3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

Journal of Integrative Neuroscience, Vol. 6, No. 3 (2007) 1–23 c Imperial College Press


Research Report GENDER DIFFERENCES IN THE USE OF EXTERNAL LANDMARKS VERSUS SPATIAL REPRESENTATIONS UPDATED BY SELF-MOTION SIMON LAMBREY∗ and ALAIN BERTHOZ LPPA, Coll`ege de France CNRS, 11 Place Marcelin Berthelot 75005 Paris, France ∗ [email protected] Received 28 March 2007 Accepted 30 July 2007 Numerous data in the literature provide evidence for gender differences in spatial orientation. In particular, it has been suggested that spatial representations of large-scale environments are more accurate in terms of metric information in men than in women but are richer in landmark information in women than in men. One explanatory hypothesis is that men and women differ in terms of navigational processes they used in daily life. The present study investigated this hypothesis by distinguishing two navigational processes: spatial updating by self-motion and landmark-based orientation. Subjects were asked to perform a pointing task in three experimental conditions, which differed in terms of reliability of the external landmarks that could be used. Two groups of subjects were distinguished, a mobile group and an immobile group, in which spatial updating of environmental locations did not have the same degree of importance for the correct performance of the pointing task. We found that men readily relied on an internal egocentric representation of where landmarks were expected to be in order to perform the pointing task, a representation that could be updated during self-motion (spatial updating). In contrast, women seemed to take their bearings more readily on the basis of the stable landmarks of the external world. We suggest that this gender difference in spatial orientation is not due to differences in information processing abilities but rather due to differences in higher level strategies. Keywords: Spatial memory; spatial orientation; spatial updating; landmark; gender differences; human.

1. Introduction According to popular belief, men have “a better sense of direction” than women. Although an experimental approach does not support the idea that men systematically outperform women on all spatial tasks, it does however suggest that gender differences exist in diverse spatial abilities [16, 18, 22, 26, 44]. The principle aim of ∗ Corresponding

author.

1

3rd Reading September 7, 2007

2

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

the present study was to assess gender differences in the use of two navigational processes, namely spatial updating by self-motion and landmark-based orientation. Spatial updating by self-motion consists of forming egocentric representations of the environment and transforming them to accommodate self-motion: the subject represents the current egocentric directions and distances — that is, metric information best described in a polar coordinate system — to significant environmental locations, e.g., the location of relevant landmarks. As they move or turn, subjects update these representations on the basis of signals generated during self-motion by a process of vector summation: the new egocentric positions of objects are computed by adding the objects’ displacement vector (relative to the subject) to their previous egocentric position vectors [46, 48]. This model is fully consistent with numerous experimental data in humans (see Ref. [25] for a review). For example, evidence for automatic updating of target locations during self-motion has been found in blind and blindfolded subjects [11, 14, 19, 24, 30, 37, 48], as well as in sighted subjects that were immersed in virtual environments where optic flow but no landmark information was available [5, 19, 23]. Other studies also suggest that spatial updating is used for object location and spatial layout memory across perspective change on the same scene [4, 40, 47]. Most likely operating in parallel with and complementing spatial updating, the second navigational process considered here is referred to as the landmark-based orientation process. A landmark may be defined as a salient and stable element belonging to the external world and capable of being used by subjects to take their bearings (e.g., [41–43]). On this view, subjects have a world-centered (i.e., allocentric) representation of the environment, that is, a representation of how the different elements of the environment are “located” relative to one or more external landmarks or, more precisely, of how these elements can be reached from a given landmark. One crucial point to be stressed is that such a landmark-based representation can contain no information about distances and directions (i.e., no metric information) but still allow efficient navigation. Indeed, one subject can remember that they have to take a given direction at the level of a specific landmark in order to reach his destination [10, 27, 41] (e.g., to go to the library, turn right when facing the big yellow statue) without any idea of the spatial relationship in metric terms between the landmark (the statue) and the destination (the library) (see Ref. [41] for a discussion). Consistent with this view, some authors have argued that navigation based on the use of landmarks helps to construct qualitative allocentric representations of the environment, such as topological networks of routes (i.e., networks of landmarks and sequences of movement decisions), as opposed to quantitative allocentric representations, that is, metric Euclidean representations [20, 38, 41]. The interactions between the two navigational processes described above are reciprocal. Indeed, metric information acquired from spatial updating by self-motion, together with quantitative memory of displacements, can gradually be incorporated into topological representations, allowing the construction of Euclidean-like allocentric representations of the environment [20, 41]. The information included in

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

3

an egocentric representation updated by self-motion can also be very useful in the event of unreliable or non-stable landmarks. Conversely, landmark information can be used to regularly make the updated internal representation fit one’s actual position in the environment. Consistent with this view, modeling the consequences of noise in the evaluation of the angular and/or the linear components of locomotion has shown that spatial updating is open to a rapid drift when the assessment of direction occurs without any external compass [2, 3]. Likewise, recent experiments in humans have emphasized the benefit of using external cues in tasks that are known to require spatial updating of an egocentric representation. For example, in perspective change paradigms, subjects are significantly more accurate at remembering object locations when memory is probed from a novel viewpoint aligned with environmental landmarks [4, 29] or with the walls of the testing room [31] compared to misaligned viewpoints. Concerning gender differences in spatial orientation, various studies have shown that mental representations of large-scale environments are different in men and women. For example, when asked to describe routes learned from a map, men preferentially use metric distances and cardinal directions, whereas women more readily use landmarks and relative right-left directions [6, 7, 9, 12, 17]. Gender differences have also been found in environmental knowledge acquired from navigation. McGuiness and Sparks [28] found that, when asked to draw a map of their campus, female students positioned more buildings than male students did. In contrast, men drew more roads and other connectors and positioned the buildings more accurately in terms of absolute metric coordinates. Using virtual reality, Astur et al. [1] showed that men are better than women at retrieving a hidden target location from different departure points in a room in which landmarks could be used to take one’s bearings. Other results suggest that, in order to perform such a task, women rely predominantly on landmark information, while men readily use both landmark and geometrical information [35]. In another study [45], subjects were asked to explore either a virtual or a real-world maze where they encountered various objects. The accuracy of the mental representation of the maze was subsequently assessed with both pointing and distance estimation tasks. The results showed that pointing to an unseen target object when facing another object in the maze was more accurate in men than in women, especially in the virtual maze condition. This was interpreted as evidence for a more accurate representation of the environment in metric terms in men than in women. More recently, Iachini et al. [13] showed that men were better than women in recalling the distance between objects positioned in a real three-dimensional environment as well as in recalling the size of the layout. In the light of the literature reviewed above, it seems clear that environmental representations are more accurate in terms of metric information in men than in women, whereas they are richer in landmark information in women than in men. But the reason for this gender difference remains to be understood. One hypothesis is that men and women basically differ in terms of navigational processes they used in daily life. Men may more readily rely on an internal egocentric representation of

