The Role of Oculomotor Proprioception in the Establishment of Long-Term Visuospatial Memory

Abstract You can point the door of the room with your eyes closed. Patients with certain brain lesions cannot, although they can point to it if they can see it. Establishing a spatially accurate memory of visual space requires solving a classic problem in cognitive neuroscience: to understand how the brain changes a retinal signal, arising from

a constantly moving eye, into a spatially accurate signal for memory, perception, and action. Two pioneering neuroscientists suggested different solutions to this problem. Hermann von Helmholtz postulated that the brain achieves spatial accuracy by feeding the motor command for an eye movement back to the visual system, to

compensate for an impending saccade, and thus maintain a spatially accurate representation of the visual world, a process now known as corollary discharge. Sir Charles Sherrington postulated that oculomotor proprioceptors could measure the position of the eyes in the orbit, and where the head is on the body, to use these data to

calculate the spatial location of visual objects. There is physiological evidence for both viewpoints: when a monkey plans a saccade, neurons in the lateral intraparietal area (LIP) respond to stimuli not in their receptive fields, but which will be brought into their receptive fields by the planned saccade, which must involve a corollary

discharge because the receptive fields change before the eye moves. In contrast, eye position modulates the visual responses in LIP, the gain fields, and from this, it is simple to calculate target position in craniotopic coordinates. When humans are asked to make memory-guided saccades after a number of intervening

saccades, their error increases linearly after the each of the first 3 saccades, implying the use of a corollary discharge mechanism. However, the error becomes stable after 6 or 7 saccades, implying that the brain has switched to a stable, craniotopic representation of space, which could use gain fields. LIP neurons have

retinotopic receptive fields, but exhibit environmental memory, showing these retinotopic receptive fields have access to a gaze invariant representation of space. There is a representation of eye position in Area 3a of somatosensory cortex, which arises from sensors in the contralateral eye. Our preliminary data suggest that

this area is necessary for both gain fields in LIP and for the establishment of long-term spatial memory, but not for the short-term corollary discharge mechanism. This proposal takes these discoveries as the starting point for three aims: 1) we will use inactivation of Area 3a to study its role in the establishment of gain fields, long-term

spatial memory, and environmental memory in LIP. 2) LIP projects to the parahippocampal gyrus (PHG), which then projects to posterior entorhinal cortex (PEC). We will categorize cell types in parahippocampal gyrus (PHG), and examine the role of oculomotor proprioception in the spatial properties of PHG neurons 3) PEC has a class

of cells, grid cells, from which the hippocampus can build its cognitive map. We hypothesize that oculomotor proprioception is necessary for construction of grid cells. These results support the novel suggestion that somatosensory deficits cause spatial memory deficits in humans and non-human primates.