Abstract

A parieto-medial temporal pathway is thought to underlie spatial navigation in humans. fMRI was used to assess the role of this pathway, including the hippocampus, in the cognitive processes likely to underlie navigation based on environmental cues. Participants completed a short-term spatial memory task in virtual space, which required no navigation but involved the recognition of a target location from a foil location based on environmental landmarks. The results showed that spatial memory retrieval based on environmental landmarks was indeed associated with increased signal in regions of the parieto-medial temporal pathway, including the superior parietal cortex, the retrosplenial cortex, and the lingual gyrus. However, the hippocampus demonstrated a signal decrease below the fixation baseline during landmark-based retrieval, whereas there was no signal change from baseline during retrieval based on viewer position. In a discussion of the origins of such negative BOLD response in the hippocampus, we consider both a suppression of default activity and an increase in activity without a corresponding boost in CBF as possible mechanisms.

INTRODUCTION

Efficient navigation requires a continuous coordination between the external environment as it is currently viewed and the internal representation of the same environment. Similarly, navigating from a new starting point or remembering a location following a viewpoint shift requires coordination between the environment as seen from the current viewpoint and the internal representation of the space acquired previously. During virtual navigation, a dorsal neural pathway, which extends medially from the posterior parietal lobe to the medial-temporal lobe via the retrosplenial cortex, is consistently shown to be active (Hartley, Maguire, Spiers, & Burgess, 2003; Grön, Wunderlich, Spitzer, Tomczak, & Riepe, 2000; Maguire et al., 1998). A similar pathway has been implicated when participants are simply required to make a reference to an environmental landmark, either to make a distance judgment or to retrieve a remembered target location (Galati, Pelle, Berthoz, & Committeri, 2010; Schmidt et al., 2007; Committeri et al., 2004), supporting a role for this pathway in the absence of actual navigation. Although the hippocampus is widely accepted to represent space in a world-centered manner and thought to represent the ultimate endpoint of the navigation pathway (Burgess, 2008; O'Keefe & Nadel, 1978), its role in landmark-based retrieval of object locations has received mixed support in healthy individuals. This study therefore aimed to investigate the role of the hippocampus, as part of the parieto-medial temporal pathway, in spatial memory retrieval based on environmental landmarks when no navigation is required.

There is little doubt that the hippocampus is important for spatial memory in humans. The discovery of place cells in the rodent and human hippocampus and demonstrations of severe spatial memory and learning deficits following hippocampal lesions in both species represent a fraction of the evidence supporting this conclusion (Ekstrom et al., 2003; Bohbot et al., 1998; Maguire, Burke, Phillips, & Staunton, 1996; Morris, 1981; Smith & Milner, 1981; O'Keefe & Nadel, 1978). Consistent with the idea that the hippocampus contributes to spatial memory by providing representations anchored in the external world, patients with hippocampal damage show a disproportionate impairment when a sudden shift to a different viewpoint forces the use of environmental landmarks to remember locations or spatial topographies (Lee et al., 2005; Hartley et al., 2003; King, Burgess, Hartley, Vargha-Khadem, & O'Keefe, 2002; Holdstock, Gutnikov, Gaffan, & Mayes, 2000; Abrahams, Pickering, Polkey, & Morris, 1997). The rationale behind such paradigms relies on the assumption that a shift in viewpoint makes the viewer position unreliable as a cue to location, which in turn encourages the use of stable landmarks in the surrounding environment to remember locations. However, it has been shown that behavioral impairment following hippocampal damage tends to be detectable only at longer delays (Holdstock et al., 2000), greater memory loads (Axmacher et al., 2007; King et al., 2002; Abrahams et al., 1997), and higher environmental complexity (Lee et al., 2005; Hartley et al., 2003). Although this suggests that an otherwise intact parieto-medial temporal pathway is sufficient at lower task demands and that the hippocampus is only critical when task demands are high, we cannot rule out that the hippocampus of healthy individuals is involved whenever landmark-based memory is required.

Evidence derived from neuroimaging in healthy individuals has provided mixed support for such a landmark-based retrieval function of the hippocampus. In a visually rich room, Lambrey, Doeller, Berthoz, and Burgess (2012) asked participants to imagine movement of their own viewpoint around an array of four objects laid out on a table or a rotation of the table itself; after which, memory for the object locations was tested in a change-detection task. Only in the viewpoint rotation condition were the spatial relations between the array and the environmental features in the room undisrupted, although the predictable imagined rotation may have allowed for the use of viewer position as an additional cue. When the viewer rotation condition was contrasted with the array rotation condition at retrieval, activation was found in the left hippocampus, in addition to other regions in the parieto-medial temporal pathway, including the retrosplenial cortex and the parahippocampal gyrus. Antonova et al. (2009) similarly encouraged the use of the cues provided by abstract patterns on the walls of a circular virtual arena by asking participants to navigate to a single remembered target location from novel start positions. When contrasted with rest, small clusters of differential activation were demonstrated in the right hippocampus of young individuals but not in old individuals. In an earlier study using the same task, however, no evidence for hippocampal activation could be demonstrated at retrieval (Parslow et al., 2004). Schmidt and colleagues (2007) also used a viewpoint shift to encourage the use of landmarks in a virtual roof garden to solve a location change-detection task of a single object. Although several of the regions in the parieto-medial temporal pathway were implicated when such landmark-based memory retrieval was compared with a high-level control condition, no differential activation could be demonstrated in the hippocampus. Similarly, in a task without a memory component, Committeri and colleagues (2004) asked participants to make viewer-, object-, and landmark-centered distance judgments in a virtual square arena in front of a building. Although the typical regions of the parieto-medial temporal pathway were active during the landmark-centered condition, including ventromedial occipito-temporal regions and the retrosplenial cortex, no differential activation was demonstrated in the hippocampus proper. In summary, despite great variability in task requirements, support for the role of the parieto-medial temporal pathway in landmark-based referencing is surprisingly consistent, whereas the same cannot be said for the hippocampus. This indicates a general involvement of the parieto-medial temporal pathway in any attempt to reference a landmark, whereas the role of the hippocampus appears more restricted.

Recent evidence derived from navigation-based spatial memory tasks has indicated that the hippocampus may be particularly important for the retrieval of environment-centered spatial representations that subsequently allow for goal-directed navigation. In a virtual Morris Water Maze, Shipman and Astur (2008) found evidence of hippocampal involvement only in the initial phase of navigation. Similarly, using a familiar large-scale virtual office environment, Xu, Evensmoen, Lehn, Pintzka, and Haberg (2010) showed that the anterior hippocampus was more active during the initial phase of wayfinding. Furthermore, by linking the brain activity associated with virtual wayfinding with postscan verbal accounts of the thought processes accompanying such wayfinding, Spiers and Maguire (2006) showed that the engagement of the hippocampus was brief and that it only occurred during the planning of a route to a new destination and not during route execution. It can therefore be proposed that the implication of the hippocampus in navigation is a reflection of its role in environment-centered memory retrieval and not in navigational movement per se. Consequently, a task that encourages the use of an environment-centered memory representation should be associated with hippocampal recruitment, independent of navigational demands. As mentioned previously, although investigations of patients with hippocampal damage have provided general support for this prediction using viewpoint-shift tasks (e.g., Hartley et al., 2003; King et al., 2002), neuroimaging investigations in healthy volunteers have yet to reach a similar consensus (e.g., Antonova et al., 2009; Schmidt et al., 2007). This study therefore aimed to further investigate the role of the hippocampus in nonnavigational environment-centered spatial memory retrieval in a neuroimaging context.

