Interrogating the Dynamic Neural Computation of the Sense of Direction

点击翻译按钮获取中文摘要

PROJECT SUMMARY / ABSTRACT Navigating dynamic environments to avoid danger and locate the necessities of life relies critically on animals keeping track of their own position and heading within the world, which cannot be sensed directly but instead must be inferred by integrating many streams of information, including vestibular, proprioceptive, and visual. Because directional heading has a straightforward mathematical definition (i.e., horizon and azimuth angles), the sense of direction represents fertile ground for quantitative study of multisensory integration and population coding in numerous areas of the brain. In particular, electrophysiological experiments studying the brain's representation of the body's spatial relationship to the environment have described a class of neurons referred to as head direction (HD) cells, which each fire when an animal is facing a specific preferred direction. The prevailing model for the generation of the signal encoded by these cells assumes that information about angular head velocity from the vestibular system drives activity changes in reciprocally-connected brainstem areas, the dorsal tegmental and lateral mammillary nuclei. These areas are modeled as a ring attractor network, which can maintain a stable pattern of activity due to excitatory connections between neurons with similar preferred directions and inhibition between neurons with opposing preferred directions. However, the main projection from the vestibular nuclei to the network of brain areas containing HD cells is via the nucleus prepositus, which principally encodes eye-related rather than head-related information. Thus, it seems likely that eye-related information plays a substantial role in the generation of the HD signal. The summation of head-in-world direction with eye-in-head direction defines gaze direction. We thus hypothesize that the "head direction" network in actuality primarily encodes gaze direction. In order to investigate this hypothesis, I will record from multiple points of the HD network in rhesus macaques during both passive motion and active locomotion using high-density neural probes while simultaneously tracking eye, body, and head movements. By analyzing eye, head, and limb movements in conjunction with recordings from the HD system at multiple stages of information processing, and in particular by recording from large numbers of neurons at multiple locations simultaneously, this project will provide new insights into fundamental questions about how the brain represents the world and the body's relationship to it.