3rd Reading September 7, 2007

4

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

where things are, which they would update during self-motion. In contrast, women may take their bearings more readily on the basis of stable landmarks of the external world. The present study assesses this hypothesis, trying to draw a distinction between gender differences in strategy versus in ability. Each volunteer entered a testing room whose the interior was perfectly symmetrical and contained three video screens. These video screens first displayed “landmark” views of the outside realworld environment, were switched off for a 30 s delay, and then switched on again. Three experimental conditions were distinguished. In the Stable-Landmarks (StableL) condition, the views displayed on the screens were the same before and after the delay. In the Rotated-Landmarks (Rot-L) condition, the views displayed after the delay were rotated with respect to their locations before darkness. In the SwitchedLandmarks (Switch-L) condition, two views out of three were switched after the delay, while the third one remained at the same location. Finally, the screens were switched off again, the light was turned on and subjects had to indicate the direction of each of nine different views in the outside environment (pointing task). Two groups were distinguished. In the immobile group, subjects had no specific instructions on what to do during the 30 s delay, whereas subjects in the mobile group had to close their eyes and rotate on the spot. The principle behind this paradigm is as follows. The only external landmarks that subjects had at their disposal to perform the pointing task were the views displayed on the screens. In order to make correct use of these landmark views in the Rot-L condition, subjects in the mobile group would need to have updated an egocentric representation of the landmark view locations during rotations in the dark. The absence of such spatial updating as well as the non-awareness of the landmarks rotation would necessarily result in major pointing errors. In the SwitchL condition, the absence of spatial updating would also result in major pointing errors but subjects could still become aware of the landmarks switch by detecting a modification in the landmarks configuration (see Tables 1 and 2). According to the literature reviewed above, women were expected to remain unaware of the landmarks rotation in the Rot-L condition as well as to make more pointing errors than men. In the Switch-L condition, no particular difference between women and men was expected concerning the fact of becoming aware of the landmarks switch. The results concerning the pointing errors should therefore allow discussing about potential gender differences in the ability to process spatial updating information. Table 1. Necessity of spatial updating for becoming aware of the landmark views transformation according to the group and to the condition. Condition

Mobile Group

Immobile Group

Stable-L Rot-L Switch-L

Not applicable Necessary Not necessary

Not applicable Not necessary Not necessary

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

5

Table 2. Necessity of spatial updating for accurate pointing according to the group and to the condition. Condition

Mobile Group

Immobile Group

Stable-L Rot-L Switch-L

Not necessary Necessary Necessary

Not necessary Not necessary Not necessary

2. Materials and Methods 2.1. Experimental set-up This study is the result of collaboration with a contemporary artist, Daniel Firman. It was conducted in the contemporary art center of Tarbes, Le Parvis, France, on the occasion of an exhibition by Daniel Firman entitled O.A.P. (Op´eration Aliments Portatifs). The experiment was approved by the ethics committee of our institution (CCPPRB 120-98). The enclosed hexagonal testing room was located in the middle of the gallery’s mezzanine floor. The mezzanine floor is an open area overlooking the shops in a shopping center [see plan in Fig. 1(A)]. The different views of the surrounding environment from the middle of the mezzanine (i.e., where the testing room was located) are shown in Fig. 1(B). The mezzanine area was open to the public and quite busy, with visitors to the gallery and shoppers crossing the gallery to get to other parts of the shopping center. Details of the experimental set-up are shown in Fig. 2. The testing room was hexagonal. Each of the six walls of the room was 250 cm high and 125 cm wide. There was a door in three of the six walls, allowing the subjects to enter the room. Each door was 125 cm high and 100 cm wide. The room had a false ceiling 125 cm high in the center of which was a 125 cm diameter circular opening enabling the subjects to stand upright. The total height inside the room from the floor to the ceiling was 250 cm. Each wall contained either a spotlight or a video screen at a height of 150 cm above the floor. The screens were called Screen 1, Screen 4 and Screen 7 and were oriented at 0◦ , 120◦ and 240◦ , respectively. A red circular scale indicating 360◦ was painted on the upper side of the false ceiling. Every degree was indicated by a line, with a longer line at 10◦ intervals. Except for this scale, the whole of the interior of the testing room was painted dark gray (above and below the false ceiling). Particular attention was paid to making the three doors light-proof and correcting all other irregularities in the interior design of the room, since the subjects could have used such imperfections as landmarks to take their bearings during the experiment. As a result, the interior of the testing room was totally symmetrical. Furthermore, the room was relatively well soundproofed and subjects could not use the faint outside sounds as landmarks, since they formed a confused homogenous noise without any spatial meaning.

3rd Reading September 7, 2007

6

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

(A)

(B)

Fig. 1. Real-world environment of the study. (A) Aerial view of the environment. The hexagonal testing room was placed in the center of the mezzanine floor. The squared zone corresponds to the lower floor. The light gray zone corresponds to the upper floor. The dark gray zones correspond to walls. The numbers indicate the direction of the nine views used for the experiment. (B) The nine views used for the pointing task.

