Essay Are Animals Autistic Savants? Giorgio Vallortigara*, Allan Snyder, Gisela Kaplan, Patrick Bateson, Nicola S. Clayton, Lesley J. Rogers

imilarities in behaviour between stimuli, as opposed to categories and animals experience unique physical autistic savants and animals higher-level concepts, and can thereby and psychological conditions that can Shave been suggested, most make performance more savant- be stressful and even distressing, their notably by Temple Grandin [1] like in both humans and animals. mental processes may be quite different in her book Animals in Translation (Editors note: See Box 1 for Grandin’s from those of wild animals. We wish (2005), and this theory seems to have response.) to address the question of animals as gained some consensus among other autistic savants in the broad context neuroscientists who are not specialists Do Animals Think like Autistic of animal behaviour in the natural in animal cognition. Here we aim to Savants? environment. discuss two specific parallels between When Grandin proposed similarities A possible relationship between the behavioural characteristics of in cognition between autistic savants the thinking of autistics and that animals and those of autistic savants and animals, she reasoned on the basis of nonhuman animals has been that have been raised in relation to that animals, like autistic humans, suggested with respect to the so-called Grandin’s book. Autistic savants show sense and respond to stimuli that “theory of mind”, i.e., the ability of extraordinary skills, particularly in nonautistic humans usually overlook. members of our species to attribute music, mathematics, and drawing. In other words, animals respond to intentional (mental) states to others, Do animals sometimes show forms of and remember the details of the world and its absence or weaker expression extreme (though, of course, different) around them, whereas nonautistic in both autistics and animals. Whether cognitive skills confined to particular humans overlook the details in favour nonhuman animals have a fully domains that resemble those shown of the overall whole. Since Grandin’s developed theory of mind is still a by autistic savants? We argue that the book Animals in Translation shows controversial issue (for a review, see extraordinary cognitive feats shown extraordinary insight into both autism [4]) and one with which we are not by some animal species can be better in humans and animal welfare, the specifically concerned here. However, understood as adaptive specialisations question deserves scrutiny from impaired social behaviour is an aspect that bear little, if any, relationship to scientists working on animal cognition of autism in humans that is not shared the unusual skills shown by savants. and comparative neuroscience. by social, nonhuman animals. It has also been argued that autistic More recently, autism has been savants “think in detail”, and that this linked to a possible deficit in the is the key to their extraordinary skills. When Grandin proposed functioning of mirror neurons in Do animals have privileged access to similarities in cognition the cortex [5,6]. Because mirror lower level sensory information before neurons are active in matching actions it is packaged into concepts, as has by autistic savants and performed by others, even when those been argued for autistic humans, or do animals, she reasoned actions are merely observed, they are they process sensory inputs according on the basis that to rules that pre-empt or filter what Citation: Vallortigara G, Snyder A, Kaplan G, Bateson is perceived even at the lowest levels P, Clayton NS, et al. (2008) Are animals autistic animals, like autistic savants? PLoS Biol 6(2): e42. doi:10.1371/journal. of sensory processing? We argue that humans, sense and pbio.0060042 animals, like nonautistic humans, Copyright: © 2008 Vallortigara et al. This is an process sensory information according respond to stimuli that open-access article distributed under the terms to rules, and that this manner of nonautistic humans of the Creative Commons Attribution License, processing is a specialised feature of which permits unrestricted use, distribution, and reproduction in any medium, provided the the left hemisphere of the brain in usually overlook. original author and source are credited. both humans and nonhuman animals. Hence, we disagree with the claim that Put in terms of pathology, the Giorgio Vallortigara is at the Centre for Mind/Brain Sciences, University of Trento, Trento, Italy. Allan animals are similar to autistic savants. question does not make sense, Snyder is at the Centre for the Mind, University of However, we discuss the possibility that and Grandin was not referring to Sydney, Sydney, New South Wales, Australia. Gisela animal models that are used to study Kaplan and Lesley J. Rogers are at the Centre for manipulations that suppress activity Neuroscience and Animal Behaviour, University of of the left hemisphere and enhance autism (e.g., rodents treated with New England, Armidale, New South Wales, Australia. control by the right hemisphere shift pharmacological agents or raised Patrick Bateson is at the Sub-Department of Animal Behaviour, University of Cambridge, Cambridge, attention to the details of individual in restricted contexts that induce United Kingdom. Nicola S. Clayton is at the repetitive behaviour [2,3]). She Department of Experimental Psychology, University was referring to the animals she has of Cambridge, Cambridge, United Kingdom. Essays articulate a specific perspective on a topic of broad interest to scientists. studied, such as farm animals and * To whom correspondence should be addressed. domestic pets. Since many of these E-mail: [email protected]

