Spatial reorientation with non-visual cues: Failure to spontaneously use auditory information

Abstract

Among the environmental stimuli that can guide navigation in space, most attention has been dedicated to visual information. The process of determining where you are and which direction you are facing (called reorientation) has been extensively examined by providing the navigator with two sources of information—typically the shape of the environment and its features—with an interest in the extent to which they are used. Similar questions with non-visual cues are lacking. Here, blindfolded sighted participants had to learn the location of a target in a real-world, circular search space. In Experiment 1, two ecologically relevant non-visual cues were provided: the slope of the floor and an array of two identical auditory landmarks. Slope successfully guided behaviour, suggesting that proprioceptive/kinesthetic access is sufficient to navigate on a slanted environment. However, despite the fact that participants could localise the auditory sources, this information was not encoded. In Experiment 2, the auditory cue was made more useful for the task because it had greater predictive value and there were no competing spatial cues. Nonetheless, again, the auditory landmark was not encoded. Finally, in Experiment 3, after being prompted, participants were able to reorient by using the auditory landmark. Overall, participants failed to spontaneously rely on the auditory cue, regardless of how informative it was.

Keywords

Navigation reorientation blind auditory cue slope or slant

The ability to navigate in an environment and reach a goal is a fundamental behaviour linked to animals’ reproductive success. The first step in navigation is reorientation, that is, determining where you are and which direction you are facing. We might be consciously aware of reorienting, for example, when coming out from an elevator in a multi-level parking lot, when we spend a few seconds looking for clues in the environment (e.g., a sign, a wall, a landmark) that help us remember where we parked our car. Reorientation is typically studied in small-scale, room-sized environments and with an emphasis on a dichotomy between two types of information that can be used: the geometric shape of the environment (e.g., the lengths of the walls of a room) and its non-geometric features (e.g., the colours of the walls) (Cheng, 1986; Gallistel, 1990; for a recent review, see Cheng, Huttenlocher, & Newcombe, 2013). A fertile line of research has been addressing the extent to which the shape of the environment is prioritised for reorientation compared with features (e.g., Hermer & Spelke, 1994; Twyman & Newcombe, 2010). However, most attention has been devoted to reorientation guided by visual cues.

Non-visual cues are important to study because they are ecologically relevant. When navigating in many natural and urban environments, we may rely, for example, on the slope of the terrain (going up or down a hill) or a sound (a river, a bell ringing, car traffic). This occurs to a greater extent under environmental conditions of low visibility, such as at night or with fog. Furthermore, non-visual reorientation is crucial for those of us with visual impairments; studying the effectiveness of different types of cues in guiding behaviour is useful for developing technological devices that can aid navigation (Giudice, 2018; Loomis, Golledge, & Klatzky, 1998; Loomis, Klatzky, & Golledge, 2001; Marston, Loomis, Klatzky, & Golledge, 2007).

Another reason to address not only visual cues is to answer this fundamental question: To what extent is the representation of space similar when the environment is learned through different sensory modalities? Growing evidence supports the theory of functional equivalence (Bryant, 1997; Loomis, Klatzky, Avraamides, Lippa, & Golledge, 2007), according to which, despite different encoding modalities, many aspects of the spatial representation and behaviour are equivalent (Avraamides, Loomis, Klatzky, & Golledge, 2004; Giudice, Betty, & Loomis, 2011; Giudice, Klatzky, & Loomis, 2009; Levine, Jankovic, & Palij, 1982; Yamamoto & Shelton, 2005). For example, in a recent study on haptic-based reorientation (Sturz, Gaskin, & Roberts, 2014), blindfolded participants had to learn the location of a target object by exploring with their hands the corners of a rectangular enclosure. The target was hidden in one of the corners, and each corner had distinct textural cues. Therefore, the task could be easily solved by remembering the texture cue associated with the target (feature information). Alternatively, participants could remember that the target was in a specific corner of the rectangle by relying on its shape (geometric information); however, this would lead to an ambiguity (50% accuracy) because each corner is geometrically equivalent to its rotational opposite. In a similar scenario but with visual cues, abundant evidence (Hermer & Spelke, 1996; Learmonth, Newcombe, & Huttenlocher, 2001; Vallortigara, Zanforlin, & Pasti, 1990; for a review, see Cheng & Newcombe, 2005) has shown that human and non-human animals tend to learn the target location with both sources of information and not with just the more informative one. Sturz and colleagues (2014) revealed that haptic-based reorientation followed the same pattern: participants learned both the target texture cue and the geometrically correct location, although the texture information alone would have sufficed.

Another source of spatial information that can give non-visual perceptual access to the environment is the slope of the terrain. A uniformly slanted, walkable surface (a hill, a ramp, etc.) can be perceived by a combination of visual, proprioceptive, and kinesthetic sensory input, and it provides a navigator with directional information that can be used to determine a heading (uphill, downhill, sideways, and every angle in between), just like a compass, and to encode a target location. Successful use of slope for reorientation has been shown in rats (Miniaci, Scotto, & Bures, 1999; Moghaddam, Kaminsky, Zahalka, & Bures, 1996) and homing pigeons (Nardi & Bingman, 2009; Nardi, Nitsch, & Bingman, 2010). In human studies, participants have been tested in a square tilted enclosure, with the target hidden in one of the corners. Although with large individual differences, both adults (Nardi, Newcombe, & Shipley, 2011) and children (Holmes, Nardi, Newcombe, & Weisberg, 2015) were able to remember the correct corner when the only strategy available was referencing the slope of the floor (e.g., “the target is uphill on the left”). However, in this line of research, participants experienced the slope with all the sensory modalities, including vision (although the visual information was poor). Therefore, the question of whether proprioceptive/kinesthetic access to a slanted environment is sufficient to guide navigation remains unanswered.

A source of information particularly useful for creating a global representation of the environment are auditory cues. Unlike visual information, sounds provide information about space even when the navigator is not facing the source. Furthermore, an auditory cue can guide navigation even when the source is far from the navigator; unlike haptic information, no physical contact is necessary with the source to perceive and encode a sound. Despite this, the literature on auditory-based reorientation is poor. With non-human animal models, to the best of our knowledge, the only two available studies reported mixed evidence. One found that, in a Morris water maze, rats did not learn the task (Rossier, Haeberli, & Schenk, 2000); however, in an analogous task, Watanabe and Yoshida (2007) reported that mice succeeded. In humans, the study closest to the classic reorientation paradigm a la Cheng (1986), in a room-sized environment, is by Viaud-Delmon and Warusfel (2014). They found that participants could use an array of three distinct auditory landmarks to learn the target location in a real-world Morris water maze adaptation. Evidence from other studies, in large-scale environments, suggests that auditory cues can be used for navigation (Blauert, 1997; Klatzky, Marston, Giudice, Golledge, & Loomis, 2006; Loomis, Klatzky, McHugh, & Giudice, 2012), although it has been reported that spatial relations may not be learned as effectively with auditory cues compared with visual cues (Sturz, Kilday, Bodily, & Kelly, 2012). In sum, how easily auditory information is used to reorient is still unclear.