A central overhead digital video camera (installed in the center of the ceiling) recorded subjects’ responses and enabled the experimenter to monitor on a surveillance screen what happened in the testing room throughout the experiment. Three videotape recorders were connected to the three video screens visible from inside the testing room and allowed the stimuli to be displayed in a synchronized manner. The surveillance screen and the videotape recorders were installed on a control table outside the testing room. 2.2. Stimuli Prior to installing the testing room, the stimuli were prepared using a digital video camera placed at a height of 150 cm (i.e., the height at which the screens were subsequently placed in the testing room) and at the point where the subjects would be standing (i.e., the center of the planned testing room). Nine video recordings

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

(A)

7

(B)

Fig. 2. Experimental set-up (diagram and real-world photo). Subjects entered through one of the three doors and then stood up inside the testing room. Three TV screens were positioned on three of the six walls of the structure. These screens displayed the landmark views.

were made from nine different viewpoints spaced at 40◦ intervals. Three of these viewpoints (0◦ , 120◦ and 240◦ ) corresponded to the planned orientation of the video screens in the testing room. The recordings corresponding to these three particular viewpoints were used for editing the video stimuli. One shot was extracted from each of the nine video recordings. The nine selected views were called Views 1 to 9 and oriented at 0◦ , 40◦ , 80◦ , 120◦ , 160◦ , 200◦ , 240◦ , 280◦ and 320◦ , respectively. They were printed on a sheet of photographic paper (17 cm wide × 14 cm high) and covered with transparent plastic. An unprinted white horizontal strip (17 cm wide × 1.5 cm high) was kept blank at the bottom of each photo and a vertical red line was drawn in the middle of the strip. The photos were used by the subjects to perform the pointing task. The red line indicated the pointed direction on the circular scale. 2.3. Procedure At the beginning, subjects were explained the experimental procedure and were clearly told that the pointing task consisted in placing the photos with respect to

3rd Reading September 7, 2007

8

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

the outside environment. They were also given a descriptive text of the instructions. The experimenter checked that the subjects had clearly understood the experiment and gave additional oral explanations if necessary. Subjects were then shown the nine photos they would be required to place on the circular scale during the pointing task. Subjects were asked to walk around the outside of the testing room, paying attention to the surrounding environment, its landmarks and its spatial organization. They were then given the nine photos, which had first been shuffled, and were indicated one of the three doors to enter the testing room. Inside the testing room, the spotlights were turned off and light came only from the three video screens that the subjects had to look at. The room space was aligned on the environmental space, that is, each screen displayed the image of one view of the environment as if this screen was a window on the environment (i.e., Screen 1 displayed View 1, Screen 4 displayed View 4, Screen 7 displayed View 7). After 90 s, the images of the surrounding environment disappeared and the screens remained dark for 30 s (darkness phase). Then the screens displayed images of the environment again and subjects had to look once more at each of the screens. After an additional 90 s, the screens were switched off and the spotlights turned on. This was the signal to perform the pointing task. During this task, subjects had to place the nine photos on the circular scale around them as accurately as possible according to the believed orientation of the view depicted on the photo. No time limit was given. Once the task had been completed, subjects had to leave the testing room. While subjects were inside the testing room, the experimenter verified on the surveillance screen that no subject placed the photos until the lights were turned on and that all of them looked at the three video screens before and after the darkness phase. After they had left the testing room, subjects were asked whether they had any remarks about the experimental trial they had just performed. Eventually, the experiment ended in a debriefing time, during which all free comments were recorded and specific questions about the awareness of landmark displacements and about the following of instructions were asked. 2.4. Experimental conditions The procedure described above was repeated in three distinct experimental conditions (Fig. 3): Stable Landmarks condition (Stable-L), Rotated Landmarks condition (Rot-L) and Switched Landmarks condition (Switch-L). In the Stable-L condition, the video screens displayed the same images before and after the darkness phase. These images corresponded to the viewpoints that subjects would have had of the environment if the screens had been windows (i.e., Screen 1 displayed View 1, Screen 4 displayed View 4, Screen 7 displayed View 7). In the Rot-L condition, the images displayed on the screens before the darkness phase were the same as in the Stable-L condition, but were shifted by 120◦ after the darkness phase in either a clockwise direction (i.e., Screen 1 displayed View 4, Screen 4 displayed View 7 and Screen 7 displayed View 1; 8 subjects) or a counter-clockwise direction (i.e., Screen 1 displayed

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

9

Fig. 3. Procedure in the Stable-L, Rot-L and Switch-L conditions. The hexagon symbolizes the testing room. The numbers outside this hexagon correspond to the scenes in the real-world environment. The numbers inside the hexagon correspond to the scenes displayed by the screens. First, each screen displayed the image of one view of the environment as if this screen were a window on the environment. Subjects were instructed to look carefully at each screen in order to take their bearings as precisely as possible. During this phase, the spotlights inside the testing room were turned off and light came only from the video screens. After 90 s, the images disappeared and the screens remained dark for 30 s. Then the screens displayed images of the environment again and subjects had to look once again at each of the screens. After an additional 90 s, the screens were switched off and the spotlights were turned on. This was the signal to perform the pointing task.

View 7, Screen 4 displayed View 1 and Screen 7 displayed View 4; 11 subjects). In the Switch-L condition, the images displayed on the screens before the darkness phase corresponded to the correct viewpoints. After the darkness phase, one screen displayed the same image as before the darkness phase, whereas the images displayed on the other two screens were switched (Stable View 1: 7 subjects; Stable View 4: 6 subjects; Stable View 7: 6 subjects). For each subject, the condition order was randomized. Between two conditions, subjects had to go outside the testing room and therefore could recalibrate their spatial representation of the surrounding environment. 2.5. Subjects Nineteen healthy volunteers (ten women, nine men) took part in the study. All of the subjects were recruited from among the visitors to the exhibition. None of them had a history of neurological disease, head injury or psychiatric disorders and none

3rd Reading September 7, 2007

10

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

complained of memory difficulties. All subjects had normal, or corrected to normal, vision. Subjects were asked to evaluate the frequency of their visits to the gallery on a scale from zero (never) to four (very often). Informed consent was obtained after the nature and possible consequences of the study were explained. Women and men did not differ significantly in terms of age (F (1, 15) = 0.07, p = 0.79), education (F (1, 15) = 0.50, p = 0.49) or frequency of visits to the gallery (F (1, 15) = 2.40, p = 0.14). See Table 3 for details. The subjects were divided into two groups: the immobile group (five women, five men) and the mobile group (five women, four men). Subjects in the mobile group were instructed to close their eyes during the darkness phase and turn round while counting aloud up to 30, then stop turning and open their eyes again. Before the beginning of the experiment, the experimenter gave the subjects an example of how quickly to count (approximately 2 per second) in order to make sure that they did not still have their eyes closed after the 30 s of darkness. Subjects in the immobile group had no specific instructions on what to do during the 30 s of darkness. The immobile and mobile groups did not differ significantly in terms of either age (F (1, 15) = 0.42, p = 0.53), education (F (1, 15) = 0.50, p = 0.49) or frequency of visits to the gallery (F (1, 15) = 0.14, p = 0.71). See Table 4 for details. 2.6. Data analysis For each view, the pointed direction was indicated by the red line at the bottom of the corresponding photo and was read directly by the experimenter on the circular Table 3. Characteristics of women and men groups in terms of age, education and frequency of visits to the study environment (gallery).

