关键词: 眼跳, 视觉显著性, 奖赏价值

Abstract: Objective: Visual attention is influenced by stimulus-driven and goal-driven processes, which could both cause different types of attentional capture and predict saccadic eye movements. Previous work has mostly focused on a trade-off between these two processes or their respective, separate influences on the time-course of eye movements by using visual search paradigm or spatial orientation tasks. Recent studies have examined how these two processes may interact with each other by using decision-making tasks. However, the results of these studies were inconsistent. It remained unclear that whether visual salience and reward value showed additive, subtractive or synergistic effects. Therefore, this study investigated how visual salience and reward value affect saccadic eye movements under cooperating and competing conditions. Methods: 60 young adults were instructed to view Gabor patch pairs with different levels of visual salience and reward value. 20 Gabor patches that varied in 4 orientations and 5 luminance contrasts were generated. Reward value was manipulated to vary with orientation. Varying the contrast of the stimuli provided a manipulation of visual salience. Gabor patch were presented in pairs, resulting in 380 possible combinations from the above-mentioned 20 Gabor patches. Participants were instructed to view the Gabor patch pairs in a natural manner or they were asked to choose the patch with higher reward value. Eye movements were recorded for both tasks using a desktop eye tracking system-EyeLink 1000. Cooperating conditions showed one of the patches having both higher value (associated with more vertical orientation) and higher visual salience (higher contrast). Competing conditions showed one patch having higher value and lower visual salience than the other patch or vice versa lower value and higher physical salience. Saccade latency and saccade capture rate were gathered and analyzed to see how fast and how likely the participants direct their saccades to the side with higher reward value under different conditions. Results: For free viewing task, visual salience showed significant effect on saccadic eye movements (p<.001). For binary choice task, repeated measure ANOVA with congruencies, reward value and visual salience as factors revealed significant main effects of congruency and reward value on saccade capture rate and saccade latency (p<.05). Results also showed significant interaction effect of congruency and visual salience on saccade capture rate and saccade latency. This revealed that reward value showed overriding effect on saccadic capture rate and saccade latency under both conditions, while the effect of visual salience showed distinct patterns under different conditions. When visual salience and reward value cooperate, visual salience had additive effect on saccadic latency (p<.05). Saccades were directed more often and faster to the Gabor patch with higher reward. When they compete, visual salience played a significant role on misleading the saccades more often and faster to the side with lower reward but higher salience (p<.05). Conclusions: This is the first study to investigate how visual salience and reward value affect saccadic eye movements under cooperating and competing conditions in humans. These findings suggest that distinct patterns of eye movements exist under cooperating and competing conditions of stimulus-driven and goal-driven processes.

Key words: saccadic eye movement, visual salience, reward value