Downloaded from http://rspb.royalsocietypublishing.org/ on December 17, 2014

rspb.royalsocietypublishing.org

Research Cite this article: Avargue`s-Weber A, Dyer AG, Ferrah N, Giurfa M. 2015 The forest or the trees: preference for global over local image processing is reversed by prior experience in honeybees. Proc. R. Soc. B 282: 20142384. http://dx.doi.org/10.1098/rspb.2014.2384

Received: 2 October 2014 Accepted: 11 November 2014

Subject Areas: cognition, behaviour, neuroscience Keywords: Apis mellifera, attention, honeybee, perceptual grouping, priming, vision

Author for correspondence: Aurore Avargue`s-Weber e-mail: [email protected]

The forest or the trees: preference for global over local image processing is reversed by prior experience in honeybees Aurore Avargue`s-Weber1,2,†, Adrian G. Dyer3,4,†, Noha Ferrah3 and Martin Giurfa1,2 1 Centre de Recherches sur la Cognition Animale, Universite´ de Toulouse; UPS, 118 route de Narbonne, Toulouse Cedex 9 31062, France 2 Centre de Recherches sur la Cognition Animale, CNRS, 118 route de Narbonne, Toulouse Cedex 9 31062, France 3 Department of Physiology, Monash University, Clayton, Victoria 3800, Australia 4 School of Media and Communication, Royal Melbourne Institute of Technology, Melbourne, Victoria 3000, Australia

Traditional models of insect vision have assumed that insects are only capable of low-level analysis of local cues and are incapable of global, holistic perception. However, recent studies on honeybee (Apis mellifera) vision have refuted this view by showing that this insect also processes complex visual information by using spatial configurations or relational rules. In the light of these findings, we asked whether bees prioritize global configurations or local cues by setting these two levels of image analysis in competition. We trained individual freeflying honeybees to discriminate hierarchical visual stimuli within a Y-maze and tested bees with novel stimuli in which local and/or global cues were manipulated. We demonstrate that even when local information is accessible, bees prefer global information, thus relying mainly on the object’s spatial configuration rather than on elemental, local information. This preference can be reversed if bees are pre-trained to discriminate isolated local cues. In this case, bees prefer the hierarchical stimuli with the local elements previously primed even if they build an incorrect global configuration. Pre-training with local cues induces a generic attentional bias towards any local elements as local information is prioritized in the test, even if the local cues used in the test are different from the pre-trained ones. Our results thus underline the plasticity of visual processing in insects and provide new insights for the comparative analysis of visual recognition in humans and animals.

1. Introduction



These authors contributed equally to this study. Electronic supplementary material is available at http://dx.doi.org/10.1098/rspb.2014.2384 or via http://rspb.royalsocietypublishing.org.

An enduring question in the study of visual perception is whether brains first analyse the details of a scene using local features, or do they prioritize global percepts integrated associated elements [1–5]? This question was raised 37 years ago by Navon in his seminal study on human perception [6], which concluded that global analysis predominates over local analysis. Not surprisingly, therefore, he entitled his work ‘Forest before trees’. Since then, many studies have confirmed the precedence of global over local processing in humans [6–13]. For instance, human subjects show faster response times in global versus local feature discrimination [7,9]. In addition, it is difficult for human subjects to ignore global configurations to process local cues (global-to-local interference), while local cues do not interfere with the recognition of global configurations [6]. The question of whether, besides humans, different species exhibit similar biases, is important to understand evolutionary and physiological constraints of visual object recognition [14]. Hierarchical compound stimuli inspired from Navon’s studies [6,7,15,16] provide experimental access to the analysis of global versus local processing. These stimuli offer two dissociated levels of analysis: a local featural level composed of small elements and a global configurational

& 2014 The Author(s) Published by the Royal Society. All rights reserved.

Downloaded from http://rspb.royalsocietypublishing.org/ on December 17, 2014

(a)

(b)

training

2 *

* *

*

80 S+

local test % of choices

learning test

S–

G– / L–

global test

G+ / L+

40

G+ / L–

conflict test

20

0

G+ / L+ G– / L–

learning test G+ / L+

G– / L+

G+ / L–

G+ / L+ G+ / L–

G+ / L+ G– / L+

local test

global test

G+ / L– G– / L+

conflict test

G– / L+

Figure 1. Global preference in visual hierarchical stimuli processing in honeybees. (a) Training and testing procedure. Each particular bee was trained with a different pair of hierarchical stimuli Sþ and S2 displaying a global shape (G) made of local shapes (L). These stimuli were randomly chosen from the set of stimuli presented in the electronic supplementary material, figure S1A. Thereafter, four non-reinforced tests were performed: a learning test, opposing the trained stimuli (GþLþ versus G2L2); a local test, leaving local information as the only discrimination cue (GþLþ versus GþL2); a global test, leaving global information as the only discrimination cue (GþLþ versus G2Lþ); and a conflict test opposing both kinds of cues (GþL2 versus G2Lþ). (b) Performance in the nonrewarded tests (% of first choices). The dashed line indicates chance level. Bees learned to discriminate the rewarded stimulus Sþ (GþLþ) from the non-rewarded stimulus S2 (G2L2). They accessed and used both the local and global cues as shown by the local and global tests, respectively, and preferred the global information in the conflict situation (conflict test). Data shown are means þ s.e. (n ¼ 13; *p , 0.05). level created by the spatial arrangement of local features (see examples in figure 1a). Studies in animals have reported that contrary to humans, species such as pigeons and monkeys exhibit local over global predominance in visual processing (baboons [9,10]; capuchins [17]; rhesus monkeys [18]; chimpanzees [11,18]; pigeons [19–23]). The causes for this difference between humans and the non-human animals studied so far remain unknown. Increasing the spectrum of species studied is required to understand the evolutionary bases of global versus local image processing. The honeybee Apis mellifera is a privileged invertebrate model for the study of visual perception [24–28]. Bees perceive and discriminate various types of visual cues and exhibit sophisticated visually driven behaviours [25,28,29]. A traditional view of insect vision posits that bees perceive only low-level, local visual cues and are incapable of global, holistic perception [30,31] (but see [26,27,29,32]). However, recent evidence suggests that after extensive training, bees use spatial configurations to recognize images of human faces [33–35] or more abstract patterns [36,37]. These results suggest that bees integrate distinct local features and their spatial interrelationships in a global representation, and they use such representations to categorize global images correctly [36,37]. Do bees, like humans, exhibit also precedence of global over local cues in visual pattern processing? Or do they prioritize local analyses? Where does the forest stand with respect to the trees in the case of the honeybee? In asking these questions, we should consider that bees may prioritize global or local analysis depending on the visual task. Perceptual strategies may differ with individual experience [38], especially if visually driven behaviours are highly plastic as in the honeybee [27,29,39]. It has been suggested that bees trained to discriminate larger bars (horizontal versus vertical) made of finer diagonal bars (458 versus 1358) prioritize global over local cues at a visual angle at which both cues are perceived [40]. Yet this

conclusion was based on tests in which a previously rewarded local orientation was set in opposition to a previously rewarded global orientation and which yielded in all cases non-significant results [40]. Whether bees prioritize global or local visual cues when both are at their disposal remains, therefore, an open question. Here, we used hierarchical compound visual stimuli [6] in which both global and local cues were perceived by bees to determine whether there is a preference for either local or global information in visual pattern recognition. We analysed whether preferences are insensitive to variations in individual experience or plastic and variable, so that the dominance of global or local cues can be adjusted depending on prior experience. In doing this, we discuss the visual strategies of bees, humans and non-human vertebrates, in order to uncover the general principles by which the species considered may achieve efficient visual processing in their respective ecological niches.

