Review

What determines our navigational

abilities?
Thomas Wolbers1 and Mary Hegarty2
1
Centre for Cognitive and Neural Systems, University of Edinburgh, Edinburgh, EH8 9JZ, UK
2
Department of Psychology, University of California, Santa Barbara, CA, 93106, USA

The ability to find one’s way in our complex environ- differ in the cues and strategies they use for navigation?
ments represents one of the most fundamental cognitive And how are these behavioral differences related to poten-
functions. Although involving basic perceptual and tial differences in the organization, functioning and integ-
memory related processes, spatial navigation is particu- rity of critical brain structures?
larly complex because it is a multisensory process in Figure 1 provides an overview of the sensory cues,
which information needs to be integrated and manipu- computational mechanisms and spatial representations
lated over time and space. Not surprisingly, humans involved in most forms of human and animal navigation.
differ widely in this ability, and recent animal and human Variability in navigational abilities can arise at multiple
work has begun to unveil the underlying mechanisms. stages, including the precision with which spatial infor-
Here, we consider three interdependent domains that mation is encoded from sensory experiences, the ability to
have been related to navigational abilities: cognitive and form spatial representations of external environments and
perceptual factors, neural information processing and the efficacy with which they are used to guide navigational
variability in brain microstructure. Together, the findings behavior. These levels are highly interdependent (i.e. exist-
converge into an emerging model of how different fac- ing mental representations can affect sensory experience),
tors interact to produce individual patterns of naviga- and many navigational tasks bridge multiple levels (i.e.
tional performance. path integration involves perceiving self-motion cues and
forming a spatial representation). Therefore, we have
Spatial navigation – a complex behavior with large adopted a broad categorization scheme with two sections,
individual differences the first of which covers findings from behavioral studies
The ability to maintain a sense of direction and location and the second addresses underlying factors with regard to
while moving about in the environment is a fundamental neural information processing. Importantly, the existing
cognitive function. Mammals rely on spatial cognitive literature does not allow for these sections to be homolo-
processes for obtaining food, avoiding prey and finding gous, as in some cases, individual differences have been
mates. In humans, spatial navigation is indispensable examined at one level but not the other (i.e. the behavioral
for finding our way in complex environments, planning work on individual differences in perspective taking ability
routes to distant locations and returning to our car after a has not yet been complemented by corresponding neuro-
walk in a new city. As a consequence, when lesions to the science experiments). Finally, a treatment of the determi-
brain impair navigational abilities, patients often experi- nants of navigational abilities might bring up questions
ence devastating effects on their everyday lives [1]. about the underlying causes (biological, environmental
Spatial navigation can be based on externalized repres- and interactional). However, most of the research to date
entations such as maps or diagrams and on internal has been concerned with identifying the cognitive and
representations derived from sensory experience. This neural components that determine navigational ability,
review focuses on internal representations, which as this analysis must precede the systematic study of
includes perceiving spatial information from multiple causal factors.
sensory cues, creating and maintaining spatial repres-
entations in short- and long-term memory, and using
and manipulating these representations to guide naviga- Glossary
tional behavior. Given the complexity of cues, representa-
Cognitive map: the term cognitive map was first coined by Tolman [96] to
tions and processes, it is not surprising that humans differ express that rats in his experiments took novel shortcuts to a goal location,
widely in their navigational abilities. We all know some thus indicating an understanding of the spatial structure of the environment.
acquaintances with an excellent sense of direction The term suggests – and is often used to suggest – that we possess spatial
representations similar to a real map, as if the world was represented from a
whereas others easily get lost. This observation has been bird’s eye perspective. However, when spatial knowledge is acquired from
confirmed under controlled laboratory conditions; how- navigation (as opposed to looking at a map), many researchers [97] prefer to
think of a cognitive map as a type of ‘see through representation’ that
ever, researchers have only started to address the under-
preserves the ground level perspective but in which distant locations can be
lying questions systematically: What are the differences in accessed through intervening points (i.e. buildings). Here, we conceive of
spatial representations and processes between people cognitive maps as a flexible internal representation of the structure of the
environment that is not associated with a specific orientation, hence spatial
with more and less navigational ability? How do people relationships (i.e. directions and distances between objects) can be inferred
from any perspective.
Corresponding author: Wolbers, T. ([email protected]).

138 1364-6613/$ – see front matter ß 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.tics.2010.01.001 Available online 6 February 2010
 Review Trends in Cognitive Sciences Vol.14 No.3

Figure 1. The complexity of spatial navigation. Schematic depiction of the various sensory cues, computational mechanisms and spatial representations that are involved
in forming spatial knowledge and in using it to guide navigational behavior. Whereas navigation based on route representations involves sequences of local views that are
associated with specific actions (i.e. turn right at the store), navigation based on cognitive maps requires an understanding of the spatial relationships between important
features (i.e. landmarks). These relationships can be inferred from a combination of self-localization and the perception of spatial attributes of the environment. Specifically,
internal and external self-motion cues can be used to maintain a sense of position and orientation (path integration), and to keep track of the locations of external features
(spatial updating). Given that these processes are not error free, familiar visual cues (i.e. salient landmarks) are often used to recalibrate estimates of position and
orientation. By contrast, visual information in particular provides us with direct information about spatial attributes of the environment (i.e. geometric layout) and the
objects therein (i.e. distances and directions between landmarks). Importantly, given that in most natural environments, not all relevant features can be seen from a single
vantage point, keeping track of one’s location and orientation is crucial if we are to integrate all relevant features into a comprehensive representation. When using spatial
knowledge to plan a route to an unseen goal location, humans either follow a familiar route or compute a novel route based on a cognitive map. In highly familiar
environments, both strategies are often simultaneously available, thus requiring efficient decision-making strategies.

