Objects with reduced visibility still contribute to size averaging

Abstract

People can rapidly judge the average size of a collection of objects with considerable accuracy. In this study, we tested whether this size-averaging process relies on relatively early object representations or on later object representations that have undergone iterative processing. We asked participants to judge the average size of a set of circles and, in some conditions, presented two additional circles that were either smaller or larger than the average. The additional circles were surrounded by four-dot masks that either lingered longer than the circle array, preventing further processing with object substitution masking (OSM), or disappeared simultaneously with the circle array, allowing the circle representation to reach later visual processing stages. Surprisingly, estimation of average circle size was modulated by both visible circles and circles whose visibility was impaired by OSM. There was also no correlation across participants between the influence of the masked circles and susceptibility to OSM. These findings suggest that relatively early representations of objects can contribute to the size-averaging process despite their reduced visibility.

Statistical extraction affects visually guided action

“…There is much debate regarding the mechanisms underlying summary statistical representations. Although there is mounting evidence to suggest such statistical representations involve a calculation of the mean of the entire set of items without the need to encode individual set members (e.g., Ariely, 2001; Choo & Franconeri, 2010; Chong, et al, 2008; Chong & Treisman, 2005; Corbett & Oriet, 2011; Joo, et al, 2009), it has been argued that perceptual averaging can be accomplished by sampling only a handful of the items in each set using focused attention (e.g., Simons & Myczek, 2008; c.f., Ariely, 2008). We have previously used the same paradigm as in the present study to demonstrate that observers are sensitive to the variance in the sizes of the dots in the adapting sets (Corbett, et al, 2012), suggesting that most, if not all of the elements in each set are included in the calculation of the mean.…”

“…Our experience of stable, thorough perception may be accomplished by integrating occasional detailed samples of the visual world with statistical summaries of the remaining areas (Ariely, 2001), such that statistical, or ensemble representations act in complement to limited capacity attentional resources needed to represent individual objects in detail (e.g., Alvarez, 2011). Along these lines, average properties of sets are extracted automatically (e.g., Oriet & Brand, 2013) and more efficiently than individual object representations (e.g., Im & Halberda, 2013), when attention is distributed broadly across the visual field (e.g., Chong & Treisman, 2005), and even when individual elements cannot be represented (e.g., Corbett & Oriet, 2011; Joo, et al, 2009) or consciously perceived (e.g., Choo & Franconeri, 2010; Parkes, et al, 2001). In addition, ensemble representations persist across eye movements and transfer between different egocentric and allocentric frames of reference (Corbett & Melcher, 2014), providing further support for the proposal that the visual system relies on statistical summaries to efficiently represent large chunks of scenes.…”

Global statistical regularities modulate the speed of visual search in patients with focal attentional deficits

“…For example, when observers distributed attention broadly over a set of items, they were at chance to recall the sizes of individual members of the set, but still retained a precise representation of the set’s mean size (Chong and Treisman, 2005). Several other studies have demonstrated that even items that are masked (Choo and Franconeri, 2010), or crowded (Parkes et al, 2001) from conscious awareness, or items that are presented when focal attentional resources are not available for detailed encoding (Joo et al, 2009; Corbett and Oriet, 2011) are nonetheless included in the calculation of average size. Overall, empirical evidence is consistent with the existence of two qualitatively different processes: one devoted to processing individual objects in detail using focused attention (with a capacity of around 3–4 elements at a time), and the other involved in representing the more global properties of sets of similar objects and the overall gist or meaning of the scene.…”

Rapid averaging? Not so fast!

“…Introducing variability in trial means, however, leads only to a larger range of possible trial means, and even if trial means occur with equal frequency, observers may nevertheless acquire a representation of the mean of these trial means (i.e., cumulative mean; Crawford, Huttenlocher, & Engebretson, 2000), which could be used to aid judgments of the trial mean. This will obviously be of benefit if trial means are determined randomly and are normally distributed (Choo & Franconeri, 2010;Corbett & Oriet, 2010) because many trial means will be similar to the cumulative mean. Even if trial means are chosen with equal frequency and follow a rectangular distribution (Chong & Treisman, 2005a, b;de Fockert & Marchant, 2008), some trial means, by definition, will be similar to the cumulative mean.…”