Prediction of complex stimuli across saccades

Abstract

The visual system can predict visual features across saccades based on learned transsaccadic associations between peripheral and foveal input. This has been shown for simple visual features such as shape, size, and spatial frequency. The present study investigated whether transsaccadic predictions are also made for more complex visual stimuli. In an acquisition phase, new transsaccadic associations were established. In the first experiment, pictures of real-world objects changed category during the saccade (fruits were changed into balls or vice versa). In the second experiment, the gender of faces was manipulated during the saccade (faces changed from male to female or vice versa). In the following test phase, the stimuli were briefly presented in the periphery, and participants had to indicate which object or face, respectively, they had perceived. In both experiments, peripheral perception was biased toward the acquired associated foveal input. These results demonstrate that transsaccadic predictions are not limited to a small set of simple visual features but can also be made for more complex and realistic stimuli. Multiple new associations can be learned within a short time frame, and the resulting predictions appear to be object specific.

What determines location specificity or generalization of transsaccadic learning?

“…Experimentally, it has been shown that, after being exposed to new transsaccadic object changes, peripheral perception is biased in the direction of the learned transsaccadic associations. This effect has been shown in various studies and with multiple object features ( Herwig & Schneider, 2014 ; Köller, Poth, & Herwig, 2020 ; Osterbrink & Herwig, 2021 ; Paeye, Collins, Cavanagh, & Herwig, 2018 ; Valsecchi & Gegenfurtner, 2016 ). One aspect that is not clear yet though, and that has also been pointed out as an open question in the review article by Stewart, Valsecchi, and Schütz (2020) is the location specificity of this learning process, because two studies that will be described in more detail hereafter have shown seemingly contradictory findings.…”

“…A learning index was computed as a measure for the presaccadic perceptual bias, thus for the effect of the transsaccadic learning on the presaccadic shape judgments ( Herwig & Schneider, 2014 ; Osterbrink & Herwig, 2021 ; Paeye et al, 2018 ). For this, the average shape judgment of the swapped object was subtracted from the shape judgment of the normal object in the circle-to-square group, and the normal shape judgment was subtracted from the swapped shape judgment in the square-to-circle group.…”

Transsaccadic object associations shape peripheral perception: The role of reliability.

“…That is, had the object’s frequency changed from low to high, they perceived the peripheral frequency higher than that of a baseline object whose frequency did not change during the acquisition phase. This object-specific, presaccadic perceptual bias has been shown in several studies and feature dimensions such as spatial frequency (Herwig et al, 2018; Herwig & Schneider, 2014), shape (Herwig et al, 2015; Köller et al, 2020; Paeye et al, 2018), size (Bosco et al, 2015, 2020; Valsecchi et al, 2020; Valsecchi & Gegenfurtner, 2016) and even more complex stimulus dimensions like the gender of a face (Osterbrink & Herwig, 2021).…”

“…There were no manipulations of reliability included in this experiment. Similar to earlier studies investigating the presaccadic perceptual bias (e.g., Herwig & Schneider, 2014; Osterbrink & Herwig, 2021) we expected participants to show biases in their peripheral perception of features in the object type that was changed during acquisition.…”

“…As a measure for the effect of the transsaccadic learning on the spatial frequency judgments, a learning index was computed for each participant by subtracting the average spatial frequency judgment of the normal object from the judgment of the swapped object for the low-to-high group and subtracting the average spatial frequency judgment of the swapped object from the judgment of the normal object for the high-to-low group. Hence, a positive value indicated a judgment bias of the changed object in the direction of previously associated foveal input and a negative value OSTERBRINK, RECKER, AND HERWIG indicated a judgment bias in the opposite direction (for a related procedure, see Herwig & Schneider, 2014;Osterbrink & Herwig, 2021;and Paeye et al, 2018). These learning indices (see Figure 3b) were significantly greater than zero for the low-contrast level, t(34) = 2.76, p = .009, d = .47, BF 10 = 4.530, medium-contrast level, t(34) = 2.20, p = .034, d = .37, BF 10 = 1.521, and highcontrast level, t(34) = 2.18, p = .036, d = .37, BF 10 = 1.468 (BF 10 is the Bayes factor from the Bayesian one-sample test of comparison of learning indices with zero).…”