This study employed an encoding and recall paradigm where participants had to encode the location of a target within a circular arena in relation to landmarks provided on the walls of the arena. Following a delay period, they had to distinguish the target location from a foil location from a new viewpoint. This required participants to refer to the positions of the landmarks to identify the target location. In addition to this environment-centered (hereafter EC) condition, the task employed two control conditions. In the viewer-centered (VC) condition, the participants were instructed to imagine that the walls of the circular arena have moved but that their own viewpoint and, therefore, the position of the target relative to themselves had not changed. The inclusion of this condition was crucial, as any processing of spatial information necessarily involves at least some form of body-centered spatial representation (e.g., location of the target stimulus and landmarks on the retina). Thus, the EC and the VC conditions were highly similar and should only differ in the need to rely on environmental landmarks to localize the target in the recall phase. Although the environmental landmarks were crucial for retrieving the spatial representation in the EC condition, these same environmental landmarks were completely noninformative as spatial references in the VC condition. Previous studies have been inconsistent in this regard with some studies comparing tightly controlled conditions (e.g., Lambrey et al., 2012), whereas other studies compared conditions that differed in multiple aspects from one another. In this study, participants were only informed about whether the viewpoint or the walls had moved at the time of the delay, which ensured identical encoding processes in the two conditions. Consequently, any differential effects between the EC and VC conditions in the recall phase should not be because of differences in preceding encoding strategies.

The second control condition in this study was also highly similar to both the EC and VC conditions in terms of visual and motor processing but differed from these two in that it did not require the encoding or retrieval of a target location and will be referred to as the no-memory (NM) condition. Instead of memorizing a target location in the virtual arena, participants saw the empty arena during the presentation phase and made a response selection based on a visual cue during the response phase (see below).

If the hippocampus is instrumental in providing a mental representation of a location relative to environmental landmarks, independently of navigational demands, subtracting any activation related to representations relative to the own body alone (VC condition) from activation related to environmental landmarks and the own body (EC condition) should maximize the specificity of the contrast in this study. Because the EC condition was expected to involve translations between body-centered and environmental reference systems as well, it was hypothesized to also recruit the retrosplenial cortex to coordinate this translation. Contrasting the EC and VC conditions with the NM condition was expected to show activations related to both the encoding and the retrieval of spatial locations in large parts of the parieto-medial temporal pathway.

METHODS

Participants

Twenty young adults (10 women, 19–33 years old, mean = 26.1 years) with no history of psychiatric or neurological illness were recruited. All participants were right-handed, as ascertained with the Edinburgh Handedness Inventory. The study was approved by the ethics committee at the Faculty of Medical Sciences at Newcastle University. Participants provided written informed consent before the study.

Experimental Task

The Northumberland Gallery Task (NGT) was developed to encourage encoding and retrieval of spatial locations in relation to environmental landmarks or the viewer. This task was developed over a series of experiments to ensure behavioral separation between conditions with above-chance performance in all cases (Nilsson, Coventry, & Ferrier, 2010). A virtual circular arena, as viewed from the top of the boundary wall, with seven paintings (animal stimuli from Rossion & Pourtois, 2004) rendered at equidistance on the wall provided the context for the task. All trials exhibited the same temporal structure (Figure 1): a 250-msec presentation of the virtual empty area followed by the presentation phase for 3000 msec, during which the to-be-remembered location is marked with a pole somewhere inside the arena. This is followed by a delay phase (4750 msec), another 250-msec display of the empty arena, and a response phase of 3500 msec, in which the target location and a foil location are shown as two colored markers on the floor of the arena. A full trial period therefore lasted exactly 11.75 sec. Each trial was triggered by a scanner pulse. Thus, individual trials were separated by a short fixation cross of 1250-msec duration (total trial-onset asynchrony of 13 sec = 5*TR) or a longer baseline fixation period of 9050 msec (total trial-onset asynchrony of 20.8 sec = 8*TR; see below).

Figure 1. 

Trial structure and example stimuli. In the EC (A) and VC (B) conditions, the target location (green pole) was presented. During the delay phase, participants were informed of the upcoming manipulation via an instruction, which was overlaid on a scrambled image of the virtual scene. In the response phase, participants had to select the target location over a foil location after a shift in viewpoint (A) or a rotation of the walls (B). In the NM condition (C), no target location was presented, and in the response phase, participants simply had to identify the color of the marker on which the pole was standing.

Figure 1. 

Trial structure and example stimuli. In the EC (A) and VC (B) conditions, the target location (green pole) was presented. During the delay phase, participants were informed of the upcoming manipulation via an instruction, which was overlaid on a scrambled image of the virtual scene. In the response phase, participants had to select the target location over a foil location after a shift in viewpoint (A) or a rotation of the walls (B). In the NM condition (C), no target location was presented, and in the response phase, participants simply had to identify the color of the marker on which the pole was standing.

There were two conditions, which differed only in terms of the manipulation that occurred during the delay, out of view. The otherwise unfilled delay consisted of a scrambled image of the environment with a one-word instruction in the upper half of the screen, informing participants of the relevant manipulation for that particular trial. The instruction “you” indicated a manipulation of the viewer position along the top of the wall, thereby encouraging the use of landmarks to locate the target (EC condition). The instruction “walls” indicated the manipulation of landmark positions by a rotation of the walls, without any movement of the arena floor or of the observer position (VC condition). In the response phase, participants were required to distinguish the target location from the foil location, as quickly and accurately as possible, by identifying its color and pressing the appropriate response button. To ensure identical task length for all participants, the scene of the response phase remained on the screen for the full 3.5 sec independently of whether a response had been recorded within that time. In the NM condition, no target location appeared during the presentation phase. The instruction “none” was shown during its delay phase, and in the response phase, a pole was presented on top of one of the usual response options and participants were simply required to respond in accordance to the color of that particular option. This NM condition was therefore identical in terms of the visual scene, and the motor response required but involved no memory for location.

Each trial started from one of four start positions and manipulations in the experimental conditions occurred in clockwise and anticlockwise rotations at magnitudes of 45°, 90°, and 135°. There were 108 trials with 36 trials in each condition and 12 trials for each manipulation magnitude. In addition, there were 36 of the abovementioned long fixation periods. The trial order was unique for each participant according to the following randomization scheme: Groups of three trials, one from each condition, along with a single long baseline period were first randomized internally and then strung together for the full experiment. Orders that put two long baseline periods next to each other were removed. These pseudorandom orders of trial types were then randomly filled with actual trial parameters (rotation angles/directions, target/foil locations). This procedure ensured that the conditions were equally distributed across time and that each condition was approximately equally likely to be followed by any of the other conditions or a baseline period. It also ensured that the maximum temporal distance between two trials of the same or different condition was limited, thereby avoiding potential signal loss because of the filtering of low-frequency noise.

Stimulus presentation and response recording were performed using the software Presentation (version 14.9, Neurobehavioral Systems, Inc., Albany, CA). The NGT took 28.4 min to complete in the scanner. The trials were split into two separate scan runs lasting approximately 14.5 min each. After completing the NGT, a brief postscan interview was implemented to ensure that the instructions had been understood and that an appropriate strategy had been used.

Prescan Training

Before the task, participants learned the landmark positions by studying a small-scale cardboard replica of the arena for 2 min. A second model, in which the frames were empty, was used at test. On the basis of a fixed order, the experimenter placed one of seven landmark cards (animal drawings) in one of the empty frames, and the participant had to place a randomly drawn card in the correct frame relative to the first. This was repeated seven times, and in the case of no errors, the NGT followed. If errors were made, the learning and testing procedure was repeated.

Participants were then given instruction about the NGT and watched an animation highlighting the manipulations of the task. Participants completed 9–18 demonstration trials before testing commenced in the scanner. These demonstration trials were designed to familiarize the participants with the trial structure and timing and with the three possible conditions. The experimenter made no strategy recommendations.

