Machine plus (EKM-KELM+) classifier.
• Extracted structural brain MRI features including gray matter volume (GMV), white matter volume, and cortical thickness.
• Utilized ensemble learning with privileged information (PI) to enhance classification accuracy.

Main Results:

• The EKM-KELM+ model achieved high accuracy in classifying hypertension grades.
• Gray matter volume (GMV) was the most discriminative feature, outperforming other metrics.
• Identified key brain regions like the olfactory cortex and orbitofrontal cortex as significantly affected by hypertension.

Conclusions:

• The EKM-KELM+ method effectively identifies brain structural changes associated with different blood pressure grades.
• Selected discriminative features align with existing neuroimaging findings.
• This approach offers a potential tool for early intervention in hypertension-related neurological conditions.