Calibration of retinal image size with distance in the Mongolian gerbil: rapid adjustment of calibrations in different contexts

Plasticity of the Association between Visual Space and Action Space in a Blind-Walking Task

Many experiments have shown that a brief visual preview provides sufficient information to complete certain kinds of movements (reaching, grasping, and walking) with high precision. This suggests that participants must possess a calibration between visual target location and the kinaesthetic, proprioceptive, and/or vestibular stimulation generated during movement towards the target. We investigated the properties of this calibration using a cue-conflict paradigm in which participants were trained with mismatched locomotor and visual input. After training, participants were presented with visual targets and were asked to either walk to them or locate them in a spatial updating task. Our results showed that the training was sufficient to produce significant, systematic miscalibrations of the association between visual space and action space. These findings suggest that the association between action space and visual space is modifiable by experience. This plasticity could be either due to modification of a simple, task-specific sensory motor association or it could reflect a change in the gain of a path integration signal or a reorganisation of the relationship between perceived space and action space. We suggest further experiments that might help to distinguish between these possibilities.

The effects of cortical lesions on recognition of object context in a visuomotor task in the Mongolian gerbil

Two experiments were carried out in order to determine the effects of either parietal or temporal lesions on performance in a depth vision task in which gerbils normally use retinal image size (RIS) as a cue to distance. In the first experiment, gerbils were trained to jump to two training targets that differed in size and which were always presented with distinctive local features and in a particular spatial location. After lesions, gerbils were presented with further training trials and sets of probe trials in which they were presented with targets that differed in width from the training targets, and sets of local features and distal cues that either matched or mismatched those presented during training. Shams and temporal animals made predictable over- or underjumps when local feature and distal information matched, and stopped using retinal image size when they did not match. Parietal animals did not use retinal image size either during the match or the mismatch conditions. In a second experiment, gerbils with parietal lesions were shown capable of using retinal image size in a simpler task that did not contain distinguishing local features or distal cues. Taken together, these results suggest that parietal lesions in gerbils disrupt object recognition, when the purpose of the recognition process is to complete a distance estimate for a visuomotor act.

The effects of lesions of anteromedial cortex on a ballistic visuomotor task in the Mongolian gerbil (Meriones unguiculatus)

Mongolian gerbils were trained to jump across a gap of randomly varying distance and then received one of three surgical treatments: Ten gerbils received aspiration lesions of anteromedial cortex (AMC), four gerbils received control lesions of a part of frontal cortex and nine gerbils received a sham procedure. Following a short recovery period, gerbils were tested in the jumping task. Gerbils with AMC lesions carried out fewer head movements than both sham and frontal gerbils. In addition, gerbils with AMC lesions were significantly less accurate in the distance estimation component of the jumping task, particularly at longer distances. These results suggest that anteromedial cortex participates in the generation of vertical head movements that are used to produce retinal motion information and confirms that these head movements are used to heighten the precision of distance estimation at longer testing distances.

Interaction of visual and tactile information in the control of chicks' locomotion in the visual cliff

The behaviour of two-day-old chicks placed on the deep side of a visual cliff was examined. With increasing depth of the floor below the chicks, latency to move over the deep side towards another chick on the shallow side increased, while speed of locomotion decreased. Chicks given the same incentive to step over a visible edge onto the deep side showed a stronger inhibition of movement at all depths, indicating that absolute depth of a surface and relative depth of an edge affect behaviour differently. At depths greater than 4 cm, the majority of chicks performed a jump from the deep to the shallow side, and the distances over which they jumped corresponded to those jumped when tested with a real gap. These results suggest that detection of the far side of a gap alone is sufficient to elicit a jump. Just before jumping, chicks adopted a head orientation which depended on the vertical distance of the cliff edge below them. Changes in head orientation did not maintain retinal fixation of the edge, and may instead be important in setting the correct direction of thrust when jumping.