A review of interactions between peripheral and foveal vision

Abstract

Visual processing varies dramatically across the visual field. These differences start in the retina and continue all the way to the visual cortex. Despite these differences in processing, the perceptual experience of humans is remarkably stable and continuous across the visual field. Research in the last decade has shown that processing in peripheral and foveal vision is not independent, but is more directly connected than previously thought. We address three core questions on how peripheral and foveal vision interact, and review recent findings on potentially related phenomena that could provide answers to these questions. First, how is the processing of peripheral and foveal signals related during fixation? Peripheral signals seem to be processed in foveal retinotopic areas to facilitate peripheral object recognition, and foveal information seems to be extrapolated toward the periphery to generate a homogeneous representation of the environment. Second, how are peripheral and foveal signals re-calibrated? Transsaccadic changes in object features lead to a reduction in the discrepancy between peripheral and foveal appearance. Third, how is peripheral and foveal information stitched together across saccades? Peripheral and foveal signals are integrated across saccadic eye movements to average percepts and to reduce uncertainty. Together, these findings illustrate that peripheral and foveal processing are closely connected, mastering the compromise between a large peripheral visual field and high resolution at the fovea.

Figure 1.

Schematic illustration of interactions between peripheral and foveal vision. The images illustrate differences between peripheral and foveal vision and the typical sequence of transsaccadic vision. Yellow, purple, green, and blue arrows indicate the direction of information flow. During fixation, peripheral vision is characterized by uncertainty in position (illustrated by spatial disarray using the Eidolon Factory (Koenderink et al., 2017); the degradation is overemphasised for the purpose of illustration), reduced spatial resolution, and increased crowding. Foveal appearance is extrapolated towards the periphery (yellow arrows; Extrapolation of foveal information to the periphery) and peripheral object recognition is supported by foveal feedback processing (purple arrow; Foveal feedback signals supporting peripheral object recognition). A saccadic eye movement (red arrow) brings an object of interest (here the castle door) to the fovea. During the saccade, vision is impaired (Dodge, 1900; for reviews see Matin, 1974; Volkmann, 1986; Wurtz, 2008; Binda & Morrone, 2018) by several factors, such as motion blur (Burr & Ross, 1982; but see Castet & Masson, 2000), masking due to the clear pre- and postsaccadic image (Matin, Clymer, & Matin, 1972; Campbell & Wurtz, 1978; Duyck et al., 2016), and active reduction of contrast sensitivity (e.g. Volkmann, 1962; Burr, Morrone, Ross, 1994; Diamond, Ross, Morrone, 2000). Differences between peripheral and foveal appearance might be compensated by transsaccadic re-calibration (blue arrow; Re-calibration of peripheral and foveal vision). Information from successive fixations might be stitched together by transsaccadic integration (green arrows; Transsaccadic integration of peripheral and foveal information).

Figure 2.

Peripheral object discrimination and foveal feedback. (A) Experimental paradigm and stimuli in the original paradigm (Williams et al., 2008). Participants had to categorize objects from three different categories. In each trial, two objects from the same or a different category were shown in the peripheral visual field and participants had to judge whether the two objects came from the same or a different category. Figure modified with permission from Williams et al. (2008). (B) Results in brain imaging (Williams et al., 2008). Region of interest (ROI) for further analysis and average correlations of brain activity elicited by objects of the same or a different category. Same-category correlations are higher than chance only in the foveal ROI, but not in peripheral ROIs outside of stimulus presentation. Figure modified with permission from Williams et al. (2008). (C) Behavioural consequences and time course of the effect (Fan et al., 2016). Presenting a noise mask in the fovea impairs peripheral object categorisation. The subpanels show conditions that require different amounts of mental rotation of the objects. The time course of the effect is modulated by the necessary amount of mental rotation. The pink dashed line indicates early detrimental effects of noise that are presumably related to the distraction of attention (Beck & Lavie, 2005). The yellow dashed line indicates the detrimental effect of noise that is related to the interference with foveal-feedback signals. Figure modified with permission from Fan et al. (2016).

