Predicting infarction growth rate II using ANFIS-based binary particle swarm optimization technique in ischemic...

Area of Science:

• Neurology
• Medical Imaging
• Artificial Intelligence

Background:

• Ischemic stroke causes brain cell death due to blocked blood flow.
• Accurate prediction of infarction growth is crucial for managing stroke progression.
• Existing AI methods like ANFIS face challenges with high dimensionality and long training times.

Purpose of the Study:

• To develop an innovative, automatic method for predicting the second infarction growth after an initial CT scan in ischemic stroke patients.
• To address the limitations of conventional ANFIS by reducing dimensionality and training time.

Main Methods:

• A novel approach combining Binary Particle Swarm Optimization (BPSO) with Adaptive Neuro-Fuzzy Inference System (ANFIS) architecture.
• Selection of clinically relevant features using Pearson correlation coefficients and P-values.
• Comparison of the proposed model against conventional ANFIS, Support Vector Regressor (SVR), shallow Neural Networks, and Linear Regression.

Main Results:

• The developed model demonstrated high accuracy in predicting infarction growth.
• Achieved a Root Mean Square Error of 0.091, Mean Squared Error of 0.0086, Mean Absolute Error of 0.064, and Cosine distance of 0.074.
• The BPSO-ANFIS model showed improved efficiency in terms of reduced dimensionality and training time compared to conventional ANFIS.

Conclusions:

• The BPSO-ANFIS model offers a more accurate and efficient solution for predicting ischemic stroke progression.
• This method holds potential for improving clinical decision-making and patient outcomes in stroke management.
• The study highlights the effectiveness of combining optimization algorithms with fuzzy inference systems for complex medical predictions.