Variability in perceptual and cognitive processing ability depends on a system equivalent to the rodent head
Although people differ in their ability to perceive spatial direction system [9].
attributes such as egocentric self-to-object distances and In addition to self-motion cues, maintaining a sense of
allocentric object-to-object distances [2], the contribution of position and orientation can also be based on close and
these differences to the variance in navigation ability distant landmarks and on environmental geometry. In
remains to be determined. Other potential sources of indi- tasks that involve active locomotion, people are equally
vidual differences include the ability to sense self-motion able to update their position and orientation on the basis of
and to maintain orientation relative to the environment. either geometric cues (e.g. room shape) or featural cues
Although people can update their position and orientation (landmarks such as a distinctive wall hanging). However,
on the basis of sensing self-motion over short distances, a they differ in the number of cues needed for accurate
process known as path integration, error accumulates over orientation, particularly when the cues are somewhat
larger distances, such that people instructed to walk a ambiguous so that they have to be integrated with body-
straight line can literally walk in circles [3]. Individual based senses [10] (Box 1). Furthermore, when humans
differences in path integration [4] are predicted by self- need to reorient to a space after disorientation, they show
reported sense of direction (SOD) [5–7], showing that substantial differences in the extent to which they rely on
people are aware of their own skills. SOD predicts ability environmental geometry versus featural cues [11]. Good
to update one’s position and orientation while walking navigators need fewer cues to remain oriented and are
blindfolded or in visually impoverished environments, more flexible in reorienting based on either geometry or
and to point to unseen orientations in a familiar environ- landmarks as the task demands.
ment. Sholl et al. [8] reported a particularly high corre- When humans acquire spatial knowledge from direct
lation between SOD and the ability to judge the direction in experience in an environment or from media such as
environment-based coordinates from which a photograph virtual environments or maps [12–15], individual differ-
of a familiar landmark had been taken. On the basis of this ences are extremely large and robust. For example, Ishi-
evidence, they proposed that self-reported SOD measures a kawa and Montello [15] led participants on the same routes
person’s evaluation of their ability to keep track of their through a novel environment once a week for 10 weeks and
facing direction relative to the environment, and that this measured their ability to estimate straight line directions

139
Review Trends in Cognitive Sciences Vol.14 No.3

Box 1. Sex differences layout from navigational experience [12,13,23], presum-

ably because adopting different perspectives is required by
There have now been several demonstrations of a human male
advantage in virtual maze tasks and in spatial learning from many navigational tasks.
navigational experience [11,14,19,53,54], somewhat paralleling sex Correlational studies are limited in that the correlation
differences in animal species [55]. Although sex differences are between two variables reflects variance specific to each of
sometimes more pronounced when tested in simulated environ- the tasks, as well as error variance. Using structural
ments [14,54], they occur with testing in both real and virtual
environments [56] and when the analyses control for video game
equation modeling to control for these sources of variance,
experience that is often greater in males than in females [19]. Hegarty et al. [14] could demonstrate that object-based
Superior performance by males is not found in all tasks at the spatial abilities and the ability to learn the layout of an
environmental scale. It is typical when people learn spatial layout environment are partially dissociated. As shown in
from direct experience, but not when they learn from maps, and is Figure 2, object-based spatial ability was significantly
also more pronounced in measures of survey knowledge than in
measures of route knowledge [53,56]. Furthermore, female perfor- more related to learning from media than to learning from
mance can vary with hormonal fluctuations, such that women can direct experience, whereas sense of direction showed a
perform as well as males during low-estrogen stages of their cycle significant difference in the opposite direction. Moreover,
[57]. Object location memory often shows an advantage in favor of learning from direct experience in an environment was
females, although this can depend on the type of objects, whether
somewhat distinct from learning from media (video and
self-motion is involved, and the degree of metric precision required
[58,59]. desktop virtual environments; [24]). The authors con-
Intriguingly, there appear to be qualitative differences in the cluded that object-based and environmental spatial abil-
environmental cues and strategies that women and men use during ities share the ability to encode spatial information from
navigation and orientation. Women typically report navigating on visual input, maintain this information in memory and
the basis of local landmarks and familiar routes, whereas men
make inferences from this information. Most importantly,
report using cardinal directions, environmental geometry and
metric distances [60,61], a result which has been supported by however, learning from direct experience also involves the
neuroimaging findings [62]. Although women do not differ from sensing of self-motion for the purpose of spatial updating
men in dependence on or ability to use landmarks, they depend less and path integration, an ability that cannot be measured
on geometry when reorienting to an environment [11] and are with simple object-based tasks in which self-motion cues
relatively more impaired at finding a target based on directional
cues (i.e. environmental slope, [60]). Women also require more
are neither available nor task-relevant.
environmental cues to remain oriented in an environment [10] and Spatial navigation can be based on separate memory
have difficulty following navigation directions based on cardinal representations. One memory system gradually acquires
directions and metric distances [21]. Thus, strategy preferences can sequences of actions (e.g. repeatedly following a fixed
reflect proficiency differences between the sexes in use of geometric
route) and provides rigid route representations, often in
cues, as well as relative cue salience.
In terms of causal factors, there is increasing evidence for the an egocentric reference frame and based on local land-
influence of sex hormones on navigational performance [25,57,63– marks. The other develops observer-independent, flexible
65], and several evolutionary theories have been proposed [66]. representations (often termed cognitive maps or survey
However, men and women also differ in navigational experience representations) that allow for planning direct paths to
[54,67] and there is some evidence that wayfinding anxiety mediates
unseen goal locations, even over unfamiliar terrain. Indi-
the differences between the sexes in navigational performance [67].
viduals can have preferences for route-based vs. survey-
based strategies that are often gender-related (Box 1), and
and distances between locations. Whereas some partici- animal models suggest that these can also depend on
pants had almost perfect configural knowledge of the hormonal fluctuations [25]. However, strategy choice can
environment after one or two learning experiences, others also depend on factors such as the demands of different
performed at chance even after 10 learning trials. The navigation tasks, the information available to the naviga-
causes of variance in navigational ability have not been tor, and the reliability of available cues [26–29]. Hence, the
studied systematically in the individual differences litera- best navigators appear to be those who switch flexibly
ture to date, although navigation ability is clearly influ- between different strategies, depending on what is optimal
enced by both genetic factors and environmental influences in a given situation [24].
such as parental guidance and exposure to maps [16,17].
There has been more attention given to causal factors in Variability in structure and function of critical brain
the literature on sex differences (Box 1). circuits
Many studies of individual differences in spatial ability In the previous section, we have discussed behavioral
have focused on smaller-scale tasks that involve simple experiments looking for factors that drive individual per-
object transformations (i.e. mental rotation), but do these formance differences in navigational abilities. Studies
abilities predict individual differences at the scale of from animal and human cognitive neuroscience have
environments in which we navigate? Some experiments attempted to link these differences to variability in neural
have shown that individuals classified as high and low in information processing and brain microstructure. For
object-based spatial ability also differ in tasks such as example, whereas the neural mechanisms that cause indi-
route retracing and learning spatial layout [18–21]. How- vidual differences in space perception remain to be
ever, a recent review of 12 studies found that the median explored, the precision of self-motion perception could be
correlation exceeded 0.3 only in two studies, and the related to differences in sensory noise. In monkeys, area
majority of correlations were not statistically significant MST, located in the medial superior temporal cortex, is
[22]. In contrast, perspective taking ability has been crucial for computing the direction of self-motion [30],
found to be more predictive of the ability to learn spatial which is partly based on incoming motion cues from

