Individual Differences in Route-Learning Strategy and Associated Working Memory Resources Carryl L. Baldwin, George Mason University, Fairfax, Virginia, and Ian Reagan, Old Dominion University, Norfolk, Virginia Objective: The current investigation examined individual differences in route-learning strategies and their relative demands on visuospatial versus verbal working memory (WM) resources in virtual environments. Background: Learning new routes is a resourcedemanding activity that must often be carried out in conjunction with other concurrent tasks. Virtual environments (VEs) are increasingly being used for training and research, pointing to the importance of determining the strategies people use to learn routes in these environments. Methods: Participants classified as having good or poor sense of direction (SOD) attempted to learn novel routes while concurrently performing either a verbal (articulatory suppression) or a visuospatial (tapping) WM interference task. Results: Different navigational strategies were observed in each SOD group. Individuals with poor SOD relied more heavily on verbal rather than visuospatial WM resources, as evidenced by greater disruption to route-learning performance from the articulatory suppression task relative to the tapping task. Conversely, individuals with good SOD exhibited more route-learning disruption from the tapping task, suggesting a greater reliance on visuospatial WM resources. Conclusion: Individuals differ from one another in the strategies they use and the WM resources they tap—verbal or visuospatial—to learn routes in VEs. Self-report measures can be used as indices of such individual differences in navigational strategy use in VE tasks. Application: Assessing SOD and associated WM resources have implications for targeted training for navigation in VEs and for the design of in-vehicle navigation systems. INTRODUCTION
Learning routes in unfamiliar environments— new buildings, cities, and so on—is a common everyday task that can be critical for gaining proficiency in many mobile occupations. Increasingly, route learning and other spatial navigation tasks must be performed in simulated, remote, and virtual environments (VEs). One example is the tasks facing operators of unmanned aerial vehicles (UAVs), who are required to navigate in a remotely located environment (Dixon & Wickens, 2006). Navigation poses significant spatial challenges, as UAV operators must maintain directional orientation using remote camera perspectives (Cooke et al., 2006) while frequently performing other essential, cognitively demanding
tasks (Dixon, Wickens, & Chang, 2005). VEs are also used to provide route-learning practice for firefighters, rescue workers, and military personnel prior to their entry into real environments. Methods of facilitating the route-learning process in these high-risk situations are of great importance (Farrell et al., 2003; Ruddle, Payne, & Jones, 1999). People differ considerably in how well they can perform such navigational tasks (Blajenkova, Motes, & Kozhevnikov, 2005; Hegarty, Montello, Richardson, Ishikawa, & Lovelace, 2006; Hugdahl, Thomsen, & Ersland, 2006). Gender, experience, age, navigational strategy, and working memory (WM) ability are some of the factors consistently found to account for significant variance in the performance of navigational
Address correspondence to Carryl L. Baldwin, George Mason University, Psychology, 4400 University Drive, MS 3F5, Fairfax, VA 22030;
[email protected]. Human Factors, Vol. 51, No. 3, June 2009, pp. 368-377. DOI: 10.1177/0018720809338187. Copyright © 2009, Human Factors and Ergonomics Society.
Route-Learning Strategy
tasks in real-world settings (Cornoldi & Vecchi, 2003; Gugerty & Brooks, 2004; Lawton, 1994; Pazzaglia, Cornoldi, & De Beni, 2000; Pazzaglia & De Beni, 2001; Thorndyke & Hayes-Roth, 1982). Such individual variability—whether attributed to heritable or environmental influences or their combination—is an important human factors issue. The workload associated with the navigational task, the impact on performance of varying display designs, and the degree of interference imposed by concurrent task performance can be expected to vary as a function of individual differences in both ability and strategy (Baldwin, 2006). In real-world environments, good navigators tend to use different and more flexible strategies than poor navigators (Kato & Takeuchi, 2003; Lawton, 1994; Saucier et al., 2002). Good navigators, often male, tend to make greater use of cardinal directions, Euclidean distances, and mental maps. These strategies have been termed survey strategies. Conversely, poor navigators tend to use strategies relying on verbal sequential lists of ego-centered or left-right directions referenced to landmarks along a route. Individuals, often female, who rely on these so-called landmark strategies, generally find it difficult or impossible to use more survey-oriented navigational strategies (MacFadden, Elias, & Saucier, 2003). However, good navigators tend to be able to use other, less preferred strategies when the situation demands (Saucier, Bowman, & Elias, 2003). For example, good navigators are able to use verbal sequential navigational directions, such as those provided by many commercial online direction services (i.e., MapQuest), even when maps and cardinal headings are not provided. At the same time that gender differences are often observed in reporting the relative use of these two different strategies, differences in opportunities to gain navigational experience are also reported (Lawton & Kallai, 2002). Use of cardinal directions increases with age (Dabbs, Chang, Strong, & Milun, 1998), and people in general, including females, who grow up in geographical locations where roads and streets are laid out in grid-like patterns, report greater use of cardinal directions (Lawton, 2001). Therefore, the relative use of these two different strategies
369