2. Material and methods (a) General procedure Free-flying honeybees (A. mellifera) were individually marked and trained one at a time to collect 1 M sucrose solution in a Y-maze [41]. The maze was covered by an ultraviolet-transparent Plexiglas ceiling and illuminated by natural daylight. The entrance of the maze led to a decision chamber, where a bee could choose between one of two arms. The hierarchic compound stimuli were presented on the back walls of the maze (20  20 cm), which were placed at a distance of 20 cm from the decision chamber. The training consisted of a differential conditioning procedure where one stimulus (Sþ) delivered a reward of sucrose solution 1 M in its centre, while the alternative stimulus (S2) offered only plain water [36,41]. If the choice of the bee was correct, it was allowed to drink sucrose solution ad libitum and then return to the hive. If the choice of the bee was incorrect, it was gently removed from the maze so that it

Proc. R. Soc. B 282: 20142384

G+ / L+

60

rspb.royalsocietypublishing.org

100

Downloaded from http://rspb.royalsocietypublishing.org/ on December 17, 2014

Hierarchical compound stimuli were constructed by arranging several identical local geometric shapes (local elements, L) into a larger geometric shape (global shape, G) (see example in figure 1a and the electronic supplementary material, S1). Local and global shapes were geometrical shapes (electronic supplementary material, figure S1). Within one particular stimulus, the local and global shapes were different [6,7,15,16]. Local shapes were 1 cm wide, thus subtending a visual angle of 38 from the decision chamber, while global arrangements were 13 cm wide (368). Prior analyses of the bees’ capacity to discriminate black-and-white gratings showed that spatial frequencies higher than 0.3 cycles deg21 cannot be resolved [42]. On the contrary, a lower frequency of 0.2 cycles deg21 determines a discrimination level of 75% [42]. These values correspond to visual angles of 38 and 58, respectively, subtended by a grating period. In other words, for a single bar (half period), corresponding to our display of unique elements instead of gratings, a visual angle of 1.58 and 2.58 would result in absence of discrimination and in discrimination, respectively. As our local elements subtended a visual angle of 38, all visual elements were above the discrimination threshold considering bee visual acuity [42] and performances in discrimination tasks involving geometric shapes [43–45]. Specific controls (local tests) were included to determine whether local cues were indeed perceived and learned by the bees.

(c) Experiment 1 In experiment 1, bees (n ¼ 13) were trained to discriminate a pair of hierarchical compound stimuli (Sþ and S2) that differed in both their global and local information. Sþ and S2 differed between bees in order to prevent potential shape preferences from influencing the results (electronic supplementary material, figure S1A; see example in figure 1a). After 36 training trials (i.e. 36 foraging bouts between the hive and the Y-maze), the bee was subjected to four non-rewarding tests offering a choice between (i) the training stimuli (Sþ and S2; figure 1a, ‘learning test’); (ii) Sþ versus a stimulus presenting local features of S2 in the Sþ global configuration (GþLþ versus GþL2; figure 1a, ‘local test’); (iii) Sþ versus a stimulus presenting local features of Sþ in the S2 configuration (GþLþ versus G2Lþ ; figure 1a, ‘global test’); and (iv) a stimulus presenting local features of Sþ in the S2 configuration versus a stimulus presenting local features of S2 in the Sþ configuration (G2Lþ versus GþL2; figure 1a, ‘conflict test’).

(d) Experiment 2 In experiment 2, bees (n ¼ 8) were first pre-trained following a differential conditioning protocol during 15 trials to discriminate two local elements, one rewarded with sucrose (Lþ) and another

(e) Experiment 3 In experiment 3, bees (n ¼ 10) were again pre-trained during 15 trials to discriminate two local elements, one rewarded with sucrose (L0 þ) and another that offered only water (L0 2), but these local elements were afterwards replaced by different ones to build the hierarchical compound stimuli used to train the bees during 36 trials (electronic supplementary material, figure S1C; see example in figure 3a). A non-rewarded ‘priming test’ presenting the primed L0 þ and L0 2 stimuli was conducted after the priming phase. After training with the two hierarchical compound stimuli made of 12 novel local elements, Lþ or L2 (electronic supplementary material, figure S1C; see example in figure 3a), the same four non-rewarded tests of experiment 1 were performed (‘learning test’, ‘global test’, ‘local test’ and ‘conflict test’; see above). All bees were trained with different shapes in order to exclude potential shape preferences.

(f ) Statistical analysis Test performance was analysed in terms of the first choice performed by each bee in a given test (two replicates per bee per test). The proportion of correct first choices was established, thus resulting in a single proportion per bee. One-sample Wilcoxon signed-rank tests were used to test the null hypothesis that this proportion was not significantly different from 50%. Comparisons between tests were performed by means of Wilcoxon nonparametric tests for dependent samples. The a-level for all analyses was 0.05.

3. Results (a) Experiment 1: bees perceive both local and global cues but prioritize global information in hierarchical stimulus discrimination To test the propensity of bees to use either local or global cues, we trained free-flying bees to collect sucrose solution on one of two hierarchical stimuli (Sþ: GþLþ). The alternative stimulus (S2: G2L2) offered water solution. After 36 conditioning trials, bees were tested with Sþ and S2 in the absence of reinforcement (‘learning test’, figure 1a) to verify that the original discrimination was learned. In all cases, bees discriminated between Sþ and S2, irrespective of the stimuli used, and preferred Sþ due to its prior association with sucrose reward (figure 1b, ‘learning test’; n ¼ 13 bees: 76.9 + 7.2% for Sþ, mean + s.e.; Z ¼ 2.37, p , 0.05).