Image Acquisition

MR scans were collected on a Philips Achieva 3-T MR scanner using an eight-channel SENSE coil. A standard T1-weighted Turbo Field Echo scan sequence (voxel size = 0.76 × 0.77 × 0.80 mm3, 225 slices, echo time [TE] = 4.6 msec) was used to acquire a structural scan for each participant. Two separate runs of functional scans were collected for the NGT using a single-shot EPI sequence (TE = 30 msec, repetition time = 2600 msec, voxel size = 2.5 × 2.5 × 3.5 mm3, 40 axial slices, tilted up approx. 20° from the AC–PC line), with 330 volumes per run. Three dummy volumes at the beginning of each run were immediately discarded. Stimulus presentation started after a further three volumes. Between both functional runs, a spoiled gradient-echo (T1-FFE) field mapping sequence (voxel size = 2 × 2 × 2 mm3, repetition time = 27 msec, TE1 = 2.6 msec, TE2 = 5.9 msec) was used to reconstruct magnetic field inhomogeneity.

fMRI Preprocessing

The MR data analysis was performed using SPM8 (www.fil.ion.ucl.ac.uk/spm) in Matlab R2010b (The MathWorks, Inc., Natick, MA). Standard preprocessing of functional images consisted of slice-time correction to the first slice, realignment and unwarping using the constructed field map, normalization to standard anatomical space using normalization parameters previously estimated from the structural scans, and spatial smoothing with an 8-mm FWHM Gaussian smoothing kernel. The first-level model consisted of three separate events per trial and condition to model the three phases of a trial: presentation, delay, and response. Because the presentation phases for EC and VC trials were indistinguishable for the participants (i.e., only during the delay were they told whether they or the walls will rotate), the onsets of the presentation phase for these two conditions were combined. The onsets were convolved with the canonical hemodynamic response function (HRF) of SPM. In addition, response times were added to the model as parametric modulators of BOLD amplitude of the response phase events. Trials in which no response was recorded (1.6% of trials) were included in the analysis as it is highly likely that participants were still engaged in the task during these trials but simply did not respond in time. This is supported by the fact that the majority of these trials were in the EC condition, which had significantly higher response times than the other two conditions (see below). Nonresponse trials were allocated a fixed value of 3.5 sec for the parametric modulator, equivalent to the maximum possible RT in the response phase. Lastly, motion parameters for each session were added to the first-level model to serve as regressors of no interest.

Parameter estimates for the various predictors were then combined across both sessions and entered into second-level models. The main analysis included the contrasts for the response phase of the three conditions along with a subject factor. Family-wise error (FWE) correction was used to correct p values for multiple comparisons to a Type-I error probability of 0.05 with an additional cluster extent threshold of 10 voxels.

BOLD signal time courses for the entire trial were based on additional first-level models using a finite-impulse basis set of order 12 for the three conditions. Here, the beginning of the encoding phase was used to define the onset of each trial. In contrast to the canonical HRF analysis described above, EC and VC conditions were therefore modeled separately from encoding phase onwards (as part of the entire trial).

RESULTS

Two participants were excluded from the analysis (one woman, one man): one because of excessive motion during scanning and one because of difficulties seeing the stimulus display, which the participant only reported after the scan.

Behavioral Performance

Sixteen of the eighteen included participants made no errors during training; one person reached training criterion in the second cycle, and one person, in the third cycle. Nonresponse trials were excluded from the analysis of behavioral performance, which made up 1.6% of trials, of which 1.2% were EC, 0.3% were VC, and 0.1% were NM trials. The percentage of trials in which participants successfully distinguished the target from the foil and RTs referring to correct responses are presented in Table 1. A repeated-measures ANOVA revealed significant differences between the three conditions in terms of accuracy (F(2, 34) = 104.66, p < .001) and RTs (F(2, 34) = 259.99, p < .001). A pairwise multiple comparison (post hoc pairwise t test) showed that participants were less accurate in the EC condition compared with the VC (p < .001) and NM (p < .001) conditions and in the VC compared with the NM condition (p < .05). One-sample t tests confirmed that performance was above chance in all conditions (all ps < .001). Similarly, RTs were longer in the EC condition compared with the VC (p < .001) and NM (p < .001) conditions, whereas there was no difference between the VC and NM conditions (p > .05). In the EC condition, angle of rotation was found to have a significant effect on accuracy (F(2, 34) = 12.76, p < .001) and RT (F(2, 34) = 30.49, p < .001), by which a greater angle of rotation was associated with lower accuracy (r = −.58, p < .001) and longer response times (r = .53, p < .001). In contrast, angle of rotation had no effect on accuracy (F(2, 34) = .90, p = .42) or on RTs (F(2, 34) = .47, p = .63) in the VC condition. This confirms that participants were taking the viewpoint shift into account when solving the EC trials, whereas they appropriately ignored and so were not affected by the shift of landmarks in the VC condition. In further support of this conclusion, all participants reported using an appropriate strategy in the postscan interview.

Table 1. 

Behavioral Data from the NGT (Mean ± Standard Deviation)

Condition
Rotation Angle (Degrees)
Accuracy (% Correct)
RT (msec)
No memory – 98% ± 3 1136 ± 236 
Environment centered average 77% ± 7 2280 ± 270 
45° 86% ± 8 2075 ± 315 
90° 77% ± 10 2300 ± 292 
135° 67% ± 15 2529 ± 287 
Viewer centered average 96% ± 4 1225 ± 207 
45° 95% ± 6 1220 ± 210 
90° 96% ± 7 1206 ± 288 
135° 97% ± 4 1252 ± 219 
Condition
Rotation Angle (Degrees)
Accuracy (% Correct)
RT (msec)
No memory – 98% ± 3 1136 ± 236 
Environment centered average 77% ± 7 2280 ± 270 
45° 86% ± 8 2075 ± 315 
90° 77% ± 10 2300 ± 292 
135° 67% ± 15 2529 ± 287 
Viewer centered average 96% ± 4 1225 ± 207 
45° 95% ± 6 1220 ± 210 
90° 96% ± 7 1206 ± 288 
135° 97% ± 4 1252 ± 219 

fMRI Results

Whole-brain Analysis

To investigate the brain regions involved in recognizing the target location based on environmental cues, independently of observer position, the signal acquired during the response phase in the EC condition was contrasted with the VC condition. A large network was observed, consisting of occipital, parietal, and mediotemporal as well as frontal regions (Figure 2). Local signal peaks were observed in regions of the parieto-medial temporal pathway, including the superior parietal lobe, the retrosplenial cortex, and the lingual gyrus (Table 2). In all clusters, the EC condition was associated with greater signal above baseline than the VC condition (Table 2). For the reverse contrast, clusters of differential signal were observed in frontal, parietal, and temporal regions (Figure 2), with local signal peaks in the posterior cingulate cortex, the medial superior frontal gyrus, and the hippocampus (Table 2). The contrast was generally characterized by the signal dropping below the baseline in the EC condition with a lesser drop or no change from baseline in the VC condition (Table 2). At encoding, the EC and VC conditions could not be differentiated and were therefore combined and contrasted with the NM condition. Clusters of differential signal associated with encoding were observed in occipital, parietal, and frontal regions (see Table 3).

Figure 2. 

Activation maps for the EC versus VC contrast at retrieval. Activations maps are shown in axial sections on the average normalized structural image computed from our sample data. Regions shown in yellow exhibited greater signal in the EC condition, whereas regions shown in red exhibited greater signal in the VC condition (p < .05, FWE, k ≥ 10). Numbers represent z coordinates in Montreal Neurological Institute (MNI) space.

Figure 2. 

Activation maps for the EC versus VC contrast at retrieval. Activations maps are shown in axial sections on the average normalized structural image computed from our sample data. Regions shown in yellow exhibited greater signal in the EC condition, whereas regions shown in red exhibited greater signal in the VC condition (p < .05, FWE, k ≥ 10). Numbers represent z coordinates in Montreal Neurological Institute (MNI) space.

Table 2. 

