Searchers adjust their eye-movement dynamics to target characteristics in natural scenes

Pathways for Naturalistic Looking Behavior in Primate I: Behavioral Characteristics and Brainstem Circuits

Organisms control their visual worlds by moving their eyes, heads, and bodies. This control of "gaze" or "looking" is key to survival and intelligence, but our investigation of the underlying neural mechanisms in natural conditions is hindered by technical limitations. Recent advances have enabled measurement of both brain and behavior in freely moving animals in complex environments, expanding on historical head-fixed laboratory investigations. We juxtapose looking behavior as traditionally measured in the laboratory against looking behavior in naturalistic conditions, finding that behavior changes when animals are free to move or when stimuli have depth or sound. We specifically focus on the brainstem circuits driving gaze shifts and gaze stabilization. The overarching goal of this review is to reconcile historical understanding of the differential neural circuits for different "classes" of gaze shift with two inconvenient truths. (1) "classes" of gaze behavior are artificial. (2) The neural circuits historically identified to control each "class" of behavior do not operate in isolation during natural behavior. Instead, multiple pathways combine adaptively and non-linearly depending on individual experience. While the neural circuits for reflexive and voluntary gaze behaviors traverse somewhat independent brainstem and spinal cord circuits, both can be modulated by feedback, meaning that most gaze behaviors are learned rather than hardcoded. Despite this flexibility, there are broadly enumerable neural pathways commonly adopted among primate gaze systems. Parallel pathways which carry simultaneous evolutionary and homeostatic drives converge in superior colliculus, a layered midbrain structure which integrates and relays these volitional signals to brainstem gaze-control circuits.

Look twice: A generalist computational model predicts return fixations across tasks and species

Primates constantly explore their surroundings via saccadic eye movements that bring different parts of an image into high resolution. In addition to exploring new regions in the visual field, primates also make frequent return fixations, revisiting previously foveated locations. We systematically studied a total of 44,328 return fixations out of 217,440 fixations. Return fixations were ubiquitous across different behavioral tasks, in monkeys and humans, both when subjects viewed static images and when subjects performed natural behaviors. Return fixations locations were consistent across subjects, tended to occur within short temporal offsets, and typically followed a 180-degree turn in saccadic direction. To understand the origin of return fixations, we propose a proof-of-principle, biologically-inspired and image-computable neural network model. The model combines five key modules: an image feature extractor, bottom-up saliency cues, task-relevant visual features, finite inhibition-of-return, and saccade size constraints. Even though there are no free parameters that are fine-tuned for each specific task, species, or condition, the model produces fixation sequences resembling the universal properties of return fixations. These results provide initial steps towards a mechanistic understanding of the trade-off between rapid foveal recognition and the need to scrutinize previous fixation locations.

Guided Search 6.0: An updated model of visual search

This paper describes Guided Search 6.0 (GS6), a revised model of visual search. When we encounter a scene, we can see something everywhere. However, we cannot recognize more than a few items at a time. Attention is used to select items so that their features can be "bound" into recognizable objects. Attention is "guided" so that items can be processed in an intelligent order. In GS6, this guidance comes from five sources of preattentive information: (1) top-down and (2) bottom-up feature guidance, (3) prior history (e.g., priming), (4) reward, and (5) scene syntax and semantics. These sources are combined into a spatial "priority map," a dynamic attentional landscape that evolves over the course of search. Selective attention is guided to the most active location in the priority map approximately 20 times per second. Guidance will not be uniform across the visual field. It will favor items near the point of fixation. Three types of functional visual field (FVFs) describe the nature of these foveal biases. There is a resolution FVF, an FVF governing exploratory eye movements, and an FVF governing covert deployments of attention. To be identified as targets or rejected as distractors, items must be compared to target templates held in memory. The binding and recognition of an attended object is modeled as a diffusion process taking > 150 ms/item. Since selection occurs more frequently than that, it follows that multiple items are undergoing recognition at the same time, though asynchronously, making GS6 a hybrid of serial and parallel processes. In GS6, if a target is not found, search terminates when an accumulating quitting signal reaches a threshold. Setting of that threshold is adaptive, allowing feedback about performance to shape subsequent searches. Simulation shows that the combination of asynchronous diffusion and a quitting signal can produce the basic patterns of response time and error data from a range of search experiments.

