In the present study, we examined the relationship between performance on a virtual navigation task and gray matter in the hippocampus in healthy older adults. With VBM, we showed that the spontaneous use of spatial memory strategies positively correlates with gray matter in the hippocampus. In contrast, the use of response strategies negatively correlates with gray matter in the hippocampus. When the peak value in the hippocampus, the area in the hippocampus which most highly correlated with spatial memory, was used as a seed voxel to correlate with gray matter in the whole brain, positive correlations were observed in regions that are anatomically and functionally linked to the hippocampus, such as the contralateral hippocampus, the right orbitofrontal cortex, bilateral amygdala, and bilateral parahippocampal cortex.
While some studies in the literature reported a significant relationship between hippocampal volume and spatial memory (Chen et al., 2010
; Head and Isom, 2010
), others have failed to demonstrate this effect (Driscoll et al., 2003
; Moffat et al., 2007
). One factor that may help shed light on this issue concerns the fact that none of these studies involved tasks that assessed spontaneous navigational strategies (when both spatial and response strategies can be used in the same task). Indeed, ours is the first structural MRI study to utilize a task in older adults that distinguishes between navigational strategies. This is a notable distinction because both human and non-human animal studies showed that with aging, there is a decreased use of spatial strategies and increased use of response strategies (Barnes et al., 1980
; Rapp et al., 1997
; Rodgers et al., 2010
; Etchamendy et al., 2011
). Therefore, older adults tested on the virtual Water Maze (e.g., Moffat et al., 2007
) may have performed well on that task despite the fact that they would have used a response strategy dependent on the caudate nucleus when tested on another task allowing for both strategies (Etchamendy and Bohbot, 2007
). If this is the case, an association between task performance and hippocampal volume would not necessarily be found. On the other hand, as mentioned above, some spatial memory tasks that did not measure spontaneous navigational strategies were sensitive to gray matter in the hippocampus (Chen et al., 2010
; Head and Isom, 2010
). Further research would be needed in order to shed light on this issue. For instance, some of these spatial tasks may have been sufficiently complex to detect a learning deficit in participants favoring response strategies, thereby inadvertently dissociating between spatial and response strategies. In support of this hypothesis, we previously reported a correlation between spatial strategies and wayfinding performance in healthy older adults (Etchamendy et al., 2012
). This would be consistent with our results that demonstrate that in fact only older adults using spatial strategies, and not those using response strategies, have more gray matter in the hippocampus.
The study by Head and Isom (2010
) reported a correlation between caudate nucleus volume and route learning. Route learning is a form of response learning that involves learning a sequence of movements in response to specific stimuli. However, we did not observe a negative correlation between performance and caudate nucleus gray matter with the current task, suggesting that it may not be sensitive enough to detect variability in the gray matter of the caudate nucleus of our population. Interestingly, Kennedy et al. (2009
) demonstrated a non-linear decline in hippocampus volume as a function of age, where the rate of hippocampus atrophy accelerates with age, while Raz et al. (2003
) showed that in the caudate nucleus, there is a linear decline in volume with aging. Therefore, within an older adult population, there may be more variability in the volume of the hippocampus than in the volume of the caudate nucleus, making it easier to detect morphological differences within the hippocampus. Recruiting participants in a wider age range may help increase sensitivity to morphological differences in the caudate nucleus. Indeed, the age range of the participants in the Head and Isom (2010
) study was 56–86 years, while in the current study it was 60–75 years. Therefore, a wider age range of participants may be needed to uncover the relationship between gray matter in the caudate nucleus and response strategies in our task.
