Building a Brain Observatory for Visual Behavior
summary
A Python-based pipeline integrates automated mouse training with the Allen Brain Observatory data flow, aiming to record activity across multiple cell types and regions rather than passive viewing. Concrete pieces: task code uses PsychoPy & PyDAQmx for stimulus and NI control, GUIs built with PyQT & Flask, live streaming via pyzmq and VEGA, analysis with pandas/matplotlib/seaborn and automated progression via transitions state machine. Teams train dozens of animals in parallel to lick on image change, shrink training time, and release data through the Allen SDK (40k+ neurons in prior releases).
AI-written summary · about
chapters (9)
- 0:00Allen Institute introduction
- 4:45Brain Observatory purpose
- 10:07Automating mouse training
- 14:25Live animal monitoring
- 15:51Behavior analysis toolkit
- 18:39Training stage automation
- 21:31Python in software
- 22:14Training efficiency gains
- 23:44Q&A
related talks
28mScience in the Browser – Orchestrating and Visualising Neural Simulations
21mVisualizing robot minds, and other fun stuff
27m"Boosting Neuroimaging Analysis and Results Digestibility 🧠"
29mMaking Neuroimage Analysis Easier to Understand
01h 00mData, visualization, and designing AI
14mMachine Hallucinations
37mUsing Exploratory Data Analysis to Discover Patterns
30mOn the Fly and On the Flight: Scientific Data Analysis Beyond the Beamline