The peripheral sensitivity profile at the saccade target reshapes during saccade preparation

Goal-directed eye movements (saccades) bring peripheral objects of interest into high-acuity foveal vision. In preparation for the incoming foveal image, the perception of the saccade target may sharpen gradually before the eye movement is executed. Indeed, previous studies suggest that pre-saccadic attention shifts enhance sensitivity to high spatial frequencies (SFs) more than sensitivity to lower SFs. This pattern, however, was observed within a narrow frequency range and may reflect local changes in the shape of a broader underlying sensitivity profile. Depending on the development of the profile's shape, SFs above the previously examined range may profit less from saccade preparation. To assess the impact of saccade preparation on the shape of a broader sensitivity profile, we prompted observers to discriminate the orientation of a sinusoidal grating (the probe) presented briefly at the target of an impending saccade, at 10 dva (degree of visual angle) eccentricity. The probe's SF ranged from 1 to 5.5 cycles per dva (cpd) and was unpredictable on a given trial. We fitted observers' response accuracies across SFs with a log-parabolic, that is, inverted U-shaped function. Long before saccade onset, the profile peaked at .6 cpd and dropped off towards lower and higher SFs with broad bandwidth. During saccade preparation, the peak of the profile increased and shifted towards higher SFs while the bandwidth of the profile decreased. As a consequence of this reshaping process, pre-saccadic enhancement increased with SF up to 2.5 cpd, corroborating previous findings. Sensitivities to higher SFs, however, profited less from saccade preparation. We conclude that the extent of pre-saccadic enhancement to a particular SF is governed by its position on a broader sensitivity profile which reshapes substantially during saccade preparation. The shift of the profile's peak towards higher SFs increases resolution at the saccade target even when the features of relevant visual information are unpredictable.

Saliency-Aware Subtle Augmentation Improves Human Visual Search Performance in VR

Visual search becomes challenging when the time to find the target is limited. Here we focus on how performance in visual search can be improved via a subtle saliency-aware modulation of the scene. Specifically, we investigate whether blurring salient regions of the scene can improve participant's ability to find the target faster when the target is located in non-salient areas. A set of real-world omnidirectional images were displayed in virtual reality with a search target overlaid on the visual scene at a pseudorandom location. Participants performed a visual search task in three conditions defined by blur strength, where the task was to find the target as fast as possible. The mean search time, and the proportion of trials where participants failed to find the target, were compared across different conditions. Furthermore, the number and duration of fixations were evaluated. A significant effect of blur on behavioral and fixation metrics was found using linear mixed models. This study shows that it is possible to improve the performance by a saliency-aware subtle scene modulation in a challenging realistic visual search scenario. The current work provides an insight into potential visual augmentation designs aiming to improve user's performance in everyday visual search tasks.

Human visual search follows a suboptimal Bayesian strategy revealed by a spatiotemporal computational model and experiment

There is conflicting evidence regarding whether humans can make spatially optimal eye movements during visual search. Some studies have shown that humans can optimally integrate information across fixations and determine the next fixation location, however, these models have generally ignored the control of fixation duration and memory limitation, and the model results do not agree well with the details of human eye movement metrics. Here, we measured the temporal course of the human visibility map and performed a visual search experiment. We further built a continuous-time eye movement model that considers saccadic inaccuracy, saccadic bias, and memory constraints. We show that this model agrees better with the spatial and temporal properties of human eye movements and predict that humans have a memory capacity of around eight previous fixations. The model results reveal that humans employ a suboptimal eye movement strategy to find a target, which may minimize costs while still achieving sufficiently high search performance.