PLoS Biology | www.plosbiology.org 0208 February 2008 | Volume 6 | Issue 2 | e42 seen as “a potential bridge between for example, can store and remember proximity tests; however, the ability to minds” [6] and certainly as essential to thousands of caches in the wild [11]. acquire new visuo-spatial information social behaviour. Since we know that Specific neurological adaptations was poorer in taxi drivers than in bus at least some nonhuman species have accompany these extraordinary drivers, apparently as a result of their mirror neurons (e.g., macaques [7]), behaviours: the hippocampus of food- showing a reduced anterior part of the the proposed impairment of mirror storing birds is enlarged compared hippocampus [19]. The extraordinary neurons as a core aspect of autism in with that of phylogenetically related ability to navigate using landmarks humans provides no support for the species of birds that have not acquired by the taxi drivers had come hypothesis that nonhuman animals developed caching abilities (reviewed at a price—acquiring new visuo-spatial may be autistic savants, at least as a in [12]). Other studies have shown information—unlike the extraordinary general characteristic of a species. This that the hippocampal enlargement cognitive adaptation of spatial memory does not preclude the possibility that, in food storers arises as a result of the in the Clark’s nutcracker, which was within a species, individual animals with specific memory demands required an addition without expense. Savant impaired mirror neuron functioning for accurately recovering previously abilities come at a cost to other aspects could be autistic. cached food [13,14]. As in the case of processing and, therefore, appear of autistic savants, the extraordinary to be unrelated to the extraordinary Savant abilities come at cognitive abilities of Clark’s species-specific adaptations seen in a cost in other aspects nutcrackers are limited to a specific some taxa. domain (that of spatial cognition). Certain animal species can of processing and, However, unlike autistic savants, reproduce sounds from sources other therefore, appear to Clark’s nutcrackers do not show than those that are species-specific, any impairment in other cognitive and do so in an extraordinarily be unrelated to the domains. The birds’ performance detailed fashion. In songbirds, this extraordinary species- when using object-specific cues (as ability to memorize and reproduce opposed to their extraordinary ability sounds depends on a set of nuclei in specific adaptations when using spatial-specific cues) is the forebrain [20]. For most humans seen in some taxa. most likely the same as that shown without any impairment, the ability to by non–food-storing species, and reproduce sounds tends to be limited Snyder and Mitchell [8] consider comparable to the performance of to some phrases. Usually, it requires the salient aspect of autistic savants to ordinary avian species [12]. This practice and repeated exposure to be their “privileged access” to lower- contrasts to autistic savants, who have reproduce a whole song or musical level sensory information before it is impaired cognitive abilities in domains passage, and to produce a complete packaged into holistic concepts and other than the one(s) in which they rendition without such repetition labels (see also [9]). Accordingly, the show extraordinary abilities [15,16]. is rare. Mozart was portrayed in the mental processes of autistic humans, In other words, the neurological film “Amadeus” as a musical savant which are less governed by rules and adaptation in food-storing birds—the because he was able to write down learned concepts that pre-empt and enlargement of their hippocampus— an entire concerto on the basis of filter the perceptions of nonautistic does not seem to be at the expense of having heard it once. Some birds may humans, are more literal and less reduced neural tissue in other parts be able to compete with this ability. categorising than those of nonautistics. of the brain [17]. It must be said, Australian magpies have been recorded This type of mental processing is often however, that it may be difficult to as mimicking a complex sequence seen as lower-order processing and locate a compensatory reduction of of sounds from a kookaburra duet hence as a possible link to the mental another brain region, and no research or even learning a whole song on processing abilities of nonhuman has focused on discovering this. a single exposure [21]. Sonograms animals. It is this interpretation that we Interestingly, domain-specific have revealed that the structure of the shall address. cognitive specialisations that result rendering was also extremely accurate. from extensive training do show some Whether or not one can compare Extraordinary Cognitive Skills form of neurological compensation: Mozart’s achievement with that of Certain animal species do show it has been reported that, in licensed a bird remains questionable, but it extraordinary skills in perceptual and London taxi drivers, an increase in the appears that the principle in each case motor behaviour when compared volume of the posterior hippocampus is the same: the entirety of the sound with humans or other animal species. is associated with a relative decrease sequence had been absorbed, encoded, Echolocation of bats or dolphins or in the anterior portion of the same and committed to memory well enough the motor performance of a hunting structure [18]. Fascinatingly, when that its faithful reproduction was feline are just two examples of such London taxi drivers were compared possible. It is not known in the case of extraordinary abilities; yet such skills with London bus drivers, matched magpies whether their ability leads to could be better characterized as for driving experience and levels of compensatory diminishment of other adaptive specialisations. The same stress but differing in that the latter areas in the brain. Their song system is probably true in a more-cognitive had to follow a constrained set of is the same as in other songbirds [22], domain. Food-storing birds show routes, the taxi drivers were found and their hyperpallium (referring to impressive memory capabilities to show better scores than the bus visual processing) is, in fact, especially (reviewed in [10]). Clark’s nutcrackers, drivers on London landmarks and large [23,24].