The main goal of this study was to address a basic question: whether auditory cues are used more or less based on how informative they are. Although there are contrasting positions (e.g., a modularity position for geometry; Cheng, 1986; Hermer & Spelke, 1994; Lee, Shusterman, & Spelke, 2006), the adaptive combination view (Newcombe & Huttenlocher, 2006; Xu, Regier, & Newcombe, 2017), an influential model of spatial learning, holds that if multiple cues are available, each one is used depending—among other factors—on how reliably it predicts the target location. This is based on evidence that spatial memory combines cues in a Bayesian fashion to reduce variance (Cheng, Shettleworth, Huttenlocher, & Rieser, 2007; Nardini, Jones, Bedford, & Braddick, 2008). Even a straight-forward associative account of reorientation would assume that a stimulus which better predicts the reward will exert greater control over behaviour than a stimulus with low predictive value (e.g., see Miller & Shettleworth, 2007). Would this generalise to reorientation with non-visual cues? We attempted to answer this by manipulating the reliability of the auditory cue: low in Experiment 1 and high in Experiment 2.

Experiment 1

Blindfolded sighted participants completed a reorientation task in which they had to learn the location of a target in a real environment. This was a circular search space on a platform, and the target could be anywhere along the circumference (Figure 1). Traditionally, reorientation studies mostly use discrete target locations in the corners of a polygonal enclosure—frequently a rectangle, square, or trapezoid (see Cheng et al., 2013). Thus, in most cases, there are only four possible responses. The circular search space was chosen because it has the advantage of providing an almost continuous dependent variable which ensures greater sensitivity; to the best of our knowledge, only one reorientation study has used it to date (Twyman, Holden, & Newcombe, 2017). In our task, participants searched for the target object and encoded its location; then, after being disoriented, they had to replace the target where it used to be.

Figure 1.

Schematic representation of the environment during training and test trials in Experiment 1. During training, the slope of the floor determined one, unambiguous, correct location, and the array of two identical auditory landmarks (the noisemakers) determined two, diametrically opposite, correct locations. In the slope-only test, the noisemakers were turned off. In the auditory-only test, the platform was horizontal (no slope). In the conflict test, the noisemakers were rotated with respect to their position during training, which determined different correct locations according to each source of information.

In Experiment 1, during an initial training phase, we provided two strategies for solving the task: auditory information and terrain slope (Figure 1). An array of two identical auditory landmarks (noisemakers generating white noise) specified the target location and its diametrical opposite; therefore, they predicted the correct location only with 50% accuracy. Furthermore, the slope of the floor determined the target location with certainty. We expected participants to learn the task and replace the target with smaller errors than chance. Subsequent test trials without feedback were carried out to ascertain which strategy was guiding behaviour. First, two single-cue tests were conducted: one with only the auditory information and one with only the slope. The purpose of these was to examine whether participants had encoded each cue and were able to use it alone for reorientation. We expected both sources of information to be encoded, as evidenced by smaller errors than chance. Particular attention was devoted to the slope-only test; if participants succeeded in this, it would suggest that perceiving slope through proprioceptive/kinesthetic input is sufficient to guide reorientation. Finally, a conflict test rotated one spatial cue with respect to the other, determining different correct locations based on the information used (Figure 1); because auditory information was ambiguous (low predictive value), we hypothesised that slope would be prioritised and control behaviour.

Method

Participants

Participants were 23 (15 females, eight males) undergraduate students enrolled in psychology courses who volunteered as a means of fulfilling course requirements or for extra credit. Participants signed up for the study using an online participation management system (SONA), in which the study was described in the following way: “This is a study on spatial abilities. Simply, you will have to find an object in a room and then remember where it is without using vision.” They were also encouraged to bring their own earphones and music. Only one participant did not complete the study because the spinning on the swivel chair made her feel uncomfortable; to avoid getting dizzy, the experimenter stopped the procedure and this participant was not included in the final sample (not part of the 23 subjects).

Apparatus

The experiment took place in a room measuring 3.8 × 3.7 m². The room was quiet and had the experimental apparatus in the centre. The experimental apparatus consisted of a square wooden platform (side of 215 cm, 10 cm tall) covered by grey carpet, with a circular PVC pipe (215 cm in diameter, 2.3 cm thick) on top. The circular pipe was the search space, and the target object (a hairclip measuring 9 cm) would be clipped on it. A swivel chair (base: 46 cm diameter; seat height: 59 cm), used to disorient participants, was in the centre of the platform, standing on a wooden base (56 cm × 56 cm, 2 cm thick). The platform could be placed horizontally on the floor or could be tilted at a 4° angle by raising one side over wooden blocks. When the platform was tilted, a wedge was placed under the base of the swivel chair such that its axis of rotation was always parallel to the force of gravity. Two small noisemakers (HoMedics, model SS-2000A, dimensions: 14 cm × 16 cm, 4.5 cm tall) were placed on the platform, in two opposite corners (distance of the speaker from the centre of the platform: 140 cm). The noisemakers generated white noise at a volume of approximately 40 dB measured from the centre. See Figure 1 for a schematic representation of the apparatus. It should be noted that the location of the target object was always on the circle at 90° away from the noisemakers; therefore, with respect to the noisemakers, there were two rotationally equivalent target locations, diametrically opposite each other.

Procedure

Training

Participants were tested individually. One experimenter (out of two male and two female experimenters) ran the study. Upon arrival in the experimental room, the platform was tilted and the noisemakers were turned on. Participants signed the consent form, and then the experimenter explained and showed the following procedure step-by-step (the participant was given time to familiarise with the apparatus and practise the procedure before starting). The participant had to wear the blindfold for the whole session. They sat down on the swivel chair and were gently spun around to lose their sense of orientation (disorientation procedure). The disorientation lasted for approximately 60 s, varying speed and changing direction at least once. Every time the participant was spun, they had to wear noise-cancelling earphones with music playing; the purpose of this was to cover the auditory cue during disorientation. The facing direction of the chair after spinning could be uphill, downhill, or a side. The facing direction changed pseudo-randomly each time the participant was spun, with the constraint that the same facing direction could not be repeated within three iterations. The sequence of facing directions was counterbalanced across participants. After spinning, the participant stood up and walked around the platform to find the target on the circle; they were instructed to always hold a hand on the back of the chair. The experimenter gave clues as to how close they were getting to the target (“cold, warm, hot”). The target was in the same location for all participants, and it was a fixed location for all training trials (reference memory paradigm). When near the target, the participant would kneel and find the target on the circular search space. They were told to remember where it was, but were not told how to do it. When ready, they would pick up the target and sit back on the chair. After being spun again, they would have to stand up and put the target back on the circle as close as possible to where it used to be. After the participant replaced the target on the circle, the location was recorded, and the experimenter gave feedback: the target was placed back in the correct location, and the participant was given clues until they found it again and picked it up. This made up the first training trial. The trial was repeated (disorientation, replacement, feedback) three more times (a total of four training trials). The only difference was that, in the last training trial, participants had to leave the room for 1 min prior to replacing the target (leaving the room was implemented to create an analogous situation to the test trials, when leaving the room was necessary so that the platform could be raised or lowered). Therefore, this is the sequence for the last training trial: Participants picked up the target, were disoriented on the swivel chair, then they removed the blindfold and left the room, waited for approximately 1 min, re-entered, re-equipped the blindfold, were disoriented again, and then had to replace the target and received feedback. The purpose of the training phase was to determine whether participants could learn the location of the target at a level of accuracy above that expected by chance (90° error), that is, whether they could successfully reorient using the available non-visual cues.