140
 Review Trends in Cognitive Sciences Vol.14 No.3

Figure 2. Relations between measures of spatial layout learning, spatial ability and self-reported sense of direction. Results of the structural equation model observed by
Hegarty et al. [14]. Boxes indicate measured variables and ovals represent latent variables derived from these, which reflect the variance shared by the different measured
variables to which they point. Participants learned three different environments from (1) direct experience, (2) watching a video and (3) interacting with a desktop virtual
environment. In each case their acquired knowledge was tested with three measures; (1) pointing to unseen locations in the environment, (2) estimating straight-line distances
to these locations and (3) drawing a map of the environment. The path coefficients (values labeling each arrow) can be interpreted as standardized regression weights indicating
the degree of relation between the predictor and predicted variables after controlling for the effects of the other variables. Note that the path coefficient linking spatial ability and
learning from direct experience is 0.5, indicating that these two abilities reflect some common variance but are also somewhat distinct. Measures of spatial ability were relatively
more predictive of learning from media than of learning from direct experience, whereas self-reported sense of direction was relatively more predictive of learning from direct
experience than from media. These results support a partial dissociation between large- and small-scale spatial abilities and suggest that self-motion perception (important in
learning from direct experience but not from media) could be a factor in this dissociation. Figure adapted, with permission, from Ref. [14].

neighboring area MT. In aging monkeys, the signal-to- ever, although the entorhinal–hippocampal circuit pro-
noise ratio in MT is significantly reduced [31,32] (Box 2), vides information about the current position of an agent,
thus conveying less informative motion signals to area keeping track of previous locations adds a working memory
MST. Although similar findings remain to be established component that involves medial prefrontal areas, particu-
in younger animals to explain individual performance larly when this information guides navigational decisions
differences, it is highly likely that the amount of sensory [35]. The visual path integration study of Wolbers et al. [33]
noise in motion processing areas is a decisive factor. Con- supports this hypothesis by revealing a tight coupling
sistent with this assumption, Wolbers et al. [33] observed a between an individual’s uncertainty about the direction
correlation between systematic pointing errors and func- to a start location and bilateral responses in the hippo-
tional magnetic resonance imaging (fMRI) responses in a campus and the medial prefrontal cortex (Figure 3). These
homologous region in the human brain (hMT+) during authors suggested that during navigation, hippocampal
visual path integration. These findings suggest that local place responses are used by the prefrontal cortex to con-
processing in motion sensitive structures affects an indi- tinuously update a homing vector during the outward
vidual’s uncertainty about the amount of self-rotation and journey. As a consequence, the degree of recruitment of
self-translation. both areas influenced the precision with which these com-
Estimates of self-motion enable us to update our pos- putations were carried out, thus having a direct impact on
ition and to keep track of previously occupied locations. In an individual’s path integration ability.
rodents, entorhinal grid cells are thought to integrate self- For most navigational tasks, estimates of self-position
motion cues for computing positional estimates, which and orientation need to be complemented by representa-
updates the firing of hippocampal place cells [34]. How- tions of the outside world. One region involved in forming

