Tadpole VR: virtual reality visualization of a simulated tadpole spinal cord

摘要

Recent advances in "developmental" approach (combining experimental study with computational modeling) of neural networks produce increasingly large data sets, in both complexity and size. This poses a significant challenge in analyzing, visualizing and understanding not only the spatial structure but also the behavior of such networks. This paper describes a virtual reality application for visualization of two biologically accurate computational models that model the anatomical structure of a neural network comprised of 1500 neurons and over 80,000 connections. The visualization enables a user to observe the complex spatiotemporal interplay between seven unique types of neurons culminating in an observable swimming pattern. We present a detailed description of the design approach for the virtual environment, based on a set of initial requirements, followed up by the implementation and optimization steps. Lastly, the results of a pilot usability study are being presented on how confident participants are in their ability to understand how the alternating firing pattern between the two sides of the tadpole's body generates swimming motion.