Brain functional network connectivity interpolation characterizes the neuropsychiatric continuum and heterogeneity

Area of Science:

• Neuroscience
• Psychiatry
• Machine Learning

Background:

• Schizophrenia (SZ) and autism spectrum disorder (ASD) present heterogeneous symptoms, often viewed as a continuum.
• Current diagnostic and neuroimaging methods struggle with individual heterogeneity and progression.
• Conventional approaches often ignore the spectrum of symptoms and connectivity patterns.

Purpose of the Study:

• To develop and validate a functional network connectivity (FNC) interpolation framework using variational autoencoders (VAEs).
• To estimate the neuropsychiatric continuum and heterogeneity in SZ and ASD.
• To capture dynamic and static FNC properties and visualize individual differences.

Main Methods:

• Utilized a variational autoencoder (VAE) framework for FNC interpolation.
• Employed static FNC (sFNC) and dynamic FNC (dFNC) data from control and patient groups (SZ, ASD).
• Compared VAE performance against linear and semi-supervised baselines.

Main Results:

• VAEs significantly outperformed baseline models in capturing FNC properties.
• Interpolated continua revealed group-wise gradients in SZ and ASD patients.
• Identified specific patterns of altered correlations within and between brain networks across the continuum.

Conclusions:

• The proposed VAE framework effectively models the FNC continuum in SZ and ASD.
• This approach offers data-driven insights into individual differences and disease staging.
• The framework enables visualization of disease progression and imputation of missing FNC data.