Neuro-VISOR

Neuro-VISOR in action
Neuro-VISOR in action.

VISOR (Virtual Interactive Simulation Of Reality) is a research and software theme developed by Temple University's Center for Computational Mathematics and Modeling (C2M2), College of Science and Technology. The overarching long-term vision is to produce novel concepts and software that enable efficient immersed virtual reality (VR) visualization and real-time interaction with simulations of real-world processes described via principled mathematical equations. Unlike traditional high performance computing (HPC) applications, the philosophy of VISOR is that (a) the simulation runs while it is visualized in a virtual environment, and (b) the simulation continues even when the user affects and/or modifies the system state or its conditions.

Neuro-VISOR experienced by conference participants
Neuro-VISOR experienced by conference participants.

VISOR (Virtual Interactive Simulation Of Reality) is a research and software theme developed by Temple University's Center for Computational Mathematics and Modeling (C2M2), College of Science and Technology. The overarching long-term vision is to produce novel concepts and software that enable efficient immersed virtual reality (VR) visualization and real-time interaction with simulations of real-world processes described via principled mathematical equations. Unlike traditional high performance computing (HPC) applications, the philosophy of VISOR is that (a) the simulation runs while it is visualized in a virtual environment, and (b) the simulation continues even when the user affects and/or modifies the system state or its conditions.

Neuro-VISOR focuses on applications in computational neuroscience. It is based on a pipeline that retrieves wire-frame neuron geometry files from the public neuron database NeuroMorpho and generates a surface meshes from it. The Neuro-VISOR software then allows the user to choose different neurons, arrange them into their virtual room, and establish synapse connections between them. Efficient numerical methods approximate the Hodgkin-Huxley model on the neuron geometries, and the voltage state of every point on each neuron is visualized in VR in real time, while the user can then interact with the surface mesh and affect the simulation while it is running via several methods, including direct voltage manipulation and placing voltage clamps.

A key use case of this framework is that is can rapidly accelerate scientific discovery by providing an immediate and very intuitive feedback to the user about how changes to the simulated system affect the system's behavior. This insight, obtained from the simple and fast models used for VISOR, can then enable the computational scientist to devise significantly more targeted (non-interactive) simulations on large-scale HPC clusters of more complex models. In addition, in the context of neuroscience, the immersed 3D environment provides a more intuitive way to navigate and comprehend complex neuron geometries (which may have lots of branching and twisting of dendrites and axons) than traditional visualizations on computer screens.

The VISOR project can serve domain scientists (e.g., neuroscientists interested in simulating their experiments), computational mathematicians (e.g., numerical analysis testing the robustness of numerical methods), and also be a great tool for education and teaching and vertical integration of research.

The Neuro-VISOR software can be run in a VR version (currently only for Oculus), as well as via a desktop version (for users without VR headset). With the exception of the 3D visualization capabilities of the VR version, the desktop version provides the same interactive real-time simulation capabilities.

Project Website and Code Releases

Neuro-VISOR project website (github)
 

Release History

Development Team

Beta Testers