appears to be based on experiential factors other than gender alone. Despite the extensive body of literature doc umenting the existence of individual differences in the strategies used to navigate and learn routes in real-world settings (Hegarty et al., 2006; Kato & Takeuchi, 2003; Lawton, 1994, 1996, 2001), to date, there has been strikingly little research on the extent to which the strategies observed in real-world settings are the same or different from those used in VEs (Chen & Stanney, 1999). Hegarty et al. (2006) observed that measures of environmental learning taken after direct experience—walking through an environment—compared with experience with a VE were associated with different factor analytic constructs. Furthermore, they observed that psychometric measures (associated with small spatial scales) were more predictive of participants’ ability to learn large-scale environments indirectly via a videotape rather than by direct experience. There are different brain mechanisms associated with different spatial scales (Di Nocera, Couyoumdjian, & Ferlazzo, 2006; Ferlazzo, Fagioli, Di Nocera, & Sdoia, 2008; Gramann, Müller, Schönebeck, & Debus, 2006; Kosslyn & Thompson, 2003; Previc, 1998; Zaehle et al., 2007). For example, egocentric versus allocentric spatial reference frames (Zaehle et al., 2007) and categorical (i.e., left and right) versus coordinate (i.e., Euclidean distances) spatial processing (Baciu et al., 1999) activate different brain mechanisms. These observations call into question the extent to which individuals are likely to use the same strategies and processes when learning routes in VEs as they do in contextually rich, real-world environments. Garden, Cornoldi, and Logie (2002) observed individual differences in WM processes associated with the strategies people used to learn routes in real environments. People who used survey strategies relied more heavily on the visuospatial sketchpad than the phonological loop of Baddeley’s WM model (Baddeley, 1992; Baddeley & Hitch, 1974). Conversely, greater demands were placed on the phonological loop relative to the visuospatial sketchpad when people used a landmark strategy. Specifically, participants learned novel routes by walking
370
through them while engaged in one of two WM interference tasks. One interference task, a pattern tapping task, was designed to disrupt visuospatial processing, whereas the other, an articulatory suppression task, was designed to disrupt use of the phonological loop for verbal processing. Individuals reporting high “maplike” representational usage on a self-report que stionnaire made more navigation errors when retraversing routes learned while engaged in the tapping task relative to the routes they learned while engaged in articulatory suppression. Conversely, individuals who reported low map-like representational usage made more errors in routes learned while engaged in articulatory suppression relative to the tapping task. A limitation of the Garden et al. (2002) investigation is that they did not examine performance on the interference tasks (i.e., accuracy and timing of articulatory rehearsal and spatial tapping). This precludes ruling out the possibility that people may have differentially shifted their attention allocation strategies between the route-learning and interference tasks. For example, rather than simply relying on a visuospatial navigation strategy, individuals with a strong map orientation could have emphasized performance of all spatial tasks. If they allocated little attention to the verbal interference task, then performance on it would be poor, but navigational disruption would be unlikely. This possibility can not be ruled out without examining interference task performance. This and other previous work leaves unanswered the question of what route-learning strategies and associated WM processes people use when navigating in VEs. The present study addressed this fundamental issue. We expected to see a pattern of individual differences in VE navigation indicating that people with good sense of direction (SOD) rely more heavily on visuospatial WM resources whereas people with poor SOD rely more heavily on verbal WM. A second aim of the current investigation was to determine whether differences in task prioritization strategies could explain the individual differences observed in the Garden et al. (2002) study that used real environments rather than VEs. Recording and analyzing performance on both the route-learning and interference tasks allowed this examination.
June 2009 - Human Factors
We chose to use an SOD questionnaire developed by Takeuchi (1992) and later refined by Kato and Takeuchi (2003) as a method of classifying individuals as either having good or poor SOD. Self-reported SOD, defined by Kozlowski and Bryant (1977, p. 590) as “people’s estimation of their own spatial orientation ability,” can be used to predict environmental spatial abilities, such as pointing accuracy and rate of acquiring accurate mental representations of unfamiliar areas (Hegarty, Richardson, Montello, Lovelace, & Subbiah, 2002; Kozlowski & Bryant, 1977). The SOD questionnaire (Kato & Takeuchi, 2003) contains 14 items designed to determine the degree to which people believe they can benefit from or remember things such as cardinal directions, maps, and landmarks. Kato and Takeuchi (2003) observed that their SOD questionnaire also predicted individual differences in strategy use. They divided participants into good- and poor-SOD groups based on their being either one standard deviation above or below, respectively, the mean of a larger sample. They observed that the good-SOD group learned novel routes quicker after walking through them and used different strategies (e.g., relying on cardinal directions) relative to the poor-SOD group. Previous research by our group (Furukawa, Baldwin, & Carpenter, 2004) indicated that the SOD questionnaire effectively distinguished between navigators who benefit more from verbal versus visual map–style navigational aids when learning routes during simulated driving. Individuals classified as having a poor versus a good SOD demonstrated more accurate recall for routes they had driven while using an egocentered verbal guidance relative to a visual map–style aid. Conversely, good-SOD individuals benefited most from an allocentric visual map format and were disrupted when the verbal directions accompanied the visual map. Thus, the SOD questionnaire was deemed well suited for the current experimental aim. We predicted that individuals with good SOD would tend to use visuospatial route-learning strategies resulting in poorer route memory (operationally defined as more navigational errors and longer traversal times) for routes learned in conjunction with visuospatial interference (spatial tapping) than with articulatory suppression.