No. in Immobile group/ No. in Mobile group Age (years) Education (years) Frequency of visits to the gallery (from 0 [never] to 4 [often])

Women (N = 10) Mean (SD)

Men (N = 9) Mean (SD)

5/5

5/4

29.5 (10.5) 14.4 (1.6) 2.8 (1.5)

28.2 (9.3) 14.9 (1.5) 1.7 (1.4)

Table 4. Characteristics of the immobile and mobile groups in terms of age, education and frequency of visits to the study environment (gallery).

No. of women/ No. of men Age (years) Education (years) Frequency of visits to the gallery (from 0 [never] to 4 [often])

Immobile Group (N = 10) Mean (SD)

Mobile Group (N = 9) Mean (SD)

5/5 30.4 (11.0) 14.9 (1.3) 2.1 (1.5)

5/4 27.2 (8.3) 14.3 (1.7) 2.4 (1.7)

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

11

scale from the video recordings of subjects’ responses. For each view, each condition and each subject, the angular deviation, which ranged from −180◦ to 180◦ , was calculated as the difference in degrees between the correct and pointed directions. Three types of error were considered for statistical analyses. For each subject and each condition, the global orientation error (in ◦ ) was calculated as the mean of angular deviations (relative values) over the nine views and indicated the shift of the 9-photos configuration with respect to the outside environment. For each view, each condition and each subject, the angular pointing error (in ◦ ) was calculated by subtracting the global orientation error from the angular deviation (relative values) and therefore corresponded to the angular accuracy of pointing independently of the global orientation of the 9-photos configuration. For each view, each condition and each subject, the categorical pointing error was equal to zero if the corresponding view was correctly placed in the configuration of views, that is, in the sequence of views when rotating on the spot (i.e., first for View 1, second for View 2, third for View 3 and so on) and it was equal to one in the opposite case. This simple scoring was appropriate for the present data, since the sequence errors always resulted from switching two successive views.

3. Results All the subjects understood and followed the instructions. During the darkness phase, all subjects of the mobile group turned round and stopped turning a few seconds before the screens inside the room were switched on again. In contrast, subjects in the immobile group either remained immobile in front of the screen they faced at the beginning of the darkness phase, or turned and faced another screen, then remaining immobile until all the screens were switched on again. At the end of the experiment, all subjects of the mobile group confirmed they closed their eyes during the darkness phase, whereas all subjects of the immobile group told the experimenter they kept their eyes open throughout the experiment. Furthermore, all subjects from both mobile and immobile groups confirmed that, during the pointing task, they aimed and tried to place the photographs as precisely as possible oriented with respect to the outside environment. 3.1. Awareness of landmark displacement In the Stable-L condition, only two (two women, both in the mobile group) of the 19 subjects had the (erroneous) feeling that the landmark views had been rotated after the darkness phase. In the Rot-L condition, 12 (two women and five men in the immobile group; two women and three men in the mobile group) of the 19 subjects noticed that the landmark views had been rotated after the darkness phase. In the Switch-L condition, all but two men in the mobile group noticed that two of the landmark views had been switched after the darkness phase. Some examples of

3rd Reading September 7, 2007

12

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

subjects’ comments, which allowed us to determine whether or not a given subject had become aware of landmark displacements, were as follows: “I have a feeling that the images have been rotated” (for the Rot-L condition), “Two of the images have been switched” (for the Switch-L condition) or “I am sure the images have been displaced”. When asked at the end of the experiment whether they sometimes perceived a discrepancy between the locations of the landmark views before and after the darkness phase, all subjects that did not spontaneously report awareness of such discrepancies answered “no”. They even expressed surprise at the question. To assess more precisely the awareness of landmark displacements in the Rot-L and Switch-L conditions, we used statistical tests. The analyses reveal that a higher proportion of men (89%) than women (40%) detected the rotation of the landmark views in the Rot-L condition (Chi2 = 4.87, p < 0.05), whereas there were statistically as many women (100%) as men (78%) that noticed the switch in the Switch-L condition (Chi2 = 2.48, p = 0.12). Significantly more women detected the switch of the landmark views in the Switch-L condition (100%) than detected the rotation of these views in the Rot-L condition (40%) (Chi2 = 8.57, p < 0.005). In contrast, there were statistically as many men who noticed the rotation in the RotL condition (89%) as men who noticed the switch in the Switch-L condition (78%) (Chi2 = 0.40, p = 0.53). These results are summarized in Fig. 4. On the other hand, there were statistically as many subjects in the mobile group as subjects in the immobile group who detected the landmark view displacement, whatever the condition (56% versus 70%, respectively, in the Rot-L condition, Chi2 = 0.42, p = 0.51; 89% versus 90%, respectively, in the Switch-L condition, Chi2 = 0.01, p = 0.94). 3.2. Global orientation error To assess the subjects’ representation of their global orientation with respect to the environment, an analysis of variance examining the absolute value of global

Fig. 4. Percentage of women and men who became aware that the landmark views had been displaced in the Rot-L and Switch-L conditions. ∗ indicates a significant difference (p < 0.05).