3

Proc. R. Soc. B 282: 20142384

(b) Stimuli

that offered only water (L2) (electronic supplementary material, figure S1B; see example in figure 2a). These local elements were afterwards used to build the hierarchical compound stimuli that were subsequently trained (electronic supplementary material, figure S1B; see example in figure 2a). In this way, we aimed at priming the bees with the local elements Lþ and L2 A non-rewarded ‘priming test’ presenting Lþ versus L2 was conducted following the priming phase to verify that the discrimination between these elements was learned. Bees were then trained during 36 trials to discriminate two hierarchical compound stimuli made of 12 local elements, Lþ or L2 (electronic supplementary material, figure S1B), and then subjected to the same four non-rewarded tests described in experiment 1 (‘learning test’, ‘global test’, ‘local test’ and ‘conflict test’; see above and figure 2a). All bees were trained with the same pair of stimuli both in the pre-training and training phases (electronic supplementary material, figure S1B) except that the rewarding contingency was inverted for half of the tested bees.

rspb.royalsocietypublishing.org

had to re-enter the maze and choose again. Only the initial choice was scored for statistical analyses. This procedure constrains the bee’s choice to the decision chamber of the Y-maze, thus allowing a control of the visual angle at the decision point. During training, the side of the rewarded stimulus (left or right) was interchanged in a pseudorandom sequence to avoid the development of side preferences [36,41]. After training, non-rewarded tests were performed with fresh stimuli. In each test, the first choice of the bees was recorded. Each test was done twice, interchanging the sides of the stimuli to exclude side biases. The pooled proportion of first choices was calculated from the two replicates of each test. The sequence of tests was randomized between bees [36,41]. Two refresher trials with the training stimuli and the presence of reward were intermingled among the non-rewarded tests to keep the bees’ motivation high.

Downloaded from http://rspb.royalsocietypublishing.org/ on December 17, 2014

(a)

pre-training

(b)

4

*

*

L–

training

80

S–

learning test

local test

60

40

20 G+ / L+

G– / L–

G+ / L+

global test

G+ / L–

conflict test 0

L+

L–

G+ / L+ G– / L–

priming test learning test G+ / L+

G+ / L–

G– / L+

G+ / L+ G+ / L–

local test

G+ / L+ G– / L+ G+ / L– G– / L+

global test conflict test

G– / L+

Figure 2. Local preference in visual processing of hierarchical compound stimuli after prior priming to the local features. (a) Pre-training, training and testing procedure. Each bee was pre-trained with a pair of local shapes Lþ and L2 (‘pre-training’), and afterwards subjected to a ‘priming test’ presenting Lþ versus L2 to verify that these shapes were learned. Each bee was then trained with a pair of hierarchical compound stimuli Sþ and S2 made of the pre-trained local elements (GþLþ versus G2L2; ‘training’). Four non-reinforced tests were finally performed: a learning test, a local test, a global test and a conflict test (see main text for details). (b) Performance in the non-rewarded tests (% of first choices). The dashed line indicates chance level. Bees learned both the local shapes ( priming test) and the hierarchical compound stimuli (learning test). They accessed and used both local and global cues as shown by the local and global tests, respectively, and preferred the local information in the conflict situation (conflict test). Data shown are means þ s.e. (n ¼ 8; *p , 0.05).

pre-training

(a)

(b)

100 L¢+

training

learning test

*

*

*

80

% of choices

S+

* *

L¢–

S–

local test

60

40

20 G+ / L+

G– / L–

global test

G+ / L+

G+ / L–

conflict test 0

L¢+

L¢–

G+ / L+ G– / L–

priming test learning test G+ / L+

G– / L+

G+ / L–

G+ / L+ G+ / L– G+ / L+ G– / L+

local test

global test

G+ / L– G– / L+

conflict test

G– / L+

Figure 3. Local preference in visual processing of hierarchical compound stimuli after prior priming to different local features. (a) Pre-training, training and testing procedure. Each bee was pre-trained with a pair of local shapes L0 þ and L0 2 (‘pre-training’) randomly chosen from the set termed ‘pre-training’ in the electronic supplementary material, figure S1C, and was afterwards subjected to a ‘priming test’ to verify that these shapes were learned. Each bee was then trained with a pair of hierarchical compound stimuli Sþ and S2 made of different local elements (GþLþ versus G2L2; ‘training’). Compound stimuli were randomly chosen from the set termed ‘training’ in the electronic supplementary material, figure S1C, taking care that their local shapes differed from those pre-trained. Four non-reinforced tests were finally performed: a learning test, a local test, a global test and a conflict test (see main text for details). (b) Performance in the non-rewarded tests (% of first choices). The dashed line indicates chance level. Bees learned both the local shapes (priming test) and the hierarchical compound stimuli (learning test). They accessed and used both local and global cues as shown by the local and global tests, respectively, and preferred the local information in the conflict situation (conflict test). Data shown are means þ s.e. (n ¼ 10; *p , 0.05).

Proc. R. Soc. B 282: 20142384

% of choices

S+

*

rspb.royalsocietypublishing.org

L+

*

*

100

Downloaded from http://rspb.royalsocietypublishing.org/ on December 17, 2014

We studied the effect of local priming on global preference in hierarchical stimulus discrimination [46,47]. We aimed at determining if prior experience with local cues would increase the weight assigned to local cues, in which case the preference observed in the previous experiment should be reversed.

(c) Experiment 3: top-down priming of local preference by prior experience In this experiment, we tested whether priming local elements during a pre-training phase determined a preference for hierarchical compound stimuli with local elements with no prior reinforcement history. The bees were pre-trained during 15 reinforced trials to discriminate two local shapes (L0 þ versus L0 2; figure 3a). After this phase, the bees were presented with the

5

Proc. R. Soc. B 282: 20142384

(b) Experiment 2: bees prioritize local information in hierarchical stimulus discrimination if this information is primed by prior pre-training

Bees were pre-trained during 15 reinforced trials to discriminate an upward from a downward triangle (Lþ versus L2; figure 2a). After this phase, the bees were presented with the pre-training triangles in a non-reinforced test (‘priming test’). Bees preferred significantly the previously reinforced triangle (upward or downward, depending on the bee), thus showing that both elements could be perfectly distinguished from each other (figure 2b; n ¼ 8; 81.3 + 12.3% for the rewarded triangle; Z ¼ 2.02, p , 0.05). These visual elements were then used to construct a pair of hierarchical stimuli whose global forms were a square and a diamond (figure 2a). Bees were then trained to discriminate this pair of hierarchical stimuli (figure 2a) during 36 reinforced trials. In the learning test performed after training, bees preferred significantly the previously rewarded hierarchical stimulus (figure 2b; 75.0 + 8.8% for Sþ; Z ¼ 2.02, p , 0.05), thus showing that they learned the training task. In the local test opposing two global squares differing only in their local triangles (upward versus downward; GþLþ versus GþL2), bees preferred the square displaying the local triangle previously rewarded (figure 2b; 87.5 + 7.7% for GþLþ; Z ¼ 2.20, p , 0.05). When in the global test, the stimuli differed only in their global shape (square versus diamond built of the same triangles; GþLþ versus G2Lþ); bees preferred the previously rewarded global shape (figure 2b; 87.5 + 7.7% for GþLþ; Z ¼ 2.20, p , 0.05), thus showing that despite their pre-training and their effective use of local information, they also successfully acquired the global information. Moreover, these results also show that during the tests, local and global cues were accessible to the bees. There was no significant difference in performance between local and global tests (Z ¼ 0.00, p ¼ 1.00). In the conflict test opposing a correct global shape built of the incorrect local triangles to an incorrect global shape built of the correct local triangles (GþL2 versus G2Lþ), the results changed dramatically with respect to those obtained in the first experiment. The primed bees did not exhibit a global preference but showed a local preference for the previously primed local triangle (figure 2b; 75.0 + 8.8% for G2Lþ; Z ¼ 2.02, p , 0.05). Thus, the visual preference of the bees was significantly influenced by the pre-training phase, which primed them to use the appropriate local triangles as discriminatory information in a conflict test situation. However, the reversion to local preference may have been the result of a pre-training that increased the salience of the local triangles through a higher number of reinforcements received along the entire experiment, compared with the number of reinforcements received on the global cues. An explanation different from this associative account consists in stating that the pre-training enhanced the attention for local elements [48–51], in which case local preference would emerge even in the case of non-pre-trained elements. The next experiment tested this hypothesis.