Route-Learning Strategy
Conversely, poor-SOD participants were expected to use strategies relying more heavily on verbal WM and so were expected to have poorer memory for routes learned in conjunction with articulatory suppression relative to the spatial tapping interference. Finally, we predicted that performance changes on the interference tasks themselves would provide support for the hypothesized individual differences in routelearning strategy use and associated WM resources rather than for the alternative explanation of differential task allocation strategies. METHODS Participants
Participants were recruited from a larger sample of 234 respondents who had previously completed the SOD questionnaire if they met prespecified criteria for being classified as having poor or good SOD (Reagan, 2005). Individuals scoring at least one standard deviation above or below the mean composite score of the larger sample were classified as having good or poor SOD, respectively. We attempted to acquire equal numbers of males and females in both SOD groups. Of 234 survey respondents, 42 scored above the criterion (19 female), and 41 scored below (31 female). Forty-two individuals completed the experiment, 20 with good SOD (11 females, 9 males, ages 18 to 44 years) and 22 with poor SOD (15 females, 7 males, ages 19 to 30 years). Procedures approved by a university institutional review board were followed, and all participants received research participation credit. All participants reported normal or corrected-to-normal visual acuity and depth perception. Materials
Route-learning environment. A commercially available computer game, the Sum of All Fears 2002, was used for the VE. Routes had 11 turn points, defined as places in the route where an individual had to change directions. Route A contained 16 choice points. Route B had 15 choice points. Choice points were defined as places where one could turn left or right or remain walking in the same direction. In other words, choice points included both correct turn points and distracter turns (where turning would result in an error). It took an average of 150.97 s
371
to traverse Route A and 154.91 s to complete Route B. Participants used a mouse and their right hand to maneuver their character from an ego-centric perspective through the VE. Length of time to traverse the route (time) and navigational errors (errors) made during the first attempt trials served as performance measures. WM interference tasks. Modeled after Garden et al. (2002), the verbal interference task required articulatory suppression in which individuals spoke an ordered series of syllables: “bay, bee, by, bo, boo, tay, tea, tie, toe, too.” Participants repeated the syllables in sequence at a rate of one syllable per second. Responses were taperecorded. The proportion of correctly uttered sequences was analyzed. The spatial interference task required sequential spatial tapping by pressing an ordered sequence of six keys on a Targus® 18-key USB keypad at a rate of one key per second. The proportion of correctly executed sequences was analyzed. Mental rotation task (MRT). Participants completed Cooper and Shepard’s (1973) MRT. The letter R, F, or L was presented either correctly or as a mirror image at varying degrees of rotation. Participants indicated as quickly as possible via key press whether the letter was forward or backward. A short practice trial of 20 stimuli was followed by the experimental block consisting of 82 letter stimuli. Participants had 5 s to respond to each stimulus. If participants failed to respond within 5 s, the next letter on the screen appeared. Response time (RT) and accuracy were analyzed. Procedure
After giving informed consent, participants practiced each task individually and then concurrently before completing the experimental blocks. Experimental task practice block. Participants practiced moving their avatar through the VE. When the participant demonstrated adequate control of the avatar, an interference task was introduced. Participants then practiced moving their avatar while engaging in each interference task, presented in counterbalanced order. Criterion performance for each was defined as the ability to walk the avatar in a continuous rectangular pattern without bumping into objects while executing five correct interference task sequences.
