The Behavior Ensemble and Neural Trajectory Observatory (Bento)
A graphical user interface for managing, exploring, and analyzing multimodal neuroscience datasets. Features:
- Synchronous display of video, audio, annotation, calcium imaging, and tracking data.
- Robust interface for manual annotation of video/other datasets.
- Built-in interfaces for PCA, nonnegative matrix factorization, and other manipulations of neural data.
Bento can also be paired with MARS to train supervised classifiers for behaviors of interest in your data.
The Mouse Action Recognition System (MARS)
A computer vision system for pose estimation and behavior classification in pairs of interacting mice, developed in collaboration with Cristina Segalin in the laboratory of Pietro Perona.
If you want to use the MARS pipeline in a setting other than the resident-intruder assay, the MARS_Developer repository lets you collect manual training data and fine-tune MARS to your own datasets.
Dynamic mushroom body model
A firing rate/spiking model of the Drosophila olfactory system, simulating the responses of ORNs, PNs, and Kenyon cells to a panel of odors. Captures the dynamics of recurrent inhibition in the PN and Kenyon cell populations, to more accurately predict Kenyon cell population representations of odors.
Other projects can be found on Github.