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

13

orientation error was conducted. A 2 × 2 × 3 ANOVA design was used, GROUP (immobile versus mobile) and GENDER (women versus men) being considered as between-subject factors, while CONDITION (Stable-L versus Rot-L versus Switch-L) was manipulated within-subjects. The ANOVA showed a significant main effect of the condition (F 2, 30 = 4.98, p < 0.05). A post-hoc Tukey test showed that the global orientation error was greater in the Rot-L condition (59◦ ) than in the Switch-L (19◦ , p < 0.05) and Stable-L (26◦ , p < 0.05) conditions. There was also a significant main effect of gender (F 1, 15 = 6.98, p < 0.05), suggesting that the global orientation error was greater in women (50◦ ) than in men (18◦ ). Interestingly, an interaction between gender and condition approached significance (F 2, 30 = 2.64, p = 0.07), suggesting that the effect of gender on the global orientation error was greatest in the Rot-L condition. In order to go further in assessing gender differences, three supplementary t-tests were conducted, considering the data in the Stable-L, Rot-L and Switch-L conditions separately. The results showed that, the global orientation error was significantly greater in women (90◦ ) than in men (25◦ ) in the Rot-L condition (p < 0.01), whereas there was no significant gender difference in the Stable-L (39◦ and 11◦ , respectively) and Switch-L (19◦ and 19◦ , respectively). The results concerning the effect of gender on the global orientation error are summarized in Fig. 5. On the other hand, there was a significant main effect of the group (F 1, 15 = 8.60, p < 0.05), suggesting that the global orientation error was greater in the mobile group (53◦ ) than in the immobile group (18◦ ). However, there was no interaction between the group and the gender (F 1, 15 = 0.51, p = 0.49) or between the group and the condition (F 2, 30 = 0.23, p = 0.80).

Fig. 5. Global orientation error (in degrees) in women and men according to the experimental condition and the group. ∗ indicates that global orientation was significantly less accurate in women than in men in the Rot-L condition (p < 0.05). Error bars represent the standard errors. Note that the global orientation was also significantly greater in the mobile group than in the immobile group, but without any interaction with the gender or the condition.

3rd Reading September 7, 2007

14

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

Fig. 6. Individual global orientation error (in degrees) in the the three experimental conditions. Each couple of letters refers to one single subject.

Of 19 subjects, 11 had a global orientation error smaller than 60◦ (M = 15%), whereas 8 (one man and four women in the mobile group; three women in the immobile group) had a global orientation error greater than 60◦ (M = 121◦ ) in the Rot-L condition. To assess whether the subjects with large global orientation errors based their pointing on the location of the landmark views after the darkness phase, the individual signed global orientation errors were considered. In five of the eight subjects, the mean pointing direction followed the rotation of the landmark views, whereas it rotated in the opposite direction in the remaining three subjects. Interestingly, only one of the former five subjects was (partially) aware of the rotation of the landmark views, compared to two of the latter three subjects. Furthermore, analysis of the individual data revealed that, in the Stable-L condition, the global orientation error was between 2◦ and 44◦ for 17 of the 19 subjects (M = 10◦ ), whereas it was 145◦ and 168◦ in two women in the mobile group, respectively. These two women were the only subjects that reported a feeling, in the Stable-L condition, that the landmark views had been displaced after the darkness phase. In the Switch-L condition, the global orientation error was between 0◦ and 17◦ for 17 of the 19 subjects (M = 7◦ ), whereas it was 132◦ and 111◦ in one woman and one man in the mobile group, respectively. These two subjects were aware that

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

15

the landmark views had been displaced but reported that they based their pointing on the location of View 1, which was one of the two switched landmark views. Individual data are summarized in Fig. 6. 3.3. Angular pointing error To assess the subjects’ accuracy in placing the photos independently of the global orientation with respect to the environment, an analysis of variance examining the mean angular pointing error over the nine views (absolute values) was conducted. A 2 × 2 × 3 ANOVA design was used, GROUP (immobile versus mobile) and GENDER (women versus men) being considered as between-subject factors, while CONDITION (Stable-L versus Rot-L versus Switch-L) was manipulated within subjects. The ANOVA showed a significant main effect of gender (F 1, 15 = 10.13, p < 0.01), suggesting that the angular pointing error was greater in women (15◦ ) than in men (10◦ ). No other effect was found. 3.4. Categorical pointing error To assess whether subjects respected the spatial configuration of the environment when performing the pointing task, an analysis of variance examining the mean categorical pointing error over the nine views was conducted. A 2×2×3 ANOVA design was used, GROUP (immobile versus mobile) and GENDER (women versus men) being considered as between-subject factors, while CONDITION (Stable-L versus Rot-L versus Switch-L) was manipulated within subjects. The ANOVA showed a significant main effect of the gender (F 1, 15 = 8.15, p < 0.05), suggesting the mean categorical pointing error was greater in women (13%) than in men (2%). No other effect was found. 3.5. Correlations between the different types of error In order to test the independence between the different types of error, correlation analyses over the 19 subjects were performed separately for the three experimental conditions. The results showed that the categorical and angular pointing errors were highly correlated whatever the condition (r 2 = 0.64 in the Stable-L condition, r 2 = 0.45 in the Rot-L condition, r 2 = 0.61 in the Switch-L condition, p < 0.05), reflecting the fact that switching two photos during the pointing task inevitably entailed an increase of the angular pointing error. By contrast, there was no significant correlation between the global orientation error and the other types of error whatever the condition. 3.6. Correlations between each type of error and the characteristics of subjects in term of age, education and gallery frequenting In order to assess the dependence of the subjects’ performances on the level of education, the level of gallery frequenting and on the age, correlation analyses over the

3rd Reading September 7, 2007

16

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

19 subjects were performed separately for the three experimental conditions and the three types of error. The results showed that the angular pointing error was negatively correlated with the level of education (r 2 = 0.52, p < 0.05). The categorical pointing error was positively correlated with the level of gallery frequenting in the Switch-L condition (r 2 = 0.25, p < 0.05) and with the age in the Rot-L condition (r 2 = 0.24, p < 0.05). There was no significant correlation concerning the global orientation error.