372
Baseline performance. After achieving criterion during practice, participants performed a baseline trial for each task. Participants walked their character in the same rectangular pattern while performing the interference task for a baseline period of 2.5 min. Route-learning practice. A route-learning practice scenario was incorporated next to reduce an expected learning effect. It contained four turn points and six choice points. Participants were instructed to “do your best to remember the route because you will be asked to complete it from memory upon reaching the destination point.” Participants followed an avatar controlled by the experimenter through the route once and then attempted to navigate the route from memory. If a mistake was made, the participant was corrected as soon as possible and asked to continue through the route. Forty of the 42 participants completed the practice route without error. Experimental trials. Next, the experimental trials began. Participants again followed the experimenter’s avatar once through the route and then attempted the route from memory. If participants made a route error, they were immediately corrected and were asked to continue toward the destination. Trials containing errors resulted in the participants being required to navigate the route again. This process continued until the route was completed without error or for a maximum of four attempt trials. The four combinations of route (A and B) and the two interference conditions were counterbalanced within each SOD group. Interference task instructions. Participants began the interference task as soon as the experimental trial began. The experimenter stressed that remembering the route was the number one goal. Participants were told that performance on each interference task was recorded and that they should try to be as accurate as possible. If the experimenter noticed that a participant stopped the interference task, the experimenter prompted the participant to resume the task with the statement, “Resume.” One participant required one prompt. Participants completed the NASA–Task Load Index (NASA-TLX; Hart & Staveland, 1988) immediately after completing the baseline trials and the first trial from memory in each interference condition. Ratings were obtained on each of the six dimensions—mental demand, physical demand, temporal demand, performance,
June 2009 - Human Factors
effort, and frustration—with the use of a scale from 1 to 10, with higher ratings indicating higher perceived workload. Each participant was instructed to “please rate how difficult you thought it was to navigate through the route just completed.” The duration of the entire experiment was between 60 and 90 min. RESULTS
A 2 (interference task: tapping or articulatory suppression) × 2 (SOD: good or poor) mixed repeated measures MANOVA was used to analyze the dependent measures, errors and time. Errors were operationalized as the number of mistakes made at choice points (turns or intersections) during the initial route traversal from memory. Time was operationalized as the time in seconds that participants took to traverse the first route from memory. Interference task was a within-subjects factor, and SOD was a betweensubjects factor. A MANOVA was implemented because of a moderately high degree of correlation between the two dependent variables: r(42) = .75, p < .001, for errors and time in the tapping interference task condition, and r(42) = .78, p < .001, for errors and time in the articulatory suppression task condition. Despite this degree of correlation, the two measures did capture some unique aspects of performance. Time increased as a function of both error recovery and indecision. Committing an error tended to be associated with having little or no memory for the correct move at that junction. Conversely, increased time at a junction in the absence of an error tended to be associated with less confident memory of the correct move. MANOVA mathematically accounts for the issue of multicollinearity, maximizing the unique variance of each measure, despite a loss of statistical power, and is therefore a more appropriate analysis than independent ANOVAs, which are likely to be highly redundant at best or misleading at worst for correlated dependent measures (Tabachnick & Fidell, 1989). A Wilks’s Lambda multivariate statistic was used and Mauchly’s test of sphericity was implemented to examine violations of sphericity. Sphericity was not violated in any of the resulting analyses. A multivariate main effect was observed for interference task, F(2, 39) = 7.65, p = .002.
Route-Learning Strategy Time to Traverse Routes 230 220
Poor SOD Good SOD
Seconds
210 200 190 180 170 160 150 Tapping
Articulatory Suppression Interference Task
Figure 1. Time to traverse routes as a function of sense of direction and interference condition. Error bars reflect standard error of the mean. Navigational Errors Made in Each Route 6 5
Poor SOD Good SOD
4 # of Errors
Univariate analyses indicated that the errors dependent measure contributed to the multivariate main effect, F(1, 40) = 7.60, p = .009. On average, participants made more navigational errors in the tapping interference task condition (M = 4.1, SD = 2.0) relative to the articulatory suppression interference task condition (M = 3.0, SD = 2.1). A multivariate main effect was also observed for the between-subjects factor SOD, F(2, 39) = 6.61, p = .003. Univariate analysis indicated that both time, F(1, 40) = 7.68, p =.008, and errors, F(1, 40) = 13.44, p = .001, contributed. As illustrated in Figures 1 and 2, individuals in the poorSOD group took longer to complete routes from memory and made more errors relative to the good-SOD group. More important, as predicted, a multivariate interaction was observed for interference task and SOD, F(2, 39) = 4.82, p = .01. Univariate analyses indicated that both time, F(1, 40) = 9.84, p = .003, and errors, F(1, 40) = 5.82, p = .02, contributed to this multivariate interaction. As can be seen in Figure 1, the time taken to traverse a route from memory that was learned while engaged in the tapping task was statistically similar for the good-SOD and poor-SOD groups, with means (and standard deviations) in seconds of 194.16 (8.3) and 202.3 (7.9), respectively. Conversely, for routes learned while engaged in the articulatory suppression task, the poor-SOD group took longer, on average 217.7 (7.59) s, relative to both their performance in the tapping condition and relative to the performance of the good-SOD group, 172.73 (7.96) s. The goodSOD group was faster in routes learned while performing the articulatory suppression task relative to the tapping task. As illustrated in Figure 2, errors revealed a somewhat similar pattern. On average, the goodSOD group demonstrated better performance, meaning fewer errors (M = 1.75, SD = 0.38), for routes learned during the articulatory suppression task relative to the tapping task (M = 3.8, SD = 0.45). The number of errors made by the poor-SOD group was higher than the goodSOD group but similar across interference tasks, M = 4.36 (SD = 0.43) for the tapping task and M = 4.23 (SD = 0.36) for the articulatory suppression task.
