If I fits I sits: A citizen science investigation into illusory contour susceptibility in domestic cats (Felis silvestris catus)

Highlights

• Domestic cats treat the Kanizsa contour illusion as they do real square contours by spontaneously sitting or standing inside.

• Utilizing cats’ natural behaviors is effective for the study of cat behavior and cognition.

• The citizen science paradigm is successful at conducting ecologically valid research into the cognition of domestic cats.

Abstract

A well-known phenomenon to cat owners is the tendency of their cats to sit in enclosed spaces such as boxes, laundry baskets, and even shape outlines taped on the floor. This investigative study asks whether domestic cats (Felis silvestris catus) are also susceptible to sitting in enclosures that are illusory in nature, utilizing cats’ attraction to box-like spaces to assess their perception of the Kanizsa square visual illusion. Carried out during the COVID-19 pandemic, this study randomly assigned citizen science participants Booklets of six randomized, counterbalanced daily stimuli to print out, prepare, and place on the floor in pairs. Owners observed and videorecorded their cats’ behavior with the stimuli and reported findings from home over the course of the six daily trials. This study ultimately reached over 500 pet cats and cat owners, and of those, 30 completed all of the study’s trials. Of these, nine cat subjects selected at least one stimulus by sitting within the contours (illusory or otherwise) with all limbs for at least three seconds. This study revealed that cats selected the Kanizsa illusion just as often as the square and more often than the control, indicating that domestic cats may treat the subjective Kanizsa contours as they do real contours. Given the drawbacks of citizen science projects such as participant attrition, future research would benefit from replicating this study in controlled settings. To the best of our knowledge, this investigation is the first of its kind in three regards: a citizen science study of cat cognition; a formal examination into cats’ attraction to 2D rather than 3D enclosures; and study into cats’ susceptibility to illusory contours in an ecologically relevant paradigm. This study demonstrates the potential of more ecologically valid study of pet cats, and more broadly provides an interesting new perspective into cat visual perception research.

Introduction

In 2017, cat fans took to Twitter to document their cats’ attraction to tight spaces by taping complete shape outlines on their floors and observing their cats sit inside, spurring over eighty-two-thousand retweets and trending hashtag #CatSquare (Fig. 1). Affectionately termed “if I fits I sits,” the urge to inhabit enclosed spaces is well-known to cat owners and has been documented to decrease stress in laboratory cats (Carlstead et al., 1993) and shelter cats given boxes in which to hide (Hawkins, 2005; Kry and Casey, 2007; Vinke et al., 2014). In fact, cats deprived of shelter resources like boxes will attempt to manufacture their own by hiding behind or underneath box-like objects like litter pans (Gourkow and Fraser, 2006). The reason for this behavior is still unknown but is clearly highly desirable.

Regardless of the reason for their attraction to enclosed spaces, this behavioral phenomenon proves to be an excellent tool to study the visual perception of shapes and contours in domestic cats. Neurological study of this phenomenon began in cats and found that these nonhuman animals’ (hereafter animals) retinal receptive fields are sensitive to contours along a luminance gradient (Redies et al., 1986). Contour comprehension is theorized to be evolutionarily critical in the understanding of physical objects and boundary interpolation (Kellman, 2003), and the study of visuo-cognitive phenomena such as illusion susceptibility offers a fascinating perspective into the effects of environmental pressures and life experience on vision (Kelley and Kelley, 2014).

The ability to perceive visual illusions is remarkable. One’s susceptibility to an illusion derives from the visual system processing an image’s features as the most likely physical reality based on learnt probabilities, a top-down incorporation of one’s preconceptions and past experience—even if it is not a true representation of reality (Gregory, 1997; Haber and Hershenson, 1973). In the real world, this processing is usually veridical or close to veridical (Palmer, 1999). Graphically rendered illusions are special cases where the same processing will result in a perception that deviates considerably from what is physically real.

The type of visual illusion considered here is subjective illusory contours, in which one mentally perceives fictitious contours connecting a shape’s inducers (modal completion) due to luminance contrast (Kanizsa, 1955). Study of the effects of age on susceptibility to the Kanizsa contour illusion in humans finds that illusory-contour perception may develop around 3–4 months and strengthens with age (Otsuka et al., 2004). Susceptibility to illusory contours has also been studied in a wide range of animal species in almost exclusively train-and-transfer testing paradigms comprising dogs (Byosiere et al., 2017), chimpanzees (Fagot and Tomonaga, 2001), bamboo sharks (Fuss et al., 2014), honeybees (Horridge et al., 1992), mice (Kanizsa et al., 1993), barn owls (Nieder and Wagner, 1999), redtail splitfin fish (Sovrano and Bisazza, 2009), and goldfish (Wyzisk and Neumeyer, 2007) (for a full reviews, see Byosiere et al., 2020; Feng et al., 2017; Kelley and Kelley, 2014; Nieder, 2002).

Previous research reveals that cats are, indeed, susceptible to certain visual illusions. De Weerd et al. (1990) found that domestic cats could discriminate illusory contour orientation via contour-inducing semicircles. In 2019, Szenczi et al. revealed that cats are susceptible to the size distorting Delboeuf illusion. Further, two studies found that both lions (Panthera leo) (Regaiolli et al., 2019) and domestic cats (Bååth et al., 2014) are susceptible to the Rotating Snake illusion, comprising a “moving” image caused by peripheral drift eliciting hunting-related behavior.

Perhaps most relevant, a study by Bravo et al. (1988) examined domestic cats’ susceptibility to subjective contours via operant response to the Kanizsa square illusion. Two young, female cats were trained to indicate where they viewed a subjective contour on an array of sectored disks in various orientations. The researchers controlled for other potential cues like luminance, temporal changes, and local patterns by introducing and modifying variables like motion and duration of stimuli exposure. They found that the cats demonstrated susceptibility to the Kanizsa illusion, indicating that cats likely perceive subjective contours as humans do (see Table 1 for a summary for illusion studies in cat species).

The present study supplements the results of Bravo et al.’s (1988) experiment with the addition of an increased sample size and a more inclusive sex and age range, in pet, rather than laboratory, cats. Moreover, rather than using standard operant conditioning procedures, the current study utilizes a more ecologically valid, real-world setting in which to evaluate spontaneous behavior. As cats transferred to novel environments can exhibit stress-related behaviors and thus not behave naturally (Amat et al., 2015), this study also offers an at-home environment to explore domestic cats’ susceptibility to Kanizsa square contours in a natural setting. Specifically, we evaluate whether cats will sit or stand within the contours of an illusory Kanizsa square more often than a control stimulus in a spontaneous choice task. Importantly, to date, cats’ attraction to enclosed spaces has been limited to 3D spaces (Carlstead et al., 1993; Gourkow and Fraser, 2006; Hawkins, 2005; Kry and Casey, 2007; Vinke et al., 2014), and thus this study also aims to formally examine the extension of this behavior to 2D shapes (such as that which was seen in the #CatSquare challenge). This study was conducted entirely remotely through citizen science engagement during the COVID-19 pandemic, and to the best of our knowledge, is the first published citizen science experiment to examine cat cognition.