rspb.royalsocietypublishing.org

Which cues did bees learn from the stimuli to solve the discrimination task? To answer this question, bees were tested with stimuli that suppressed either the global cues (‘local test’), leaving only the local cues as discriminatory information, or the local cues (‘global test’), leaving only the global cues as discriminatory information (figure 1a). Bees learned both levels of cue as they significantly preferred the option with the correct local stimulus (GþLþ over GþL2) in the local test (figure 1b; 73.1 + 9.2 for GþLþ; Z ¼ 2.20, p , 0.05) and the option with the correct global stimulus (GþLþ over G2Lþ) in the global test (figure 1b; 69.2 + 7.0 for GþLþ; Z ¼ 2.02, p , 0 0.05), irrespective of the stimuli used. Performance in both the Local and the global test was comparable (Z ¼ 0.34, p ¼ 0.74). These results show, therefore, that bees achieved the discrimination using either the global (GþLþ versus G2Lþ) or the local information (GþLþ versus GþL2) (i.e. that both kinds of cues were accessible to the bees during the tests). To determine which level of cues were preferred, we tested the bees in a conflict situation in which the correct global shape was built of the incorrect local elements and the incorrect global shape of the correct local elements (GþL2 versus G2Lþ; figure 1a, ‘conflict test’). In this test, bees should choose randomly if global and local cues are equally weighed or else prefer the stimulus displaying the dominant cue. Bees significantly preferred the stimulus with the correct global configuration (GþL2) despite the presence of incorrect local cues and irrespective of the global stimulus tested (figure 1b; 73.1 + 9.2% for GþL2; Z ¼ 2.20, p , 0.05). This result indicates that global cues dominate over local cues in hierarchical stimulus discrimination in honeybees. We also tested if the bee’s preference for global over local cues depended on the density of the local element as higher local density favours global processing [9]. Reducing the number of local elements constituting the global shape (electronic supplementary material, figure S2A) did not affect the performance shown in figure 1b. Bees still preferred the Sþ over the S2 stimulus in the learning test (electronic supplementary material, figure S2B; ‘learning test’: 75.0 + 9.7% for Sþ; n ¼ 12; Z ¼ 2.20, p , 0.05) and chose the stimulus with the correct global shape and the stimulus with the correct local elements in the global and the local tests, respectively (electronic supplementary material, figure S2B; global test: 70.8 + 7,4 for GþLþ; Z ¼ 2.02, p , 0.05; local test: 75.0 + 7.5 for GþLþ; Z ¼ 2.20, p , 0.05; there was no significant difference in performance between local and global tests: Z ¼ 0.40, p ¼ 0.69). In the conflict test, they also used predominantly the global information (electronic supplementary material, figure S2B; conflict test: 70.8 + 7.4% for GþL2; Z ¼ 2.02, p , 0.05). Global preference is thus resistant to a form of stimulus manipulation that could have favoured local processing [8].

Downloaded from http://rspb.royalsocietypublishing.org/ on December 17, 2014

Our results demonstrate that honeybees exhibit preference of global over local visual information (experiment 1) although they are clearly capable of perceiving and resolving the local elements of the hierarchical compound stimuli used in our study. This global preference is unaffected by a decrease of stimulus density that may favour local processing. Yet the observed preference is not hardwired as it can be modulated by prior experience priming local information (experiments 2 and 3). Under these circumstances, bees still use both local and global cues, but prioritize local information, irrespective of the nature of the local cues previously pre-trained (experiment 3). They exhibit, therefore, a form of top-down processing in which prior experience with local elements seems to induce an attentional bias towards local information processing. Our study demonstrates the importance of global configurations in the visual recognition strategies employed by bees. In addition to highlighting bees’ impressive capacity to extract specific features when explicitly trained to do so (e.g. [52–55]), our results suggest that focusing on global configurations could be a naturally preferred way of analysing complex pictures. Both humans and honeybees exhibit a preference of global over local cues in image processing, despite obvious differences in the architecture of their visual systems and their evolutionary histories. Yet experiments performed in nonhuman primates provide results that indicate instead a preference for local/featural information (baboons [9,10]; capuchins

6

Proc. R. Soc. B 282: 20142384

4. Discussion

[17]; rhesus monkeys [18]; chimpanzees [11,18]). For instance, chimpanzees find it difficult to mentally link separate elements into a global shape [11,13], a task that can be solved both by humans [11,13] and bees (electronic supplementary material, figures S1E and S3) [40,52,56]. In our study, bees were capable of transferring their choice from the global configuration made of separate elements to the corresponding plain shape (electronic supplementary material, figures S1E and S3), thus showing a capacity to link local elements into a global shape. Besides the bee, the only other animal species for which a global preference was shown is, to our knowledge, a teleost fish, Xenotoca eiseni [57]. Interestingly, in our work as well as in the fish study [57], the animals were free to move and to decide based on visual information. By contrast, harnessed animals and fixed visual fields viewed with foveal vision are typically used in the primate studies (e.g. [9,17,58]). This difference may bias primates towards local processing, while the possibility to move and sample images from different viewpoints granted to bees and fishes may promote global processing. In the case of a moving animal, previous experience of successful visually guided foraging may render global information more robust to movement disturbance or variations in view angle, among other factors. Additionally, constructing a global landscape from local landmarks is an important navigation strategy for both bees and ants [32,59–61]. Interestingly, arbitrating between local and global processing of the visual panorama, among other available cues [62–64], has revealed a crucial factor of navigation success [32,60]. Global or local precedence depends indeed on the individual experience about which cues are more relevant in a given position [32,60]. The mechanisms underpinning the global preference of bees remain unknown. A first hypothesis may be a bottom-up explanation based on the physiological properties of the bees’ visual system. In prior experiments on grating discrimination in free-flying bees, dominance of local or global orientation depended on the visual angle subtended by the stimuli: local cues dominated at larger visual angles, while global cues dominated at smaller visual angles, a fact that can be simply explained by the optic resolution of the bees’ visual system [40]. In our case, stimuli were conceived to be resolvable at the decision point of the maze at both the local and the global level, and both kinds of cue were learned and discriminated at the same level by the bees, as demonstrated by the global and local tests (figures 1b, 2b and 3b). In addition, increasing the visual angles subtended by the local elements by placing the stimuli at 15 cm instead of 20 cm from the decision chamber did not influence the observed global preference in the conflict situation (electronic supplementary material, figure S4). The global preference may rather arise from a precedence to process low spatial frequencies. Indeed, low spatial frequencies mediate global information of visual images, while high spatial frequencies are associated with image details [65]. Psychophysical and physiological evidences in human subjects suggest that the processing of low spatial frequencies precedes and dominates high-spatial-frequency processing, thus explaining the reported global advantage [65–72]. The fact that this global preference may be reversed to local preference through prior priming with local cues underlines the plasticity of the bee’s visual system, which may adopt different processing strategies depending on individual experience [49,73,74]. Feature priming is an implicit memory effect in which prior exposure to a specific feature influences the response to the same feature when presented in conjunction