373
3 2 1 0 Tapping
Articulatory Suppression
Interference Task Condition
Figure 2. Errors as a function of sense-of-direction group and interference condition. Error bars reflect 95% confidence intervals.
The good-SOD individuals also required fewer attempt trials before they were able to traverse the route error free. Refer to Figure 3 for a graphical depiction of the cumulative percentage of people in each SOD group who could traverse the route error free after one, two, three, or four attempts as a function of the interference task condition. NASA-TLX. A mixed repeated measures ANOVA was used to examine differences between SOD groups (between subjects) as a
374
June 2009 - Human Factors
Percentage Who Could Complete Routes Without Error After Each Attempt Trial
Cummulative Percentage
100 90 80 70 60 50 40 30 20 10 0
1
2
3
4
Attempt Trial Poor SOD Tapping Good SOD Tapping
Poor SOD Articulatory Suppression Good SOD Articulatory Suppression
Figure 3. The percentage of participants in each senseof-direction group who could traverse the routes error free after the first, second, third, and fourth attempt as a function of interference task.
function of interference task (within subjects) on perceived mental workload as assessed by the unweighted average of ratings obtained on the NASA-TLX after each condition. A main effect was observed for interference task, F(1, 40) = 5.36, p =.026. The tapping task condition resulted in higher perceived workload ratings (M = 4.56) than did the articulatory suppression task condition (M = 3.89). A main effect was also observed for SOD group, F(1, 40) = 4.75, p = .035. The poor-SOD group (M = 4.71) perceived the workload of the route-learning task to be higher than did the good-SOD group (M = 3.69). The interaction between SOD group and interference task was not significant. Interference task performance. A mixed rep eated measures ANOVA indicated a significant main effect for interference condition as measured by accuracy (the proportion of correctly executed sequences), F(1, 39) = 38.80, p < .001. Participants were more accurate at performing the articulatory suppression task (M = .98) than the tapping task (M = .85). Neither the effect for SOD group nor the interaction effect between SOD group and interference condition was significant. Second, analysis of difference scores
between baseline and experimental conditions for accuracy revealed a main effect for task, F(1, 38) = 12.49, p < .01. Participants’ difference scores were greater with the tapping task (M = .09) than with articulatory suppression (M = .02). Mental rotation and SOD. Neither accuracy nor RT for the MRT was significantly correl ated with score on the SOD questionnaire, r(40) = –.175, p = .28, ns. No significant differences were observed between the average RTs of the poor SODs and good SODs (M = 1,545 ms, SD = 386, versus M = 1,514 ms, SD = 300, respectively) or for the mean accuracy calculated as proportion correct (M = .88, SD = .15, versus M = .95, SD = .09, for the poor and good SODs, respectively) on the MRT. MRT accuracy and RT were negatively correlated with each other, r(40) = –.413, p = .008. DISCUSSION
We examined the strategies people use to learn routes in a VE and the WM resources such strategies tap. As hypothesized, individuals classified as having good SOD tended to rely more on visuospatial WM than on verbal WM when learning routes. This was evidenced by the good-SOD group traversing routes faster and with fewer errors when learned under conditions of articulatory suppression (verbal) interference relative to the visuospatial tapping task (see Figures 1 and 2, respectively). Conversely, those with poor SOD were faster at traversing routes learned under spatial interference relative to verbal interference. Also consistent with our predictions, individuals with good SOD learned routes in fewer trials (refer to Figure 3) than did those with poor SOD. This was particularly evident when the good-SOD individuals learned the routes while engaged in the verbal interference task. In this condition, 25% of the good-SOD individuals were able to traverse the route from memory without error on their first attempt after following the avatar. In contrast, this level was not achieved in the poor-SOD group in either interference condition even after two attempt trials. Attempt trials included correction, thus providing additional opportunity to learn the route. People with poor SOD still had greater
Route-Learning Strategy