141
 Review Trends in Cognitive Sciences Vol.14 No.3

Box 2. Individual differences in aging subjects

Individual differences in navigational abilities are particularly promi- and pattern separation in the hippocampus [73], which could in
nent at later stages in life. For example, when Monacelli et al. [68] turn be related to the level of hippocampal neurogenesis [74]. In
tested various groups of subjects on their ability to navigate in a addition, place cells can exhibit differential firing behavior upon
familiar hospital, both healthy and demented elderly subjects scored repeated exposure to the same environment [75], suggesting
worse than a young control group on various performance measures. deficits with forming stable links between environmental cues
Importantly, even among the healthy aged subjects, some partici- and spatial representations. Finally, genes that control synaptic
pants had no navigation difficulties whereas others made substantial transmission and morphology in the hippocampus are associated
navigational errors. In the extant literature on age-related changes in with age-related spatial learning impairments [76]. Taken
spatial memory, such individual deficits have predominantly been together, these findings suggest that the efficiency of information
linked to sensory, mnemonic and executive processes. processing and storage within the hippocampus is an important
(i) Motion processing. The accuracy of visual motion processing determinant of age-related changes of spatial abilities.
varies across aging humans, and individual motion thresholds (iii) Executive functions. Moffat and Resnick [77] demonstrated that
are predictive of navigational abilities [69]. Given that aging aging humans often show inefficient search strategies in a virtual
monkeys show reduced signal-to-noise ratios in the motion Morris water maze. Specifically, some elderly individuals took
sensitive area MT [31,32], difficulties with extracting the speed longer to find the platform and traveled longer distances in the
and direction of self-motion from optic flow could impair path very first trial when spatial representations were yet to be formed.
integration [70] and spatial updating, processes that are impor- These subjects were reluctant to disengage from locations that
tant for self and object localization [34,71] and that can enable us had been adequately explored, suggesting problems with
to link views of scenes or places with accurate spatial positions. strategic control mechanisms. A follow-up study revealed that
(ii) Hippocampal processing. Although the firing patterns of hippo- individual performance was correlated with gray and white
campal place cells in rats with age-related memory impairments matter volumes in various extrahippocampal structures, includ-
do not differ from the ones of unimpaired or young animals, ing the prefrontal cortex [52]. As a consequence, a reduced
these cells often fail to encode changes in the environment and to integrity of brain structures involved in executive control could
create new representations for novel environments [72]. These further impinge upon the individual navigational abilities of aging
findings could reflect an imbalance between pattern completion humans.

Figure 3. Variability in neural information processing. (i) Neural correlates of individual variability in path integration performance. During fMRI scanning, subjects were
virtually moved along two legs of a triangle before pointing towards the origin of travel. Hippocampal and medial prefrontal activation during encoding of the outbound path
was greater in those participants who showed higher response consistency. This suggests that a consistent engagement of both structures is crucial for updating self-positions
and maintaining previously occupied locations in working memory. (ii) Upper panels: in the study of Wolbers et al. [38], participants were repeatedly moved throughout a
complex virtual environment, thereby encountering 12 distinct landmarks. Subjects were instructed to infer the spatial layout of the environment and the correct landmark
locations, knowledge that was tested with a subsequent retrieval task. Lower panels: anterior hippocampal recruitment during cognitive map formation predicts the speed of
learning. Hippocampal activation (shown in red) in subject 11 was strongest in the initial learning stage and decayed rapidly after performance had reached level in session 3. In
contrast, the slower learning process in subject 03 was paralleled by stronger hippocampal activation in the second half of the experiment. As a consequence, anterior
hippocampal activation appears to be most prominent whenever substantial performance improvements are observable. (iii) Hippocampal activation during wayfinding in a
familiar virtual town correlated with performance across subjects, demonstrating a greater engagement in better navigators than in poorer navigators. Similar results were
observed in the head of right caudate during route following (not shown), suggesting a dissociation between the retrieval of a cognitive map and the retrieval of habitual routes.
Figure adapted, with permission, from Refs. [33,38,44].