Figure 3.

Visual illusions cancelling or inducing differences between foveal and peripheral appearance. (A) Uniformity illusion. Texture statistics are different in the central (only circles) and peripheral (squares, penta- and hexagons) part of the stimulus. After maintaining fixation on the centre for a few seconds, the peripheral part of the stimulus appears identical to the centre. Figure modified with permission from Otten, Pinto, Paffen, Seth, & Kanai (2017). (B) Extinction illusion. Each crossing of the grey bands contains a small white disk, but these disks are only visible in the fovea. Figure modified with permission from Ninio and Stevens (2000).

Figure 4.

Relevant aspects of the main paradigms used in the literature to test the effects of sensorimotor learning on perception. The joined screens represent the pre- and postsaccadic physical situation (blue and light red contours, respectively). The smaller screens surrounded by dashed lines represent the supposed appearance after learning. Top Row: saccade-contingent colour perception. Notice that, in this case, the observer never sees any peripheral stimuli, and saccade direction alters postsaccadic foveal appearance. In the test phase the perceptual effect is assessed by having the observer compare the colour of the pre- and postsaccadic (foveal) stimuli. Center Row: shape re-calibration. Notice that, in this case, we only represent the case of a swapped object, which turns from a square into a circle transsaccadically in the training phase. After training the peripheral stimulus appears more circular. In the test phase, the perceptual effect is assessed by having the observer compare the shape of the presaccadic (peripheral) stimulus with the shape of the postsaccadic (foveal) stimulus. Bottom Row: size re-calibration. In this case, the observer experiences a transsaccadic increase in size, which increases the perceived size of the peripheral stimulus. Notice that in this paradigm there are no distinct training and test phases. Within the same trial, size perception is assessed first by having the observer compare two stimuli that are presented simultaneously. In the second part of the trial, the observer executes the saccade and experiences the transsaccadic change.

Figure 5.

Overview of differing mechanisms of transsaccadic stimulus comparison and combination, from low-level pattern overlay (fusion) to high-level feature averaging and uncertainty reduction. When a saccade is made to a stimulus (this figure shows an example of a saccade being made to the door of Rauischholzhausen Castle), the presaccadic, peripheral percept of the stimulus must be reconciled with the postsaccadic, foveal percept of the stimulus. At the lowest level, this can occur as an overlay of pre- and postsaccadic patterns (fusion). Note that fusion only occurs under specific conditions: a task that does not require precise spatial alignment, and a weak foveal stimulus (Fusion: low level pattern overlay). On a higher level of feature combination, the pre- and postsaccadic stimuli are encoded into visual working memory, and then the pre- and postsaccadic stimulus feature information can be combined (Integration of feature information across saccades). The ultimate percept of the stimulus may be a result of both pre- and postsaccadic stimulus features (feature averaging); furthermore, the transsaccadic percept may be more reliable than either the pre- or postsaccadic percept alone, as a result of transsaccadic integration (reducing uncertainty).

Figure 6.

(A) Presaccadic information can affect postsaccadic perception. Fabius et al. (2019) used the High Phi illusion to measure transsaccadic information transfer (Presaccadic information alters postsaccadic perception). The paradigm is shown on the left, and the results are shown on the right. An inducer stimulus presented before the saccade created an illusory jump in spatiotopically matched stimulus texture after the saccade; this jump was greater when a saccade was executed than in a spatially matched fixation condition, or when the presaccadic inducer was static. Figure modified with permission from Fabius et al. (2019). (B) Optimal transsaccadic integration is tested by measuring the reliability of the pre- and postsaccadic percepts alone, and the reliability of the combined transsaccadic percept. If information is integrated, the reliability in the transsaccadic trials should be greater than in either pre- or postsaccadic trials alone.