Testing

After the training trials, three test trials were carried out. During the testing phase, participants did not receive feedback. The sequence of test trials was the following and was the same for all participants, except that the first two test trials (slope-only and auditory-only) were presented in counterbalanced order across participants. For each test trial, after leaving the room for 1 min (during which the platform was raised or lowered), participants came back inside and had to replace the target as usual; basically, each test trial followed the same procedure as the last training trial, with the following exceptions.

Slope-only test

The noisemakers were turned off before the participant re-entered the room; therefore, the only available cue for replacing the target where it used to be was the slope of the platform (see Figure 1). This determined one unambiguous, correct location. The purpose of this test was to assess whether participants had encoded the target location with respect to the slope of the floor and could use that cue alone to reorient.

Auditory-only test

Before re-entering the room, the platform was placed horizontally on the floor; therefore, the only available cue for replacing the target where it used to be was the auditory information provided by the two noisemakers (see Figure 1). This determined two diametrically opposite correct locations. The purpose of this test was to assess whether participants had encoded the target location with respect to the auditory landmarks and could use them alone to reorient. (The order of presentation of this trial was counterbalanced with the slope-only test.)

Conflict test

The platform was tilted; however, this time the noisemakers were moved to the other two opposite corners of the platform, that is, the noisemakers rotated 90° with respect to their learned position (see Figure 1). This created a situation of conflict between the correct replacement locations predicted by the available cues: there was one correct location according to the slope (slope-correct location; stable with respect to the incline) and two different correct locations according to the auditory landmarks (auditory-correct locations; stable with respect to the noisemakers, rotated 90° with respect to the incline).

Follow-up tests

After the conflict test, participants were disoriented and then they were told to choose one noisemaker and to place the hairclip on the circle in the direction from which the white noise was coming (referred to as noise localization test). This served to determine a baseline of accuracy for noise source localization and completion of a simple sensorimotor task. Subsequently, participants were disoriented and then they were told to place the hairclip on the circle in the direction of uphill (referred to as uphill identification test). This served to determine a baseline of accuracy for the identification of uphill and completion of a simple sensorimotor task. After this, the experimental session was over and participants took off the blindfold.

Debriefing

Participants were asked what information they used to remember the location of the target and to replace it where it used to be. Then, they were debriefed about the details and purpose of the study. None of the phases of the study were timed, and participants were told to take as much time as they needed. The overall duration of the experimental session was on average 50 min.

Data analysis

The circular PVC pipe had 160 equally spaced marks that were used to measure where participants replaced the target object (the hairclip), and this was converted into angles. Because we had no interest in directionality, the dependent variable was the absolute angular error, that is, the unsigned angle subtended by the participant’s response and the correct target location; this can be analysed with linear statistics (Batschelet, 1981). One-sample t-tests were used to assess whether the average accuracy was above chance (for similar analyses of behavioural responses, see, for example, Ishikawa & Montello, 2006; Weisberg, Schinazi, Newcombe, Shipley, & Epstein, 2014).

The slope of the platform predicted the target location with certainty. When testing whether participants were using this cue, the absolute error was coded such that it could range from 0° (replacement at the exact correct location) to 180° (largest possible error), with performance at chance yielding on average 90° error. Successful use of slope to encode the target would be supported by an average error significantly smaller than 90°.

The auditory landmarks (the two noisemakers) predicted the target location with rotational ambiguity because they specified both the correct location and its diametrically opposite location. When testing whether participants were using this cue, the absolute error was coded as the angular distance from the closest correct location, such that it could range from 0° (replacement at either of the exact correct locations) to 90° (largest possible error), with performance at chance yielding on average 45° error. Successful use of auditory information to encode the target would be supported by an average error significantly smaller than 45°.

Although the sample size of male subjects was very limited and does not warrant consideration of gender differences, for the sake of completeness, the gender factor was included as a variable in the analyses.

Results

Training

Because the slope predicted the target location with certainty, participants’ average replacement error during the training trials was compared with a chance level of 90° (Figure 2). A one-sample t-test revealed that the average error during the first training trial (M = 67.7, standard error of the mean [SEM] = 11.0) was marginally smaller (not significantly) than chance, t(22) = −2.019, p = .056, d = 0.42. However, in the remaining training trials, the errors were significantly smaller than chance: Trial 2 (M = 62.5, SEM = 11.5), t(22) = −2.378, p = .027, d = 0.50; Trial 3 (M = 64.4, SEM = 11.0), t(22) = −2.318, p = .030, d = 0.48; and Trial 4 (M = 52.2, SEM = 10.0), t(22) = −3.771, p = .001, d = 0.79. When testing for the use of auditory landmarks in the first training trial, the average error (considering diametrically opposite correct locations) was not significantly smaller than chance (45°), t(22) = −0.615, p = .545, d = 0.13. As expected, these results suggest that participants learned to replace the target at a level above chance.

Figure 2.

Average absolute error during training and test trials in Experiment 1 (±SEM). For training, the slope-only test, and the conflict test, the error shown is relative to the slope-correct location and chance is 90°. Conversely, for the auditory-only test, the error shown is relative to the auditory-correct locations (two diametrically opposite locations), and chance is 45°. The error was significantly smaller than expected by chance for the training trials (starting from Trial 2), the slope-only test, and the conflict test.

A 2 (sex) × 4 (trials) mixed analysis of variance (ANOVA) was conducted on the replacement errors. There was no significant main effect of sex, F(1, 21) = 2.576, p = .123, $η_{p}^{2} = . 11$ , no significant main effect of trial, F(3, 63) = 0.193, p = .900, $η_{p}^{2} = . 01$ , and no significant interaction, F(3, 63) = 0.623, p = .603, $η_{p}^{2} = . 03$ .

Slope-only test

The use of slope to replace the target was tested by comparing the average error to a chance level of 90°. A one-sample t-test revealed that the average error (M = 58.9, SEM = 12.6) was significantly smaller than chance, t(22) = −2.452, p = .023, d = 0.51 (Figure 2). There was no significant difference between sexes, t(21) = 0.782, p = .443, d = 0.33. As expected, these results suggest that participants encoded the slope and could use it alone.

Auditory-only test

The use of auditory information to replace the target was tested by comparing the average error (considering diametrically opposite correct locations) to a chance level of 45°. A one-sample t-test revealed that the average error (M = 35.5, SEM = 6.1) was not significantly different from chance, t(22) = −1.559, p = .133, d = 0.33 (Figure 2). There was no significant difference between sexes, t(21) = 0.276, p = .785, d = 0.12. Contrary to expectations, these results suggest that participants did not encode the auditory array and could not use it alone.

