Pier Luigi Dragotti

This paper introduces a deep learning method to improve calcium imaging by reducing scattering and increasing throughput in light-field microscopy.

This paper develops a physics-driven deep learning method to reconstruct neuronal activity from light field microscopy images, outperforming existing methods in quality and speed.

This study shows that neural heterogeneity improves task performance and learning stability in spiking neural networks, matching experimental parameter distributions.

This study demonstrates sub-cellular resolution 3D light field imaging of voltage signals in acute mouse brain slices using genetically encoded voltage indicators.

The paper introduces CosMIC, a new metric for evaluating spike inference accuracy in calcium imaging, showing it better discriminates precision and recall than existing metrics.

This paper introduces an adaptive spiral scanning algorithm for multi-photon microscopy to improve temporal resolution in neural activity imaging.

ABLE detects cell locations in two-photon calcium imaging data using a level set method, achieving 67.5% success on a manually labelled mouse in vivo dataset.

This paper extends the FRI framework for calcium transient detection to include GCaMP6s data, achieving 90% spike detection accuracy on real data.

The paper compares spike detection performance using FRI methods with new GECIs and a synthetic dye, calculating Cramer-Rao bounds for spike position uncertainty.