Verbal working and long-term episodic memory associations with white matter microstructure in normal aging investigated using tract-based spatial statistics

Acknowledgement: This work was funded by Research into Ageing, UK (227, H.S.M.; 259, T.R.B.). The authors report no conflict of interest.

In normal aging, cognitive decline has been linked to degradation of cerebral white matter attributable to vascular damage (Charlton, Barrick, Lawes, Markus, & Morris, 2010; Davis et al., 2009; Kennedy & Raz, 2009b; O’Sullivan et al., 2004; Perry et al., 2009; Wozniak & Lim, 2006). Indeed, loss of white matter microstructure has been interpreted as causing “disconnection,” in which dynamic interaction between key cortical and subcortical brain regions become less efficient, accounting for decline in particular types of cognitive functions (O’Sullivan et al., 2001; Charlton et al., 2008).

The development of diffusion tensor imaging (DTI), including quantitative measures such as mean diffusivity (MD) and fractional anisotropy (FA), has enabled more accurate detection of age-related loss of white matter microstructure (Chabriat et al., 1999; O’Sullivan et al., 2004). Prominent changes in white matter microstructure have been noted in periventricular regions, the central semiovale, corona radiata and pericallosal frontal and cortical regions as well as in association fibers (e.g., Abe et al., 2002; Bhagat & Beaulieu, 2004; Furutani, Harada, Minato, Morita, & Nishitani, 2005; O’Sullivan et al., 2001; Stadlbauer, Salomonowitz, Strunk, Hammen, & Ganslandt, 2008). Furthermore, specific regions of white matter damage have been associated with decline in particular cognitive abilities, for example, executive functioning (Gunning-Dixon & Raz, 2003), processing speed (Prins et al., 2005), and memory function (Charlton et al., 2006; Charlton, Barrick, Markus, & Morris, 2010).

White matter deterioration has been found to reduce working memory (WM) and long-term episodic memory (LTM), mnemonic functions hypothesized to be core deficits in normal aging, that undermine everyday cognition (Naveh-Benjamin, Moscovitch, & Roediger, 2002; Reuter-Lorenz & Sylvester, 2005). Although multiple memory systems have been hypothesized, it has also been acknowledged that these systems interact and therefore may rely on both unique and overlapping brain regions (Henson & Gagnepain, 2010; McDonald, Devan, & Hong, 2004). Results from lesion studies have found LTM to be particularly reliant on temporal lobe regions, with fMRI studies indicating frontal involvement in encoding and retrieval processes; in contrast, WM relies primarily on frontal and parietal interactions (Daselaar, Veltman, Rombouts, Raaijmakers, & Jonker, 2003; Ranganath, Johnson, & D’Esposito, 2003; Moscovitch et al., 2005). The integrity of connectivity between the associated gray matter regions has been investigated by exploring changes in white matter microstructure and how this relates to performance on different memory functions (Charlton et al., 2006; Charlton et al., 2010; Kennedy & Raz, 2009a; Persson et al., 2006; Zahr, Rohlfing, Pfefferbaum, & Sullivan, 2009).

Region of interest studies have identified white matter microstructure in frontal brain regions as being associated with multiple memory abilities. Specifically, areas of the prefrontal cortex, the centrum semiovale, and the genu of the corpus callosum are shown to correlate with both WM and LTM task performance among older adults (Bucur et al., 2008; Charlton et al., 2006; Charlton et al., 2010; Kennedy & Raz, 2009a; Persson et al., 2006). One study by Kennedy and Raz (2009a) conducted a direct comparison of WM and LTM in relation to white matter associations using small ROI to focus on predefined regions. The authors identified correlations between frontal white matter (prefrontal cortex, genu of the corpus callosum) and verbal and nonverbal WM tasks. Additional correlations with WM performance were identified with ROI in the parietal lobe. In contrast, associations with verbal LTM were in ROI in midbrain regions—in the anterior and posterior internal capsule—and the temporal lobe; no correlations were observed between LTM and prefrontal cortex ROI.