rspb.royalsocietypublishing.org

pre-training elements in a non-reinforced test (‘priming test’) in which they preferred the previously reinforced local shape L0 þ (figure 3b; n ¼ 10; 75.0 + 8.3% of correct choices; Z ¼ 2.02, p , 0.05), thus showing that they had learned the priming information. Bees were then trained during 36 reinforced trials to discriminate a pair of hierarchical stimuli Sþ and S2 (figure 3a) in which both local and global shapes differed from the shapes used in the priming phase. In the subsequent learning test, the bees preferred the previously rewarded stimulus Sþ (figure 3b; 75.0 + 8.3% for Sþ; Z ¼ 2.02, p , 0.05), thus showing that they had learned the training task. In the local test (GþLþ versus GþL2), the bees preferred the stimulus displaying the rewarded local element Lþ (figure 3b; 80.0 + 8.2% for GþLþ; Z ¼ 2.20, p , 0.05), thus showing that they had learned and used local information. In the global test (GþLþ versus G2Lþ), the bees preferred the previously rewarded global shape (figure 3b; 80.0 + 8.2% for GþLþ; Z ¼ 2.20, p , 0.05), thus showing that they had also acquired and used the global information. There was no significant difference in performance between local and global tests (Z ¼ 0.0, p ¼ 1). In the conflict test opposing the rewarded global shape built of the incorrect local elements to the incorrect global shape built of the correct local elements (GþL2 versus G2Lþ), the bees preferred the hierarchical compound stimulus with the local element that was rewarded during the training (figure 3b; 75.0 + 8.3% for G2Lþ; Z ¼ 2.02, p ¼ 0.05). Despite the fact that pre-training primed different local elements, global preference was again reversed to local preference, even if global and local elements appearing in the conflict test had received exactly the same number of reinforcements. Thus, local priming seems to induce a bias towards local information, irrespective of its specific nature.

Downloaded from http://rspb.royalsocietypublishing.org/ on December 17, 2014

Acknowledgements. We thank three anonymous reviewers for useful remarks and experiment suggestion as well as Vivek Nityananda for helpful comments on a previous version of this work. Thanks are also due to David Reser and Marcello Rosa for support and discussions during the project. Funding statement. A.A.-W. and M.G. thank the French Research Council (CNRS), the National Research Agency (ANR: Project MINICOG) and the Human Frontier Science Program (HFSP) for generous support. A.A.-W. was supported by a Travelling Fellowship of The Company of Biologists, and by the University Paul Sabatier (ATUPS fellowship). A.G.D. acknowledges the ARC DP0878968 and DP0987989, and the Alexander von Humboldt Foundation for funding support.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

Logothetis NK, Sheinberg DL. 1996 Visual object recognition. Annu. Rev. Neurosci. 19, 577–621. (doi:10.1146/annurev.ne.19.030196.003045) Altmann CF, Bu¨lthoff HH, Kourtzi Z. 2003 Perceptual organization of local elements into global shapes in the human visual cortex. Curr. Biol. 13, 342–349. (doi:10.1016/s0960-9822(03)00052-6) Peterson MA, Rhodes G. 2003 Perception of faces, objects, and scenes: analytic and holistic processes. New York, NY: Oxford University Press. Palmeri J, Gauthier I. 2004 Visual object understanding. Nat. Rev. Neurosci. 5, 291 –304. (doi:10.1038/nrn1364) Oliva A, Torralba A. 2006 Building the gist of a scene: the role of global image features in recognition. Progress Brain Res. 155, 23 –36. Navon D. 1977 Forest before trees: the precedence of global features in visual perception. Cogn. Psychol. 9, 353 –383. (doi:10.1016/0010-0285 (77)90012-3) Navon D. 1981 The forest revisited: more on global precedence. Psychol. Res. 43, 1– 32. (doi:10.1007/ BF00309635) Kimchi R. 1992 Primacy of wholistic processing and global/local paradigm: a critical review. Psychol. Bull. 112, 24 –38. (doi:10.1037/0033-2909.112. 1.24) Fagot J, Deruelle C. 1997 Processing of global and local visual information and hemispheric specialization in humans (Homo sapiens) and baboons (Papio papio). J. Exp. Psychol. Hum.

10.

11.

12.

13.

14.

15.

16.

17.

Percept. Perform. 23, 429–442. (doi:10.1037/ 0096-1523.23.2.429) Deruelle C, Fagot J. 1998 Visual search for global/ local stimulus features in humans and baboons. Psych. Bull. Rev 5, 476 –481. (doi:10.3758/ BF03208825) Fagot J, Tomonaga M. 1999 Global and local processing in humans (Homo sapiens) and chimpanzees (Pan troglodytes): use of a visual search task with coumpound stimuli. J. Comp. Psychol. 113, 3–12. (doi:10.1037/0735-7036.113.1.3) Love BC, Rouder JN, Wisniewski EJ. 1999 A structural account of global and local processing. Cogn. Psychol. 38, 291–316. (doi:10.1006/cogp.1998.0697) Fagot J, Tomonaga M. 2001 Effects of element separation on perceptual grouping by humans (Homo sapiens) and chimpanzees (Pan troglodytes): perception of Kanizsa illusory figures. Anim. Cogn. 4, 171 –177. (doi:10.1007/s100710100109) Lea SEG, Goto K, Osthaus B, Ryan CME. 2006 The logic of the stimulus. Anim. Cogn. 9, 247 –256. (doi:10.1007/s10071-006-0038-3) Navon D. 1983 How many trees does it take to make a forest? Perception 12, 239–254. (doi:10. 1068/p120239) Navon D. 1991 Testing a queue hypothesis for the processing of global and local information. J. Exp. Psychol. 120, 173–189. (doi:10.1037/0096-3445. 120.2.173) Spinozzi G, De Lillo C, Truppa V. 2003 Global and local processing of hierarchical visual stimuli in

18.

