Ryan Field

Kernel Flow meets Virtual Reality

This week we ran our first sessions combining Kernel Flow and Virtual Reality (VR). We used Pico Interactive’s Neo 2 Eye VR headset and added Kernel Flow modules to cover the prefrontal cortex.

On the customized headset, we ran a modified version of Virtuleap’s Magic Deck designed to produce robust hemodynamic responses. The game design is based upon Dr. Barbara Sahakian and the University of Cambridge’s research "Innovative methods for improving cognition, motivation and wellbeing in schizophrenia", specifically targeting episodic memory.

The pairing of Kernel Flow—the world’s first full-head coverage TD-fNIRS system—and VR is of particular interest to us as we build a scalable software and hardware ecosystem for both organizations and individuals.  

VR is increasingly popular in both animal and human neuroscience: VR enables the methodical and systematic gathering of otherwise annoying and sometimes study-breaking environmental confounds, which could include:

  • Eliminating variable lighting conditions, peripheral noise, or room size
  • Creating perceived, naturalistic motion without too much actual motion
  • Enabling safe, remote and simultaneous social interaction
  • Improving generalizability to real-world tasks, beyond what is possible in the laboratory
  • Standardizing task/stimulus design and hardware calibrations to be "plug and play" 

In addition to minimizing confounds, VR has the potential to expand the reach of neuroscience through enabling real-time neurofeedback in a fully immersive environment. This could take training, education, and general self-improvement to new heights. 

All of these factors contribute to increased replicability, study power, and ecological relevance. Coupled with extremely high-dimensional behavioral data (eye-tracking, motion tracking, etc, with the possibility of nearly 100,000 events logged from the VR headset every second), Kernel Flow + VR offers exciting new discovery potential. We believe the Flow-VR ecosystem will be a significant step forward over what exists today. 

We are in the early stages, but as a proof-of-concept, we show differential activation patterns during the working memory task as compared to a control condition in a single subject doing a memory task in VR:

Imagining what Kernel Flow could do if paired with VR? Apply for one of our first 50 Kernel Flow systems early next year here.