Predicting the cognitive function status in end-stage renal disease patients at a functional subnetwork scale

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Biology

Background:

• Functional magnetic resonance imaging (fMRI) reveals brain functional networks crucial for understanding cognition.
• Global network topological parameters offer a broad view but lack detailed insights into regional interactions.
• Existing methods struggle to capture the complex relationships and information transfer within brain networks.

Purpose of the Study:

• To develop a framework for predicting cognitive function status in end-stage renal disease (ESRD) patients at a functional subnetwork scale (CFSFSS).
• To enhance the accuracy of cognitive function prediction by examining subnetworks rather than global networks.
• To provide a more precise tool for clinical decision-making and intervention for cognitive impairment in ESRD patients.

Main Methods:

• Proposed the Cognitive Function Framework at a Functional Subnetwork Scale (CFSFSS) by combining nodes from different network indicators.
• Extracted topological attribute parameters of functional subnetworks as features, selected using minimal Redundancy Maximum Relevance (mRMR).
• Optimized support vector regression (SVR) parameters using an enhanced whale optimization algorithm (E-WOA) for global optimization.

Main Results:

• CFSFSS demonstrated superior predictive performance compared to other methods.
• Achieved low error metrics: Mean Absolute Error (MAE) of 0.5951, Mean Absolute Percentage Error (MAPE) of 0.0281, and Root Mean Square Error (RMSE) of 0.9994.
• Identified active brain regions associated with cognitive function status, yielding more precise predictive features.

Conclusions:

• The functional subnetwork approach provides more precise features for predicting cognitive function status in ESRD patients.
• CFSFSS enhances the accuracy of cognitive function prediction, aiding clinical decision-making and intervention for cognitive impairment.
• This method offers valuable insights into the complex brain network alterations associated with cognitive decline in ESRD.