19.

20.

21.

22.

23.

24.

tufted capuchin monkeys (Cebus apella). J. Comp. Psychol. 117, 15 –23. (doi:10.1037/0735-7036. 117.1.15) Hopkins W, Washburn D. 2002 Matching visual stimuli on the basis of global and local features by chimpanzees (Pan troglodytes) and rhesus monkeys (Macaca mulatta). Anim. Cogn. 5, 27 –31. (doi:10. 1007/s10071-001-0121-8) Cerella J. 1980 The pigeon’s analysis of pictures. Pattern Recogn. 12, 1– 6. (doi:10.1016/00313203(80)90048-5) Aust U, Huber L. 2001 The role of item- and category-specific information in the discrimination of people versus nonpeople images by pigeons. Anim. Learn. Behav. 29, 107 –119. (doi:10.3758/ bf03192820) Cavoto KK, Cook RG. 2001 Cognitive precedence for local information in hierarchical stimulus processing by pigeons. J. Exp. Psychol. Anim. Behav. Process. 27, 3 –16. (doi:10.1037/0097-7403.27.1.3) Goto K, Lea SEG. 2003 Discrimination of direction of movements in pigeons following previous experience of motion/static discrimination. J. Exp. Anal. Behav. 80, 29–42. (doi:10.1901/jeab.2003.80-29) Kelly DM, Cook RG. 2003 Differential effects of visual context on pattern discrimination by pigeons (Columba livia) and humans (Homo sapiens). J. Comp. Psychol. 117, 200–208. (doi:10.1037/ 0735-7036.117.2.200) Giurfa M, Menzel R. 1997 Insect visual perception: complex abilities of simple nervous systems. Curr.

7

Proc. R. Soc. B 282: 20142384

This study reveals the importance of global configurations in visual image processing in honeybees, and additionally confirms the capacity of the bee brain to extract and combine features and their interrelation in order to successfully discriminate complex pictures [37,87–89]. This conclusion might be surprising as similar capacities are thought to require specialized neural circuits and extensive experience in humans and mammals in general [90–96]. However, these performances are possible with reduced computational resources, as suggested by analyses with artificial neural networks [97], in particular if the brain possesses the flexibility to allocate resources towards the analysis of local or global cues through top-down processing to accommodate various contexts.

rspb.royalsocietypublishing.org

with other stimuli [47,75,76]. In our case, the reversion to local preference was not purely associative (i.e. based on a higher number of rewards associated with the local cues in the whole experiment; see experiment 2), as local preference was maintained if the local features were changed between the priming and the training phase (experiment 3). It thus seems that the local priming biased perceptual attention towards local features, irrespective of their nature and reinforcement history, in a process akin to the so-called attentional weighting [77]. Attention describes our ability to focus perception on one stimulus (or group of related stimuli), while filtering out other simultaneous stimuli that are less relevant at any moment [78]. Our results suggest that visual processing in bees can be related to attentional-like processes, which, depending on the training procedure, may enhance the relevance of specific features while filtering other cues. Attentional-like mechanisms have been shown to mediate visual performances in insects [79–83]. In fruitflies, Drosophila melanogaster, attention in a visual tracking task [79,84–86] could be retraced to the level of the mushroom bodies, which are higher-order structures of the insect brain. In honeybees walking stationarily in the middle of a visual arena, neural signatures underlying attentional tracking of moving bars were found in different visual neuropiles of the brain [83]. Yet the neural mechanisms underlying both the precedence of global configurations and the capacity to revert this precedence to local analysis in visual image processing remain to be determined.

Downloaded from http://rspb.royalsocietypublishing.org/ on December 17, 2014

26.

27.

29.

30.

31. 32.

33.

34.

35.

36.

37.

38.

39.

40.

58. Hopkins WD. 1997 Hemispheric specialization for local and global processing of hierarchical visual stimuli in chimpanzees (Pan troglodytes). Neuropsychologia 35, 343–348. (doi:10.1016/ S0028-3932(96)00089-9) 59. Graham P, Cheng K. 2009 Ants use the panoramic skyline as a visual cue during navigation. Curr. Biol. 19, R935 –R937. (doi:10.1016/j.cub.2009.08.015) 60. Pahl M, Zhu H, Tautz J, Zhang S. 2011 Large scale homing in honeybees. PLoS ONE 6, e19669. (doi:10. 1371/journal.pone.0019669) 61. Wystrach A, Beugnon G, Cheng K. 2011 Landmarks or panoramas: what do navigating ants attend to for guidance? Front. Zool. 8, 21. (doi:10.1186/17429994-8-21) 62. Collett M. 2012 How navigational guidance systems are combined in a desert ant. Curr. Biol. 22, 927–932. (doi:10.1016/j.cub.2012.03.049) 63. Wystrach A, Graham P. 2012 What can we learn from studies of insect navigation? Anim. Behav. 84, 13– 20. (doi:10.1016/j.anbehav.2012.04.017) 64. Wystrach A, Schwarz S, Baniel A, Cheng K. 2013 Backtracking behaviour in lost ants: an additional strategy in their navigational toolkit. Proc. R. Soc. B 280, 20131677. (doi:10.1098/rspb.2013.1677) 65. Morrison D, Schyns P. 2001 Usage of spatial scales for the categorization of faces, objects, and scenes. Psych. Bull. Rev. 8, 454 –469. (doi:10.3758/ bf03196180) 66. Northmore DPM, Dvorak CA. 1979 Contrast sensitivity and acuity of the goldfish. Vis. Res. 19, 255–261. (doi:10.1016/0042-6989(79)90171-8) 67. Shulman GL, Wilson J. 1987 Spatial frequency and selective attention to local and global information. Perception 16, 89 –101. (doi:10.1068/p160089) 68. Badcock JC, Whitworth FA, Badcock DR, Lovegrove WJ. 1990 Low-frequency filtering and the processing of local-global stimuli. Perception 19, 617–629. (doi:10.1068/p190617) 69. Lamb M, Yund E. 1993 The role of spatial frequency in the processing of hierarchically organized stimuli. Atten. Percept. Psychophys. 54, 773 –784. (doi:10. 3758/bf03211802) 70. Hughes HC, Nozawa G, Kitterle F. 1996 Global precedence, spatial frequency channels, and the statistics of natural images. J. Cogn. Neurosci. 8, 197–230. (doi:10.1162/jocn.1996.8.3.197) 71. Boeschoten MA, Kemner C, Kenemans JL, Engeland HV. 2005 The relationship between local and global processing and the processing of high and low spatial frequencies studied by event-related potentials and source modeling. Cogn. Brain Res. 24, 228 –236. (doi:10.1016/j.cogbrainres.2005.01. 021) 72. Goffaux V, Hault B, Michel C, Vuong QC, Rossion B. 2005 The respective role of low and high spatial frequencies in supporting configural and featural processing of faces. Perception 34, 77 –86. (doi:10. 1068/p5370) 73. Giurfa M, Schubert M, Reisenman C, Gerber B, Lachnit H. 2003 The effect of cumulative experience on the use of elemental and configural visual discrimination strategies in honeybees. Behav. Brain

8

Proc. R. Soc. B 282: 20142384

28.

