Development of virtual environments to investigate path integration in mice
MetadataShow full item record
Path integration is the ability to navigate to a goal location without using external cues, thus relying entirely on self-motion information. To do so, two components of a path need to be encoded: orientation and distance. While the ability to estimate distance, called linear path integration, is well established in humans, it is unclear whether rodents are equally capable of doing so and the underlying neural circuit mechanisms are only poorly understood. This thesis discusses the development of a virtual reality system and behavioural task to investigate linear path integration in mice, and the results obtained from experiments carried out with this system. The setup provides full control over visual input while de-correlating vestibular and olfactory signals from location. Manipulations of the translation from physical to virtual movement can thus be used to probe relative influences of motor related and visual signals. Chapter 1 reviews the current literature on path integration and provides a background to the technical setup of the system. Chapter 2 describes the design and construction of the virtual reality system, its individual components and the software created to run experiments. It discusses how 3-d modelling and 3-d printing have successfully been combined to allow rapid development and production of custom components in different materials. Chapter 3 discusses the development of behavioural tasks designed to investigate linear path integration. It shows that by using a simple virtual track design and a carefully monitored food-deprivation regime, mice can be trained to successfully associate a visually indicated location with a reward. Chapter 4 describes behavioural experiments carried out using this virtual linear track. I obtained evidence that mice can estimate the distance to the rewarded zone reliably using path integration strategies. To test whether mice rely on motor information or optic flow, I manipulated the gain between physical movement and virtual movement. My results suggest that mice primarily rely on motor information for linear path integration. In the final chapter the results are discussed in the context of other recent work and areas for further development of the system are identified.