Spatial cell firing during virtual navigation of open arenas by head-restrained mice

Guifen Chen, John A. King, Yi Lü, Francesca Cacucci, Neil Burgess

TL;DR

Researchers have developed a novel virtual reality (VR) system for head-restrained mice that allows for simultaneous neural recording and spatial navigation. This system successfully replicates the characteristic firing patterns of place and grid cells observed in real environments, opening new avenues for studying the neural basis of spatial cognition with advanced imaging techniques.

What Problem Does This Paper Address?

Investigating the neural mechanisms of spatial navigation and memory in rodents has traditionally relied on freely moving animals. While this allows for naturalistic behavior, it poses significant challenges for advanced neural recording techniques like multi-photon microscopy, which require head-fixation. Existing VR systems for head-fixed animals often limit movement to one dimension or prevent full head rotation, thus failing to capture the full complexity of 2D spatial firing patterns observed in real environments. This paper addresses this gap by presenting a VR system that enables mice to navigate in 2D open arenas while being head-restrained, making it compatible with high-resolution imaging and detailed neural analysis.

Methods

The core innovation is a VR system designed for head-restrained mice. Key components include:

  • Head restraint: A custom head-plate and holder system allows for horizontal rotation of the head while keeping it fixed in place, crucial for multi-photon imaging.
  • Immersive environment: Multiple screens and projectors create a 360-degree visual environment, including a floor projection, simulating an open arena.
  • Locomotion tracking: Optical mice mounted on a floating sphere detect and translate the mouse's movements into virtual locomotion.
  • Behavioral tasks: Mice were trained on a "fading beacon" task, similar to a Morris Water Maze, to learn and navigate to a hidden reward location, demonstrating allocentric spatial memory.
  • Electrophysiology: Tetrodes were implanted in the hippocampus (CA1) and dorsomedial entorhinal cortex (dmEC) to record the activity of place cells, grid cells, and head-direction cells.

This setup allows for precise control over visual stimuli and the decoupling of motor and sensory inputs, while still enabling the expression of complex spatial behaviors and neural activity.

Datasets

The study did not introduce new benchmark datasets for training or evaluation. Instead, it generated novel electrophysiological data from mice navigating in both their custom-built virtual reality (VR) environments and comparable real-world (R) environments. These datasets were used to compare the spatial firing properties of neurons under different conditions.

Key Insights and Ideas

  • VR as a viable platform for 2D spatial navigation: The system successfully elicits 2D spatial firing patterns (place and grid cells) in head-restrained mice, a significant step beyond previous 1D VR paradigms.
  • Broader spatial tuning in VR: Place and grid cells exhibited broader spatial tuning in VR compared to real environments. This might be a compensatory mechanism for the lack of vestibular and other distal cues, leading to increased reliance on visual and proprioceptive information.
  • Directionality differences: Place cells showed increased directionality in VR, potentially due to the absence of omnidirectional local cues that normally support less directional firing in real arenas. Grid cells, on the other hand, did not show this increased directionality, suggesting they might be more strongly tied to self-motion cues.
  • Theta frequency and speed: The relationship between running speed and LFP theta frequency was attenuated in VR, likely due to the absence of vestibular acceleration signals. However, neuronal firing rates still increased with running speed, indicating a dissociation between LFP theta and neuronal firing rate modulation by speed.
  • Theta phase precession is preserved: A crucial finding is that theta phase precession, a key temporal coding mechanism in spatial navigation, is maintained in VR, even with reduced theta frequency and absent vestibular cues. This suggests that visual and proprioceptive inputs are sufficient for this phenomenon.

Key Results

  • Navigation: Mice successfully learned to navigate to a hidden reward location in the VR environment, demonstrating allocentric spatial memory.
  • Place cells: Showed broader spatial tuning and increased directionality in VR compared to real environments. Spatial information content was lower in VR.
  • Grid cells: Also exhibited broader spatial tuning in VR, with a larger spatial scale. Spatial information content was lower in VR. Gridness scores were comparable, but some older mice showed lower gridness in VR.
  • Head-direction cells: Showed similar firing properties and tuning in VR and real environments, suggesting they are less affected by the VR setup's limitations.
  • Theta frequency: Increased less with running speed in VR compared to real environments.
  • Firing rates: Increased with running speed similarly in both VR and real environments.
  • Theta phase precession: Was comparable in VR and real environments for both place and grid cells.

Limitations and Future Work

  • Spatial scale and directionality: The broader spatial tuning and increased directionality of place cells in VR suggest that the absence of local cues and potentially vestibular signals influences their firing. Future work could explore how different types of visual cues or the introduction of artificial vestibular cues might modulate these properties.
  • Grid cell sensitivity: The observation that some grid cells showed reduced gridness in VR, particularly in older mice, suggests that translational self-motion cues are important for their firing. Further investigation into the specific sensory inputs that drive grid cell activity in VR is warranted.
  • Multi-photon imaging compatibility: While the system is designed for head-fixation, the current setup with multiple screens might require further optimization for seamless integration with advanced imaging techniques like two-photon microscopy, especially concerning potential image distortions from in-plane rotation.
  • Future directions: The authors highlight the potential of this VR system for future studies involving manipulations of sensory inputs, genetic tools, and advanced imaging to dissect the neural mechanisms of spatial cognition with unprecedented precision. The ability to decouple sensory and motor signals in a controlled 2D environment is a powerful tool for future research.

Summary updated

Abstract

We present a mouse virtual reality (VR) system which restrains head-movements to horizontal rotations, compatible with multi-photon imaging. This system allows expression of the spatial navigation and neuronal firing patterns characteristic of real open arenas (R). Comparing VR to R: place and grid, but not head-direction, cell firing had broader spatial tuning; place, but not grid, cell firing was more directional; theta frequency increased less with running speed, whereas increases in firing rates with running speed and place and grid cells' theta phase precession were similar. These results suggest that the omni-directional place cell firing in R may require local-cues unavailable in VR, and that the scale of grid and place cell firing patterns, and theta frequency, reflect translational motion inferred from both virtual (visual and proprioceptive) and real (vestibular translation and extra-maze) cues. By contrast, firing rates and theta phase precession appear to reflect visual and proprioceptive cues alone.
BibTeX
@inproceedings{chen2018spatial,
  author = {Guifen Chen and John A. King and Yi Lü and Francesca Cacucci and Neil Burgess},
  title = {{Spatial cell firing during virtual navigation of open arenas by head-restrained mice}},
  booktitle = {eLife},
  year = {2018},
  doi = {10.7554/elife.34789},
}
Figure