We observed an association between the aging hippocampus and response strategies. However, the causal relationship cannot be determined in this study, i.e., the use of response strategies may be the consequence of decreased gray matter in the hippocampus or it may be the cause of a decrease gray matter in the hippocampus, i.e., “use it or loose it.” In other words, biological factors such as genes (Banner et al., 2011
) may have a negative impact on the hippocampus leading to an increase use in response strategies. Alternatively, environmental factors such as repetition (which leads to habit) (Iaria et al., 2003
), stress (Schwabe et al., 2008
), or reward (Del Balso et al., 2010
) could promote response strategies at the expense of spatial strategies associated with the hippocampus. It is most likely a combination of both. Importantly, the fact that we found an association between response strategies and the aging hippocampus opens new possibilities for spatial memory-based cognitive interventions toward healthy aging (Bohbot et al., 2011
). We have previously shown that strategy use can alter gray matter in the mouse hippocampus (Lerch et al., 2011
). In that study, mice trained to use spatial strategies showed increased gray matter in the hippocampus relative to mice trained on the response strategy. Preliminary findings in older adults replicated these findings, demonstrating increased gray matter in the hippocampus after the administration of spatial memory training (Fouquet et al., 2011
The relationship between navigational strategy and the hippocampus in young adults was previously investigated in Iaria et al. (2003
) and Bohbot et al. (2007
). Iaria et al. (2003
) found that in a task where both strategies can be used, 50% of young adults spontaneously used a spatial strategy while 50% used a response strategy. Furthermore, individuals who used a spatial strategy showed significant fMRI activity in the hippocampus, while those who used a response strategy showed significant fMRI activity in the caudate nucleus, relative to baseline. Bohbot et al. (2007
) found that the spontaneous use of spatial strategies was associated with increased gray matter in the hippocampus. The same technique of VBM was used to analyze the results in both Bohbot et al. (2007
) and the current study. However, a different navigation task was used to assess navigational strategy. In Bohbot et al. (2007
), the 4-on-8 virtual maze task was used while in the current study the CSDLT was used. Consistent results were obtained in the two studies.
Similar to Bohbot et al. (2007
), a number of brain areas anatomically linked to the hippocampus correlated with gray matter values extracted from our peak voxel in the hippocampus in older adults, such as the orbitofrontal cortex, the parahippocampal cortex, and the amygdala. It is possible that the co-activity of this network of regions during navigation leads to increased gray matter. It has been suggested that the orbitofrontal cortex in rodents is involved in spatial memory (Vafaei and Rashidy-Pour, 2004
) and learning reward expectancy in a spatial context (Young and Shapiro, 2011
). Indeed, orbitofrontal cortex neurons respond when a rodent enters an arm that is expected to contain a reward. The parahippocampal cortex is known to be involved in navigation and its functions involve scene processing and spatial learning (Bohbot et al., 1998
; Epstein, 2008
). In rodents, the amygdala is associated with the hippocampus in place preference learning when it is associated with reward (Gaskin and White, 2006
We also investigated the relationship between neuropsychological performance and gray matter in the hippocampus. Previous studies have found a correlation between the RAVLT and volume of the left hippocampus (Hackert et al., 2002
; Chen et al., 2010
), however we did not find this effect in our study. The inconsistency from the literature may arise from a cohort effect. The current sample population is composed of high functioning older adults with no detectable cognitive impairment. The RAVLT and RO may therefore not be sensitive enough to detect differences in gray matter in a more homogeneous population. These results suggest that spatial memory strategies are more sensitive to structural differences in the hippocampus than standard neuropsychological tests. In fact, a prospective study showed that patients with dementia had deficits in spatial cognition that preceded conversion to dementia by 3 years. In the same cohort, verbal memory, and working memory deficits were detected only 1 year before patients were diagnosed with dementia (Johnson et al., 2009
), further demonstrating the sensitivity of spatial memory.
In conclusion, our present findings are consistent with the literature reporting that in healthy older adults there is a positive correlation between spatial memory and gray matter in the hippocampus. Furthermore, our results showed that those who have lower hippocampal gray matter are actually using another form of navigational strategy, the response strategy. With whole brain analysis, we observed that individuals with more gray matter in the hippocampus had more gray matter in anatomically and functionally connected cortical areas, namely the contralateral hippocampus, the right orbitofrontal cortex, bilateral amygdala, and bilateral parahippocampal cortex. Future studies will be needed to examine whether an intervention method focusing on spatial memory training can increase gray matter in the hippocampus and anatomically connected areas. Promoting the use of spatial strategies can be a potential avenue for intervention methods against hippocampal atrophy (Fotuhi et al., 2012
). As lower hippocampal volume is a risk factor for cognitive decline as well as a number of disorders including Alzheimer's disease (Convit et al., 1997
; Rusinek et al., 2003
; Tapiola et al., 2008
), promoting the use of spatial memory can potentially have protective effects against cognitive deficits in normal aging as well as risks of degenerative disorders of the hippocampus.
Conflict of interest statement
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.