Few studies of aging have examined associations between memory function and microstructure along white matter tracts. One study has performed tractography to extract selected tracts, namely the genu and splenium of the corpus callosum, the cingulum, the uncinate and inferior longitudinal fascicule in a sample of 20 older adults (Mean age: 68.89 years; SD: 5.3 years), finding that microstructure in these tracts was correlated with WM and LTM (Davis et al., 2009). Visuospatial WM was associated with the integrity of the uncinate fasciculus; whereas visuospatial LTM was significantly associated with white matter microstructure in the cingulum and the inferior longitudinal fasciculus (Davis et al., 2009). No areas of white matter were significantly associated with both mnemonic functions. Given the known functional involvement of temporal and frontal regions in LTM (Daselaar et al., 2003; Kircher et al., 2008), it is rather surprising that the uncinate fasciculus—which connects these regions—was not associated with LTM abilities in the Davis et al. study. One may also expect fronto-parietal tracts to be associated with WM performance given their functional connectivity (Narayanan et al., 2005), however the superior longitudinal fasciculus was not measured and no significant correlations were observed with the cingulum.

Similar patterns of results have been observed in studies of clinical populations using tract-based spatial statistics (TBSS), a technique which provides three-dimensional analysis of white matter microstructure across the whole brain. For example, a study of young (age range 31–56 years) multiple sclerosis patients (n = 43) investigated white matter tracts associated with WM and LTM. Using TBSS analysis, both mnemonic abilities were associated with portions of the corpus callosum, the cingulum, inferior longitudinal fasciculus, and the parietal portion of the superior longitudinal fasciculus (Dineen et al., 2009). LTM was uniquely associated with the fornix. A separate TBSS study of young adults with schizophrenia also identified the superior longitudinal fasciculus (bilaterally) as being significantly associated with verbal WM (Karlsgodt et al., 2008).

In summary, the studies reviewed above suggest that types of memory function may be differentially associated with loss of white matter microstructure, depending on the importance of specific tracts to a particular neuronal network (Charlton et al., 2010; Davis et al., 2009; Kennedy & Raz, 2009a). DTI methods have begun to identify associations with cognitive performance and how these are affected by cognitive aging (Charlton et al., 2010; Zahr et al., 2009). To date, DTI studies have extracted specific tracts that are hypothesized a priori to be involved in memory function, which has the significant advantage of specificity, but an alternative strategy used here is to use analysis that makes no a priori selection and so potentially identifies tracts that might be overlooked by previous analyses. Accordingly, in this study we use TBSS, a technique that has been successful in other fields such as multiple sclerosis (Dineen et al., 2009), to explore white matter microstructure across the whole brain. Although studies have used TBSS to investigate age-related changes in white matter microstructure (Barrick, Charlton, Clark, & Markus, 2010; Burzynska et al., 2010; Damoiseaux et al., 2009), studies have not yet investigated associations with cognitive performance in aging.

In this study we explored the relationship between white matter microstructure and WM and LTM in a large sample of middle aged and older adults, using TBSS to explore these two memory functions. We hypothesize that WM performance will be associated with white matter microstructure in fronto-parietal white matter tracts such as the fronto-parietal fasciculus, cingulum, arcuate fasciculus, and superior longitudinal fasciculus. In contrast, we hypothesize that LTM will be associated with microstructure in temporo-frontal pathways such as the fornix and uncinate fasciculus. Frontal white matter will be important for both mnemonic abilities.

A population sample of 106 healthy adults (55 males, 51 females; age range 50 to 90 years; mean age = 69 years) were recruited via a local United Kingdom (U.K.) National Health Service family practice by random sampling. Because of the manner in which U.K. people are normally registered with a family doctor (general practitioner), this provides an approximate population sample. The sample was part of the GENIE study, as described elsewhere (Charlton et al., 2006; Charlton et al., 2010). Exclusion criteria included any previous psychiatric or neurological disorder including prior stroke; participants were required to have English as their first language and be suitable for a MRI scan; presence of diabetes or hypertension was not an exclusion criteria. One individual had diabetes (controlled by diet); 22 individuals had treated hypertension, and 35 individuals had high blood pressure at testing and were referred to their general practitioner. Given that cardiovascular risk factors are common in aging and are likely to reflect individual differences, these factors were not controlled for in the analysis. Of the 106 participants, 99 had DTI scans that were deemed good quality for analysis; 98 participants had complete cognitive and imaging data. Individuals who were excluded from this analysis ranged in age from 66–89 years old; the group mean for these individuals was significantly older than the sample as a whole, F(1, 105) = 7.56, p = .007 but they did not differ on any other demographics or vascular risk factors. Participant demographics are shown in Table 1.