Conflict test

Because the array of auditory landmarks moved relative to the slope, each source of information specified different correct locations. First, we considered the use of the slope. The replacement error was calculated with respect to the slope-correct location and was compared with a chance level of 90°. A one-sample t-test revealed that the average error (M = 52.6, SEM = 9.6) was significantly smaller than chance, t(22) = −3.884, p = .001, d = 0.81 (Figure 2). Participants were making errors smaller than expected by chance with respect to the slope-correct location. There was not a significant difference between sexes, t(21) = 1.379, p = .182, d = 0.58.

Then we considered the use of auditory landmarks. The replacement error was calculated with respect to the closest of the diametrically opposite auditory-correct locations and was compared with a chance level of 45°. A one-sample t-test revealed that the average error (M = 52.2, SEM = 5.7) was not significantly different from chance, t(22) = 1.290, p = .211, d = 0.26. Participants were making errors not different from chance with respect to the auditory-correct locations. There was not a significant difference between sexes, t(21) = 0.434, p = .669, d = 0.20.

As expected, this set of results suggests that participants were relying on the slope more than the auditory information.

Follow-up tests

One participant’s data (female) for the follow-up tests were lost. In the noise localization test, participants’ average error was 10.3° (SEM = 1.9), which is significantly smaller than chance (90° error), t(21) = −42.403, p < .001, d = 8.99. There was no significant sex difference, t(20) = −1.497, p = .150, d = 0.61.

In the uphill identification test, participants’ average error was 18.9° (SEM = 3.4), which is significantly smaller than chance (90° error), t(21) = −20.493, p < .001, d = 4.40. There was no significant sex difference, t(20) = 0.553, p = .586, d = 0.26.

The errors in the noise localization test and uphill identification test were compared with a paired t-test. The error was significantly smaller in the noise localization test, t(21) = 2.295, p = .032, d = 0.66.

Correlations

To better understand individual differences in performance, we looked into correlations among errors during training (last trial), slope-only test, auditory-only test, conflict test, noise localization test, and slope identification test. The replacement error during training correlated significantly with that during the slope-only test, r(23) = .674, p < .001, and with the error to the slope-correct location during the conflict test, r(23) = .611, p = .002. Furthermore, the error during the slope-only test correlated significantly with that to the slope-correct location during the conflict test, r(23) = .568, p = .005. Other correlations were not statistically significant. Participants who reoriented better were those who encoded the slope more accurately and used it to guide their behaviour.

Self-reported strategy use

When asked what information they were using to remember the location of the target and to replace it, participants overwhelmingly mentioned more than a single source of information. Eighteen participants (78%) mentioned the slope/floor, and only four mentioned the white noise (17%). Of the remaining non-effective strategies, the one that was mentioned the most was keeping track of spinning on the swivel chair (nine participants; 39%).

Discussion

In a slanted environment with an array of two identical auditory landmarks, blindfolded and disoriented participants learned the location of a target and were able to replace it where it used to be. The subsequent test trials revealed two main findings. First, when slope was presented alone (slope-only test), accuracy was greater than chance. The fact that blindfolded participants were able to reorient suggests that visual sensing of slope is not necessary. To the best of our knowledge, this is the first clear study to isolate the proprioceptive/kinesthetic sensory contribution to slope perception and to indicate that this information alone is able to guide navigation on a geographical slant for humans.

Second, the test trials showed that when the auditory information was presented alone (auditory-only test), errors were not different from chance; in other words, participants were disoriented. Furthermore, when the two possible strategies were set in conflict (conflict test), the slope of the floor guided behaviour.

These results support our hypotheses only in part. We expected participants to learn the task and perform better than chance during training. We also expected them to show successful encoding of slope and to use slope over the auditory information during the conflict test because the slope was more informative. The array of two identical noisemakers determined the target location with ambiguity because it specified the correct location and its diametrical opposite. Nonetheless, we hypothesised that participants would encode the auditory landmarks even if they were less reliable of a predictor. Our expectation was based on previous studies on visual-based reorientation, which overall reveal that a source of spatial information—especially a geometric cue—tends to be encoded even when not necessary or sufficient (e.g., Cheng, 1986; Cheng et al., 2013). This has been shown even with haptic cues (Sturz et al., 2014). In contrast, in Experiment 1, there was no evidence that the auditory information was encoded at all.

The white noise was disregarded despite evidence suggesting that it is an easier signal to localise compared with other auditory sources, for example, a pure tone or noises with smaller bandwidth (Butler, 1986; Shigeno & Oyama, 1983; Tonning, 1975). Furthermore, in Experiment 1, the noise localization test revealed that participants were able to localise the auditory source, and the localization error was significantly smaller than that in the uphill identification test; this is an indication that a noisemaker has the potential of being a more accurate source of information for reorientation than the slope of the floor.

A plausible reason why auditory information was not encoded is because it was not very useful for the task at hand. The array of two auditory landmarks was not a good predictor of the target location. Furthermore, there was an additional cue, the slope of the floor, which also determined the target location—even with better accuracy. It is possible that this cue competition led to the slope gaining all the associative strength and overshadowing the auditory information. To test this hypothesis, in Experiment 2 we created a situation in which the auditory cue was a much more useful source of spatial information. Would it be encoded then?

Experiment 2

Experiment 2 addressed—for the first time, to the best of our knowledge—reorientation with a single auditory landmark. Compared with Experiment 1, the floor of the platform was horizontal, and now there was only one noisemaker. This was the only useful source of information for the task, and it determined with certainty (100% accuracy) the target location (see Figure 3). Because of this, we hypothesised that auditory information would be encoded, as revealed by replacement errors smaller than chance during training.

Figure 3.

Schematic representation of the environment in Experiments 2 and 3 during training. Unlike Experiment 1, the floor was horizontal (no slope) and there was only one noisemaker. Therefore, the only spatial cue predictive of the target was the auditory landmark, which determined the correct location with certainty.

As a secondary goal, after training, Experiment 2 included a control test trial with the noisemaker turned off. The purpose of this was to ensure that there were no other unwanted sources of information in the environment that could have been used to reorient. We hypothesised that, with no slope and no auditory cues, participants would replace the target with an error not different from chance.

Finally, Experiment 2 also included a test in which participants had to replace the target without wearing the blindfold (visual replacement test). The purpose was to assess whether adding vision would improve participants’ replacement accuracy.

Method

Participants

Sixteen new participants (12 females, four males) were collected. The study was advertised in the same way as Experiment 1. Data collection for the two experiments overlapped in time during the semester.

Apparatus

The same apparatus of Experiment 1 was used, with these two notable exceptions: (a) the platform was placed horizontally on the floor (no slope information), and (b) there was only one noisemaker (see Figure 3). Therefore, the target location was predicted only by the auditory cue, and because the target object was always on the circle at a fixed angle (90° counterclockwise) away from the noisemaker, the auditory cue this time determined the target location without ambiguity.

