DisQ-HNet: A Disentangled Quantized Half-UNet for Interpretable Multimodal Image Synthesis Applications to Tau-PET Synthesis from T1 and FLAIR MRI

Researchers developed DisQ-HNet, a new AI framework that synthesizes tau-PET brain scans from MRI data to detect Alzheimer's disease pathology. The method uses advanced neural network architectures to generate cost-effective alternatives to expensive PET imaging while maintaining diagnostic accuracy.

• →DisQ-HNet combines vector-quantized encoders with Half-UNet decoders to synthesize tau-PET scans from T1 and FLAIR MRI data.
• →The framework provides interpretable analysis by showing how each MRI modality contributes to tau-PET predictions.
• →Testing shows the method maintains reconstruction quality while preserving disease-relevant signals for Alzheimer's diagnosis.
• →The approach offers a more accessible alternative to expensive tau-PET imaging for detecting Alzheimer's pathology.
• →Partial Information Decomposition enables modality-specific attribution of synthesized brain uptake patterns.