The participants completed tests selected as robust measures of verbal WM and verbal LTM (sensitive to amnesic memory syndromes; Squire, 1992). We use the term long-term episodic memory (LTM; Tulving, 1983) to distinguish it from episodic WM (Baddeley, Eysenck, & Anderson, 2009). To measure WM we used the Digit Span Backwards and Letter-Number Sequencing subtests from the Wechsler Memory Scale III (WMS III; Wechsler, Wycherley, Benjamin, Callanan, Lavender, Crawford, Mockler, 1998). LTM was measured using two subtests from the WMS-III: Logical Memory and Verbal Paired Associates for immediate recall conditions. Raw scores from each subtest were transformed into z-scores (using means and standard deviations for the whole group). Separate composite scores for WM and LTM were then obtained by averaging the z-scores across the mnemonic domains. Cronbach’s alphas were calculated to assess how well each variable measured a latent construct; values were in the moderate range (WM, α = .590; LTM, α = .654).

Diffusion-weighted images were acquired using a diffusion-sensitized spin-echo planar imaging (EPI) sequence on a 1.5T General Electric Signa MRI system (running software version 8x) equipped with magnetic field gradients of up to 22 mTm−1 (GE Electric, Milwaukee, WI) and a proprietary head coil. Fifty interleaved slices, each 2.8 mm thick were acquired with a field of view of 240 mm × 240 mm with a 96 × 96 matrix, providing contiguous whole brain coverage (TE 80ms; TR 7s). After an acquisition without diffusion sensitization (b = 0 s mm−2), images were acquired with diffusion gradients applied (b = 1000 s mm−2) in 12 unique directions to eliminate diffusion-imaging gradient cross terms (Barrick & Clark, 2004). The acquisitions were repeated four times to improve signal to noise ratio.

Images were realigned to remove eddy current distortions (Woods, Grafton, Holmes, Cherry, & Mazziota, 1998; Woods, Grafton, Holmes, Cherry, & Mazziotta, 1998). Diffusion tensor elements were computed at each voxel as described by Basser et al. (1994) and diagonalized to determine eigenvalues and eigenvectors in FSL4 from which FA maps were computed (Pierpaoli & Basser, 1996). Images were transformed to standard space using TBSS software (Smith et al., 2006). To generate a study-specific template in standard space each FA map was coregistered using a nonlinear transformation to every other FA map to determine the most representative FA map in the dataset. This target image was then normalized by 12-parameter affine transformation to the Montreal Neurological Institute space standard (MNI152) space. After this step, every FA map was transformed into the MNI152 space by combining the nonlinear transformation to the target FA map with the affine transformation from that target to MNI152 space. Each normalized image had isotropic 1 mm3 voxels and the skull and dura were removed (Brain extraction tool, BET; http://www.fmrib.ox.ac.uk/fsl/; Smith et al., 2002). All normalized images were visually inspected to assure no gross errors had occurred during the normalization procedure.

As part of the TBSS software the normalized individual FA maps were averaged to produce a group averaged FA map. This image was used to generate a group-wise (one pixel wide medial trajectory) skeleton of white matter tracts. The individual FA maps were then projected onto the group-wise skeleton to account for residual misalignment among individual white matter tracts. This is performed because FA values vary rapidly perpendicular to the tract direction but slowly along the direction of the tract. In particular, the maximum FA value was obtained perpendicular to each voxel of the skeleton surface and assigned to the appropriate skeleton voxel. Normalized FA maps and their corresponding skeletons are shown for normalized axial slices of a single subject in Figures 1a and 1b. For each individual, the FA voxel projections defined above were applied to project MD values to the group-wise white matter skeleton.

Pearson’s correlations were used to explore the cognitive data. Associations between age and both verbal WM and LTM and between the two types of memory were examined.