Procedure

The general procedure was analogous to Experiment 1, with only the following points worth emphasising.

Training

The same procedure of Experiment 1 was followed, for a total of four training trials (see Figure 3). The target was in the same location for all participants, and it was a fixed location for all training trials (reference memory paradigm). Just like in Experiment 1, participants were not told what strategy to use to solve the task.

Control test

After the training trials, a control test was carried out. This followed the same procedure of the last training trial (analogous to testing in Experiment 1), with the only difference being that the noisemaker was turned off before the participant re-entered the room. Therefore, there were no cues available to replace the target. No feedback was provided. The purpose of this trial was to ensure that there were no unwanted cues for reorientation and that, on a horizontal platform without auditory information, participants were disoriented.

Noise localization test

After the control test, participants were disoriented and then they were told to place the hairclip on the circle in the direction from which the white noise was coming. No feedback was provided.

Visual replacement test

For the last trial, after being disoriented, participants took off the blindfold and were told to replace the target where it used to be. The purpose of this test was to assess whether adding vision would improve participant’s accuracy. No feedback was provided.

Debriefing

Participants were asked what information they used to remember the location of the target and to replace it. Then, they were debriefed about the details and purpose of the study. None of the phases of the study were timed, and participants were told to take as much time as they needed. The overall duration of the experimental session was on average 30 min.

Data analysis

Because the auditory cue predicted the target location with certainty, the absolute error was coded such that it could range from 0° (replacement at the exact correct location) to 180° (largest possible error), with performance at chance yielding on average 90° error. Successful use of the auditory cue to encode the target would be supported by an average replacement error significantly smaller than chance.

Results

Training

Participants’ average replacement error during the training trials was compared with a chance level of 90° (Figure 4). One-sample t-tests revealed that the average error during each trial was not significantly different from chance: Trial 1 (M = 91.6, SEM = 13.9), t(15) = 0.126, p = .901, d = 0.03; Trial 2 (M = 64.3, SEM = 15.5), t(15) = −1.660, p = .118, d = 0.42; Trial 3 (M = 69.7, SEM = 13.4), t(15) = −1.506, p = .153, d = 0.38; and Trial 4 (M = 73.5, SEM = 14.7), t(15) = −1.118, p = .281, d = 0.28. Contrary to expectations, these results suggest that participants did not learn the task.

Figure 4.

Average absolute replacement error during training in Experiments 2 and 3 and during test trials (in Experiment 2 only). Error bars represent ±1 SEM. In both experiments, the auditory landmark determined one unambiguous, correct location; however, Experiment 2 was an open task, whereas in Experiment 3 participants were explicitly told to use the auditory cue. The difference between the errors in Experiments 2 and 3 was not significant. However, in Experiment 2, the average error was not significantly different from chance (90°) for all the trials, whereas in Experiment 3 errors were significantly smaller than chance starting from Trial 2.

A 2 (sex) × 4 (trials) mixed ANOVA was conducted on the replacement errors. There was no significant main effect of sex, F(1, 14) = 0.024, p = .879, $η_{p}^{2} = . 01$ , no significant main effect of trial, F(3, 42) = 0.908, p = .445, $η_{p}^{2} = . 06$ , and no significant interaction, F(3, 42) = 0.400, p = .754, $η_{p}^{2} = . 03$ .

Control test

Participants’ average replacement error during the control test was 89.6° (SEM = 15.9), which is not significantly different from chance (90°), t(15) = −0.016, p = .988, d = 0.01 (Figure 4). There was no significant difference between sexes, t(14) = 0.254, p = .803, d = 0.14. As expected, these results suggest that participants were disoriented when no cues for reorientation were provided.

Noise localization test

Participants’ average error was 9.4° (SEM = 3.0), which is significantly smaller than chance (90° error), t(15) = −26.532, p < .001, d = 6.62. There was no significant sex difference, t(14) = 1.960, p = .070, d = 0.87.

Visual replacement test

When replacing the target without the blindfold, participant’s average error was 77.2° (SEM = 14.2), which is not significantly different from chance (90°), t(15) = −0.894, p = .386, d = 0.23 (Figure 4). There was no significant difference between sexes, t(14) = 0.288, p = .777, d = 0.19. We compared the average error between the visual replacement test and the last training trial with a paired t-test; there was no statistically significant difference, t(15) = 0.215, p = .832, d = 0.06. Being able to see did not improve participants’ accuracy.

Correlations

We looked into correlations between errors during training (last trial) and noise localization test, r(16) = −.143, p = .597, between errors during training and visual replacement test, r(16) = .295, p = .268, and between noise localization test and visual replacement test, r(16) = −.004, p = .990. None of these correlations were statistically significant.

Self-reported strategy use

When asked what information they were using to remember the location of the target and to replace it, participants mentioned one or two sources of information. Nine participants (56%) mentioned the white noise. Of the remaining non-effective strategies, the one that was mentioned the most was keeping track of spinning on the swivel chair (eight participants; 50%).

Discussion

Following an analogous procedure to Experiment 1, Experiment 2 included a control test trial, during which participants replaced the target with errors not different from chance. The fact that they were disoriented reassures us that, in Experiment 1, besides the slope and auditory cues, there was no other source of spatial information in the environment that could be used to successfully encode the target location.

Surprisingly, participants failed to reorient during the training trials, when there was an auditory landmark in a consistent position relative to the target location. The replacement error was not different from chance in any of the four trials. This suggests that it was the addition of slope in Experiment 1 that had allowed participants to solve the task. In conclusion, the auditory information was not used even when it was the only cue available and it predicted the target location with certainty. Why did this occur?

First, it is worth mentioning that performance did not improve even when participants removed their blindfold (visual replacement test). This alleviates the doubt that the large replacement errors observed might have been due to the potentially awkward and uncomfortable situation of having to walk on a platform and pick up and replace the target deprived of vision. If this were the case, the task should have been completed more effectively without the blindfold. This did not happen, and it suggests that the reason for the large replacement errors was a failure to encode the auditory cue.

There are two broad factors that might be responsible. Experiments 1 and 2 were both open tasks in which participants were not told which strategy to adopt. It is possible that participants failed to solve Experiment 2 because they did not spontaneously use the auditory cue and attempted to rely on other ineffective strategies. This is confirmed by the fact that, at the end of the experimental session, a large proportion of the sample reported trying to keep track of how much they were being spun on the swivel chair during disorientation (path integration), a strategy that would lead to performing at chance. However, another explanation that cannot be ruled out at this point is that participants were not able to use the auditory cue because of the properties of the stimulus, the experimental setup, or an inability of the participants themselves. Experiment 3 addressed this possibility.

Experiment 3

Experiment 3 replicated the four training trials of Experiment 2, but with the crucial difference that, at the beginning of the session, attention was drawn to the white noise and participants were explicitly told to use it for the task. Therefore, Experiment 3 tested more specifically the ability to encode the auditory cue, rather than the likelihood to spontaneously rely on it above other ineffective strategies. We reasoned that if a single auditory landmark cannot be used to reorient or if there is an issue with the specific auditory signal that we provided, participants should perform at chance even if prompted to rely on this source of information. In contrast, if the problem was the propensity to use the auditory cue (choosing this over other ineffective strategies), now that they are instructed on how to solve the task, they should perform above chance.