142
Review Trends in Cognitive Sciences Vol.14 No.3

these representations is the parahippocampal cortex, the posterior hippocampus to allow for this representation
because it supports the learning of places from individual to be elaborated, a finding that has also been reported in
views. In good navigators, parahippocampal responses other species (i.e. pigeons, [50]).
show greater differences between novel and familiar places Finally, creating and retrieving spatial representa-
and between novel and familiar views compared with bad tions also relies upon executive functions such as plan-
navigators [36]. These findings indicate that individuals ning and spatial working memory, and lesions to the
vary as to how useful parahippocampal representations prefrontal cortex, a key player for the control of executive
are for distinguishing between different places and differ- functions, can severely impair navigational abilities [51].
ent views. Importantly, even in healthy humans, preliminary evi-
In order for place representations to be useful for navi- dence suggests that regional brain volumes in prefrontal
gation, we need to integrate them into a representation areas are associated with learning performance in a
that preserves some degree of spatial relationships (i.e. virtual version of the Morris water maze [52]. This
topological, metric, action based, etc.). Whereas action- indicates that the efficiency which with prefrontal areas
based route representations have been linked to the dorsal carry out executive functions can further determine over-
striatum, observer-independent cognitive maps critically all navigational proficiency.
depend on retrosplenial cortex and the hippocampus
[27,37,38]. As a consequence, interpretations of neural Concluding remarks
data require a precise control or identification of the type Spatial navigation involves multiple sensory cues, inter-
of representation used in a given task. Wolbers et al. acting processes and representations, and performance
[38,39] developed one such task (Figure 3) and their results differences can arise at various stages. On the sensory
showed that anterior hippocampal recruitment determines side, people differ with regard to the accuracy with which
the speed with which an individual forms a cognitive map.
These and other results that have linked memory conso-
Box 3. The impact of genetic factors
lidation in the hippocampus to navigational skills [40] are
supported by animal findings, suggesting that the learn- The structural and functional integrity of neuronal circuits is jointly
ing-related hippocampal circuits of fast learners are better determined by environmental and physiological factors, the latter
including genetic predispositions. Genetic association studies in
suited to solve spatial tasks than those of slow learners.
animals have demonstrated various genetic influences on hippo-
Specifically, the latter require structural reorganization to campal processes involved in spatial navigation [78]. Specific
form spatial memories as opposed to the more economic examples include the brain derived neurotrophic factor (BDNF) that
mechanism of altering synaptic efficacy used by the former is known for its role in activity-dependent plasticity and hippocam-
[41]. pal long-term potentiation. Both processes are thought to underlie
the formation of new learning and memories, and suppression of
Hippocampal activation is not only observed during BDNF synthesis impairs spatial learning in rodents [79]. Although
cognitive map formation but also during retrieval, with direct effects of BDNF on human navigational learning remain to be
some studies suggesting an anterior–posterior functional established, BDNF modulation of hippocampal engagement is a key
division in the hippocampus [27,42], reminiscent of the process in the initial acquisition of information about novel indoor
dorsoventral differentiation in rodents [43]. At retrieval, and outdoor scenes [80]. In addition, polymorphisms of the BDNF
gene have also been associated with hippocampal volume [81],
more accurate navigators show greater hippocampal acti- which could contribute to preferences for specific strategies in a
vation when navigating to an unseen goal in a familiar navigational task [46].
environment [44]. Importantly, the same individuals show A second route for genetic predispositions to affect hippocampal
greater caudate responses when following a well-learned processing and hence navigational abilities involves pattern separa-
tion. To distinguish between environments or regions within an
route, whereas poor navigators show the reverse pattern.
environment, hippocampal subfields create orthogonal representa-
Hence, the relative engagement of the hippocampus and tions [82]. This ability to pattern separate is directly related to
the caudate seems to determine the individual proficiency neurogenesis in the dentate gyrus, which is in turn controlled by
of forming and retrieving cognitive maps and route-based several genes [83]. Given that ablation of pattern separation in mice
representations. In addition, given that a chosen strategy induces deficits in spatial learning in a radial arm maze [84], it
appears probable that individual genetic predispositions that
might not be optimal for the task at hand, the ability to
control hippocampal neurogenesis can have direct effects on
flexibly switch between hippocampal and striatal repres- navigational abilities via differences in pattern separation.
entations is a further characteristic of successful naviga- Finally, as spatial navigation also involves executive control
tors [45]. processes that involve subdivisions of the prefrontal cortex
Although preliminary reports of associations between [33,85], genes that regulate prefrontal functioning should have the
potential to influence navigational abilities. For example, given the
brain microstructure and navigational abilities [46,47] dopaminergic metabolism in the prefrontal cortex, the gene
need to be qualified by studies that control for unspecific producing catechol-O-methyltransferase (COMT) is thought to have
variables that could also account for the anatomical varia- a major impact on functions such as the manipulation of informa-
bility, extended navigational experience induces plastic tion [86] and the resolution of uncertainty [87], both of which are
changes in the brain. Maguire et al. [48,49] demonstrated involved in spatial navigation. Moreover, COMT polymorphisms
also affect prefrontal–hippocampal coupling [88], which is crucial for
that the posterior hippocampal volume of London taxi but navigational planning [35].
not bus drivers – the latter experiencing similar levels of Taken together, although the existing animal findings strongly
stress, driving and self-motion – correlates with driving suggest genetic influences on navigational abilities, a direct
experience. Hence, the extensive use of a flexible spatial demonstration remains to be established in humans. Given the
complexity of spatial navigation, genetic variability is likely to affect
representation in the taxi drivers – as opposed to the fixed
navigational functions at multiple processing stages.
routes that the bus drivers follow – enlarges the volume of

143
Review Trends in Cognitive Sciences Vol.14 No.3

Box 4. Outstanding questions and future directions

 Age-related differences in navigational abilities have only been  Spatial navigation experiments have employed a large variety of
studied in cross-sectional experiments. This makes it difficult to paradigms that differ with regard to how spatial information is
determine the underlying causes: Was this variability already acquired (i.e. locomotion, learning from maps or virtual animations,
present at a younger age or is it specifically linked to different etc.) and retrieved (i.e. estimating distances and directions,
developmental trajectories? If the latter, what mechanisms account navigating to a goal, map drawing, etc.). In addition, the scale of
for those trajectories, both on the neural and the behavioral level? the environment differs widely: whereas in tasks such as the Morris
And is there a potential to alleviate or even stop age-related decline water maze (and virtual renditions thereof) most of the relevant
in spatial abilities? spatial information can be apprehended from a single vantage
 Extensive navigational experience leads to structural changes in the point, learning complex environments is much more demanding as
hippocampus, but does this pave the way for acquiring new spatial observers need to infer spatial relationships between distant
representations more easily and more accurately? The finding that locations. To identify sources of individual variability common to
anterior hippocampal volume was negatively correlated with driving different studies, we need to develop a comprehensive taxonomy of
experience in the taxi drivers studied by Maguire et al. [48,49] is not in spatial navigation tasks and the participating processes [93].
line with this conjecture, but it remains to be tested directly.  Is it possible to improve navigational functions? People with a poor
 Psychiatric conditions such as pathological anxiety, depression or sense of direction experience daily problems with navigation,
post-traumatic stress disorder can have navigational consequences, sometimes even in familiar environments [94]. With the widespread
which have been related to functional changes in the hippocampus use of GPS-based navigation aids, these problems become even
and to reduced hippocampal volume [89–92]. Moreover, anxiety has more severe when the technology fails [95] or is unavailable (i.e.
also been proposed as a mediating factor in sex differences in indoors). Given that many expert navigators can use spatial cues
navigational ability [67]. These issues require further investigation to that are ignored by the majority of us, it is important to investigate if
identify the specific navigational processes that are impaired and to and how efficient navigational strategies can be learned to
characterize how potential treatments affect navigational behavior. overcome individual difficulties.

