Diffusion tensor tractography in children with sensory processing disorder: Potentials for devising machine learning...

Area of Science:

• Neuroscience
• Developmental Neuroscience
• Medical Imaging

Background:

• Sensory Processing Disorder (SPD) affects brain organization of sensory input.
• Understanding the neurobiological underpinnings of SPD is crucial for diagnosis and intervention.
• Diffusion Tensor Imaging (DTI) offers insights into white matter microstructure and connectivity.

Purpose of the Study:

• To determine DTI microstructural and connectivity correlates of SPD.
• To apply machine learning for identifying children with SPD using DTI/tractography metrics.
• To investigate specific white matter tract alterations in children with SPD.

Main Methods:

• Recruited 44 children with SPD and 41 typically developing children (TDC).
• Acquired DTI data, including fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), edge density (ED), and track density (TD).
• Employed machine learning algorithms (naïve Bayes, random forest, support vector machine, neural networks) for classification.

Main Results:

• Children with SPD exhibited lower FA, ED, and TD, and higher MD and RD, primarily in posterior white matter tracts.
• Reduced track density (TD) in the splenium of the corpus callosum was the most significant distinguishing feature.
• Random forest models using tract-based TD achieved 77.5% accuracy in classifying SPD.

Conclusions:

• SPD is associated with impaired microstructural and connectivity integrity in posterior white matter tracts.
• Reduced TD in the splenium of the corpus callosum is a key indicator of SPD.
• DTI-derived connectivity metrics, analyzed with machine learning, show potential as imaging biomarkers for neurodevelopmental disorders like SPD.