4. Discussion The present study assessed gender differences by distinguishing two navigational processes: spatial updating by self-motion and landmark-based orientation. Subjects were asked to perform a pointing task in three experimental conditions, which differed in terms of the reliability of the external landmarks. Two groups of subjects were distinguished, a mobile group and an immobile group (each with a comparable number of women and men), in which spatial updating of environmental locations by self-motion did not have the same importance in enabling subjects to perform the pointing task correctly. Three types of error were distinguished in order to describe the performance of subjects at the pointing task: the global orientation error, the angular pointing error and the categorical pointing error. However, correlation analyses revealed that the categorical and angular pointing errors were highly correlated whatever the condition, reflecting that switching two photos during the pointing task inevitably entailed an increase of the angular pointing error. It is therefore not possible to rule out that the angular pointing error only mirror the categorical pointing error in the present study. Accordingly, only the results concerning the global orientation error and the categorical pointing error will be discussed below. First of all, the potential limitations to our study due to the characteristics of the experimental environment have to be examined. Indeed, the experiment was run in a complex real-world environment, so that some parameters such as the ambient noise or the number of persons present in the gallery could not be fully controlled. Accordingly, we cannot totally rule out that different subjects perceived the environment differently. Furthermore, most of the subjects already came to the gallery before the experiment and were therefore more or less familiar to the environment. Consequently, we cannot totally rule out that the perception of the environment differed from one subject to the other and we are aware that this may be a bias in the results. However, it is worth noting that the environmental parameters such as noise or people in the gallery, if hardly controllable, were naturally set at random for each subject. On the other hand, correlation analyses showed that the level of gallery frequenting across the subjects could explain a small part of the variance of the categorical pointing error but not the global orientation error or the angular pointing error. Accordingly, some limitations due to the characteristics of the experimental environment exist but have to be relativized. Nevertheless, we

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

17

think that the findings we discuss below will have to be replicated in further studies using an unknown and fully controlled laboratory environment. The first result to be discussed is that the global orientation error was significantly greater in the mobile group than in the immobile group, whatever the condition. Such a result was expected. Indeed, subjects in the immobile group kept their eyes open throughout the experimental trial, including during the darkness phase and were not instructed to move. They were consequently considered unlikely to lose their bearings. In contrast, subjects in the mobile group had to close their eyes and turn round between the two presentations of the landmark views, and were therefore potentially disoriented. So as not to lose their bearings, subjects in the mobile group could, as they turned, update the locations of the landmark views in an internal egocentric representation. However, because spatial updating by self-motion, is subject to a considerable degree of drift when no external compass is possible [2, 3], subjects in the mobile group were nevertheless expected to make greater global orientation errors than subjects in the immobile group. For the same reason, subjects in the mobile group were also expected to be less likely to become aware of the rotation of landmarks in the Rot-L condition than subjects in the immobile group. Indeed, because of the drift in spatial updating, one may hypothesize that, in the mobile group, the expected location of the landmark views was not accurate and so did not really conflict with their actual rotated locations after the darkness phase, resulting in a non-awareness of the rotation. However, the results did not show any significant effect of the group on landmark displacement awareness in the Rot-L condition. One explanation for this result could be that even the subjects in the mobile group noticed the landmark rotation because it was very large (120◦ ). On the contrary, however, the data suggest that more subjects in the immobile group than expected remained unaware of the landmark rotation. This point will be discussed below. Concerning the Switch-L condition, there was no difference between the mobile and the immobile group in the awareness of the landmark switch. This finding was expected since subjects could detect that two views had been switched simply by remembering the relative positions of the views, i.e. their spatial configuration (categorical information), without the need for information from spatial updating (metric information). Subjects of the immobile group, therefore, did not have any particular advantage over subjects of the mobile group in the Switch-L condition. Another important result of the present study is that the global orientation error was greater in the Rot-L condition than in both the Stable-L and Switch-L conditions, whereas there was no difference between the latter two conditions. Furthermore, even if individual data suggest that the performances of some women were similar to those of men and vice versa, group analyses indicate that, overall, women made significantly greater global orientation errors than men in the Rot-L condition but not in the Stable-L or Switch-L conditions. One explanation for these results may be that women and men differ in their ability to use the spatial updating

3rd Reading September 7, 2007

18

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

process [4, 36], which was crucial to make correct use of the landmark views when performing the pointing task in the Rot-L condition. However, this process was crucial in the Rot-L condition only for subjects in the mobile group. Furthermore, although categorical information about landmark configuration was sufficient for subject to become aware of the landmark switch in the Switch-L condition, knowing which of the three views had been switched and which one remained at the same location required metric information only provided by the spatial updating process. Accordingly, in the event of gender differences in the ability to use the spatial updating process, no particular interaction between the condition and the gender would be expected but there should be a significant interaction between the gender and the group. In fact, the results show quite the opposite, suggesting that the present findings cannot be explained by gender differences in the ability to use the spatial updating process. Another explanation of the data must therefore be found. Becoming aware of the landmark rotation in the Rot-L condition required a comparison of the actual rotated locations of the landmarks with metric information from an updateable egocentric representation of where these landmarks were expected to be. In contrast, becoming aware of the landmark switch in the Switch-L condition required a comparison of the actual landmark configuration with categorical information from an allocentric representation of what the landmark configuration should be. Interestingly, the results indicate that significantly more women than men remained unaware of the landmark rotation, whereas no such gender difference was found concerning the landmark switch. Furthermore, significantly fewer women noticed the landmark rotation than those who noticed the landmark switch. In the light of these findings, an explanation for the results concerning the global orientation error may therefore be that women and men differ in their strategic use of an egocentric representation of landmark locations rather than in their ability to update such a representation to accommodate self-motion. Indeed, in the Rot-L condition, women seem to have trusted the external landmarks at their rotated locations, without referring to an internal representation of the expected egocentric landmark locations. Accordingly, women likely performed the pointing task on the basis of the rotated landmarks, resulting in large global orientation errors. In contrast, men were aware of the landmark rotation and made small global orientation errors, suggesting that they referred to an egocentric representation of where the landmarks were expected to be. In the Switch-L condition, men were also aware of the landmark displacement and made small global orientation errors. Likewise, women did not consider the landmark views at their new locations reliable and made small global orientation errors. As already discussed above, reference to an egocentric representation of landmark locations was indispensable in order to know which views had been switched and thus correctly perform the pointing task in the Switch-L condition. These results therefore suggest that, unlike men, women did not spontaneously use such a representation to take their bearings (cf. Rot-L condition) but could refer to it if need be (cf. Switch-L condition) and that navigational differences between the sexes may not be due to differences in the way spatial information is processed