Peak Activations for the Whole-brain Analyses When Contrasting the EC and VC Conditions (p < .05, FWE, k ≥ 10)

Contrast
Region
Local Peak
Left
Right
Cluster (Voxels)
t
x, y, z (MNI)
Diff. Baseline (EC/VC)
Cluster (Voxels)
t
x, y, z (MNI)
Diff. Baseline (EC/VC)
EC > VC Parieto-occipital-temporal Retrosplenial cortex 10,830 13.48 −16, −70, 10 +/+ – 12.20 18, −62, 16 +/+ 
Lingual gyrus 10.48 −12, −61, 0 +/+ – 8.15 26, −86, −10 +/+ 
Fusiform gyrus 6.55 −38, −44, −20 +/+ – 6.74 38, −42, −22 +/+ 
Inf. occipital gyrus 7.90 −38, −76, −12 +/+ – 7.16 38, −76, −12 +/+ 
Precuneus 8.88 −4, −64, 52 +/+ – 9.93 12, −66, 52 +/+ 
Sup. parietal lobe 8.84 −16, −70, 54 +/+ – 13.24 20, −72, 48 +/+ 
Mid. temporal gyrus – – +/+ 41 6.49 46, −70, 14 +/+ 
Frontal Insula 197 8.93 −28, 24, −2 +/+ 119 6.69 32, 26, −4 +/+ 
Medial frontal gyrus 671 8.33 −6, 10, 50 +/+ – 7.83 2, 14, 50 +/+ 
Mid. frontal gyrus 442 9.39 −24, 0, 52 +/+ 319 7.74 34, 0, 50 +/+ 
Inf. frontal gyrus 64 6.85 50, 32, 24 +/+ – – – +/+ 
Precentral gyrus 402 8.74 −38, 8, 30 +/+ 18 6.02 46, 12, 30 +/+ 
Other Brain stem 302 7.91 −4, −28, −4 +/+ – – –  
BG 48 7.28 −12, 0, −2, +/+ – – –  
Cerebellum – – –  34 6.65 8, −72, −26 +/+ 
VC > EC Parietal Posterior cingulate cortex 217 7.76 −10, −50, 26 −/− 18 6.39 6, −50, 24 −/− 
Frontal Insula 1,624 9.34 −38, −10, 14 0/+ 35 7.02 38, 6, 10 0/+ 
Angular gyrus 11 6.86 −42, −64, 28 −/− 14 6.28 52, −58, 28 −/− 
Sup. medial frontal gyrus 143 6.70 −8, 54, 0 −/− 24 6.34 8, 58, 18 −/− 
Sup. frontal gyrus 88 7.68 −14, 54, 24 −/− – – –  
Precentral gyrus – – –  31 7.01 44, −20, 52 −/0 
Supramarginal gyrus 51 6.42 58, −24, 22 0/0 – – –  
Temporal Temporal pole – – –  96 7.98 42, 10, −34 −/− 
Mid. temporal gyrus – – –  275 7.08 54, −38, 0 −/0 
– – –  21 6.55 52, 2, −20 −/− 
Hippocampus 11 6.31 −28, −14, −22 −/0 86 7.07 30, −16, −18 −/0 
Contrast
Region
Local Peak
Left
Right
Cluster (Voxels)
t
x, y, z (MNI)
Diff. Baseline (EC/VC)
Cluster (Voxels)
t
x, y, z (MNI)
Diff. Baseline (EC/VC)
EC > VC Parieto-occipital-temporal Retrosplenial cortex 10,830 13.48 −16, −70, 10 +/+ – 12.20 18, −62, 16 +/+ 
Lingual gyrus 10.48 −12, −61, 0 +/+ – 8.15 26, −86, −10 +/+ 
Fusiform gyrus 6.55 −38, −44, −20 +/+ – 6.74 38, −42, −22 +/+ 
Inf. occipital gyrus 7.90 −38, −76, −12 +/+ – 7.16 38, −76, −12 +/+ 
Precuneus 8.88 −4, −64, 52 +/+ – 9.93 12, −66, 52 +/+ 
Sup. parietal lobe 8.84 −16, −70, 54 +/+ – 13.24 20, −72, 48 +/+ 
Mid. temporal gyrus – – +/+ 41 6.49 46, −70, 14 +/+ 
Frontal Insula 197 8.93 −28, 24, −2 +/+ 119 6.69 32, 26, −4 +/+ 
Medial frontal gyrus 671 8.33 −6, 10, 50 +/+ – 7.83 2, 14, 50 +/+ 
Mid. frontal gyrus 442 9.39 −24, 0, 52 +/+ 319 7.74 34, 0, 50 +/+ 
Inf. frontal gyrus 64 6.85 50, 32, 24 +/+ – – – +/+ 
Precentral gyrus 402 8.74 −38, 8, 30 +/+ 18 6.02 46, 12, 30 +/+ 
Other Brain stem 302 7.91 −4, −28, −4 +/+ – – –  
BG 48 7.28 −12, 0, −2, +/+ – – –  
Cerebellum – – –  34 6.65 8, −72, −26 +/+ 
VC > EC Parietal Posterior cingulate cortex 217 7.76 −10, −50, 26 −/− 18 6.39 6, −50, 24 −/− 
Frontal Insula 1,624 9.34 −38, −10, 14 0/+ 35 7.02 38, 6, 10 0/+ 
Angular gyrus 11 6.86 −42, −64, 28 −/− 14 6.28 52, −58, 28 −/− 
Sup. medial frontal gyrus 143 6.70 −8, 54, 0 −/− 24 6.34 8, 58, 18 −/− 
Sup. frontal gyrus 88 7.68 −14, 54, 24 −/− – – –  
Precentral gyrus – – –  31 7.01 44, −20, 52 −/0 
Supramarginal gyrus 51 6.42 58, −24, 22 0/0 – – –  
Temporal Temporal pole – – –  96 7.98 42, 10, −34 −/− 
Mid. temporal gyrus – – –  275 7.08 54, −38, 0 −/0 
– – –  21 6.55 52, 2, −20 −/− 
Hippocampus 11 6.31 −28, −14, −22 −/0 86 7.07 30, −16, −18 −/0 

Differences from baseline for the two conditions are marked as + when positive, as − when negative, and as 0 when not significant. Diff = difference; inf = inferior; sup = superior; mid = middle.

Table 3. 

Peak Activations for the Whole-brain Analyses When Contrasting Encoding (EC and VC Conditions Combined) with the NM Condition (p < .05, FWE, k ≥ 10)

Contrast
Region
Local Peak
Left
Right
Cluster (Voxels)
t
x, y, z (MNI)
Diff. Baseline (Enc/NM)
Cluster (Voxels)
t
x, y, z (MNI)
Diff. Baseline (Enc/NM)
Encod > NoMem Parietal Sup. parietal lobe 1,635 13.94 −30, −46, 42 +/0 169 9.72 42, −40, 48 +/0 
– – –  534 10.91 18, −66, 54 +/+ 
Angular gyrus – – –  65 8.64 32, −54, 42 +/+ 
Occipital Mid. occipital gyrus – – –  1,727 12.94 24, −88, 8 +/+ 
Inf. occipital gyrus 1,904 14.71 −38, −78, −4 +/+ – – –  
Cuneus 27 8.75 −10, −94, 16 +/+ – – –  
Frontal Inf. frontal gyrus 385 11.42 −38, 6, 26 +/0 – – –  
Mid. frontal gyrus 127 9.57 −22, 4, 48 +/0 – – –  
55 9.22 −46, 2, 48 +/+ – – –  
– – –  48 9.61 28, 8, 56 +/0 
Medial frontal gyrus 199 9.88 −5, 10, 50 +/0 – – –  
Other Putamen 75 9.06 −14, 10, 4 +/0 – – –  
NoMem > Encod Frontal Medial frontal gyrus 47 8.30 0, 62, 16 −/0 – – –  
Other Subgyral temporal 32 8.95 −30, −54, 2 −/0     
Contrast
Region
Local Peak
Left
Right
Cluster (Voxels)
t
x, y, z (MNI)
Diff. Baseline (Enc/NM)
Cluster (Voxels)
t
x, y, z (MNI)
Diff. Baseline (Enc/NM)
Encod > NoMem Parietal Sup. parietal lobe 1,635 13.94 −30, −46, 42 +/0 169 9.72 42, −40, 48 +/0 
– – –  534 10.91 18, −66, 54 +/+ 
Angular gyrus – – –  65 8.64 32, −54, 42 +/+ 
Occipital Mid. occipital gyrus – – –  1,727 12.94 24, −88, 8 +/+ 
Inf. occipital gyrus 1,904 14.71 −38, −78, −4 +/+ – – –  
Cuneus 27 8.75 −10, −94, 16 +/+ – – –  
Frontal Inf. frontal gyrus 385 11.42 −38, 6, 26 +/0 – – –  
Mid. frontal gyrus 127 9.57 −22, 4, 48 +/0 – – –  
55 9.22 −46, 2, 48 +/+ – – –  
– – –  48 9.61 28, 8, 56 +/0 
Medial frontal gyrus 199 9.88 −5, 10, 50 +/0 – – –  
Other Putamen 75 9.06 −14, 10, 4 +/0 – – –  
NoMem > Encod Frontal Medial frontal gyrus 47 8.30 0, 62, 16 −/0 – – –  
Other Subgyral temporal 32 8.95 −30, −54, 2 −/0     

Differences from baseline for the two conditions are marked as + when positive, as − when negative, and as 0 when not significant. Enc/Encod = encoding; NoMem = no memory.