they perceive spatial properties of the environment. In 2 Norman, J.F. et al. (2005) The perception of distances and spatial
relationships in natural outdoor environments. Perception 34, 1315–
addition, variability in self-motion perception, probably
1324
related to signal-to-noise ratios in cortical areas involved 3 Souman, J.L. et al. (2009) Walking straight into circles. Curr. Biol. 19,
in (self-)motion processing, could influence the accuracy 1538–1542
with which people keep track of their orientation and pos- 4 Loomis, J.M. et al. (1993) Nonvisual navigation by blind and sighted:
ition relative to the environment. This ability to path inte- assessment of path integration ability. J. Exp. Psychol. Gen. 122,
73–91
grate further depends on efficient hippocampal recruitment.
5 Hegarty, M. et al. (2002) Development of a self-report measure of
The second level of individual differences pertains to the environmental spatial ability. Intelligence 30, 425–447
cues and strategies we use when acquiring spatial infor- 6 Kozlowski, L.T. and Bryant, K.J. (1977) Sense-of-direction, spatial
mation: whereas some people prefer featural cues to main- orientation, and cognitive maps. J. Exp. Psychol. Hum. Percept.
tain orientation and to infer spatial relationships, others Perform. 3, 590–598
7 Sholl, M.J. (1988) The relationship between sense of direction and
focus on geometric properties such as the layout of an mental geographic updating. Intelligence 12, 299–314
environment. Furthermore, humans differ with regard to 8 Sholl, M.J. et al. (2006) Allocentric-heading recall and its relation to
their preferred strategies for acquiring spatial information, self-reported sense-of-direction. J. Exp. Psychol. Learn. Mem. Cogn. 32,
with route-based strategies depending on the dorsal stria- 516–533
tum and cognitive mapping involving the hippocampus. 9 Taube, J.S. (2007) The head direction signal: origins and sensory-motor
integration. Annu. Rev. Neurosci. 30, 181–207
Because these systems allow for different types of naviga- 10 Kelly, J.W. et al. (2009) Individual differences in using geometric and
tional behavior, the ability to choose the adequate strategy featural cues to maintain spatial orientation: cue quantity and cue
for a given situation appears to be a key characteristic of ambiguity are more important than cue type. Psychon. Bull. Rev. 16,
successful navigators. In combination with planning oper- 176–181
11 Sandstrom, N.J. et al. (1998) Males and females use different distal
ations and working memory processes, such executive func-
cues in a virtual environment navigation task. Brain Res. Cogn. Brain
tions critically depend on efficient processing in prefrontal Res. 6, 351–360
circuits. 12 Allen, G.A. et al. (1996) Predicting environmental learning from spatial
Individual differences in navigational abilities are abilities: an indirect route. Intelligence 22, 327–355
particularly prominent at later stages in life and can 13 Fields, A.W. and Shelton, A.L. (2006) Individual skill differences and
large-scale environmental learning. J. Exp. Psychol. Learn. Mem.
determine how independently elderly people can go about
Cogn. 32, 506–515
their lives. Thus, even though much progress has been 14 Hegarty, M. et al. (2006) Spatial abilities at different scales: individual
made, it is important to arrive at a more comprehensive differences in aptitude-test performance and spatial-layout learning.
characterization of the factors that influence our naviga- Intelligence 34, 151–176
tional abilities. This will need to include a more detailed 15 Ishikawa, T. and Montello, D.R. (2006) Spatial knowledge acquisition
from direct experience in the environment: individual differences in the
understanding of the impact of biological predispositions development of metric knowledge and the integration of separately
(Box 3) and could ultimately help us understand how to learned places. Cogn. Psychol. 52, 93–129
achieve optimal developmental trajectories for successful 16 Szechter, L.E. and Liben, L.S. (2004) Parental guidance in
aging (see Box 4 for a list of outstanding questions and preschoolers’ understanding of spatial-graphic representations.
priorities for further research). Child Dev. 75, 869–885
17 Uttal, D.H. (2000) Seeing the big picture: map use and the development
of spatial cognition. Dev. Sci. 3, 247–286
References 18 Blajenkova, O. et al. (2005) Individual differences in the
1 Aguirre, G.K. and D’Esposito, M. (1999) Topographical disorientation: representations of novel environments. J. Environ. Psychol. 25,
a synthesis and taxonomy. Brain 122, 1613–1628 97–109