PLoS Biology | www.plosbiology.org 0209 February 2008 | Volume 6 | Issue 2 | e42 Do Animals Think in Details? are shown collections of a particular It has been argued that autistic people letter arranged to form a global think in details and have difficulty configuration that is also a letter (e.g., grouping attributes into holistic an “H” made up of “S”s). Identification pictures and forming labels. One responses are usually faster for the autistic boy, for example, learned the global configurations than for the local concept of a giraffe through a detail— letters. Some studies have pointed the pattern of the coat—not the long to a local precedence effect in both neck or overall shape. As a result, he pigeons [29] and nonhuman primates also identified a leopard as a giraffe [30–32]. However, more recent work [15,25–27]. has shown that local information does Grandin states animals think like not always take precedence over global autistic savants. Normal humans are information in nonhuman animals. good at seeing the big picture but For instance, in certain conditions, bad at what Grandin calls “all the tiny pigeons can acquire a global-feature little details that go into that picture”. categorisation more rapidly than a local-feature categorisation, exactly as For normal humans, the big picture is doi:10.1371/journal.pbio.0060042.g001 not created by accumulating sensory do humans [33]. It seems that the local- details, but rather by filtering out feature precedence observed in these Figure 1. The Kanizsa Triangle details [27]. Like autistic people—and studies simply results from the specific Animals do experience the illusion, although some species may be more inclined to attend unlike normal humans—animals, stimuli and procedures employed (e.g., the subtended visual angle and to details, of the pac-man before extracting Grandin argues (as does Snyder for the overall triangular shape (see text). autistic savants) have direct access to the density of local elements, see also the raw sensory data that an ordinary [34]). details and parts of stimuli and events, human brain would discard by filtering Similar arguments apply to the of higher-order categories and concepts out. Grandin also argues that humans celebrated studies (e.g., [35]) showing on the basis of past experience. These evolved larger association areas than that pigeons showed no decrement in mechanisms are attentive and affected nonhuman animals and that, in order recognizing stimuli that were spatially by overall knowledge and beliefs. They to avoid overloading these association scrambled, and therefore suggesting are top-down mechanisms. areas, access to raw sensory data should that pigeons’ discriminative responding As to the early processing be strictly limited. is controlled by local features alone. In mechanisms of segregation and contrast, more recent work has shown grouping, here sometimes differences that pigeons use both global and local between species can be observed, and Very generally, the left aspects of the stimuli, with different certain animals may indeed show an hemisphere sets up rules mixtures of these types of information inclination to attend to details in some based on experience and depending on the particular perceptual contexts, probably because of specific context [36]. For instance, pigeons adaptations. Pigeons, for instance, the right hemisphere trained to discriminate cartoon people when faced with illusory figures like avoids rules in order on a variety of background scenes the famous Kanizsa’s triangle [38] attend to global aspects of cartoon (Figure 1) show a tendency to attend to detect details and people; on the other hand, using line to the inducing pac-man rather than unique features that drawings of cartoon faces as stimuli and extracting the overall triangular figure. examining the effects of scrambling In a study on subjective contours allow it to decide what and deletion of components showed with Kanizsa’s triangles and squares, is familiar and what is that a set of components (eyes and Prior and Güntürkün [39] found that eyebrows) appear to exert strong only four out of 14 animals tested novel. This is true for control over behaviour and scrambling succeeded. As indicated by control human and nonhuman only moderately suppresses responding tests, pigeons responding to subjective [37]. contours were attending to the global animals, likely reflecting The mechanisms that determine pattern of the stimuli, whereas pigeons ancient evolutionary prevalence of processing details or not responding to subjective contours origins of the underlying the big picture seem to be similar in were attending to the local details human and nonhuman species. There of the stimuli. Pigeons also showed brain mechanisms. could be two senses in which “details” similar difficulties when dealing are assembled into a big picture. with recognition of partly occluded We believe, however, that there One is the operating of preattentive objects, to which they tend to respond is little evidence in support of these mechanisms of perceptual grouping literally to the visible parts instead of views. It is true that when humans to segregate figures from ground and mentally completing them [40,41]. It process visual stimuli, global parse objects in visual scenes. These is important to note, however, that this information often takes precedence mechanisms are largely bottom-up and clearly shows relative preference for over local information. For instance, data driven. The other is the extraction, pigeons to attend to some aspects of in Navon’s experiments [28], people from already segregated and parsed a visual scene rather than an absolute