Time Course Analyses

Time course analyses were performed in voxels of local peak activation in the superior parietal lobe, retrosplenial cortex, hippocampus, and posterior cingulate cortex (Figure 3). Given the typical time course of the canonical HRF implemented in SPM, signal related to the encoding phase should peak around 5 sec after trial onset, signal related to the delay phase should peak between 8 and 13 sec after trial onset, and signal related to the recall phase should peak around 13–15 sec after trial onset. In the early part of the time course (0–10 sec), BOLD signal changes in the EC and VC conditions appeared similar in all depicted regions, reflecting the necessarily identical encoding process for these two conditions. Later in the time series (13–16 sec), however, the change in BOLD signal in the EC condition was observed to be more substantial compared with the VC condition, which was reflected in the greater increase above the fixation baseline in the superior parietal lobe and the retrosplenial cortex and a greater drop below the baseline in the posterior cingulate cortex and the hippocampus. In this part of the time course, as opposed to mirroring the BOLD signal change of the EC condition, the signal change associated with the VC condition was observed to converge toward that of the NM control condition. Across the time course, the NM control condition consistently produced a BOLD signal closer to baseline compared with the other two conditions.

Figure 3. 

Plots of BOLD signal time course changes in the superior parietal lobe (A), retrosplenial cortex (B), hippocampus (C), and posterior cingulate cortex (D). Time course changes are shown in sagittal sections on the average normalized structural image computed from our sample data (activation maps: p < .05, FWE, k ≥ 10). The voxels selected for analysis were the ones with peak differences in the EC versus VC contrasts in each of the regions. Signal changes in the EC (blue, solid), VC (green, dashed), and NM (red, dotted) conditions were modeled from the onset of the trial.

Figure 3. 

Plots of BOLD signal time course changes in the superior parietal lobe (A), retrosplenial cortex (B), hippocampus (C), and posterior cingulate cortex (D). Time course changes are shown in sagittal sections on the average normalized structural image computed from our sample data (activation maps: p < .05, FWE, k ≥ 10). The voxels selected for analysis were the ones with peak differences in the EC versus VC contrasts in each of the regions. Signal changes in the EC (blue, solid), VC (green, dashed), and NM (red, dotted) conditions were modeled from the onset of the trial.

DISCUSSION

This study investigated whether the parieto-medial temporal pathway, with a particular focus on the hippocampus, is important for STM retrieval of target locations based on environmental landmarks when no navigation is required. To this end, the NGT was used to separate and contrast spatial memory based on environmental cues or viewer position. The EC condition was found to be associated with a substantial cluster of activation extending medially from the superior part of the posterior parietal lobe to the medial-temporal lobe, via the retrosplenial cortex, indicating that the full extent of the parieto-medial temporal pathway is important for the type of processing likely to precede spatial navigation based on environmental cues (Kravitz, Saleem, Baker, & Mishkin, 2011; Burgess, 2008; Shipman & Astur, 2008).

The posterior parietal involvement in the EC condition may appear contradictory to its predominantly viewer-centered role (Burgess, 2008). It is consistent, however, with studies linking this region to the coordination of body knowledge with sensory maps of space (Galati et al., 2010; Committeri et al., 2004); although retinotopic and body-centered representations largely overlap in the VC condition, the shift in virtual position of the observer in the EC condition would have reduced this overlap, necessitating more substantial remapping between several viewer-centered coordinate systems in the EC condition. A posterior parietal involvement is also consistent with recent proposals extending the function of this region from solely involving transformations between different body-centered representations to also covering transformations between body-centered and environment-centered representations (Calton & Taube, 2009; Save & Poucet, 2009; Byrne, Becker, & Burgess, 2007). Independent of transformation type, a greater reliance on spatial transformations in the EC condition is likely to account for the posterior parietal involvement.

The strongest effect associated with the EC condition was observed in the retrosplenial cortex, which adds to existing evidence supporting its involvement in transformations between body-centered and environment-centered representations (Burgess, 2008; Byrne et al., 2007; Maguire, 2001). It also supports a role for the retrosplenial cortex extending beyond navigation (Vann, Aggleton, & Maguire, 2009) to also include environmental referencing following an instantaneous shift in viewpoint (Galati et al., 2010). Furthermore, the time series analysis revealed an early and a late peak in BOLD signal change in the retrosplenial cortex, indicating that environmental referencing is likely to have taken place at encoding as well as at retrieval. The precuneus was also found to play a part in the EC condition, which may be linked to the imagery formed of the retrieved material (Burgess, Maguire, Spiers, & O'Keefe, 2001; Fletcher, Shallice, Frith, Frackowiak, & Dolan, 1996). Specifically, following the viewpoint shift, the precuneus could have provided an updated image of the target location and of obstructed parts of the scene. Furthermore, the EC condition was associated with activation in the lingual gyrus, which likely reflects the reliance on the orientation value of landmarks in this condition (Aguirre & D'Esposito, 1999; Aguirre, Zarahn, & D'Esposito, 1998). The absence of differential signal in the parahippocampal place area is likely to be accounted for by passive viewing of scenes, which took place in both conditions (Epstein & Kanwisher, 1998). From these results, it is evident that large parts of the parieto-medial temporal pathway feature when a reference to environmental landmarks is required, supporting its prenavigational role.

In stark contrast to the regions discussed above, the hippocampus showed a greater signal in the VC condition compared with the EC condition. Relative to the fixation baseline, however, this effect was characterized by a large drop in signal below baseline in the EC condition and little change from baseline in the VC condition. This effect also appeared to be relatively exclusive to the retrieval phase of the task, as indicated by the time series analysis. Although contradictory to the predictions, a negative BOLD response in the hippocampus is not an uncommon finding. Such a response has been associated with virtual versions of traditional spatial memory tasks, including the Morris Water Maze (Shipman & Astur, 2008) and the Radial Arm Maze (Astur et al., 2005), as well as with increasing goal distance in a route-planning task (Viard, Doeller, Hartley, Bird, & Burgess, 2011), autobiographical spatial judgments of long-term memories (Rekkas et al., 2005), and detection of location changes following an imagined shift in viewpoint (Lambrey et al., 2012). Such reports of negative hippocampal BOLD responses in spatially demanding and arguably hippocampus-relevant task conditions not only indicate the reliability of the finding but also highlight the need for interpretation.

Although the positive BOLD response has been related to increased neuronal activity (Logothetis, Pauls, Augath, Trinath, & Oeltermann, 2001), the mechanisms underlying the negative BOLD response are much less clear (Hayes & Huxtable, 2012). Three theoretical accounts are generally offered: The negative deflection could be the result (a) of vascular steal by which oxygenated blood is diverted away from less active areas to more active areas, (b) of an active neuronal suppression in the region, or (c) of a contradictory increase in neuronal activity without a corresponding boost in blood flow (Wade, 2002). Although vascular steal is unlikely because of the small changes in CBF accompanying cognition and the substantial hemodynamic reserve of the brain (Gusnard & Raichle, 2001), the two latter accounts are relevant in the present context.

It is possible that the EC condition, contrary to the prediction, requires a functional suppression of the hippocampus. Such an interpretation was favored by Reas, Gimbel, Hales, and Brewer (2011), who demonstrated a negative BOLD response in the hippocampus during elaborate associative recall, which was greater for poorly remembered than for strongly remembered items. It was argued that the longer memory search accompanying poorly remembered items required a greater suppression of encoding-related activity in the hippocampus, in favor of retrieval-related processes taking place elsewhere. Compared with the VC condition, the EC condition indeed required a longer memory search, in addition to potential retrieval of obscured relevant landmarks. However, no evidence was found of an increased BOLD response in the anterior hippocampus at encoding, contradicting the idea that signal in this region is reflective of encoding processes in the task. In addition, in the case of a functional suppression, one would expect greater suppression to be associated with better performance. Contrary to this, the generally poorer performance in the EC condition relative to the VC condition supports an inverse relationship between the level of negative BOLD signal and performance, similar to the results of Reas et al. (2011).