144
Review Trends in Cognitive Sciences Vol.14 No.3

19 Moffat, S.D. et al. (1998) Navigation in a ‘‘virtual’’ maze: sex differences 46 Bohbot, V.D. et al. (2007) Gray matter differences correlate with
and correlation with psychometric measures of spatial ability in spontaneous strategies in a human virtual navigation task. J.
humans. Evol. Hum. Behav. 19, 73–87 Neurosci. 27, 10078–10083
20 Pazzaglia, F. and DeBeni, R. (2006) Are people with high and low 47 Iaria, G. et al. (2008) Navigational skills correlate with hippocampal
mental rotation abilities differently susceptible to the alignment effect? fractional anisotropy in humans. Hippocampus 18, 335–339
Perception 35, 369–383 48 Maguire, E.A. et al. (2000) Navigation-related structural change in the
21 Saucier, D.M. et al. (2002) Are sex differences in navigation caused by hippocampi of taxi drivers. Proc. Natl. Acad. Sci. U. S. A. 97, 4398–4403
sexually dimorphic strategies or by differences in the ability to use the 49 Maguire, E.A. et al. (2006) London taxi drivers and bus drivers: a
strategies? Behav. Neurosci. 116, 403–410 structural MRI and neuropsychological analysis. Hippocampus 16,
22 Hegarty, M. and Waller, D. (2005) Individual differences in spatial 1091–1101
abilities. In The Cambridge Handbook of Visuospatial Thinking (Shah, 50 Cnotka, J. et al. (2008) Navigational experience affects hippocampus
P. and Miyake, A., eds), Cambridge University Press, pp. 121–169 size in homing pigeons. Brain Behav. Evol. 72, 233–238
23 Kozhevnikov, M. et al. (2006) Perspective-taking vs. mental rotation 51 Ciaramelli, E. (2008) The role of ventromedial prefrontal cortex in
transformation and how they predict spatial navigation performance. navigation: a case of impaired wayfinding and rehabilitation.
Appl. Cogn. Psychol. 20, 397–417 Neuropsychologia 46, 2099–2105
24 Liben, L.S. et al. Identifying locations and directions on field and 52 Moffat, S.D. et al. (2007) Extrahippocampal contributions to age
representational mapping tasks: predictors of success. Spat. Cogn. differences in human spatial navigation. Cereb. Cortex 17, 1274–1282
Comput. (2010) 53 Montello, D.R. et al. (1999) Sex-related differences and similarities in
25 Korol, D.L. et al. (2004) Shifts in preferred learning strategy across the geographic and environmental spatial abilities. Ann. Assoc. Am. Geogr.
estrous cycle in female rats. Horm. Behav. 45, 330–338 89, 515–534
26 Etchamendy, N. and Bohbot, V.D. (2007) Spontaneous navigational 54 Waller, D. (2000) Individual differences in spatial learning from
strategies and performance in the virtual town. Hippocampus 17, 595– computer-simulated environments. J. Exp. Psychol. Appl. 6, 307–321
599 55 Jonasson, Z. (2005) Meta-analysis of sex differences in rodent models of
27 Doeller, C.F. et al. (2008) Parallel striatal and hippocampal systems for learning and memory: a review of behavioral and biological data.
landmarks and boundaries in spatial memory. Proc. Natl. Acad. Sci. U. Neurosci. Biobehav. Rev. 28, 811–825
S. A. 105, 5915–5920 56 Coluccia, E. and Louse, G. (2004) Gender differences in spatial
28 Foo, P. et al. (2005) Do humans integrate routes into a cognitive map? orientation: a review. J. Environ. Psychol. 24, 329–340
Map- versus landmark-based navigation of novel shortcuts. J. Exp. 57 Chabanne, V. et al. (2004) Sex differences and women’s hormonal cycle
Psychol. Learn. Mem. Cogn. 31, 195–215 effect on spatial performance in a virtual environment navigation task.
29 Hölscher, C. (2009) Adaptivity of wayfinding strategies in a multi- Curr. Psychol. Cogn. 22, 351–375
building ensemble: the effects of spatial structure, task requirements 58 Voyer, D. et al. (2007) Gender differences in object location memory: a
and metric information. J. Environ. Psychol. 29, 208–219 meta-analysis. Psychon. Bull. Rev. 14, 23–38
30 Britten, K.H. (2008) Mechanisms of self-motion perception. Annu. Rev. 59 Burgess, N. et al. (2004) Orientational manoeuvres in the dark:
Neurosci. 31, 389–410 dissociating allocentric and egocentric influences on spatial memory.
31 Liang, Z. et al. Aging affects the direction selectivity of MT cells in Cognition 94, 149–166
rhesus monkeys. Neurobiol. Aging.(2010) 60 Chai, X.J. and Jacobs, L.F. (2009) Sex differences in directional cue use
32 Yang, Y. et al. (2009) Aging affects the neural representation of speed in in a virtual landscape. Behav. Neurosci. 123, 276–283
Macaque area MT. Cereb. Cortex 19, 1957–1967 61 Lawton, C.A. (1994) Gender differences in wayfinding strategies:
33 Wolbers, T. et al. (2007) Differential recruitment of the hippocampus, relationship to spatial ability and spatial anxiety. Sex Roles 30,
medial prefrontal cortex, and the human motion complex during path 765–779
integration in humans. J. Neurosci. 27, 9408–9416 62 Grön, G. et al. (2000) Brain activation during human navigation:
34 Jeffery, K.J. (2007) Self-localization and the entorhinal-hippocampal gender-different neural networks as substrate of performance. Nat.
system. Curr. Opin. Neurobiol. 17, 684–691 Neurosci. 3, 404–408
35 Jones, M.W. and Wilson, M.A. (2005) Theta rhythms coordinate 63 Bell, S. and Saucier, D.M. (2004) Relationship among environmental
hippocampal-prefrontal interactions in a spatial memory task. PLoS pointing accuracy, mental rotation, sex, and hormones. Environ.
Biol. 3, e402 Behav. 36, 251–265
36 Epstein, R.A. et al. (2005) Learning places from views: variation in 64 Burkitt, J. et al. (2007) Evidence for the influence of testosterone in the
scene processing as a function of experience and navigational ability. J. performance of spatial navigation in a virtual water maze in women
Cogn. Neurosci. 17, 73–83 but not in men. Horm. Behav. 51, 649–654
37 Burgess, N. (2008) Spatial cognition and the brain. Ann. N. Y. Acad. 65 Driscoll, I. et al. (2005) Virtual navigation in humans: the impact of age,
Sci. 1124, 77–97 sex, and hormones on place learning. Horm. Behav. 47, 326–335
38 Wolbers, T. and Büchel, C. (2005) Dissociable retrosplenial and 66 Jones, C.M. et al. (2003) The evolution of sex differences in spatial
hippocampal contributions to successful formation of survey ability. Behav. Neurosci. 117, 403–411
representations. J. Neurosci. 25, 3333–3340 67 Lawton, C.A. and Kallai, J. (2002) Gender differences in wayfinding
39 Wolbers, T. et al. (2004) Neural foundations of emerging route strategies and anxiety about wayfinding: a cross cultural comparison.
knowledge in complex spatial environments. Brain Res. Cogn. Brain Sex Roles 47, 389–401
Res. 21, 401–411 68 Monacelli, A.M. et al. (2003) Spatial disorientation in Alzheimer’s
40 Janzen, G. et al. (2008) Memory consolidation of landmarks in good disease: the remembrance of things passed. Neurology 61, 1491–1497
navigators. Hippocampus 18, 40–47 69 Duffy, C.J. (2009) Visual motion processing in aging and Alzheimer’s
41 Sandi, C. et al. (2004) Neurobiological and endocrine correlates of disease: neuronal mechanisms and behavior from monkeys to man.
individual differences in spatial learning ability. Learn. Mem. 11, Ann. N. Y. Acad. Sci. 1170, 736–744
244–252 70 Mahmood, O. et al. (2009) Age differences in visual path integration.
42 Iaria, G. et al. (2007) Retrosplenial and hippocampal brain regions in Behav. Brain Res. 205, 88–95
human navigation: complementary functional contributions to the 71 Wolbers, T. et al. (2008) Spatial updating: how the brain keeps track of
formation and use of cognitive maps. Eur. J. Neurosci. 25, 890–899 changing object locations during observer motion. Nat. Neurosci. 11,
43 Moser, M.B. and Moser, E.I. (1998) Functional differentiation in the 1223–1230
hippocampus. Hippocampus 8, 608–619 72 Wilson, I.A. et al. (2003) Place cell rigidity correlates with impaired
44 Hartley, T. et al. (2003) The well-worn route and the path less traveled. spatial learning in aged rats. Neurobiol. Aging 24, 297–305
Distinct neural bases of route following and wayfinding in humans. 73 Wilson, I.A. et al. (2006) Neurocognitive aging: prior memories hinder
Neuron 37, 877–888 new hippocampal encoding. Trends Neurosci. 29, 662–670
45 Iaria, G. et al. (2003) Cognitive strategies dependent on the 74 Drapeau, E. et al. (2003) Spatial memory performances of aged rats in
hippocampus and caudate nucleus in human navigation: variability the water maze predict levels of hippocampal neurogenesis. Proc. Natl.
and change with practice. J. Neurosci. 23, 5945–5952 Acad. Sci. U. S. A. 100, 14385–14390

