Douglas Lorne Arnold

The paper quantifies bias in proportionality-based association metrics for count data and proposes a logit-normal multinomial model to improve accuracy.

This paper develops a deep learning method with uncertainty estimation to detect MRI artifacts in large imbalanced datasets, achieving high accuracy and outperforming previous approaches.

This paper tracks the conversion of diffusely abnormal white matter to T2 lesions in MS patients, showing it occurs independently of inflammation and gadolinium enhancement.

The paper segments and tracks white matter lesions in MS patients, finding that transformation of diffuse abnormal white matter to focal lesions correlates with clinical progression.

This paper automates the separation of focal white matter lesions and diffusely abnormal white matter in MS using MRI intensity thresholds and finds MTR values are lower in focal lesions compared to diffusely abnormal white matter.