PLoS Biology | www.plosbiology.org 0210 February 2008 | Volume 6 | Issue 2 | e42 of details for the formation of a big picture: a specialisation of the left hemisphere. We see what we expect. And what we expect is encapsulated in our mindsets—mental templates formed largely by experience. Brains are wired to present us with the best hypothesis, the best guess based on our prior experience, especially experience in early life [27,64]. But autistic savants, according to some conceptions, would be literal in the sense that, overwhelmed by details, they would be impaired in their ability to extract general, higher-order concepts on the basis of past experience. This means that they would have less access than nonautistics to left hemisphere processing. Animals, of course, cannot be autistic in this sense. They are not literal, but have mindsets allowing them to be fast at executing their lifestyle within their doi:10.1371/journal.pbio.0060042.g002 particular niche. Unlike autistic people, Figure 2. Lateralization in the Avian Brain, Illustrated Using the Australian Magpie, animals are not overwhelmed by detail. Gymnorhina tibicen Quite interestingly, they also show Behaviour controlled by the left and right hemispheres is listed. As in humans, functions of the left brain specialisations similar to those hemisphere are consistent with those of autistic humans. revealed in humans in terms of the (Photograph: G. Kaplan, Centre for Neuroscience and Animal Behaviour, University of New England, Australia) roles that the two cerebral hemispheres play in assembling categories and inability, because more recent evidence favour examination of fine stimulus thus discarding idiosyncratic details has revealed that pigeons are able both details (as demonstrated in use of the in favour of invariant properties when to see subjective contours [42] and frontal, binocular field in selecting analyzing stimuli and events (reviewed to complete partly occluded objects between food types [52]) and may be in [58,65–67]). [43,44]. Also, this is clearly not a responsible for the local advantages The savant syndrome is often general characteristic of the avian brain observed in most experiments that used associated with some left brain as opposed to the mammalian brain, frontal presentations of visual stimuli. dysfunction, together with right brain because other species of birds show no The lateral visual fields, in contrast, compensation [68]. Snyder has shown attributes similar to those of pigeons may be more concerned with the larger [69–73] that it is possible to induce (e.g., domestic chicks, zebra finches scale integration of scene and flight savant-like skills in normal healthy and owls; see [45–49]). control information [53], thus showing people by inhibiting the left anterior It seems likely that there could more sensitivity to global information temporal lobe with magnetic pulses be ecological differences favouring (see [54]). at the rate of 1 pulse per second—a perception based on response to Specialisation of the left and right process known as transcranial magnetic parts or reconstruction of whole sides of the brain (lateralisation) may stimulation, or TMS. But why does this objects on the basis of interpolation be instrumental in attending to detail inhibition of the left anterior temporal between these parts. Linking function or the big picture in the first of the two lobe induce savant-like skills? to structure, it is interesting to senses mentioned above. As in humans Gazzaniga [74] refers to the observe that in experiments using [55–57], gestalt-like properties of visual function of the left hemisphere as the conditioning procedures, such as scenes seem to be associated with “interpreter”. The left hemisphere those performed with pigeons, the right hemispheric function in animals tends to create a belief system stimuli were presented in the frontal (reviewed in [58–60]. In chicks, or mindset and to fold any new binocular visual field of the animals, amodal completion has been proven information into the existing belief a portion of the visual field that to be more likely to occur under right system. The hemispheres are in is mainly represented within the hemispheric control [61], exactly as competition with one another, with the tectofugal pathway in pigeons [50,51]. it is in humans [62]. Moreover, the hypothesis-imposing left hemisphere The frontal visual field seems to be right hemisphere of the chick attends dominating or inhibiting the right specialized for (myopic) foraging for to global, spatial cues as opposed hemisphere, which is more open to food on the ground, whereas the lateral to landmark cues, to which the left novelty [27]. The right hemisphere, visual field seems to be specialized hemisphere attends [63]. on the other hand, tends to play, as for predator detection and flight There is however another, more Ramachandran [75] has put it, the role control. Near-sighted acuity would top-down, route to the integration of the “Devil’s advocate”, looking for