145
Review Trends in Cognitive Sciences Vol.14 No.3

75 Barnes, C.A. et al. (1997) Multistability of cognitive maps in the 86 Goldman, D. et al. (2009) The role of COMT Val158Met in cognition.
hippocampus of old rats. Nature 388, 272–275 Biol. Psychiatry 65, e1–e2 author reply e3–e4
76 Burger, C. et al. (2007) Changes in transcription within the CA1 field of 87 Frank, M.J. et al. (2009) Prefrontal and striatal dopaminergic genes
the hippocampus are associated with age-related spatial learning predict individual differences in exploration and exploitation. Nat.
impairments. Neurobiol. Learn. Mem. 87, 21–41 Neurosci. 12, 1062–1068
77 Moffat, S.D. and Resnick, S.M. (2002) Effects of age on virtual 88 Bertolino, A. et al. (2006) Prefrontal-hippocampal coupling during
environment place navigation and allocentric cognitive mapping. memory processing is modulated by COMT val158met genotype.
Behav. Neurosci. 116, 851–859 Biol. Psychiatry 60, 1250–1258
78 Ruiz-Opazo, N. and Tonkiss, J. (2006) Genome-wide scan for 89 Bremner, J.D. et al. (2008) Structural and functional plasticity of the
quantitative trait loci influencing spatial navigation and social human brain in posttraumatic stress disorder. Prog. Brain Res. 167,
recognition memory in Dahl rats. Physiol. Genomics 26, 145–151 171–186
79 Mizuno, M. et al. (2000) Involvement of brain-derived neurotrophic 90 Herrero, A.I. et al. (2006) Individual differences in anxiety trait are
factor in spatial memory formation and maintenance in a radial arm related to spatial learning abilities and hippocampal expression of
maze test in rats. J. Neurosci. 20, 7116–7121 mineralocorticoid receptors. Neurobiol. Learn. Mem. 86, 150–159
80 Hariri, A.R. et al. (2003) Brain-derived neurotrophic factor val66met 91 Kalisch, R. et al. (2006) Anxiety and hippocampus volume in the rat.
polymorphism affects human memory-related hippocampal activity Neuropsychopharmacology 31, 925–932
and predicts memory performance. J. Neurosci. 23, 6690– 92 Sheline, Y.I. et al. (2002) The hippocampus and depression. Eur.
6694 Psychiatry 17 (Suppl. 3), 300–305
81 Bueller, J.A. et al. (2006) BDNF Val66Met allele is associated with 93 Wiener, J.M. et al. (2009) Towards a taxonomy of wayfinding tasks: a
reduced hippocampal volume in healthy subjects. Biol. Psychiatry 59, knowledge-based approach. Spat. Cogn. Comput. 9, 152–165
812–815 94 Iaria, G. et al. (2009) Developmental topographical disorientation: case
82 Colgin, L.L. et al. (2008) Understanding memory through hippocampal one. Neuropsychologia 47, 30–40
remapping. Trends Neurosci. 31, 469–477 95 Aporta, C. and Higgs, E. (2005) Satellite culture: global positioning
83 Kempermann, G. et al. (2006) Natural variation and genetic covariance systems, Inuit wayfinding, and the need for a new account of
in adult hippocampal neurogenesis. Proc. Natl. Acad. Sci. U. S. A. 103, technology. Curr. Anthropol. 46, 729–753
780–785 96 Tolman, C.E. (1948) Cognitive maps in rats and man. Psychol. Rev. 55,
84 Clelland, C.D. et al. (2009) A functional role for adult hippocampal 189–208
neurogenesis in spatial pattern separation. Science 325, 210–213 97 Thorndyke, P.W. and Hayes-Roth, B. (1982) Differences in spatial
85 Spiers, H.J. and Maguire, E.A. (2007) A navigational guidance system knowledge acquired from maps and navigation. Cogn. Psychol. 14,
in the human brain. Hippocampus 17, 618–626 560–589

146