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

19

(in terms of ability) but rather due to differences in higher level strategy. More precisely, it can be hypothesized that women do not spontaneously use a representation of where the landmarks are expected to be relative to themselves but rather take their bearings relative to the actual locations of the landmarks as they appear in the environment. In contrast, men would expect the landmarks at given locations according to an updateable egocentric spatial representation and would take their bearings relative to both the actual and expected landmark locations. It is worth noting that the data reported by James and Kimura in a study assessing object location memory [15] are very consistent with our results. Indeed, these authors found that women were significantly better than men at detecting object exchanges (when objects change location with each other, as in our Switch-L condition), but not at detecting that objects had been shifted from their original location to new locations (as in our Rot-L condition). How does this hypothesis relate to the numerous data in the literature [16, 21, 28, 45], that suggest that environmental representations are more accurate in terms of metric information in men than in women but are richer in landmark information in women than in men? In fact, spatial updating of an egocentric representation requires the use of metric movement information (i.e., angles and distances). It is therefore likely that subjects who spontaneously use the spatial updating process to take their bearings continuously manipulate such metric information, which can at the same time be incorporated in environmental representations. On the other hand, subjects who preferentially take their bearings relative to external landmarks do not need to process metric information about self motion. Accordingly, the gender differences in orientation strategies would entail greater processing of metric information in men than in women. In other words, the fact that the environmental representations are more accurate in terms of metric information in men than in women may be due to the fact that male orientation strategies rely on quantitative information about self motion. Taken together, the results concerning the global orientation error and the angular pointing error therefore suggest that the men used the spatial updating process (a process providing metric information) more readily than the women, whereas the women used categorical information, namely information about the spatial configuration of the landmark views, at least as readily as men. This view is consistent with the findings reviewed in the introduction about gender differences in environmental representations constructed from navigation. It is also consistent with other results on non-navigational spatial tasks. For example, men have been found to be more accurate than women when asked to remember the absolute metric locations of objects in a tabletop-like paradigm [32, 33]. Conversely, Silverman and Eals [8, 39] demonstrated that women outperformed men when asked to indicate which objects in a previously presented array had changed their relative location. Men were also found to answer faster than women when calibrating distances between the elements of a visual stimulus but slower when making judgements about the relative position of these same elements [34].

3rd Reading September 7, 2007

20

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

The question is then to explain why the categorical pointing error was greater in women than in men in the present study. Indeed, such a result seems to go against the findings of Silverman and Eals referred to above as well as against the fact that women were more numerous than men to become aware of the modification of the landmark views configuration (categorical information) in the switch-L condition. One explanation is that the configuration of objects used by Silverman and Eals was two-dimensional and intelligible from a single point of view, whereas, in the present study, the configuration of views was three-dimensional and required different viewpoints on the environment to be apprehended. It should also be noted that it was obviously much more difficult to distinguish between two views in the present study than between two of the objects used by Silverman and Eals. One other possibly important point is that views differ from objects in that there is a degree of continuity between views, while a clear distinction between two items is probably crucial to judge their mutual spatial relation in terms of categories. Consequently, our categorical pointing error may refer not only to categorical processing but also to the ability to make continuity judgments. Further experiments will be needed in order to elucidate this point. As a conclusion, the results of the present study showed a difference between the sexes in the use of two navigational processes: spatial updating from self-motion versus landmark-based orientation. Furthermore, it can be hypothesized from the present data that this gender difference may not be due to differences in specific abilities or in the way spatial information is processed but rather due to differences in higher level strategy. Even if these findings will have to be replicated with larger numbers of subjects in the different groups as well as using a fully controlled laboratory environment, we think that they may help to better understand many findings about gender differences in spatial orientation that have been reported in the literature. Acknowledgments Simon Lambrey was supported by the Lilly Institute. The project was supported by the European way finding process contract. The equipment was provided by the contemporary art center of Tarbes, France. The authors wish to thank Daniel Firman for constructing the experimental set up used in the study and France Maloumian for the graphs presented in this article. References [1] Astur RS, Ortiz ML, Sutherland RJ, A characterization of performance by men and women in a virtual Morris water task: A large and reliable sex difference, Behav Brain Res 93:185–190, 1998. [2] Benhamou S, Sauv´e JP, Bovet P, Spatial memory in large scale movements: Efficiency and limitation of the egocentric coding process, J Theor Biol 145:1–12, 1990.

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

21

[3] Bovet J, Mammals, in Papi F (ed.), Animal Homing, Chapman & Hall, London, pp. 321–361, 1992. [4] Burgess N, Spiers HJ, Paleologou E, Orientational manoeuvres in the dark: Dissociating allocentric and egocentric influences on spatial memory, Cogn 94:149–166, 2004. [5] Chance SS, Gaunet F, Beall AC, Loomis JM, Locomotion mode affects the updating of objects encountered during travel: The contribution of vestibular and proprioceptive inputs to path integration, Presence 7:168–178, 1998. [6] Choi J, Silverman I, Sexual dimorphism in spatial behaviors: Applications to route learning, Evo Cogn 2:165–171, 1996. [7] Dabbs JM, Chang EL, Strong RA, Milun R, Spatial ability, navigation strategy, and geographic knowledge among men and women, Evo Hum Behav 19:89–98, 1998. [8] Eals M, Silverman I, The hunter-gatherer theory of spatial sex differences: Factors mediating the female advantage in location memory, Etho Sociobio 15:95–105, 1994. [9] Galea LAM, Kimura D, Sex differences in route-learning, Pers Indiv Differ 14:53–65, 1993. [10] Gillner S, Mallot HA, Navigation and acquisition of spatial knowledge in a virtual maze, J Cogn Neurosci 10:445–463, 1998. [11] Glasauer S, Amorim MA, Vitte E, Berthoz A, Goal-directed linear locomotion in normal and labyrinthine-defective subjects, Exp Brain Res 98:323–335, 1994. [12] Gwinn HM, Fernando S, James S, Wilson JF, Do landmarks help or hinder women in route learning? Percept Mot Skills 95:713–718, 2002. [13] Iachini T, Sergi I, Ruggiero G, Gnisci A, Gender differences in object location memory in a real three-dimensional environment, Brain Cogn 59:52–59, 2005. [14] Ivanenko Y, Grasso R, Israel I, Berthoz A, Spatial orientation in humans: Perception of angular whole-body displacements in two-dimensional trajectories, Exp Brain Res 117:419–427, 1997. [15] James TW, Kimura D, Sex differences in remembering the locations of objects in an array: Location-shifts versus location-exchanges, Evo Hum Behav 18:155–163, 1997. [16] Jones CM, Braithwaite VA, Healy SD, The evolution of sex differences in spatial ability, Behav Neurosci 117:403–411, 2003. [17] Joshi MS, MacLean M, Carter W, Children’s journey to school: Spatial skills, knowledge and perceptions of the environment, Brit J Dev Psychol 17:125–139, 1999. [18] Kimura D, Sex, sexual orientation and sex hormones influence human cognitive function, Curr Opin Neurobiol 6:259–263, 1996. [19] Klatzky RL, Loomis JM, Beall AC, Chance SS, Golledge RG, Spatial updating of selfposition and orientation during real, imagined, and virtual locomotion, Psychol Sci 9:293–298, 1998. [20] Kuipers B, The “map in the head” metaphor, Environ Behav 14:202–220, 1982. [21] Lawton CA, Gender differences in way-finding strategies: Relationship to spatial ability and spatial anxiety, Sex Roles 30:765–779, 1994. [22] Linn MC, Petersen AC, Emergence and characterization of sex differences in spatial ability: A meta-analysis, Child Dev 56:1479–1498, 1985. [23] Loomis JM, Beall AC, Klatzky RL, Golledge RG, Evaluating the sensory inputs to path integration, Meeting of the Psychonomic Society, Los Angeles, 1995.