As opposed to a task-specific effect, the link between negative BOLD signal and poor performance may instead indicate a general effect of task difficulty. In support of such an interpretation, at rest, the hippocampus demonstrates functional correlations with the default network, which consistently shows deactivations during active compared with passive baseline conditions (Shulman et al., 1997) and during difficult compared with easy task conditions (Gimbel & Brewer, 2011; McKiernan, Kaufman, Kucera-Thompson, & Binder, 2003). Furthermore, it is noteworthy that several of the studies demonstrating hippocampal deactivations in arguably hippocampus-relevant conditions reported worse behavioral performance in that particular condition (Lambrey et al., 2012; Rodriguez, 2010; Shipman & Astur, 2008; Rekkas et al., 2005). However, the coupling of the hippocampus with other default regions during memory retrieval appears to vary according to task condition (Gimbel & Brewer, 2011; Huijbers, Pennartz, Cabeza, & Daselaar, 2011; Reas et al., 2011), which indicates that the areas deactivated during memory retrieval may only partially overlap with those deactivated during nonmemory tasks (Israel, Seibert, Black, & Brewer, 2010). In relation to this, although the more difficult EC condition was associated with deactivations in default regions such as the medial pFC, the posterior cingulate cortex, and the inferior parietal lobule, it was associated with strong activations in another default region, the retrosplenial cortex (Buckner, Andrews-Hanna, & Schacter, 2008). More importantly, when BOLD signal is not used as the measure of neuronal activity, previous studies tend to contradict a suppression of hippocampal activity during spatial memory tasks. For example, when contrasted to a visuomotor control condition, goal-directed navigation has been found to be associated with increased CBF to the hippocampus, as measured by PET (Maguire, Frackowiak, & Frith, 1997), and with increased theta activity in the hippocampus, as measured by magnetoencephalography (Cornwell, Johnson, Holroyd, Carver, & Grillon, 2008). Furthermore, contrary to the characteristics of the default network, neurons in the hippocampus have been found to show no or very low activity during baseline conditions while firing selectively in response to different categories of visual stimuli (Kraskov, Quiroga, Reddy, Fried, & Koch, 2007; Quiroga, Reddy, Kreiman, Koch, & Fried, 2005). Thus, although a task-independent effect of difficulty cannot be excluded in this study, a general suppression of the default network may not fully account for the negative BOLD response in the hippocampus.

The negative BOLD response in the EC condition may not be a reflection of suppression of neuronal activity but of an increase of neuronal activity. This is possible because of the relative nature of the BOLD signal, which depends on a complex interplay between changes in CBF, cerebral blood volume, and oxygen metabolism (cerebral metabolic rate of oxygen, CMRO2) that results from neuronal activity (Buxton, 2012; Buxton, Uludag, Dubowitz, & Liu, 2004; Logothetis & Wandell, 2004). In fact, measurable increases in BOLD signal rely on a relatively greater increase in CBF compared with CMRO2 (Ogawa, Lee, Kay, & Tank, 1990). Consequently, if neuronal activity causes a greater increase in CMRO2 relative to the increase in CBF, a decreased BOLD signal could theoretically result (Buxton, 2012).

Ekstrom (2010) proposed that the negative BOLD response in the hippocampus during memory encoding and retrieval tasks could be explained by such a neurovascular account. In support of this account, the coupling between CBF and CMRO2 in the hippocampus appears more complex than that traditionally observed in the cortex (Restom, Perthen, & Liu, 2008; Leontiev, Dubowitz, & Buxton, 2007), possibly as a result of a more limited vascular capacity in the hippocampus compared with the cortex (Borowsky & Collins, 1989). For example, whereas BOLD changes in the parahippocampus were found to be positively correlated with local field potential (LFP) power changes in a sample of epilepsy patients during a spatial navigation task, BOLD changes in the hippocampus showed a weak or no correlation with LFP power changes (Ekstrom, Suthana, Millett, Fried, & Bookheimer, 2009). Furthermore, Schridde et al. (2008) found that induced seizures in the rat resulted in marked increases in LFP activity across the entire brain but that such increases were associated with negative BOLD responses in the hippocampus and with positive BOLD responses in the cortex. Importantly, the coupling between CMRO2 and CBF was found to account for the negative BOLD response; the increase in CRMO2 nearly matched the increase in CBF in the hippocampus, whereas the normal CBF/CMRO2 overshoot was observed in the cortex. On the basis of such results, Ekstrom (2010) proposed that demanding memory tasks may be associated with an increase in CMRO2 that is just matched or even undershot by the increase in CBF, which, when contrasted with a resting baseline condition, results in a negative BOLD signal. Considering the lack of signal change relative to the baseline in the VC condition, such a scenario could account for the present findings: The demand of the EC condition could have caused the oxygen consumption in the hippocampus to exceed its supply, resulting in a negative BOLD signal in the face of increased neuronal activity.

The discussion above has highlighted two valid but mutually exclusive accounts of the negative BOLD signal demonstrated in the hippocampus, which is the direct consequence of the indirect relationship between BOLD signal and neuronal activity. Although an independent baseline allowed a more comprehensive interpretation of the unpredicted pattern of BOLD signal in this study, a more direct measure of neuronal activity will be required to characterize the precise relationship between BOLD signal and neuronal activity in the human hippocampus. We argue, in agreement with previous recommendations (Hayes & Huxtable, 2012; Ekstrom, 2010), that techniques such as calibrated fMRI and multimodal imaging have the potential of disambiguating the mechanisms underlying the BOLD response in the hippocampus, which will be critical for experimental interpretations in both basic and applied research settings.

Conclusion

In line with the predictions, this study has demonstrated the importance of the parieto-medial temporal pathway in short-term spatial memory that relies on environmental landmarks, even when no navigation is required. The hippocampus was found to play a differential role in the EC condition, but in contrast to the prediction, this role was characterized by a substantial negative BOLD response. As such, this study builds on previous demonstrations of a negative BOLD response in the hippocampus in spatially demanding task conditions. Although a suppression of default activity is a valid account, this would be in stark contrast to a long and robust tradition of linking the hippocampus to spatial memory, both at an electrophysiological (e.g., O'Keefe & Nadel, 1978) and a pathophysiological level (e.g., King et al., 2002). We therefore suggest that the alternative account, by which the negative BOLD response is a reflection of increased neuronal activity without a corresponding boost in CBF, is explored in future investigations.

Reprint requests should be sent to Jonna Nilsson, Newcastle University, Institute for Ageing and Health, Campus for Ageing and Vitality, NE4 5PL, Newcastle upon Tyne, United Kingdom, or via e-mail: jonna.nilsson@ncl.ac.uk.

