improv

improv enables adaptive neuroscience experiments by integrating real-time data analysis with experimental control to support dynamic, data-driven adjustments during data collection.

Key Features:

• Integration of custom modeling and analysis: Combines custom modeling with real-time data collection and experimental control to support adaptive experimental designs.
• Real-time data processing: Processes neuroscience data, including 2-photon calcium imaging, in real time to provide immediate information about current brain state.
• Support for diverse experimental designs: Supports both in silico and in vivo studies across different model organisms and data types.
• Advanced analytical techniques: Implements real-time machine learning for dimension reduction and predictive modeling of latent neural and behavioral features.
• Custom behavioral analysis and functional typing: Performs real-time behavioral analyses and rapidly classifies neural responses from large populations using calcium microscopy.
• Optimal stimulus selection: Enables optimal visual stimulus selection driven by ongoing data analysis.
• Interactive imaging and optogenetic stimulation: Enables model-driven interactive imaging and simultaneous optogenetic photostimulation, exemplified with larval zebrafish expressing GCaMP6s and the opsin rsChRmine.

Scientific Applications:

• Adaptive experimental design: Enables experiments that adapt stimuli and recording parameters in real time based on incoming data.
• Real-time hypothesis generation and testing: Facilitates on-the-fly exploration and refinement of hypotheses through integrated modeling and analysis during data collection.
• Neural circuit mapping and manipulation in model organisms: Supports mapping and manipulation of neural activity and behavior, including optogenetic perturbations in larval zebrafish.

Methodology:

Integrates custom modeling with real-time data collection and experimental control; performs real-time processing of 2-photon calcium imaging and calcium microscopy data; applies real-time machine learning for dimension reduction and predictive modeling; conducts real-time behavioral analysis, rapid population-level neural classification, optimal visual stimulus selection, and model-driven interactive imaging with simultaneous optogenetic photostimulation.