PLoS Biology | www.plosbiology.org 0211 February 2008 | Volume 6 | Issue 2 | e42 novelty and inconsistencies. Because of details (see Figure 2). (Note that by pebbles, each as a separate and novel its duty to recognize novelty, the right attention to “details” here, we do entity [81]. hemisphere is forced to maintain a not mean raw sensory information, A general theoretical framework more literal record of the events. but rather the information on for these findings has been proposed, It is quite fascinating that similar parts available after applying based on the idea of “functional specialisations of the two hemispheres gestalt mechanisms of perceptual incompatibility” [67]. When assessing have been reported for the nonhuman processing at the preattentive level.) a novel stimulus—an event likely to be animal brain (reviewed in [76]). In the A noticeable example is the role faced quite commonly even by the most domestic chick, for example, which of the two hemispheres in social primitive vertebrates—an organism is probably the most investigated recognition associated with imprinting, must carry out two different types of animal species with respect to brain in which the left hemisphere builds analyses. First, it must rely on previous lateralisation [65,77,78], evidence up a representation of the general comparable experiences to estimate the has been collected that the right properties of the social partner degree of novelty of the stimulus, and hemisphere mainly attends to novelty (without including in it individual to do so it must recall stored memories and focuses on those unique and recognition), whereas the right and then elaborate on them for future idiosyncratic properties of visual stimuli hemisphere encodes the peculiar and use. Secondly, certain appropriate that may help to establish whether a unique properties of a social partner cues, based on past experience (or on particular stimulus is or is not the same as an individual [79, 80]. Similarly, phylogenetically based information), as the one encountered before. The when presented with grains scattered must be used to try to assign the left hemisphere, in contrast, tends to on a background of inedible pebbles, stimulus to a category, and so to decide attribute stimuli to categories in order a chick using its left hemisphere what sort of response (if any) should to provide fast and efficient responses. categorises the array of stimuli as food/ be given. Categorisation must be In so doing, the left hemisphere tends nonfood and avoids pecking at the made on the basis of selected stimulus to attend to the general characteristics pebbles, whereas a chick using the right properties, despite variation in many of the stimuli, rather than to the hemisphere pecks at both grain and other properties. All this is reminiscent

Box 1. Response by Temple Grandin to the Essay “Are Animals Autistic Savants?” Editors note: We asked Temple Grandin to respond to the essay “Are I was most interested to learn that the skills that taxi drivers Animals Autistic Savants?”, which presents a critique, based on studies had developed to navigate by using land markers “had come at of comparative animal cognition, of Grandin’s premise that animals and the price of acquiring new visual spatial information” according autistic savants share similarities in cognition. to the authors. Further experiments need to be done with birds to either confirm or disprove Vallortigara et al.’s hypothesis The Essay “Are Animals Autistic Savants?” by Giorgio that birds such as the Clark’s nutcracker, which has savant-like Vallortigara et al. provides a fascinating overview of the memory for food storage, has retained good cognition in other most recent research on animal cognition. I think the basic domains. My hypothesis is that birds that have savant-like skills disagreement between the authors and me arises from the for food storage sites or remembering migration routes may concept of details—specifically how details are perceived by be less flexible in their cognition. It is well known that people humans, who think in language, compared with animals, who with autism do poorly on the Wisconsin card sorting task, where think in sensory-based data. Since animals do not have verbal colors and shapes have to be sorted into different categories. language, they have to store memories as pictures, sounds, or The person with autism is slower to respond correctly when the other sensory impressions. Sensory-based information by its very category is switched from a shape to a color. nature is more detailed than word-based memories. As a person An operant conditioning task could be used to compare with autism, all my thoughts are in photo-realistic pictures. I can flexible problem solving in migratory and food-storing bird search my own brain, like using Google, for images. As I read species to species that do not have savant-like skills. The task about the cognition experiments, I saw the birds performing could be to peck a lever when a light comes on. After the bird in my imagination like a virtual reality computer system. The had learned this, the stimulus would be switched to a sound. I main similarity between animal thought and my thought