Triploid genetic algorithm for convolutional neural network-based diagnosis of mild cognitive impairment

Area of Science:

• Neuroimaging and Artificial Intelligence
• Computational Neuroscience
• Medical Diagnostics

Background:

• Mild cognitive impairment (MCI) is an early stage of dementia, crucial for timely intervention.
• Accurate classification of MCI converts and non-converts is vital for predicting dementia progression.
• Structural magnetic resonance imaging (sMRI) provides valuable data for neurodegenerative disease analysis.

Purpose of the Study:

• To propose a novel deep learning approach for classifying MCI converts and non-converts.
• To investigate the impact of activation functions and hyper-parameters on model performance.
• To enhance automated diagnostic capabilities for MCI using advanced algorithms.

Main Methods:

• A deep learning model incorporating a triploid genetic algorithm (a variant of genetic algorithms) was developed.
• Structural magnetic resonance imaging (sMRI) data was utilized for classification.
• The influence of different activation functions and hyper-parameter tuning was explored.

Main Results:

• The proposed deep learning method achieved a maximum classification accuracy of 0.97961.
• Empirical studies demonstrated the superiority of the proposed method compared to existing approaches.
• The model's performance was significantly influenced by the choice of activation functions and hyper-parameters.

Conclusions:

• The developed deep learning approach offers a promising tool for the effective diagnosis of MCI.
• This method has the potential to significantly aid clinicians in identifying individuals at risk of dementia.
• The findings contribute to the advancement of automated diagnostic systems in clinical settings.