difficulty learning the routes and reported significantly higher perceived mental workload for the task relative to the good-SOD group. These results support the conclusion that individuals differ from one another in the strategies they use and the WM resources they tap—verbal or visuospatial—to learn routes in VEs. Self-report measures can be used as indices of such individual differences in navigational strategy use in VE tasks. Garden et al. (2002) found similar strategy differences in a real environment. However, they did not assess interference task performance in their dual-task study, a limitation that weakened their conclusion because of the possibility of differential attention allocation strategies in low- and high-spatial-ability groups. The current investigation controlled for this possibility by assessing both route learning and interference task performance. Both groups found the tapping task to be more difficult than the articulatory suppression task (on the basis of performance and subjective workload), but both had equivalent performance levels on the two interference tasks. Despite equivalent interference task performance between the good- and poor-SOD groups, the tapping task, relative to the articulatory suppression task, interfered more with route learning for the good-SOD group. This observation strengthens the conclusion that individuals with high spatial ability rely more heavily on visuospatial WM when learning routes relative to low-spatial-ability individuals, who place more reliance on verbal WM, and extends the hypothesis to navigation in VEs. The pattern of individual differences obs erved in this study could not have been predicted with standard psychometric tests of spatial ability (i.e., MRT). Our SOD groups did not differ significantly in this respect. Dissociation between spatial abilities involving mental rotation from an object-based framework (required in standard MRTs) and perspective taking (selfmotion relative to objects) is supported by both behavioral (Kozhevnikov & Hegarty, 2001) and neuroscience evidence (Burgess, Maguire, & O’Keefe, 2002; Burgess & O’Keefe, 2003; Ekstrom et al., 2003; O’Keefe & Nadel, 1978). Our results are consistent with this dissociation
375
and with previous empirical work demonstrating that self-reported SOD is only moderately, if at all, correlated with standard psychometric tests of spatial abilities (Hegarty et al., 2002). VEs are increasingly relied on for training and research because of their ever-improving graphics as well as the significant safety and monetary benefits of virtual relative to real environments (Arthur & Hancock, 2001). Route learning and other navigational tasks must often be carried out in VEs while individuals are performing other concurrent tasks of a visuospatial or verbal nature (Chen & Terrence, 2008). The current results indicate that SOD will affect the strategies people use to perform these tasks and the relative associated workload and can predict the types of concurrent tasks that are likely to be most detrimental to navigational performance. Maintaining orientation is difficult in VEs (Bowman, Davis, Hodges, & Badre, 1999), and people with poor SOD can be expected to experience even greater difficulty than those with good SOD. Using navigational instructions that support individual WM strategies (Baldwin, in press; Furukawa et al., 2004) and eliminating extraneous demands on the same processes may reduce disorientation and training time while improving efficiency. Individuals with poor SOD can be expected to particularly benefit from verbal guidance instructions and elimination of extraneous verbal interference while navigating (Saucier et al., 2003). The performance of those with good SOD could be more robust to verbal interference—but potentially more disrupted by concurrent visuospatial interference during navigation. Furthermore, individuals with poor SOD might benefit from targeted strategy training, as our results indicate that relative to those with good SOD, they are less adept at all navigational cues and strategies. Previous research indicates that performance on many spatial tasks is amenable to training (Gugerty & Brooks, 2004; Lawton & Morrin, 1999). Training that targets development of specific spatial strategies might particularly benefit individuals with poor SOD. The current results also have implications for the design of in-vehicle navigation systems. First, simulators and VEs can be used to examine navigational strategy differences and
376
June 2009 - Human Factors
their impact on interface design issues. Second, brief self-report SOD questionnaires provide a reliable method for assessing strategy choice. Third, individuals with good SOD will benefit from navigational aids that capitalize on their preference for visuospatial strategies (i.e., use of cardinal directions and maplike representations). Conversely, those with poor SOD can be expected to benefit from verbal guidance, particularly if it makes reference to landmarks (Reagan & Baldwin, 2006). Our findings suggest that interfaces that reduce all forms of verbal distraction during the navigational task would be beneficial for these drivers. REFERENCES Arthur, E. J., & Hancock, P. A. (2001). Navigation training in virtual environments. International Journal of Cognitive Ergonomics, 5(4), 387–400. Baciu, M., Koenig, O., Vernier, M.