is the predict that the more savant-like birds will require more trials to lack of verbal language. Verbal language narrates the images switch back and forth between the light and the sound cue. when I “surf the Internet” inside my own brain. If you give me I am pleased that my book has stimulated so much a “keyword,” such as “peanut,” I start to see images like a series discussion, and I hope it stimulates more research on animal of slides shown one after the other. The first image was the cognition. Planter’s Peanut logo, the second was a Western restaurant that serves peanuts, and the third was a bag of peanuts on a plane. References The Essay by Vallortigara et al. clearly showed that cognition 1. Miller BL, Boone K, Cummings JL, Read SL, Mishkin F (2000) Functional correlates of musical and visual ability in frontal temporal dementia. Br J Psych is very real in animals. In normal humans, higher brain processes 176: 458-463. cover up the sensory-based processing that we share with 2. Miller BL, Cummings J, Mishkin F, Boone K, Prince F, et al. (1998) Emergence of animals. In Animals in Translation, I discussed the work of art talent in frontal temporal dementia. Neurology 51: 978-981. Bruce Miller, who studied patients with frontal-temporal lobe dementia [1,2]. As the disease destroyed higher brain function, Temple Grandin is the author of Animals in Translation and is Professor of Animal art and music talent emerged. The sensory-based, more-detailed Science, Colorado State University, Fort Collins, Colorado, United States of America. thinking we share with animals was unmasked. E-mail: [email protected]

PLoS Biology | www.plosbiology.org 0212 February 2008 | Volume 6 | Issue 2 | e42 of the functional incompatibility References Melbourne: Natural History Series, University among logical demands, a condition 1. Temple G (with Johnson C) (2005) Animals of New South Wales Press, Sydney and CSIRO. in translation. Using the mysteries of autism to 24. Kaplan G (2008) Alarm calls and referentiality hypothesized to underlie the evolution decode animal behavior. New York: Scribner in Australian Magpies: between midbrain and of multiple memory systems by Sherry Book Company 368 p. ISBN 0-7432-4769-8. forebrain, can a case be made for complex 2. Lewis MH, Tanimura Y, Lee LW, Bodfish JW cognition? Brain Research Bull, In press. and Schachter [82]. To categorise (2007) Animal models of restricted repetitive 25. Rimland B (1963) Infantile autism. New York: events (or stimuli), the organism behavior autism. Behav Brain Res 176: 66-74. Appleton-Century-Crofts. must recognise, and memorise, 3. Klauck SM, Poutska A. (2007) Animal models 26. De Clercq H (2003) Mum, is that a human of autism. Drug Discovery Today: Disease being or an animal? Bristol (United Kindgom): those features of an experience (or Models 3: 313-318. Lucky Duck Publishing. of a stimulus) that recur in different 4. Tomasello M, Call J, Hare B (2003) 27. Snyder A, Bossomaier T, Mitchell DJ (2004) episodes (or stimuli) and, at the same Chimpanzees understand psychological states – Concept formation: ‘Object’ attributes the question is which ones and to what extent. dynamically inhibited from conscious time, ignore or discard unique and Trends Cogn Sci 7: 153-156. awareness. J Integ Neurosci 3: 19-34 idiosyncratic features that do not recur 5. Iacoboni M, Dapretto M (2006) The mirror 28. Navon D (1977) Forest before trees: the neuron system and the consequences of its precedence of global features in visual and thus are not essential to learning. dysfunction. Nature Rev Neurosci 7: 942-951. perception. Cogn Psychol 9:353–383. The selective attention that results is 6. Williams JHG, Whiten A, Suddendorf T, Perrett 29. Cavoto KK, Cook RG (2001) Cognitive one of the brain’s main functions, as DI (2001) Imitation, mirror neurons and precedence for local information in autism. Neurosci Biobehav Rev 25: 287-295. hierarchical stimulus processing by pigeons. J it enables the smooth, and eventually 7. Rizzolatti G, Fadiga L, Gallese V, Fogassi L Exp Psychol Anim Behav Process 27:3–16. automatic, execution of skilled motor (1996) Premotor cortex and the recognition of 30. Fagot J, Deruelle C (1997) Processing of global behaviour, performed in response to motor actions. Cogn Brain Res 3: 131-141. and local visual information and hemispheric 8. Snyder A, Mitchell D J (1999) Is integer specialization in humans (Homo sapiens) and certain invariant features of episodes. arithmetic fundamental to mental processing? baboons (Papio papio). J Exp Psychol Hum In contrast, to detect novelty and to The mind’s secret arithmetic. Proc R Soc Lond Percept Perform 32:429–442 Ser B 266: 587-592. 31. Fagot J, Tomonaga M (1999) Global and local build up a detailed record of episodic 9. Humphrey N (1998) Cave art, autism and the processing in humans (Homo sapiens) and experiences, the organism must attend evolution of the human mind. Cambridge chimpanzees (Pan troglodytes): use of a visual to the contextual details that mark Archaeol J 8: 165-191. search task with compound stimuli. J Comp 10. De Kort SR, Tebbich JM, Dally NJ, Emery Psychol 113:3–12 individual experiences uniquely, i.e., NJ, Clayton NS (2006) The comparative 32. Spinozzi G, De Lillo C, Truppa V (2003) Global to recognize variance across episodes cognition of caching. In Comparative and local processing of hierarchical visual rather than invariance. Processing cognition. Experimental explorations of anima stimuli in tufted capuchin monkeys (Cebus intelligence, Wasserman EA, Zentall TR, Eds., apella). J Comp Psychol 117:15–23 of information about invariance and Oxford: Oxford University Press. pp 602-618. 