A Deep Neural Network-Based Model for Screening Autism Spectrum Disorder Using the Quantitative Checklist for Autism...

Area of Science:

• Neuroscience
• Developmental Psychology
• Artificial Intelligence

Background:

• Autism spectrum disorder (ASD) is a complex neurodevelopmental condition affecting communication and behavior.
• Early diagnosis and intervention are crucial for improving outcomes in children with ASD.
• Current screening methods, like the Quantitative Checklist for Autism in Toddlers (QCHAT), are manual and can lead to diagnostic delays.

Purpose of the Study:

• To apply deep neural network (DNN) algorithms for identifying ASD using QCHAT datasets.
• To evaluate the efficacy of DNNs in improving the speed and accuracy of ASD screening.
• To compare DNN performance against contemporary methods for ASD detection.

Main Methods:

• Utilized two datasets: QCHAT and QCHAT-10.
• Applied deep neural network (DNN) algorithms for pattern recognition within the datasets.
• Trained and validated DNN models for ASD classification.

Main Results:

• The proposed DNN-based method demonstrated superior performance compared to existing techniques.
• Achieved accurate identification of ASD-related behavioral traits from QCHAT data.
• Indicated potential for reduced diagnostic delays through automated screening.

Conclusions:

• Deep neural networks offer a powerful tool for enhancing ASD screening accuracy and efficiency.
• Automated analysis of QCHAT data using DNNs can significantly expedite the diagnostic process.
• This AI-driven approach holds promise for improving early intervention and patient outcomes in ASD.