Analysis of the relationship between FA and MD 3D TBSS maps and both WM and LTM was performed using randomize software (http://www.fmrib.ox.ac.uk/fsl/). Statistical analyses were performed using linear correlation designs with WM and LTM in turn as the dependent variable. The number of permutations in the analysis was set to 5000, and correction for multiple comparisons was achieved using threshold free cluster enhancement (TFCE; Smith & Nichols, 2009). TFCE is a technique that does not need arbitrary prestatistical smoothing of images and does not depend on an initial cluster-forming threshold such as t-statistic thresholding. However, TFCE requires several parameters to be set. Specifically these relate to the cluster extent (E) and peak height (H) of the statistic at a given voxel. As suggested by Smith and Nichols (2009), these were set to E = 0.5 and H = 2.0 to provide a statistic that is sensitive to all levels of the signal. After TFCE voxel clusters were deemed significant at p < .05. Overlap between significant clusters for WM and LTM analyses were examined. A binary mask was created for each set of results and the proportion of overlap between results was calculated. Binary files were used to visualize areas of overlap between the results. Because our interest is age-related associations, the focus of this study is the TBSS results described above; however, to satisfy readers interested in brain-cognition associations independent of age, TBSS linear correlations were repeated including age as covariate of interest. To examine regions where microstructure independently associates with one function, TBSS partial correlations were performed for WM covarying for LTM, and for LTM covarying for WM.

For associations between age and memory see Table 1 and Figure 2. Age correlated significantly with both WM (r = −.321, p = .001) and LTM (r = −.456, p < .001). A significant correlation between the two types of memory was also observed (r = .416, p < .001).

FA

The TBSS analysis revealed significant positive correlations between FA and WM and LTM (see Figure 3a and b for t-statistics). WM was associated with left parietal lobe white matter microstructure, likely to include the anterior segment of the arcuate fasciculus and white matter in the inferior parietal lobule (see Figure 3a and red areas in Figure 3c). 2.5% of the total skeleton was associated with both WM and LTM; an additional 2.6% of the skeleton was associated with WM performance and 4.5% with LTM. Significant voxels associated with LTM were located in bilateral temporal regions, specifically the uncinate fasciculus, the fronto-occipital fasciculus and the anterior limb of the internal capsule, see Figure 3b and blue areas in Figure 3c. Areas of white matter common to both mnemonic functions (from the overlapping binary masks, colored green in Figure 3c) were in bilateral frontal regions including the genu of the corpus callosum and the forceps minor, left dorsolateral prefrontal white matter, and left frontal callosal fibers. For ease of viewing, areas where FA is associated with one or both types of mnemonic abilities are displayed as overlapping images in Figure 3c. No negative correlations between memory performance and FA values were observed.

The above analysis was repeated including age as a covariate of interest. When age is included in this analysis, no white matter voxels are associated with either WM or LTM.

When partial correlations were performed for one function while controlling for the other, no unique regions of white matter were found.

MD

Analysis showed that only 0.56% of the skeleton white matter was associated with WM performance and only 0.26% for LTM. For WM these voxels were within white matter in the left frontal and parietal lobes, likely to include the corona radiata, corticospinal tracts, and superior longitudinal fasciculus as well as the genu and the splenium of the corpus callosum. For LTM, significant voxels were found bilaterally in the uncinate and in the anterior limb of the internal capsule. Overlap between FA and MD results was minimal and represented 0.02% of the skeleton for WM and 0.13% of the skeleton for LTM. Overlapping voxels were located in the left forceps minor, the center of the genu of the corpus callosum, and the left superior longitudinal fasciculus. See Figure 4 for representation of MD results.