Modeling the effects of perisaccadic attention on gaze statistics during scene viewing

How we perceive a visual scene depends critically on the selection of gaze positions. For this selection process, visual attention is known to play a key role in two ways. First, image-features attract visual attention, a fact that is captured well by time-independent fixation models. Second, millisecond-level attentional dynamics around the time of saccade drives our gaze from one position to the next. These two related research areas on attention are typically perceived as separate, both theoretically and experimentally. Here we link the two research areas by demonstrating that perisaccadic attentional dynamics improve predictions on scan path statistics. In a mathematical model, we integrated perisaccadic covert attention with dynamic scan path generation. Our model reproduces saccade amplitude distributions, angular statistics, intersaccadic turning angles, and their impact on fixation durations as well as inter-individual differences using Bayesian inference. Therefore, our result lend support to the relevance of perisaccadic attention to gaze statistics.

On the relative (un)importance of foveal vision during letter search in naturalistic scenes

The importance of high-acuity foveal vision to visual search can be assessed by denying foveal vision using the gaze-contingent Moving Mask technique. Foveal vision was necessary to attain normal performance when searching for a target letter in alphanumeric displays, Perception & Psychophysics, 62 (2000) 576-585. In contrast, foveal vision was not necessary to correctly locate and identify medium-sized target objects in natural scenes, Journal of Experimental Psychology: Human Perception and Performance, 40 (2014) 342-360. To explore these task differences, we used grayscale pictures of real-world scenes which included a target letter (Experiment 1: T, Experiment 2: T or L). To reduce between-scene variability with regard to target salience, we developed the Target Embedding Algorithm (T.E.A.) to place the letter in a location for which there was a median change in local contrast when inserting the letter into the scene. The presence or absence of foveal vision was crossed with four target sizes. In both experiments, search performance decreased for smaller targets, and was impaired when searching the scene without foveal vision. For correct trials, the process of target localization remained completely unimpaired by the foveal scotoma, but it took longer to accept the target. We reasoned that the size of the target may affect the importance of foveal vision to the task, but the present data remain ambiguous. In summary, the data highlight the importance of extrafoveal vision for target localization, and the importance of foveal vision for target verification during letter-in-scene search.

Task-dependence in scene perception: Head unrestrained viewing using mobile eye-tracking

Real-world scene perception is typically studied in the laboratory using static picture viewing with restrained head position. Consequently, the transfer of results obtained in this paradigm to real-word scenarios has been questioned. The advancement of mobile eye-trackers and the progress in image processing, however, permit a more natural experimental setup that, at the same time, maintains the high experimental control from the standard laboratory setting. We investigated eye movements while participants were standing in front of a projector screen and explored images under four specific task instructions. Eye movements were recorded with a mobile eye-tracking device and raw gaze data were transformed from head-centered into image-centered coordinates. We observed differences between tasks in temporal and spatial eye-movement parameters and found that the bias to fixate images near the center differed between tasks. Our results demonstrate that current mobile eye-tracking technology and a highly controlled design support the study of fine-scaled task dependencies in an experimental setting that permits more natural viewing behavior than the static picture viewing paradigm.

Disentangling bottom-up versus top-down and low-level versus high-level influences on eye movements over time

Bottom-up and top-down as well as low-level and high-level factors influence where we fixate when viewing natural scenes. However, the importance of each of these factors and how they interact remains a matter of debate. Here, we disentangle these factors by analyzing their influence over time. For this purpose, we develop a saliency model that is based on the internal representation of a recent early spatial vision model to measure the low-level, bottom-up factor. To measure the influence of high-level, bottom-up features, we use a recent deep neural network-based saliency model. To account for top-down influences, we evaluate the models on two large data sets with different tasks: first, a memorization task and, second, a search task. Our results lend support to a separation of visual scene exploration into three phases: the first saccade, an initial guided exploration characterized by a gradual broadening of the fixation density, and a steady state that is reached after roughly 10 fixations. Saccade-target selection during the initial exploration and in the steady state is related to similar areas of interest, which are better predicted when including high-level features. In the search data set, fixation locations are determined predominantly by top-down processes. In contrast, the first fixation follows a different fixation density and contains a strong central fixation bias. Nonetheless, first fixations are guided strongly by image properties, and as early as 200 ms after image onset, fixations are better predicted by high-level information. We conclude that any low-level, bottom-up factors are mainly limited to the generation of the first saccade. All saccades are better explained when high-level features are considered, and later, this high-level, bottom-up control can be overruled by top-down influences.