-P., Bedoin, N., Rubin, C., & Segebarth, C. (1999). Categorical and coordinate spatial relations: fMRI evidence for hemispheric specialization. Neuroreport: For Rapid Communication of Neuroscience Research, 10(6), 1373–1378. Baddeley, A. D. (1992). Working memory. Science, 255, 556–559. Baddeley, A. D., & Hitch, G. (1974). Working memory. In G. H. Bower (Ed.), The psychology of learning and motivation (Vol. 8, pp. 47–89). Orlando, FL: Academic Press. Baldwin, C. L. (2006). User-centered design of in-vehicle route guidance systems. In D. de Waard, K. Brookhuis, & A. Toffetti (Eds.), Developments in human factors in transportation, design, and evaluation (pp. 43–49). Maastricht, Netherlands: Shaker. Baldwin, C. L. (in press). Individual differences in navigational strategy: Implications for display design. Theoretical Issues in Ergonomic Science. Blajenkova, O., Motes, M. A., & Kozhevnikov, M. (2005). Individual differences in the representations of novel environments. Journal of Environmental Psychology, 25(1), 97–109. Bowman, D. A., Davis, E. T., Hodges, L. F., & Badre, A. N. (1999). Maintaining spatial orientation during travel in an immersive virtual environment. Presence: Teleoperators and Virtual Environments, 8(6), 61–631. Burgess, N., Maguire, E. A., & O’Keefe, J. (2002). The human hippocampus and spatial and episodic memory. Neuron, 35(4), 625–641. Burgess, N., & O’Keefe, J. (2003). Neural representations in human spatial memory. Trends in Cognitive Sciences, 7(12), 517–519. Chen, J. L., & Stanney, K. M. (1999). A theoretical model of wayfinding in virtual environments: Proposed strategies for navigational aiding. Presence: Teleoperators and Virtual Environments, 8(6), 671–685. Chen, J. Y. C., & Terrence, P. I. (2008). Effects of tactile cueing on concurrent performance of military and robotics tasks in a simulated multitasking environment. Ergonomics, 51(8), 1137–1152. Cooke, N. J., Pringle, H. L., Pedersen, H. K., Connor, O., Cooke, N. J., Pringle, H. L., et al. (2006). Human factors of remotely operated vehicles. Amsterdam, Netherlands: Elsevier. Cooper, L. A., & Shepard, R. N. (1973). The time required to prepare for a rotated stimulus. Memory and Cognition, 1(3), 246-250.
Cornoldi, C., & Vecchi, T. (2003). Visuo-spatial working memory and individual differences. New York: Psychology. Dabbs, J. M., Chang, E. L., Strong, R. A., & Milun, R. (1998). Spatial ability, navigation strategy, and geographic knowledge among men and women. Evolution and Human Behavior, 19(2), 89–98. Di Nocera, F., Couyoumdjian, A., & Ferlazzo, F. (2006). Crossing the pillars of Hercules: The role of spatial frames of reference in error making. Quarterly Journal of Experimental Psychology, 59(1), 204–221. Dixon, S. R., & Wickens, C. D. (2006). Automation reliability in unmanned aerial vehicle control: A reliance-compliance model of automation dependence in high workload. Human Factors, 48, 474–486. Dixon, S. R., Wickens, C. D., & Chang, D. (2005). Mission control of multiple unmanned aerial vehicles: A workload analysis. Human Factors, 47, 479–487. Ekstrom, A. D., Kahana, M. J., Caplan, J. B., Fields, T. A., Isham, E. A., Newman, E. L., et al. (2003). Cellular networks underlying human spatial navigation. Nature, 425(6954), 184–188. Farrell, M. J., Arnold, P., Pettifer, S., Adams, J., Graham, T., & MacManamon, M. (2003). Transfer of route learning from virtual to real environments. Journal of Experimental Psychology: Applied, 9(4), 219–227. Ferlazzo, F., Fagioli, S., Di Nocera, F., & Sdoia, S. (2008). Shifting attention across near and far spaces: Implications for the use of hands-free cell phones while driving. Accident Analysis & Prevention, 40, 1859–1864. Furukawa, H., Baldwin, C. L., & Carpenter, E. M. (2004). Supporting drivers’ area-learning task with visual geo-centered and auditory ego-centered guidance: Interference or improved performance? In D. A. Vincenzi, M. Mouloua, & P. A. Hancock (Eds.), Human performance, situation awareness and automation: Current research and trends, HPSAA II (pp. 124–129). Daytona Beach, FL: Lawrence Erlbaum. Garden, S., Cornoldi, C., & Logie, R. H. (2002). Visuo-spatial working memory in navigation. Applied Cognitive Psychology, 16(1), 35–50. Gramann, K., Müller, H. J., Schönebeck, B., & Debus, G. (2006). The neural basis of ego- and allocentric reference frames in spatial navigation: Evidence from spatio-temporal coupled current density reconstruction. Brain Research, 1118(1), 116–129. Gugerty, L., & Brooks, J. (2004). Reference-frame misalignment and cardinal direction judgments: Group differences and strategies. Journal of Experimental Psychology: Applied, 10(2), 75–88. Hart, S. G., & Staveland, L. E. (1988). Development of NASA-TLX (Task Load Index): Results of empirical and theoretical research. In P. A. Hancock & N. Meshkati (Eds.), Human Mental Workload (pp. 239–250). Amsterdam: North Holland Press. Hegarty, M., Montello, D. R., Richardson, A. E., Ishikawa, T., & Lovelace, K. (2006). Spatial abilities at different scales: Individual differences in aptitude-test performance and spatiallayout learning. Intelligence, 34(2), 151–176. Hegarty, M., Richardson, A. E., Montello, D. R., Lovelace, K., & Subbiah, I. (2002). Development of a self-report measure of environmental spatial ability. Intelligence, 30(5), 425–447. Hugdahl, K., Thomsen, T., & Ersland, L. (2006). Sex differences in visuo-spatial processing: An fMRI study of mental rotation. Neuropsychologia, 44(9), 1575–1583. Kato, Y., & Takeuchi, Y. (2003). Individual differences in way finding strategies. Journal of Environmental Psychology, 23(2), 171–188. Kosslyn, S. M., & Thompson, W. L. (2003). When is early visual cortex activated during visual mental imagery? Psychological Bulletin, 129(5), 723–746.
