Adaptive network based fuzzy inference system efficientnet for autism spectrum disorder detection with optimization...

Area of Science:

• Neuroscience
• Artificial Intelligence
• Medical Imaging

Background:

• Autism Spectrum Disorder (ASD) is linked to abnormal neural activity and premature brain development.
• Traditional ASD diagnostic methods are time-consuming and yield unclear results, impacting social and communication skills.
• There is a need for advanced, efficient, and accurate diagnostic tools for ASD.

Purpose of the Study:

• To develop and evaluate a novel hybrid approach for Autism Spectrum Disorder (ASD) detection.
• To improve the accuracy and efficiency of ASD diagnosis using neuroimaging data.
• To combine Adaptive Network Based Fuzzy Inference (ANFIS) and EfficientNet for enhanced ASD identification.

Main Methods:

• A hybrid model, Adaptive Network Based Fuzzy Inference System EfficientNet (ANFIS-EffNet), was developed by modifying ANFIS and EfficientNet layers.
• Image pre-processing involved Anisotropic diffusion and Region of Interest (ROI) extraction.
• Functional connectivity-based pivotal region extraction used Adam War Strategy Optimization (AWSO), combining Adam optimization and War Strategy Optimization (WSO).
• Feature extraction incorporated Learned Invariant Feature Transform (LIFT) and statistical methods.

Main Results:

• The ANFIS-EffNet model demonstrated high performance in ASD detection.
• Achieved an accuracy of 93.219%.
• Achieved a sensitivity of 93.670% and specificity of 93.840%.

Conclusions:

• The ANFIS-EffNet model presents a highly effective and accurate method for Autism Spectrum Disorder detection.
• This hybrid approach significantly outperforms traditional methods in terms of accuracy and efficiency.
• The developed model shows promise for improving early diagnosis and intervention for ASD.