Sensitivity of a harbor seal (Phoca vitulina) to coherent visual motion in random dot displays

Abstract

Motion vision is one of the fundamental properties of the visual system and is involved in numerous tasks. Previous work has shown that harbor seals are able to perceive visual motion. Tying in with this experimental finding, we assessed the sensitivity of harbor seals to visual motion using random dot displays. In these random dot displays, either all or a percentage of the dots plotted in the display area move into one direction which is referred to as percent coherence. Using random dot displays allows determining motion sensitivity free from form or position cues. Moreover, when reducing the lifetime of the dots, the experimental subjects need to rely on the global motion over the display area instead of on local motion events, such as the streaks of single dots. For marine mammals, the interpretation of global motion stimuli seems important in the context of locomotion, orientation and foraging. The first experiment required the seal to detect coherent motion directed upwards in one out of two stimulus displays and psychophysical motion coherence detection thresholds were obtained ranging from 5% to 35% coherence. At the beginning of the second experiment, which was conducted to reduce the differential flickering of the motion stimulus as secondary cue, the seal was directly able to transfer from coherent motion detection to a discrimination of coherent motion direction, leftward versus rightward. The seal performed well even when the duration of the local motion event was extremely short in the last experiment, in which noise was programmed as random position noise. Its coherence threshold was determined at 23% coherence in this experiment. This motion sensitivity compares well to the performance of most species tested so far excluding monkeys, humans and cats. To conclude, harbor seals possess an effective global motion processing system. For seals, the interpretation of global and coherent motion might e. g. play a role in the interpretation of optic flow information or when breaking the camouflage of cryptic prey items.

Keywords: Global motion; Motion coherence; Motion vision; Pinnipeds; Random dot display; Vision.

Figure 1

Random dot displays as used to assess motion sensitivity in a harbor seal. Schematic representation of random dot displays at different levels of motion coherence (columns) and with either random direction (upper line) or random position noise (lower line). For explanation of random direction and random position noise please see Material and methods. All dots move in different directions at 0% coherence (left). At 50% coherence, one half of the dots move to the right, while the other dots move in random directions (middle). At 100% coherence all dots move to the right (right). Each random dot display as displayed in this figure represents the coordinates of all dots over the time course of 10 frames rendering the trajectories of the dots visible. Please note that, during the coherent motion detection experiment, the coherent signal was directed upwards (not displayed in this figure) whereas, during the coherent motion direction discrimination experiments, it was either leftwards or rightwards.

Figure 2

Experimental setup and procedure. a Schematic drawing of the projection apparatus depicting the arrangement of projector (P), mirrors (M), frame (F) and projection screen (PS) that was used for the underwater projection. The experimenter sat on the platform above the projection screen out of sight of the experimental animal and operated the computer (C) controlled stimuli. b The animal watched the stimulus display by placing its head through a hole in the frontal wall of the experimental chamber resting at a stationary target (view of the experimenter). c It communicated its decision, thereby indicating the position of the stimulus area containing the coherent stimulus (coherent motion detection experiment) in case of a correct choice, by moving its head to one out of two response targets (only right response target shown for clarity in A) to the right and left of the stationary target. In the coherent motion direction discrimination experiments (not shown), it indicated the direction of stimulus motion in one stimulus area by again moving its head to one of the two response targets.

Figure 3

Sensitivity to coherent motion in random dot displays for a harbor seal. Results are depicted as psychometric functions (a , d , e) plotting the performance of the seal in % correct choices as a function of % coherence. Threshold performance of 75% correct choices is indicated by a dashed horizontal line, and arrows pointing to the x-axis mark the thresholds. In b, c results are plotted as % coherence at the 75% threshold as a function of dot lifetime (in ms) or dot density (in dots/deg²). a Performance of the seal during the initial determination of coherence threshold (coherent motion detection experiment). Each data point represents the mean performance of 90 presentations; due to human and equipment error the seal’s performance at 20% and 40% coherence was averaged over 89 presentations only. The 75% threshold was determined at 33.7%. b Thresholds assessed during the estimation of the effect of dot lifetime (coherent motion detection experiment phase 1). A marked decrease in coherence at threshold with increasing dot lifetime is visible in the performance. c Thresholds assessed during the estimation of the effect of dot density (coherent motion detection experiment phase 2). Very low coherence thresholds down to 4.7% coherence were obtained in this experiment, and performance increased with increasing stimulus parameter. d Performance of the seal during coherent motion direction discrimination with random direction noise. Each data point represents the mean performance of 30, in the threshold range from 10–40% coherence of 60 presentations. The seal could discriminate motion direction with a threshold of 21.0% coherence. e Performance of the sea during coherent motion direction discrimination with random position noise. The 75% threshold for discriminating motion direction with random position noise of 22.5% coherence compares favorably with the direction threshold obtained in the coherent motion direction discrimination experiment with random direction noise.