Method

Participants

Sixteen new participants (12 females, four males) were collected. The study was advertised in the same way as Experiments 1 and 2, and data collection was completed in the following semester. The mean age of the participants was 19.0 years (SD = 1.3), and the range of ages was 18 to 22 years.

Procedure

The same apparatus and procedure of Experiment 2 were used, but only four training trials were run (see Figure 3). Unlike Experiment 2, to ensure that participants could not use any cue other than the white noise to reorient, rather than adding a control test trial, the location of the noisemaker (and the related correct target location) alternated between two diametrically opposite locations throughout the trials. Crucially, unlike Experiment 2, during the explanation of the procedure, participants were told, “The white noise will help you complete the task. In order to solve the task, you will have to remember where the target is with respect to the white noise.”

Results

Participants’ average replacement error during the four training trials was compared with a chance level of 90° (Figure 4). A one-sample t-test revealed that the average error during the first trial (M = 70.5, SEM = 15.1) was not significantly different from chance, t(15) = −1.293, p = .215, d = 0.32. However, the average errors for the following trials were significantly smaller than chance: Trial 2 (M = 51.6, SEM = 13.2), t(15) = −2.913, p = .011, d = 0.73; Trial 3 (M = 44.7, SEM = 15.9), t(15) = −2.840, p = .012, d = 0.71; and Trial 4 (M = 45.6, SEM = 13.9), t(15) = −3.189, p = .006, d = 0.80. Participants learned to replace the target at a level above chance.

A 2 (sex) × 4 (trials) mixed ANOVA was conducted on the replacement errors. There was no significant main effect of sex, F(1, 14) = 0.057, p = .815, $η_{p}^{2} = . 01$ , no significant main effect of trial, F(3, 42) = 0.955, p = .423, $η_{p}^{2} = . 06$ , and no significant interaction, F(3, 42) = .185, p = .906, $η_{p}^{2} = . 01$ .

We compared the average replacement error in the training trials with that in Experiment 2. A 2 (Experiments 2 and 3) × 4 (training trials) mixed factorial ANOVA revealed that there was no significant main effect of experiment, F(1, 30) = 2.874, p = .100, $η_{p}^{2} = . 09$ , of trial, F(3, 90) = 1.550, p = .207, $η_{p}^{2} = . 05$ , and no significant interaction, F(3, 90) = .131, p = .941, $η_{p}^{2} = . 01$ .

Discussion

Unlike Experiments 1 and 2 in which participants had to figure out which strategy to use, in Experiment 3 they were told how to solve the task. As a result, although performance did not increase significantly with respect to Experiment 2, crucially the replacement errors were significantly smaller than chance. Previous studies have indicated that auditory information can guide navigation in large-scale environments (Blauert, 1997; Klatzky et al., 2006; Loomis et al., 2012) and that an array of distinct auditory landmarks can be used in a small-scale place learning task (Viaud-Delmon & Warusfel, 2014). The present finding extends this evidence and suggests that even a single auditory landmark can guide reorientation in a circular search space.

This excludes the possibility that the specific auditory signal we presented in this study could not be used to solve the task. The failure to encode the stimulus in Experiment 2 was probably not related to a difficulty in the ability to reorient, but in the propensity to spontaneously rely on the auditory cue above other ineffective strategies.

General discussion

This study investigated reorientation by providing access to the environment through two non-visual spatial sources of information: the slope of the floor and an auditory cue. Participants successfully used the slope of the floor even if unprompted. However, they did not rely on the auditory cue unless they were explicitly instructed to do so; this occurred regardless of how informative the auditory cue was for solving the task. We focus our discussion on three main points.

First, this study shows that proprioceptive/kinesthetic access to the slope of the floor is sufficient to guide navigation. Previous studies on rats have been carried out in darkness (Miniaci et al., 1999; Moghaddam et al., 1996), but in human studies the visual information derived from the slope was available (Holmes et al., 2015; Nardi et al., 2011, 2013) A terrain can be perceived as slanted from a number of potential visual cues, for example: (a) trees and buildings are aligned with gravity; thus, their angle of incidence with the slanted ground is acute on one side and obtuse on the other; (b) when keeping your head upright, objects on the uphill side appear higher than on the downhill side. In this study, when only the slope was available (slope-only test trial of Experiment 1), blindfolded participants replaced the target at a level of performance better than chance. This indicates that a slanted floor can be used for reorientation not just in a square environment with discrete target locations in the corners (Holmes et al., 2015; Nardi et al., 2011, 2013) but even in an almost continuous, circular search space. More importantly, this suggests that navigation is possible on a geographical slant with only a combination of the sense of balance, angles of the joints, and muscular effort derived from walking and standing on the tilted floor. Visual access to the environment is not necessary. Is it sufficient? According to a navigation study in a virtual environment (Weisberg & Newcombe, 2014) and a location memory task in a real-world environment (Weisberg, Nardi, Newcombe, & Shipley, 2014), the answer seems to be yes. Altogether, evidence suggests that slope is a multisensory cue providing redundant information for navigation. This redundancy may be a reason why slope, at least at steep inclinations and in pigeon studies, dominates over other competing spatial cues (Nardi & Bingman, 2009; Nardi et al., 2010). There are two dissociable channels to access sloped environments that are ecologically relevant—observing a slanted surface while standing on a horizontal floor, or sensing a slanted floor by walking on it without vision—and each can guide navigation independently. This opens up the possibility that there might be functional equivalence (Bryant, 1997; Loomis et al., 2007) between these channels, leading perhaps to a common representation and supporting similar performance. Future studies will have to address this by using a similar setup and dependent variable (we used an almost continuous variable compared with previous studies which used four discrete options only). Furthermore, the present study did not address gender differences because of the limited amount of male participants in our sample (in none of our experiments were the gender differences statistically significant). It should be noted that remarkable sex differences in favour of men have been found in slope-based reorientation, including visual access to the environment (Nardi et al., 2011). It is possible that the female disadvantage might be reduced when participants are visually deprived because there is less distraction by ineffective visual cues. Follow-up studies should test this using larger samples of men and women.

Second, this study provides the first evidence that a single auditory landmark can be used to solve a small-scale reorientation task. Previous research on auditory-based reorientation is scarce, and only one similar study provided multiple, distinct, auditory cues (Viaud-Delmon & Warusfel, 2014). An auditory environment with an array of sources provides geometric information. Conversely, our study (Experiment 3) shows that a single auditory source can be used for reorientation. This is not surprising. Just like with a visual landmark, the participant only had to encode a specific spatial relation (an angle) between the auditory landmark and the correct location on the circular search space. This is something that in a rectangular enclosure humans can accomplish with vision starting between 3 and 4 years old (Learmonth, Nadel, & Newcombe, 2002). Our findings extend the evidence in support of similar visual-accessed and auditory-accessed landmark-guided navigation.