41. Giurfa M, Hammer M, Stach S, Stollhoff N, Mullerdeisig N, Mizyrycki C. 1999 Pattern learning by honeybees: conditioning procedure and recognition strategy. Anim. Behav. 57, 315 –324. (doi:10.1006/ anbe.1998.0957) 42. Srinivasan MV, Lehrer M. 1988 Spatial acuity of honeybee vision and its spectral properties. J. Comp. Physiol. A 162, 159 –172. (doi:10.1007/BF0060 6081) 43. Campan R, Lehrer M. 2002 Discrimination of closed shapes by two species of bee, Apis mellifera and Megachile rotundata. J. Exp. Biol. 205, 559–572. 44. Hempel de Ibarra N, Giurfa M. 2003 Discrimination of closed coloured shapes by honeybees requires only contrast to the long wavelength receptor type. Anim. Behav. 66, 903–910. (doi:10.1006/anbe. 2003.2269) 45. Lehrer M, Campan R. 2005 Generalization of convex shapes by bees: what are shapes made of? J. Exp. Biol. 208, 3233 –3247. (doi:10.1242/jeb.01790) 46. Ward LM. 1982 Determinants of attention to local and global features of visual forms. J. Exp. Psychol. 8, 562– 581. (doi:10.1037/0096-1523.8.4.562) 47. Schwarzkopf DS, Rees G. 2011 Interpreting local visual features as a global shape requires awareness. Proc. R. Soc. B 278, 2207–2215. (doi:10.1098/rspb.2010.1909) 48. Kinchla R, Wolfe J. 1979 The order of visual processing: ‘top-down,’ ‘bottom-up,’ or ‘middleout’. Atten. Percept. Psychophys. 25, 225–231. (doi:10.3758/bf03202991) 49. Zhang S, Srinivasan M. 1994 Prior experience enhances pattern discrimination in insect vision. Nature 368, 330–332. (doi:10.1038/368330a0) 50. Weissman DH, Mangun GR, Woldorff MG. 2002 A role for top-down attentional orienting during interference between global and local aspects of hierarchical stimuli. NeuroImage 17, 1266–1276. (doi:10.1006/nimg.2002.1284) 51. Bar M et al. 2006 Top-down facilitation of visual recognition. Proc. Natl Acad. Sci. USA 103, 449 –454. (doi:10.1073/pnas.0507062103) 52. van Hateren JH, Srinivasan MV, Wait PB. 1990 Pattern recognition in bees: orientation discrimination. J. Comp. Physiol. A 167, 649–654. (doi:10.1007/BF00192658) 53. Horridge GA, Zhang SW. 1995 Pattern vision in honeybees (Apis mellifera): flower-like patterns with no predominant orientation. J. Insect Physiol. 41, 681 –688. (doi:10.1016/0022-1910(95)00021-L) 54. Giurfa M, Eichmann B, Menzel R. 1996 Symmetry perception in an insect. Nature 382, 458 –461. (doi:10.1038/382458a0) 55. Horridge A. 1997 Pattern discrimination by the honeybee: disruption as a cue. J. Comp. Physiol. A 181, 267 –277. (doi:10.1007/s003590050113) 56. Horridge GA, Zhang S-W, O’Carroll D. 1992 Insect perception of illusory contours. Phil. Trans. R. Soc. Lond. B 337, 59 –64. (doi:10.1098/rstb.1992.0083) 57. Truppa V, Sovrano VA, Spinozzi G, Bisazza A. 2010 Processing of visual hierarchical stimuli by fish (Xenotoca eiseni). Behav. Brain Res. 207, 51 –60. (doi:10.1016/j.bbr.2009.09.039)

rspb.royalsocietypublishing.org

25.

Opin. Neurobiol. 7, 505–513. (doi:10.1016/S09594388(97)80030-X) Srinivasan MV. 2010 Honey bees as a model for vision, perception, and cognition. Annu. Rev. Entomol. 55, 267 –284. (doi:10.1146/annurev.ento. 010908.164537) Avargue`s-Weber A, Mota T, Giurfa M. 2012 New vistas on honey bee vision. Apidologie 43, 244–268. (doi:10.1007/s13592-012-0124-2) Dyer AG. 2012 The mysterious cognitive abilities of bees: why models of visual processing need to consider experience and individual differences in animal performance. J. Exp. Biol. 215, 387–395. (doi:10.1242/jeb.038190) Zhang S, Si A, Pahl M. 2012 Visually guided decision making in foraging honeybees. Front. Neurosci. 6, 88. (doi:10.3389/fnins.2012.00088) Avargue`s-Weber A, Deisig N, Giurfa M. 2011 Visual cognition in social insects. Annu. Rev. Entomol. 56, 423–443. (doi:10.1146/annurev-ento-120709144855) Dafni A, Lehrer M, Keyan PG. 1997 Spatial flower parameters and insect spatial vision. Biol. Rev. 72, 239– 282. (doi:10.1111/j.1469-185X.1997.tb00014.x) Horridge A. 2009 What does an insect see? J. Exp. Biol. 212, 2721–2729. (doi:10.1242/jeb.030916) Lent David D, Graham P, Collett Thomas S. 2013 Visual scene perception in navigating wood ants. Curr. Biol. 23, 684– 690. (doi:10.1016/j.cub.2013. 03.016) Dyer AG, Neumeyer C, Chittka L. 2005 Honeybee (Apis mellifera) vision can discriminate between and recognise images of human faces. J. Exp. Biol. 208, 4709–4714. (doi:10.1242/jeb.01929) Dyer AG, Vuong QC. 2008 Insect brains use image interpolation mechanisms to recognise rotated objects. PLoS ONE 3, e4086. (doi:10.1371/journal. pone.0004086) Tibbetts EA, Dyer AG. 2013 Good with faces. Sci. Am. 309, 62 –67. (doi:10.1038/scientificamerican 1213-62) Stach S, Benard J, Giurfa M. 2004 Local-feature assembling in visual pattern recognition and generalization in honeybees. Nature 429, 758–761. (doi:10.1038/nature02594) Avargue`s-Weber A, Portelli G, Benard J, Dyer AG, Giurfa M. 2010 Configural processing enables discrimination and categorization of face-like stimuli in honeybees. J. Exp. Biol. 213, 593–601. (doi:10. 1242/jeb.039263) Avargue`s-Weber A, Giurfa M. 2014 Cognitive components of color vision in honey bees: how conditioning variables modulare color learning and discrimination. J. Comp. Physiol. A 200, 449 –461. (doi:10.1007/s00359-014-0909-z) Dyer AG, Paulk AC, Reser DH. 2011 Colour processing in complex environments: insights from the visual system of bees. Proc. R. Soc. B 278, 952–959. (doi:10.1098/rspb.2010.2412) Zhang SW, Srinivasan MV, Horridge GA. 1992 Pattern recognition in honeybees: local and global analysis. Proc. R. Soc. Lond. B 248, 55 –61. (doi:10. 1098/rspb.1992.0042)

Downloaded from http://rspb.royalsocietypublishing.org/ on December 17, 2014

75.

