Marcus A. Triplett

This paper shows that recurrent neural networks can learn and represent causal structures from observational data, forming implicit beliefs about the full graphical structure even when trained only on pairwise queries.

The paper develops computational tools for rapid monosynaptic connectivity mapping in neural circuits using holographic stimulation and model-based compressed sensing.

This paper introduces a computational method to reduce off-target stimulation in two-photon holographic optogenetics by optimizing stimulation sites using adaptive non-negative basis function regression.

The paper introduces PhoRC, a method to remove photocurrent artifacts in optogenetic connectivity mapping data using constrained matrix factorization, improving the accuracy of synaptic connection detection.

This paper introduces Bayesian target optimisation to reduce off-target stimulation in optogenetics, showing improved precision in simulations and in vitro experiments.

The paper introduces a statistical model to separate evoked and spontaneous neural activity in calcium imaging data, providing quantitative insights into their contributions.