What is surprising—and makes our third discussion point—is that participants failed to spontaneously use the auditory cue. Without being told in advance which strategy to use to solve the task, there is no evidence that the auditory cue was encoded at all. This occurred when there was a competing spatial cue and the auditory information was ambiguous because it was provided by an array of two identical sources (Experiment 1) and even when the only useful spatial information for the task was a single auditory signal that determined the target location with certainty (Experiment 2). Therefore, increasing its predictive value did not lead to greater associative strength for guiding behaviour. The fact that, when prompted (Experiment 3), participants actually used the same auditory stimulus above chance rules out a general inability in auditory-based reorientation and the possibility that our auditory signal could not guide reorientation. Overall, participants were able to reorient with the auditory information, but if not explicitly told to use it, they seemed to struggle relying on this as opposed to other cues that might (slope) or might not (keeping track of the spinning during disorientation) be effective. Why was the auditory cue not encoded spontaneously?

It is possible that the specific stimulus was not very salient perceptually. To someone who is already directing attention to it, a 40-dB white noise stands out in a quiet room when blindfolded. Previous studies have shown that white noise is easier to localise compared with other types of simple signals with smaller bandwidth, including a pure tone (Butler, 1986; Shigeno & Oyama, 1983; Tonning, 1975). This is why white noise was chosen for this study. Indeed, in Experiments 1 and 2, we found that participants had no problem localising the signal, and their error was comparable to what was shown in the literature (7°; Blauert, 1997). Furthermore, in Experiment 1, the accuracy for localising the auditory source was superior to that for localising the direction of the slope. However, localising a cue that is pointed out is one thing, using it spontaneously to navigate is another thing. In this sense, the auditory cue might have been difficult to be perceived/attended, leading participants to focus more on other strategies. The low salience of our stimulus may be attributed to its properties: being a continuous signal and not loud. In Viaud-Delmon and Warusfel (2014), the auditory cues presented were a cicada, a melody played on a piano, and a text read by a male voice; however, they did not systematically test the extent to which participants had encoded the association between target and each of the cues individually, as opposed to the array of sources. Future studies will have to vary the auditory stimuli, including rhythmic and louder sounds, and signals that are perhaps more frequently associated with reorientation in everyday life (e.g., a clock tower), and test if even in these cases people fail to spontaneously encode them. In addition, the auditory cue might not have been encoded spontaneously because of low motivation from the participants (not enough effort trying to figure out the correct strategy) or too short training session. While we do not doubt that with greater incentive and more training trials participants would have succeeded, these explanations simply move the question to the next step: Why did participants have enough motivation and acquisition to use the slope but not the auditory cue? An ecological explanation is that, in general, auditory information might be assigned lower weight for reorientation compared with other non-visual sources of spatial information. At the present state, this is just a speculation. However, there is some indication from previous research that auditory cues might not be effective for learning spatial relations among locations (humans: Sturz et al., 2012; rats: Rossier et al., 2000). If future studies on blind navigation provide equivalent access to the environment with auditory cues and other non-visual modalities and replicate our findings using different auditory stimuli, it would raise the interesting possibility that humans have a lower propensity to rely on auditory information. For example, our study hints at the fact that people might be more likely to use the slope of the floor.

In conclusion, taking into account the present results and previous studies on slope (Holmes et al., 2015; Nardi et al., 2011), a common finding is that sighted participants have difficulty relying on a non-visual cue for reorientation if blindfolded or the visual information is impoverished. When they have to figure out which strategy to use on their own (open task), many participants are grasping at straws and end up relying on intuitively more difficult or ineffective strategies. As an anecdote, in a pilot phase of this study, instead of using the “obvious”—we thought—slope or auditory cue to reorient, a participant reported relying on the sunshine sensed on their legs when walking in a particular spot of the platform (this is when we decided to completely block the window). A take-home message of our study is that different sensory cues may each theoretically provide sufficient spatial information for solving a task, but this does not mean that they will give equal ease of access to the environment and be equally preferred. This question has important implications for blind people, and given the scarcity of research on navigation mediated by non-visual information (see Giudice, 2018; Schinazi, Thrash, & Chebat, 2016), it deserves more attention.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Avraamides

M. N.

Loomis

J. M.

Klatzky

R. L.

Golledge

R. G.

(2004). Functional equivalence of spatial representations derived from vision and language: Evidence from allocentric judgments. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30, 801-814.

Batschelet

(1981). Circular statistics in biology. London, England: Academic Press.

Blauert

(1997). Spatial hearing: The psychophysics of human sound localization. Cambridge, MA: The MIT Press.

Bryant

D. J.

(1997). Representing space in language and perception. Mind & Language, 12, 239-264.

Butler

R. A.

(1986). The bandwidth effect on monaural and binaural localization. Hearing Research, 21, 67-73.

Cheng

(1986). A purely geometric module in the rat’s spatial representation. Cognition, 23, 149-178.

Cheng

Huttenlocher

Newcombe

N. S.

(2013). 25 years of research on the use of geometry in spatial reorientation: A current theoretical perspective. Psychonomic Bulletin & Review, 20, 1033-1054.

Cheng

Newcombe

N. S.

(2005). Is there a geometric module for spatial orientation? Squaring theory and evidence. Psychonomic Bulletin & Review, 12(1), 1-23.

Cheng

Shettleworth

S. J.

Huttenlocher

Rieser

J. J.

(2007). Bayesian integration of spatial information. Psychological Bulletin, 133, 625-637.

10.

Gallistel

C. R.

(1990). The organization of learning. Cambridge, MA: The MIT Press.

11.

Giudice

N. A.

Betty

M. R.

Loomis

J. M.

(2011). Functional equivalence of spatial images from touch and vision: Evidence from spatial updating in blind and sighted individuals. Journal of Experimental Psychology: Learning, Memory, and Cognition, 37, 621-634.

12.

Giudice

N. A.

Klatzky

R. L.

Loomis

J. M.

(2009). Evidence for amodal representations after bimodal learning: Integration of haptic-visual layouts into a common spatial image. Spatial Cognition & Computation, 9, 287-304.

13.

Giudice

N. A.

(2018). Navigating without vision: Principles of blind spatial cognition. In Montello

D. R.

(Ed.), Handbook of behavioral and cognitive geography (Chapter 15, pp. 260-290). Cheltenham, UK: Edward Elgar Publishing.

14.

Hermer

Spelke

(1994). A geometric process for spatial representation in young children. Nature, 370, 57-59.

15.

Hermer

Spelke

(1996). Modularity and development: The case of spatial reorientation. Cognition, 61, 195-232.

16.

Holmes

C. A.

Nardi

Newcombe

N. S.

Weisberg

S. M.

(2015). Children’s use of slope to guide navigation: Sex differences relate to spontaneous slope perception. Spatial Cognition & Computation, 15, 170-185.

17.

Ishikawa

Montello

D. R.

(2006). Spatial knowledge acquisition from direct experience in the environment: Individual differences in the development of metric knowledge and the integration of separately learned places. Cognitive Psychology, 52, 93-129.

