James Kozloski

Neuralign identifies shared neural patterns across heterogeneous datasets without requiring matched neurons or trials, revealing common structures in motor cortex and mouse brain recordings.

This paper introduces DeepHM, a method combining deep learning and biophysical modeling to infer ion channel conductance changes in Alzheimer's mouse models.

The paper evaluates a large-scale brain model's functional requirements for coordinated learning between cortical and sub-cortical systems and discusses its relevance to Huntington's disease.

This paper uses global sensitivity analysis to evaluate model-derived metrics for predicting drug-induced TdP risk, finding that these metrics are not strongly correlated with EAD generation and can be explained by direct features.