3rd Reading September 7, 2007

22

11:10

WSPC/179-JIN

00157

Lambrey & Berthoz

[24] Loomis JM, Klatzky RL, Golledge RG, Cicihelli JG, Pellegrino JW, Fry PA, Nonvisual navigation by blind and sighted: Assessment of path integration ability, J Exp Psychol Gen 122:73–91, 1993. [25] Loomis JM, Klatzky RL, Golledge RG, Philbeck JW, Human navigation by path integration, in Golledge RG (ed.), Wayfinding Behavior, The Johns Hopkins University Press, pp. 125–151, 1999. [26] Maguire EA, Burgess N, O’Keefe J, Human spatial navigation: Cognitive maps, sexual dimorphism, and neural substrates, Curr Opin Neurobiol 9:171–177, 1999. [27] Mallot HA, Gillner S, Route navigating without place recognition: What is recognised in recognition-triggered responses? Perception 29:43–55, 2000. [28] McGuiness D, Sparks J, Cognitive style and cognitive maps: Sex differences in representations of a familiar terrain, J Ment Imag 7:91–100, 1983. [29] McNamara TP, Rump B, Werner S, Egocentric and geocentric frames of reference in memory of large-scale space, Psychon Bull Rev 10:589–595, 2003. [30] Mittelstaedt ML, Glasauer S, Idiothetic navigation in gerbils and humans, Zoologische jahrbucher Abteilungen f¨ ur algemeine Zoologie und Physiologie der Tiere 95:427–435, 1991. [31] Mou W, McNamara TP, Intrinsic frames of reference in spatial memory, J Exp Psychol Learn Mem Cogn 28:162–170, 2002. [32] Postma A, Izendoorn R, de Haan EH, Sex differences in object location memory, Brain Cogn 36:334–345, 1998. [33] Postma A, Winkel J, Tuiten A, van Honk J, Sex differences and menstrual cycle effects in human spatial memory, Psychoneuroendocrinology 24:175–192, 1999. [34] Rybash JM, Hoyer WJ, Hemispheric specialization for categorical and coordinate spatial representations: A reappraisal, Mem Cognit 20:271–276, 1992. [35] Sandstrom NJ, Kaufman J, Huettel SA, Males and females use different distal cues in a virtual environment navigation task, Brain Res Cogn Brain Res 6:351–360, 1998. [36] Saucier DM, Green SM, Leason J, MacFadden A, Bell S, Elias LJ, Are sex differences in navigation caused by sexually dimorphic strategies or by differences in the ability to use the strategies? Behav Neurosci 116:403–410, 2002. [37] Sholl MJ, The relation between horizontality and rod-and-frame and vestibular navigational performance, J Exp Psychol Learn Mem Cogn 15:110–125, 1989. [38] Siegel AW, White SH, The development of spatial representations of large-scale environments, Adv Child Dev Behav 10:9–55, 1975. [39] Silverman I, Eals M, Spatial sex differences: Evolutionary theory and data, in Barkow J, Cosmides L, Tooby J (eds.), The Adapted Mind: Evolutionary Psychology and the Generation of Culture, Oxford University Press, New York, pp. 487–503, 1992. [40] Simons DJ, Wang RF, Perceiving real-world viewpoint changes, Psychol Sci 9:315–320, 1998. [41] Trullier O, Wiener SI, Berthoz A, Meyer JA, Biologically based artificial navigation systems: Review and prospects, Prog Neurobiol 51:483–544, 1997. [42] Tversky B, Levels and structure of spatial knowledge, in Kitchin R, Freundschuh S (eds.), Cognitive Mapping: Past, Present and Future, Routledge, London, pp. 24–43, 2000.

3rd Reading September 7, 2007

11:10

WSPC/179-JIN

00157

External Landmarks versus Spatial Representations

23

[43] Tversky B, Functional significance of visuospatial representations, in Shah P, Miyake A (eds.), Handbook of Higher-Level Visuospatial Thinking, Cambridge University Press, Cambridge, 2003. [44] Voyer D, Voyer S, Bryden MP, Magnitude of sex differences in spatial abilities: A meta-analysis and consideration of critical variables, Psychol Bull 117:250–270, 1995. [45] Waller D, Knapp D, Hunt E, Spatial representations of virtual mazes: The role of visual fidelity and individual differences, Hum Factors 43:147–158, 2001. [46] Wang R, Spelke E, Human spatial representation: Insights from animals, Trends Cogn Sci 6:376, 2002. [47] Wang RF, Simons DJ, Active and passive scene recognition across views, Cognition 70:191–210, 1999. [48] Wang RF, Spelke ES, Updating egocentric representations in human navigation, Cognition 77:215–250, 2000.