33. Goto K, Wills AJ, Lea SEG (2004) Global- variance between experiences requires 11. Vander Wall SB (1982) An experimental feature classification can be acquired more analysis of cache recovery in Clark’s nutcracker. rapidly than local-feature classification in both the use of mutually incompatible Anim Behav 30: 84-94. humans and pigeons Anim Cogn 7: 109–113. processes that might best be handled by 12. Shettleworth SJ (2003) Memory and 34. Hopkins WD, Washburn DA (2002) Matching functionally separate systems. hippocampal specialization in food-storing visual stimuli on the basis of global and local birds: Challenges for research on comparative features by chimpanzees (Pan troglodytes) and Overall, and very generally, the left cognition. Brain Behav Evol 62: 108-116. rhesus monkeys (Macaca mulatta). Anim Cogn hemisphere sets up rules based on 13. Clayton N S (1995) Development of memory 5: 27–31. experience, and the right hemisphere and the hippocampus: comparison of food- 35. Cerella J (1980) The pigeon’s analysis of storing and non-storing birds on a one-trial pictures. Pattern Recogn 12: 1-6. avoids rules in order to detect details associative memory task. J Neurosci 15: 2796- 36. Kirkpatrick-Steger K, Wasserman EA, and unique features that allow it to 2807. Biederman I (1998) Effects of geon deletion, 14. Clayton NS, Krebs JR (1994) Hippocampal scrambling, and movement on picture decide what is familiar and what is growth and attrition in birds affected by recognition in pigeons. J Exp Psychol Anim novel. This is true for human and experience. Proc Natl Acad Sci U S A 91: 7410- Behav Proc 24: 34–46. nonhuman animals, likely reflecting 7414. 37. Matsukawa A, Inoue S, Jitsumori M (2004) 15. Frith U, Hill E (2003) Autism: mind and brain. Pigeon’s recognition of cartoons: effects of ancient evolutionary origins of the Oxford: Oxford University Press. p. 9. fragmentation, scrambling, and deletion of underlying brain mechanisms. 16. Treffert DA (2005) The savant syndrome elements. Behav Process 65: 25-34. None of the evidence mentioned in autistic disorder, Recent developments 38. Kanizsa G (1979) Organization in vision. New in autism research. Cassanova MF, Ed. York: Praeger. above favours the idea that animals Nauppauge (New York): Science Publishers. 39. Prior H, Güntürkün O (1999) Patterns of visual attend to details more than humans pp. 27-55. lateralization in pigeons: seeing what is there 17. Harvey PH, Krebs JR (1990) Comparing brains. and beyond. Perception, Suppl 28: 22. do. The balance between attending Science 249: 140-146. 40. Fujita K (2001) Perceptual completion in to details or the big picture may, of 18. Maguire EA, Gadian DG, Johnsrude IS, Good rhesus monkeys (Macaca mulatta) and pigeons course, vary between species and CD, Ashburner J, et al. (2000) Navigation- (Columba livia). Percep Psychophysics 63: 115- related structural change in the hippocampi 125. within the same species depending on of taxi drivers. Proc Natl Acad Sci U S A 97: 41. Sekuler AB, Lee JAJ, Shettleworth SJ (1996) ecological constraints and perceptual 4398–4403. Pigeons do not complete partly occluded contexts, but it is performed by 19. Maguire EA, Woollett K, Spiers HJ (2008) figures. Perception 25: 1109-1120. London taxi drivers and bus drivers: a 42. Niu Y-Q, Xiao Q, Liu R-F, Wu L-Q, Wang S-R similar brain mechanisms, without any structural MRI and neuropsychological (2006) Response characteristics of the pigeon’s clear-cut divide between human and analysis. Hippocampus, In press. pretectal neurons to illusory contours and 20. Zeigler PP, Marler P (2008) The neuroscience motion. J Physiol 577: 805–813. nonhuman species. ◼ of birdsong. Cambride (United Kingdom): 43. Nagasaka Y, Hori K, Osada Y (2005) Perceptual Cambridge University Press. grouping in pigeons. Perception 34: 625-632. Acknowledgments 21. Kaplan G (2000) Song structure and 44. Di Pietro NT, Wasserman EA, Young ME The ideas presented in this paper arose function of mimicry in the Australian magpie (2002) Effects of occlusion on pigeon’s visual (Gymnorhina tibicen) compared to the lyrebird object recognition. Perception 31:1299-1312. out of a workshop on “Higher Cognition (Menura ssp.). Int J Comp Psychol 12: 219-241. 45. Forkman B, Vallortigara G (1999) in Animals” funded by the Forum 22. Deng C, Kaplan G and Rogers LJ (2001) Minimization of modal contours: an essential for European-Australian Science and Similarity of the song control nuclei of male cross-species strategy in disambiguating relative Technology and held in New South Wales, and female Australian magpies (Gymnorhina depth. Anim Cogn 2: 181--185. tibicen). Behav Brain Res 123: 89-102. 46. Nieder A, Wagner H (1999) Perception and Australia, in May 2007, with support from 23. Kaplan G (2004) Australian magpie: biology neuronal coding of subjective contours in the the Australian Academy of Science. and behaviour of an unusual songbird. owl. Nature Neurosci 2:660–663.