18.

Klatzky

R. L.

Marston

J. R.

Giudice

N. A.

Golledge

R. G.

Loomis

J. M.

(2006). Cognitive load of navigating without vision when guided by virtual sound versus spatial language. Journal of Experimental Psychology: Applied, 124, 223-232.

19.

Learmonth

A. E.

Nadel

Newcombe

N. S.

(2002). Children’s use of landmarks: Implications for modularity theory. Psychological Science, 13, 337-341.

20.

Learmonth

A. E.

Newcombe

N. S.

Huttenlocher

(2001). Toddlers’ use of metric information and landmarks to reorient. Journal of Experimental Child Psychology, 80, 225-244.

21.

Lee

S. A.

Shusterman

Spelke

E. S.

(2006). Reorientation and landmark-guided search by young children: Evidence for two systems. Psychological Science, 17, 577-582.

22.

Levine

Jankovic

I. N.

Palij

(1982). Principles of spatial problem solving. Journal of Experimental Psychology: General, 111, 157-175.

23.

Loomis

J. M.

Golledge

R. G.

Klatzky

R. L.

(1998). Navigation system for the blind: Auditory display modes and guidance. Presence: Teleoperators and Virtual Environments, 7, 193-203.

24.

Loomis

J. M.

Klatzky

R. L.

Avraamides

Lippa

Golledge

R. G.

(2007). Functional equivalence of spatial images produced by perception and spatial language. In Mast

F. W.

Jäncke

(Eds.), Spatial processing in navigation, imagery and perception (pp. 29-48). New York, NY: Springer.

25.

Loomis

J. M.

Klatzky

R. L.

Golledge

R. G.

(2001). Navigating without vision: Basic and applied research. Optometry and Vision Science, 78, 282-289.

26.

Loomis

J. M.

Klatzky

R. L.

McHugh

Giudice

N. A.

(2012). Spatial working memory for locations specified by vision and audition: Testing the amodality hypothesis. Attention, Perception, & Psychophysics, 74, 1260-1267.

27.

Marston

J. R.

Loomis

J. M.

Klatzky

R. L.

Golledge

R. G.

(2007). Nonvisual route following with guidance from a simple haptic or auditory display. Journal of Visual Impairment & Blindness, 101, 203-211.

28.

Miller

N. Y.

Shettleworth

S. J.

(2007). Learning about environmental geometry: An associative model. Journal of Experimental Psychology: Animal Behavior Processes, 33, 191-212.

29.

Miniaci

M. C.

Scotto

Bures

(1999). Place navigation in rats guided by a vestibular and kinesthetic orienting gradient. Behavioural Neuroscience, 113, 1115-1126.

30.

Moghaddam

Kaminsky

Y. L.

Zahalka

Bures

(1996). Vestibular navigation directed by the slope of terrain. Proceedings of the National Academy of Sciences of the United States of America, 93: 3439-3443.

31.

Nardi

Bingman

V. P.

(2009). Pigeon (Columba livia) encoding of a goal location: The relative importance of shape geometry and slope information. Journal of Comparative Psychology, 123, 204-216.

32.

Nardi

Newcombe

N. S.

Shipley

T. F.

(2011). The world is not flat: Can people reorient using slope? Journal of Experimental Psychology: Learning, Memory, and Cognition, 37, 354-367.

33.

Nardi

Newcombe

N. S.

Shipley

T. F.

(2013). Reorienting with terrain slope and landmarks. Memory & Cognition, 41, 214-228.

34.

Nardi

Nitsch

K. P.

Bingman

V. P.

(2010). Slope-driven goal location behavior in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 36, 430-442.

35.

Nardini

Jones

Bedford

Braddick

(2008). Development of cue integration in human navigation. Current Biology, 18, 689-693.

36.

Newcombe

N. S.

Huttenlocher

(2006). Development of spatial cognition. In Damon

Lerner

R. M.

Kuhn

Siegler

R. S.

(Eds.), Handbook of child psychology: Cognition, perception, and language (vol. 2, pp. 737-776, 6th ed.). Hoboken, NJ: John Wiley & Sons.

37.

Rossier

Haeberli

Schenk

(2000). Auditory cues support place navigation in rats when associated with a visual cue. Behavioural Brain Research, 117, 209-214.

38.

Schinazi

V. R.

Thrash

Chebat

D. R.

(2016). Spatial navigation by congenitally blind individuals. Wiley Interdisciplinary Reviews: Cognitive Science, 7, 37-58.

39.

Shigeno

Oyama

(1983). Localization of speech and non-speech sounds. Japanese Psychological Research, 25, 112-117.

40.

Sturz

B. R.

Gaskin

K. A.

Roberts

J. E.

(2014). Incidental encoding of enclosure geometry does not require visual input: Evidence from blindfolded adults. Memory & Cognition, 42, 935-942.

41.

Sturz

B. R.

Kilday

Z. A.

Bodily

K. D.

Kelly

D. M.

(2012). No evidence that consistent auditory cues facilitate learning of spatial relations among locations. Behavioural Processes, 90, 198-203.

42.

Tonning

F. M.

(1975). Auditory localization and its clinical applications. Audiology, 14, 368-380.

43.

Twyman

A. D.

Holden

M. P.

Newcombe

N. S.

(2017). First direct evidence of cue integration in reorientation: A new paradigm. Cognitive Science. Advance online publication. doi: 10.1111/cogs.12575.

44.

Twyman

A. D.

Newcombe

N. S.

(2010). Five reasons to doubt the existence of a geometric module. Cognitive Science, 34, 1315-1356.

45.

Vallortigara

Zanforlin

Pasti

(1990). Geometric modules in animals’ spatial representations: A test with chicks (Gallus gallus domesticus). Journal of Comparative Psychology, 104, 248-254.

46.

Viaud-Delmon

Warusfel

(2014). From ear to body: The auditory-motor loop in spatial cognition. Frontiers in Neuroscience, 8, 283.

47.

Watanabe

Yoshida

(2007). Auditory cued spatial learning in mice. Physiology & Behavior, 92, 906-910.

48.

Weisberg

S. M.

Nardi

Newcombe

N. S.

Shipley

T. F.

(2014). Up by upwest: Is slope like north? The Quarterly Journal of Experimental Psychology, 67, 1959-1976.

49.

Weisberg

S. M.

Newcombe

N. S.

(2014). A slippery directional slope: Individual differences in using slope as a directional cue. Memory & Cognition, 42, 648-661.

50.

Weisberg

S. M.

Schinazi

V. R.

Newcombe

N. S.

Shipley

T. F.

Epstein

R. A.

(2014). Variations in cognitive maps: Understanding individual differences in navigation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 40, 669-682.

51.

Regier

Newcombe

N. S.

(2017). An adaptive cue combination model of human spatial reorientation. Cognition, 163, 56-66.

52.

Yamamoto

Shelton

A. L.

(2005). Visual and proprioceptive representations in spatial memory. Memory & Cognition, 33, 140-150.