Route-Learning Strategy Kozhevnikov, M., & Hegarty, M. (2001). A dissociation between object manipulation spatial ability and spatial orientation ability. Memory & Cognition, 29, 745–756. Kozlowski, L. T., & Bryant, K. J. (1977). Sense of direction, spatial orientation, and cognitive maps. Journal of Experimental Psychology: Human Perception and Performance, 3(4), 590–598. Lawton, C. A. (1994). Gender differences in way-finding strategies: Relationship to spatial ability and spatial anxiety. Sex Roles, 30(11/12), 765–779. Lawton, C. A. (1996). Strategies for indoor wayfinding: The role of orientation. Journal of Environmental Psychology, 16(2), 137–145. Lawton, C. A. (2001). Gender and regional differences in spatial referents used in direction giving. Sex Roles, 44(5/6), 321–337. Lawton, C. A., & Kallai, J. (2002). Gender differences in wayfinding strategies and anxiety about wayfinding: A cross-cultural comparison. Sex Roles, 47(9/10), 389–401. Lawton, C. A., & Morrin, K. A. (1999). Gender differences in pointing accuracy in computer-simulated 3D mazes. Sex Roles, 40(1), 73–92. MacFadden, A., Elias, L., & Saucier, D. (2003). Males and females scan maps similarly, but give directions differently. Brain and Cognition, 53(2), 297–300. O’Keefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map. Oxford, UK: Oxford University Press. Pazzaglia, F., Cornoldi, C., & De Beni, R. (2000). Individual differences in spatial representation and in orientation ability: Presentation of a self-report questionnaire. Giornale Italiano di Psicologia, 27(3), 627–650. Pazzaglia, F., & De Beni, R. (2001). Strategies of processing spatial information in survey and landmark-centred individuals. European Journal of Cognitive Psychology, 13(4), 493–508. Previc, F. H. (1998). The neuropsychology of 3-D space. Psychological Bulletin, 124(2), 123–164. Reagan, I. (2005). Using working memory interference tasks to test for verbal and spatial route learning strategies. Norfolk, VA: Old Dominion University. Reagan, I., & Baldwin, C. L. (2006). Facilitating route memory with auditory route guidance systems. Journal of Environmental Psychology, 26(2), 146–155. Ruddle, R. A., Payne, S. J., & Jones, D. M. (1999). The effects of maps on navigation and search strategies in very-large-scale
377 virtual environments. Journal of Experimental Psychology: Applied, 5(1), 54–75. Saucier, D., Bowman, M., & Elias, L. (2003). Sex differences in the effect of articulatory or spatial dual-task interference during navigation. Brain and Cognition, 53(2), 346–350. Saucier, D., Green, S. M., Leason, J., MacFadden, A., Bell, S., & Elias, L. J. (2002). Are sex differences in navigation caused by sexually dimorphic strategies or by differences in the ability to use the strategies? Behavioral Neuroscience, 116(3), 402–410. Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics (2nd ed.). New York: HarperCollins. Takeuchi, Y. (1992). Sense of direction and its relationship with geographical orientation, personality traits and mental ability. Japanese Journal of Educational Psychology, 40, 47–53. Thorndyke, P. W., & Hayes-Roth, B. (1982). Differences in spatial knowledge acquired from maps and navigation. Cognitive Psychology, 14(4), 560–589. Zaehle, T., Jordan, K., Wüstenberg, T., Baudewig, J., Dechent, P., & Mast, F. W. (2007). The neural basis of the egocentric and allocentric spatial frame of reference. Brain Research, 1137, 92–103.
Carryl L. Baldwin is an assistant professor of psychology and member of the Arch Lab at George Mason University in Fairfax, Virginia, and an adjunct associate professor at Old Dominion University in Norfolk, Virginia. She received her PhD in human factors psychology at the University of South Dakota in Vermillion in 1997. Ian Reagan is a doctoral student at Old Dominion University in Norfolk, Virginia. He received his MS in psychology from Old Dominion University in 2005. Date submitted: January 4, 2008 Date accepted: March 30, 2009