PLoS Biology | www.plosbiology.org 0213 February 2008 | Volume 6 | Issue 2 | e42 47. Okanoya K, Takahashi M (2000) Shikaku- development. In Slater PJB, Rosenblatt J, 69. Snyder A (2001) Paradox of the savant mind. teki hokan e no seitaigaku-teki apuroochi Snowdon C, Roper T, editors. Advances in the Nature 413: 251-252. [Ecological approach to visual completion]. study of behavior 31. pp. 107-162. 70. Snyder A (2004) Autistic genius? Nature 428: Kokoro no hattatsu: ninchi-teki seicho no 59. Vallortigara G. (2004) Visual cognition and 470-471. kikoo 1999 (Reports of the Grant-in-aid for representation in birds and primates. In: 71. Snyder A, Thomas M (1997) Autistic artists give Scientific Research for Priority Areas). pp. Vertebrate comparative cognition: are primates clues to cognition. Perception 26: 93-96. 34–41. superior to non-primates? Rogers LJ, Kaplan G, 72. Snyder A, Mulcahy E, Taylor JL, Mitchell D 48. Regolin L, Vallortigara G (1995) Perception of editors. New York: Kluwer Academic/Plenum J, Sachdev P, et al. (2003) Savant-like skills partly occluded objects by young chicks. Percep Publishers. pp 57–94. exposed in normal people by suppressing the Psychophysics 57: 971-976. 60. Vallortigara G (2006) The cognitive left front-temporal lobe. J Integ Neurosci 2: 49. Zanforlin M (1981) Visual perception of chicken: visual and spatial cognition in a 149-158. complex forms (anomalous surfaces) in chicks. non-mammalian brain. In: Comparative 73. Snyder A, Bahramali H, Hawker T, Mitchell D J Ital J Psychol 1:1–16. cognition: experimental explorations of animal (2006) Savant-like numerosity skills revealed in 50. Güntürkün O, Hahmann U (1999) Functional intelligence. Wasserman EA, Zentall TR, normal people by magnetic pulses. Perception subdivisions of the ascending visual pathways in editors. Oxford: Oxford University Press. pp. 35: 837-845. the pigeon. Behav Brain Res 98: 193-201. 41-58. 74. Gazzaniga MS (1998) The mind’s past. 51. Hellmann B, Güntürkün O (1999) Visual field 61. Regolin L, Marconato F, Vallortigara G (2004). Berkeley: University of California Press. specific heterogeneity within the tectofugal Hemispheric differences in the recognition 75. Ramachandran VS, Blakeslee S (1998) projection of the pigeon. Eur J Neurosci 11: of partly occluded objects by newly-hatched Phantoms in the brain. London: Fourth Estate 1-18. domestic chicks (Gallus gallus). Anim Cogn 7: Limited. 52. Prior H, Wilzeck C (2008) Selective feeding 162-170. 76. Vallortigara G, Rogers LJ (2005) Survival in birds depends on combined processing 62. Corballis PM, Fendrich R, Shapley RM, with an asymmetrical brain: advantages and in the left and right brain hemisphere. Gazzaniga M (1999) Illusory contour disadvantages of cerebral lateralisation. Behav Neuropsychologia., In press. perception and amodal boundary completion: Brain Sci 28: 575-633. 53. Martinoya C, Rivaud S, Bloch S (1984) evidence of a dissociation following 77. Andrew RJ (1991) The nature of behavioural Comparing frontal and lateral viewing in callosotomy. J Cogn Neurosci 11:459–466. lateralization in the chick. In: Neural and pigeons: II. Velocity thresholds for movement 63. Tommasi L, Vallortigara G (2001) Encoding behavioural plasticity: The use of the chick as discrimination. Behav Brain Res 8: 375-385. of geometric and landmark information in the a model. Oxford, Oxford University Press. pp. 54. Cook RG (2001) Hierarchical stimulus left and right hemispheres of the avian brain. 536-554. processing by pigeons. In Cook RG, editor, Behav Neurosci 115: 602-613. 78. Rogers LJ (1997) Minds of their own. St. Avian visual cognition. Available: http://www. 64. Blakemore C, Cooper GF (1970) Development Leonards (Australia): Allen & Unwin. pigeon.psy.tufts.edu/avc/cook/. Accessed 15 of the brain depends on the visual 79. Vallortigara G, Andrew RJ (1991) January 2008. environment. Nature 228: 477-478. Lateralization of response by chicks to change 55. De Renzi E, Spinnler H (1966) Visual 65. Rogers LJ, Andrew RJ, editors (2002) in a model partner. Anim Behav 41: 187-194. recognition in patients with unilateral cerebral Comparative vertebrate lateralization. 80. Deng C, Rogers LJ (2002) Social recognition disease. J Nerv Ment Dzs 142: 515-525. Cambridge (United Kingdom): Cambridge and approach in the chick: lateralization and 56. Levy-Agresti J, Sperry RW (1968) Differential University Press. effect of visual experience. Anim Behav 63: perceptual capacities in major and minor 66. Vallortigara G (2000) Comparative 697-706. hemispheres. Proc Natl Acad Sci U S A neuropsychology of the dual brain: A stroll 81. Rogers LJ, Andrew RJ, Johnston ANB (2007) 61:1151. through left and right animals’ perceptual Light experience and the development of 57. Wasserstein J, Zappulla R, Rosen J, Gerstman L, worlds. Brain Lang 73: 189-219. behavioural lateralization in chicks. III. Rock D (1987) In search of closure: subjective 67. Vallortigara G, Rogers LJ, Bisazza A (1999) Learning to distinguish pebbles from grain. contour illusions, Gestalt completion tests, and Possible evolutionary origins of cognitive brain Behav Brain Res 177: 61-69. implications. Brain Cogn 6:1-14. lateralization. Brain Res Rev 30: 164-175. 82. Sherry DF Schachter DL (1987) The evolution 58. Rogers LJ (2002) Lateralization in vertebrates: 68. Treffert DA Christensen DD (2005) Inside the of multiple memory systems. Psychol Rev 94: Its early evolution, general pattern and mind of a savant. Sci Amer 293: 6. 439-454.

PLoS Biology | www.plosbiology.org 0214 February 2008 | Volume 6 | Issue 2 | e42