REFERENCES

Abrahams
,
S.
,
Pickering
,
A.
,
Polkey
,
C. E.
, &
Morris
,
R. G.
(
1997
).
Spatial memory deficits in patients with unilateral damage to the right hippocampal formation.
Neuropsychologia
,
35
,
11
24
.
Aguirre
,
G. K.
, &
D'Esposito
,
M.
(
1999
).
Topographical disorientation: A synthesis and taxonomy.
Brain
,
122
,
1613
1628
.
Aguirre
,
G. K.
,
Zarahn
,
E.
, &
D'Esposito
,
M.
(
1998
).
An area within human ventral cortex sensitive to “building” stimuli: Evidence and implications.
Neuron
,
21
,
373
383
.
Antonova
,
E.
,
Parslow
,
D.
,
Brammer
,
M.
,
Dawson
,
G. R.
,
Jackson
,
S. H.
, &
Morris
,
R. G.
(
2009
).
Age-related neural activity during allocentric spatial memory.
Memory
,
17
,
125
143
.
Astur
,
R.
,
Germain
,
S. A. S.
,
Baker
,
E. K.
,
Calhoun
,
V.
,
Pearlson
,
G. D.
, &
Constable
,
R. T.
(
2005
).
fMRI hippocampal activity during a virtual radial arm maze.
Applied Psychophysiology and Biofeedback
,
30
,
307
317
.
Axmacher
,
N.
,
Mormann
,
F.
,
Fernandez
,
G.
,
Cohen
,
M. X.
,
Elger
,
C. E.
, &
Fell
,
J.
(
2007
).
Sustained neural activity patterns during working memory in the human medial temporal lobe.
Journal of Neuroscience
,
27
,
7807
7816
.
Bohbot
,
V. D.
,
Kalina
,
M.
,
Stepankova
,
K.
,
Spackova
,
N.
,
Petrides
,
M.
, &
Nadel
,
L.
(
1998
).
Spatial memory deficits in patients with lesions to the right hippocampus and to the right parahippocampal cortex.
Neuropsychologia
,
36
,
1217
1238
.
Borowsky
,
I. W.
, &
Collins
,
R. C.
(
1989
).
Metabolic anatomy of brain—A comparison of regional capillary density, glucose-metabolism, and enzyme-activities.
Journal of Comparative Neurology
,
288
,
401
413
.
Buckner
,
R. L.
,
Andrews-Hanna
,
J. R.
, &
Schacter
,
D. L.
(
2008
).
The brain's default network—Anatomy, function, and relevance to disease.
Year in Cognitive Neuroscience 2008
,
1124
,
1
38
.
Burgess
,
N.
(
2008
).
Spatial cognition and the brain.
Annals of the New York Academy of Sciences
,
1124
,
77
97
.
Burgess
,
N.
,
Maguire
,
E. A.
,
Spiers
,
H. J.
, &
O'Keefe
,
J.
(
2001
).
A temporoparietal and prefrontal network for retrieving the spatial context of lifelike events.
Neuroimage
,
14
,
439
453
.
Buxton
,
R. B.
(
2012
).
Dynamic models of BOLD contrast.
Neuroimage
,
62
,
953
961
.
Buxton
,
R. B.
,
Uludag
,
K.
,
Dubowitz
,
D. J.
, &
Liu
,
T. T.
(
2004
).
Modeling the hemodynamic response to brain activation.
Neuroimage
,
23
,
S220
S233
.
Byrne
,
P.
,
Becker
,
S.
, &
Burgess
,
N.
(
2007
).
Remembering the past and imagining the future: A neural model of spatial memory and imagery.
Psychological Review
,
114
,
340
375
.
Calton
,
J. L.
, &
Taube
,
J. S.
(
2009
).
Where am I and how will I get there from here? A role for posterior parietal cortex in the integration of spatial information and route planning.
Neurobiology of Learning and Memory
,
91
,
186
196
.
Committeri
,
G.
,
Galati
,
G.
,
Paradis
,
A. L.
,
Pizzamiglio
,
L.
,
Berthoz
,
A.
, &
LeBihan
,
D.
(
2004
).
Reference frames for spatial cognition: Different brain areas are involved in viewer-, object-, and landmark-centered judgments about object location.
Journal of Cognitive Neuroscience
,
16
,
1517
1535
.
Cornwell
,
B. R.
,
Johnson
,
L. L.
,
Holroyd
,
T.
,
Carver
,
F. W.
, &
Grillon
,
C.
(
2008
).
Human hippocampal and parahippocampal theta during goal-directed spatial navigation predicts performance on a virtual Morris water maze.
Journal of Neuroscience
,
28
,
5983
5990
.
Ekstrom
,
A.
(
2010
).
How and when the fMRI BOLD signal relates to underlying neural activity: The danger in dissociation.
Brain Research Reviews
,
62
,
233
244
.
Ekstrom
,
A.
,
Suthana
,
N.
,
Millett
,
D.
,
Fried
,
I.
, &
Bookheimer
,
S.
(
2009
).
Correlation between BOLD fMRI and theta-band local field potentials in the human hippocampal area.
Journal of Neurophysiology
,
101
,
2668
2678
.
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
,
184
188
.
Epstein
,
R.
, &
Kanwisher
,
N.
(
1998
).
The parahippocampal place area: A cortical representation of the local visual environment.
Neuroimage
,
7
,
S341
.
Fletcher
,
P. C.
,
Shallice
,
T.
,
Frith
,
C. D.
,
Frackowiak
,
R. S. J.
, &
Dolan
,
R. J.
(
1996
).
Brain activity during memory retrieval—The influence of imagery and semantic cueing.
Brain
,
119
,
1587
1596
.
Galati
,
G.
,
Pelle
,
G.
,
Berthoz
,
A.
, &
Committeri
,
G.
(
2010
).
Multiple reference frames used by the human brain for spatial perception and memory.
Experimental Brain Research
,
206
,
109
120
.
Gimbel
,
S. I.
, &
Brewer
,
J. B.
(
2011
).
Reaction time, memory strength, and fMRI activity during memory retrieval: Hippocampus and default network are differentially responsive during recollection and familiarity judgments.
Cognitive Neuroscience
,
2
,
19
26
.
Grön
,
G.
,
Wunderlich
,
A. P.
,
Spitzer
,
M.
,
Tomczak
,
R.
, &
Riepe
,
M. W.
(
2000
).
Brain activation during human navigation: Gender-different neural networks as substrate of performance.
Nature Neuroscience
,
3
,
404
408
.
Gusnard
,
D. A.
, &
Raichle
,
M. E.
(
2001
).
Searching for a baseline: Functional imaging and the resting human brain.
Nature Reviews Neuroscience
,
2
,
685
694
.
Hartley
,
T.
,
Maguire
,
E. A.
,
Spiers
,
H. J.
, &
Burgess
,
N.
(
2003
).
The well-worn route and the path less traveled: Distinct neural bases of route following and wayfinding in humans.
Neuron
,
37
,
877
888
.
Hayes
,
D. J.
, &
Huxtable
,
A. G.
(
2012
).
Interpreting deactivations in neuroimaging.
Frontiers in Psychology
,
3
,
27
.
Holdstock
,
J. S.
,
Gutnikov
,
S. A.
,
Gaffan
,
D.
, &
Mayes
,
A. R.
(
2000
).
Perceptual and mnemonic matching-to-sample in humans: Contributions of the hippocampus, perirhinal and other medial temporal lobe cortices.
Cortex
,
36
,
301
322
.
Huijbers
,
W.
,
Pennartz
,
C. M. A.
,
Cabeza
,
R.
, &
Daselaar
,
S. M.
(
2011
).
The hippocampus is coupled with the default network during memory retrieval but not during memory encoding.
PLoS One
,
6
,
e17463
.
Israel
,
S. L.
,
Seibert
,
T. M.
,
Black
,
M. L.
, &
Brewer
,
J. B.
(
2010
).
Going their separate ways: Dissociation of hippocampal and dorsolateral prefrontal activation during episodic retrieval and post-retrieval processing.
Journal of Cognitive Neuroscience
,
22
,
513
525
.
King
,
J. A.
,
Burgess
,
N.
,
Hartley
,
T.
,
Vargha-Khadem
,
F.
, &
O'Keefe
,
J.
(
2002
).
Human hippocampus and viewpoint dependence in spatial memory.
Hippocampus
,
12
,
811
820
.
Kraskov
,
A.