76.

78.

79.

80.

81.

82.

84.

85.

86.

87.

88.

89.

90. Gauthier I, Tarr MJ. 1997 Becoming a ‘Greeble’ expert: exploring mechanisms for face recognition. Vis. Res. 37, 1673 –1682. (doi:10.1016/S00426989(96)00286-6) 91. Gauthier I, Skudlarski P, Gore CG, Anderson AW. 2000 Expertise for cars and birds recruits brain areas involved in face recognition. Nat. Neurosci. 3, 191–197. (doi:10.1038/72140) 92. Kanwisher N. 2000 Domain specificity in face perception. Nat. Neurosci. 3, 759–763. (doi:10. 1038/77664) 93. Peirce JW, Leigh AE, Kendrick KM. 2000 Configurational coding, familiarity and the right hemisphere advantage for face recognition in sheep. Neuropsychologia 38, 475–483. (doi:10.1016/ S0028-3932(99)00088-3) 94. Sigala N, Logothetis NK. 2002 Visual categorization shapes feature selectivity in the primate temporal cortex. Nature 415, 318–320. (doi:10.1038/ 415318a) 95. Tsao DY, Freiwald WA, Tootell RBH, Livingstone MS. 2006 A cortical region consisting entirely of faceselective cells. Science 311, 670–674. (doi:10.1126/ science.1119983) 96. Lobmaier JS, Klaver P, Loenneker T, Martin E, Mast FW. 2008 Featural and configural face processing strategies: evidence from a functional magnetic resonance imaging study. Neuroreport 19, 287–291. (doi:10.1097/WNR.0b013e3282f556fe) 97. Aitkenhead MJ, McDonald AJS. 2003 A neural network face recognition system. Eng. Appl. Artif. Intell. 16, 167– 176. (doi:10.1016/S0952-1976 (03)00042-3)

9

Proc. R. Soc. B 282: 20142384

77.

83.

bumblebees. J. Exp. Biol. 215, 2515 –2523. (doi:10. 1242/jeb.066399) Paulk AC, Stacey JA, Pearson TWJ, Taylor GJ, Moore RJD, Srinivasan MV, van Swinderen B. 2014 Selective attention in the honeybee optic lobes precedes behavioral choices. Proc. Natl Acad. Sci. USA 111, 5006–5011. (doi:10.1073/pnas.1323297111) van Swinderen B, McCartney A, Kauffman S, Flores K, Agrawal K, Wagner J, Paulk A. 2009 Shared visual attention and memory systems in the Drosophila brain. PLoS ONE 4, e5989. (doi:10.1371/journal. pone.0005989) van Swinderen B. 2011 Attention in Drosophila. Int. Rev. Neurobiol. 99, 51 –85. (doi:10.1016/B978-012-387003-2.00003-3) Miller SM, Ngo TT, van Swinderen B. 2012 Attentional switching in humans and flies: rivalry in large and miniature brains. Front. Hum. Neurosci. 5, 188. (doi:10.3389/fnhum.2011.00188) Avargue`s-Weber A, Dyer AG, Giurfa M. 2011 Conceptualization of above and below relationships by an insect. Proc. R. Soc. B 278, 898 –905. (doi:10. 1098/rspb.2010.1891) Avargue`s-Weber A, Dyer AG, Combe M, Giurfa M. 2012 Simultaneous mastering of two abstract concepts with a miniature brain. Proc. Natl Acad. Sci. USA 109, 7481 –7486. (doi:10.1073/pnas. 1202576109) Dyer AG, Griffiths DW. 2012 Seeing near and seeing far; behavioural evidence for dual mechanisms of pattern vision in the honeybee (Apis mellifera). J. Exp. Biol. 215, 397–404. (doi:10.1242/jeb. 060954)

rspb.royalsocietypublishing.org

74.

Res. 145, 161–169. (doi:10.1016/S0166-4328 (03)00104-9) Stach S, Giurfa M. 2005 The influence of training length on generalization of visual feature assemblies in honeybees. Behav. Brain Res. 161, 8– 17. (doi:10.1016/j.bbr.2005.02.008) Fremouw T, Herbranson WT, Shimp CP. 1998 Priming of attention to global and local levels of visual analysis. J. Exp. Psychol. Anim. Behav. Process. 24, 278–290. (doi:10.1037/0097-7403.24.3.278) Gauthier I. 2000 Visual priming: the ups and downs of familiarity. Curr. Biol. 10, R753 –R756. (doi:10. 1016/S0960-9822(00)00738-7) Petrov AA, Dosher BA, Lu ZL. 2005 The dynamics of perceptual learning: an incremental reweighting model. Psychol. Rev. 112, 715–743. (doi:10.1037/ 0033-295X.112.4.715) Posner MI, Snyder CR, Davidson BJ. 1980 Attention and the detection of signals. J. Exp. Psychol. 109, 160–174. (doi:10.1037/0096-3445. 109.2.160) van Swinderen B, Greenspan RJ. 2003 Salience modulates 20 – 30 Hz brain activity in Drosophila. Nat. Neurosci. 6, 579–586. (doi:10.1038/nn1054) Spaethe J, Tautz J, Chittka L. 2006 Do honeybees detect colour targets using serial or parallel visual search? J. Exp. Biol. 209, 987– 993. (doi:10.1242/ jeb.02124) Sareen P, Wolf R, Heisenberg M. 2011 Attracting the attention of a fly. Proc. Natl Acad. Sci. USA 108, 7230–7235. (doi:10.1073/pnas.1102522108) Morawetz L, Spaethe J. 2012 Visual attention in a complex search task differs between honeybees and

The forest or the trees: preference for global over local image processing is reversed by prior experience in honeybees.

Traditional models of insect vision have assumed that insects are only capable of low-level analysis of local cues and are incapable of global, holist...
558KB Sizes 1 Downloads 5 Views