Extrafoveal Analysis of Categorically Defined Stereometric Shapes

The article outlines the possibilities of extrafoveal analysis during a conceptual visual search, when the target stimulus represents a geometric concept with an unknown form. The previous works have shown that when the target and distractors are widely diverging, the very first saccade is headed towards the target or the answer is given without any saccades, suggesting that the target is recognized extrafoveally. Our experimental design and data analysis intend to reveal extrafoveal processing of visualized concepts while saccades planning. We made the stimuli harder varying two factors: the difficulty of the target shape and its similarity to the distractors to determine the limit of complexity when the stimulus becomes too complex to be processed extrafoveally. Twelve participants with normal or corrected-to-normal vision took part in Experiment 1 and eighteen in Experiment 2. Four images of stereometric shapes were spaced at an equal distance from the center. In the first experiment, we got 2x2 design: the targets were (1) either prisms or pyramids with 3 to 6 base angles (shapes with each amount of angles were presented in an every probe) (2) among either prisms or pyramids as distractors. We received drastically different results for each of four types of the task. When searching for a target prism among pyramids, the observers made a single saccade or even no saccades towards the target. In the task with the target pyramid among prisms, the performance was a bit more difficult. A target prism among other prisms was a yet harder but still a solvable task. At last, searching for the target pyramid among other pyramids was proved to be so challenging that the amount of saccades evidenced the random search for the target shape by most participants. In the second experiment, we investigated the aforementioned last condition; so, 144 sets of pyramids were presented. Additionally, in the last 16 probes the participants were required to avoid any saccades away from the center. The findings revealed similar tendencies with the first experiment; additionally, it has been found that participants have never shown any significant decrease of saccade amount before the target is reached during the whole experiment. Moreover, one participant demonstrated a mean quantity of saccades that was significantly lower than the random one, but this parameter tended to increase from the beginning to the end of the experiment. However, in the last probes in case of fixed eye movements, most participants gave significantly more right answers comparatively to random guessing, the use of extrafoveal analysis was possible for them. Our findings indicate that extrafoveal processing as a part of conceptual visual search appears to be one of the possible behavioral acts in the complex system of perceptual action, rather than a merely low-level mechanism of saccade planning. The choice to use or not to use extrafoveal analysis is based upon comparing the required energy cost, necessity of foveal confirmation and the estimating subjective cost of a mistake.

Keywords

Authors

References