,
Quiroga
,
R. Q.
,
Reddy
,
L.
,
Fried
,
I.
, &
Koch
,
C.
(
2007
).
Local field potentials and spikes in the human medial temporal image lobe are selective to image category.
Journal of Cognitive Neuroscience
,
19
,
479
492
.
Kravitz
,
D. J.
,
Saleem
,
K. S.
,
Baker
,
C. I.
, &
Mishkin
,
M.
(
2011
).
A new neural framework for visuospatial processing.
Nature Reviews Neuroscience
,
12
,
217
230
.
Lambrey
,
S.
,
Doeller
,
C.
,
Berthoz
,
A.
, &
Burgess
,
N.
(
2012
).
Imagining being somewhere else: Neural basis of changing perspective in space.
Cerebral Cortex
,
22
,
166
174
.
Lee
,
A. C. H.
,
Buckley
,
M. J.
,
Pegman
,
S. J.
,
Spiers
,
H.
,
Scahill
,
V. L.
,
Gaffan
,
D.
,
et al
(
2005
).
Specialization in the medial temporal lobe for processing of objects and scenes.
Hippocampus
,
15
,
782
797
.
Leontiev
,
O.
,
Dubowitz
,
D. J.
, &
Buxton
,
R. B.
(
2007
).
CBF/CMRO2 coupling measured with calibrated BOLD fMRI: Sources of bias.
Neuroimage
,
36
,
1110
1122
.
Logothetis
,
N. K.
,
Pauls
,
J.
,
Augath
,
M.
,
Trinath
,
T.
, &
Oeltermann
,
A.
(
2001
).
Neurophysiological investigation of the basis of the fMRI signal.
Nature
,
412
,
150
157
.
Logothetis
,
N. K.
, &
Wandell
,
B. A.
(
2004
).
Interpreting the BOLD signal.
Annual Review of Physiology
,
66
,
735
769
.
Maguire
,
E. A.
(
2001
).
The retrosplenial contribution to human navigation: A review of lesion and neuroimaging findings.
Scandinavian Journal of Psychology
,
42
,
225
238
.
Maguire
,
E. A.
,
Burgess
,
N.
,
Donnett
,
J. G.
,
Frackowiak
,
R. S. J.
,
Frith
,
C. D.
, &
O'Keefe
,
J.
(
1998
).
Knowing where and getting there: A human navigation network.
Science
,
280
,
921
924
.
Maguire
,
E. A.
,
Burke
,
T.
,
Phillips
,
J.
, &
Staunton
,
H.
(
1996
).
Topographical disorientation following unilateral temporal lobe lesions in humans.
Neuropsychologia
,
34
,
993
1001
.
Maguire
,
E. A.
,
Frackowiak
,
R. S. J.
, &
Frith
,
C. D.
(
1997
).
Recalling routes around London: Activation of the right hippocampus in taxi drivers.
Journal of Neuroscience
,
17
,
7103
7110
.
McKiernan
,
K. A.
,
Kaufman
,
J. N.
,
Kucera-Thompson
,
J.
, &
Binder
,
J. R.
(
2003
).
A parametric manipulation of factors affecting task-induced deactivation in functional neuroimaging.
Journal of Cognitive Neuroscience
,
15
,
394
408
.
Morris
,
R. G. M.
(
1981
).
Spatial localization does not require the presence of local cues.
Learning and Motivation
,
12
,
239
260
.
Nilsson
,
J.
,
Coventry
,
K.
, &
Ferrier
,
N.
(
2010
).
Distinguishing location memory based on allocentric and egocentric spatial systems
.
Presented at the Spatial Cognition Conference, Portland, Oregon
.
Ogawa
,
S.
,
Lee
,
T. M.
,
Kay
,
A. R.
, &
Tank
,
D. W.
(
1990
).
Brain magnetic-resonance-imaging with contrast dependent on blood oxygenation.
Proceedings of the National Academy of Sciences, U.S.A.
,
87
,
9868
9872
.
O'Keefe
,
J.
, &
Nadel
,
L.
(
1978
).
The hippocampus as a cognitive map.
Oxford, UK
:
Clarendon Press
.
Parslow
,
D. M.
,
Rose
,
D.
,
Brooks
,
B.
,
Fleminger
,
S.
,
Gray
,
J. A.
,
Giampietro
,
V.
,
et al
(
2004
).
Allocentric spatial memory activation of the hippocampal formation measured with fMRI.
Neuropsychology
,
18
,
450
461
.
Quiroga
,
R. Q.
,
Reddy
,
L.
,
Kreiman
,
G.
,
Koch
,
C.
, &
Fried
,
I.
(
2005
).
Invariant visual representation by single neurons in the human brain.
Nature
,
435
,
1102
1107
.
Reas
,
E. T.
,
Gimbel
,
S. I.
,
Hales
,
J. B.
, &
Brewer
,
J. B.
(
2011
).
Search-related suppression of hippocampus and default network activity during associative memory retrieval.
Frontiers in Human Neuroscience
,
5
,
112
.
Rekkas
,
P. V.
,
Westerveld
,
M.
,
Skudlarski
,
P.
,
Zumer
,
J.
,
Pugh
,
K.
,
Spencer
,
D. D.
,
et al
(
2005
).
Neural correlates of temporal-order judgments versus those of spatial-location: Deactivation of hippocampus may facilitate spatial performance.
Brain and Cognition
,
59
,
103
113
.
Restom
,
K.
,
Perthen
,
J. E.
, &
Liu
,
T. T.
(
2008
).
Calibrated fMRI in the medial temporal lobe during a memory-encoding task.
Neuroimage
,
40
,
1495
1502
.
Rodriguez
,
P. F.
(
2010
).
Human navigation that requires calculating heading vectors recruits parietal cortex in a virtual and visually sparse water maze task in fMRI.
Behavioral Neuroscience
,
124
,
532
540
.
Rossion
,
B.
, &
Pourtois
,
G.
(
2004
).
Revisiting Snodgrass and Vanderwart's object pictorial set: The role of surface detail in basic-level object recognition.
Perception
,
33
,
217
236
.
Save
,
E.
, &
Poucet
,
B.
(
2009
).
Role of the parietal cortex in long-term representation of spatial information in the rat.
Neurobiology of Learning and Memory
,
91
,
172
178
.
Schmidt
,
D.
,
Krause
,
B. J.
,
Weiss
,
P. H.
,
Fink
,
G. R.
,
Shah
,
N. J.
,
Amorim
,
M. A.
,
et al
(
2007
).
Visuospatial working memory and changes of the point of view in 3D space.
Neuroimage
,
36
,
955
968
.
Schridde
,
U.
,
Khubchandani
,
M.
,
Motelow
,
J. E.
,
Sanganahalli
,
B. G.
,
Hyder
,
F.
, &
Blumenfeld
,
H.
(
2008
).
Negative BOLD with large increases in neuronal activity.
Cerebral Cortex
,
18
,
1814
1827
.
Shipman
,
S. L.
, &
Astur
,
R. S.
(
2008
).
Factors affecting the hippocampal BOLD response during spatial memory.
Behavioural Brain Research
,
187
,
433
441
.
Shulman
,
G. L.
,
Fiez
,
J. A.
,
Corbetta
,
M.
,
Buckner
,
R. L.
,
Miezin
,
F. M.
,
Raichle
,
M. E.
,
et al
(
1997
).
Common blood flow changes across visual tasks: 2. Decreases in cerebral cortex.
Journal of Cognitive Neuroscience
,
9
,
648
663
.
Smith
,
M. L.
, &
Milner
,
B.
(
1981
).
The role of the right hippocampus in the recall of spatial location.
Neuropsychologia
,
19
,
781
793
.
Spiers
,
H. J.
, &
Maguire
,
E. A.
(
2006
).
Thoughts, behaviour, and brain dynamics during navigation in the real world.
Neuroimage
,
31
,
1826
1840
.
Vann
,
S. D.
,
Aggleton
,
J. P.
, &
Maguire
,
E. A.
(
2009
).
What does the retrosplenial cortex do?
Nature Reviews Neuroscience
,
10
,
792
802
.
Viard
,
A.
,
Doeller
,
C. F.
,
Hartley
,
T.
,
Bird
,
C. M.
, &
Burgess
,
N.
(
2011
).
Anterior hippocampus and goal-directed spatial decision making.
Journal of Neuroscience
,
31
,
4613
4621
.
Wade
,
A. R.
(
2002
).
The negative BOLD signal unmasked.
Neuron
,
36
,
993
995
.
Xu
,
J. A.
,
Evensmoen
,
H. R.
,
Lehn
,
H.
,
Pintzka
,
C. W. S.
, &
Haberg
,
A. K.
(
2010
).
Persistent posterior and transient anterior medial temporal lobe activity during navigation.
Neuroimage
,
52
,
1654
1666
.