Extensive T1-weighted MRI Preprocessing Improves Generalizability of Deep Brain Age Prediction Models

Area of Science:

• Neuroimaging
• Artificial Intelligence
• Biomarkers

Background:

• Brain age estimation from T1w MRI is a key biomarker for brain aging and diseases.
• Current brain age prediction accuracy is within 2-3 years, but inter-study comparisons are difficult due to varied preprocessing.
• Deep learning models are increasingly used for brain age prediction.

Approach:

• Investigated the impact of four T1w MRI preprocessing pipelines on four deep learning brain age models.
• Evaluated variations in registration, grayscale correction, and software implementation.
• Assessed model performance using mean absolute error (MAE) on T1w images.

Key Points:

• Preprocessing choices significantly affect prediction error (up to 0.7 years MAE increase).
• Affine registration improved MAE compared to rigid registration.
• 3D isotropic 1 mm³ models were less sensitive to preprocessing variations than 2D or downsampled 3D models.
• Offset correction is crucial for generalizing model performance across datasets with different preprocessing.

Conclusions:

• Extensive T1w preprocessing can enhance MAE, particularly for new datasets, contrary to some literature.
• Offset correction is essential for robust brain age prediction generalization, irrespective of preprocessing.
• Standardizing preprocessing or implementing offset correction is vital for reliable brain age estimation.