Using TBSS, this study demonstrated associations between integrity of white matter tracts and two verbal mnemonic functions in aging. FA was associated with performance on both WM and LTM tasks predominately in the inferior frontal lobe bilaterally. These common areas included the genu of the corpus callosum and the forceps minor (the frontal radiations of the corpus callosum) bilaterally, as well as left dorsolateral and frontal callosal white matter. FA correlated with WM only in the left parietal white matter including the arcuate fasciculus, whereas bilateral temporal FA including the uncinate fasciculus was associated with LTM. MD values demonstrated significant correlations in the splenium of the corpus callosum with WM but for LTM were less widespread than FA results. Despite this, both MD and FA demonstrated the same leftward-WM and bilateral-LTM patterns. Given that the memory measures included only verbal tasks, this may be related to the specificity of left parietal white matter for verbal WM, whereas the left dorsolateral prefrontal white matter is important for both verbal WM and LTM. It is necessary to note that if regressions were performed for WM covarying for LTM (and vice versa), no regions reached significance. This would be expected because all cognitive functions share variance and in this instance correlate significantly (r = .416) and therefore such an analysis may be too stringent to expect significant results. In comparison with studies that have assessed white matter microstructure with WM and LTM, our results show consistencies and divergences.

The areas of white matter in the left parietal lobe where FA was associated with WM included the anterior segment of the arcuate fasciculus (Catani & Mesulam, 2008; Schmahmann et al., 2007) and the inferior parietal lobule; these have previously been described as important for WM by our group and others (Catani, Howard, Pajevic, & Jones, 2002; Charlton et al., 2010). The arcuate fasciculus connects frontal and parietal regions (but also temporal and parietal cortices), which are critical for verbal WM function (Mencl et al., 2000; Owen, McMillan, Laird, & Bullmore, 2005), and damage to these fascicule has been associated with reduced WM performance in normal aging (Charlton et al., 2010) and multiple sclerosis (Audoin et al., 2007; Dineen et al., 2009). Furthermore, a TBSS study of patients with multiple sclerosis identified parietal regions including both the arcuate and the superior longitudinal fascicule as being associated with WM performance (Dineen et al., 2009). However, these results differ substantially from a study of extracted tracts, where visuospatial WM was associated with the integrity of the uncinate fasciculus (Davis et al., 2009)—a tract associated with LTM in our study.

In the current analysis only LTM was associated with bilateral temporal FA in the anterior limb of the internal capsule, the uncinate, and the fronto-occipital fascicule. Few studies have investigated associations between LTM and white matter tracts. A recent study identified associations between implicit sequence learning and FA values in tracts passing between dorsolateral prefrontal cortex and both caudate in left hemisphere and the hippocampus in the right hemisphere (Bennett, Madden, Vaidya, Howard, & Howard, in press). Unfortunately, other fascicules were not measured in this study. In contrast to the current predominantly temporo-frontal FA associations with LTM, Davis et al., (2009) study of extracted tracts demonstrated significant associations between visuospatial LTM and microstructure of the cingulum and the inferior longitudinal fasciculus (Davis et al., 2009). Previous studies support the importance of a network of white matter pathways sustaining LTM. Studies using WMH measurements to reflect white matter damage have demonstrated that regardless of location of damage, presence of WMH correlates with reduced LTM performance (Gunning-Dixon & Raz, 2000; Van Petten et al., 2004). Furthermore, age-related changes in functional connectivity between the prefrontal cortex and the medial temporal lobe, thought to rely on the integrity of the uncinate fasciculus, have been demonstrated (Grady, McIntosh, & Craik, 2003). It is also worth noting that this study did not identify correlations between FA in the fornix and LTM, in contrast to some recent studies (Metzler-Baddeley, Jones, Belaroussi, Aggleton, & O’Sullivan, 2011; Zahr et al., 2009). The fornix is especially difficult to identify using TBSS because of partial volume effects; issues related to TBSS limitations are discussed further below.

In this study we also identified regions where FA was associated with both WM and LTM, in bilateral frontal white matter including the genu of the corpus callosum. Some studies have not found this pattern (Davis et al., 2009; Kennedy & Raz, 2009a), whereas others have demonstrated associations between integrity of frontal white matter and mnemonic abilities (Persson et al., 2006; Zahr et al., 2009). In one study where white matter tracts were extracted, the integrity of the genu of the corpus callosum was associated with WM performance (Zahr et al., 2009). Moreover, in a longitudinal study, lower FA values in the genu were identified in individuals whose LTM performance had declined over the preceding decade (Persson et al., 2006). fMRI studies have also demonstrated that bilateral frontal activity is associated with both episodic and WM performance (Cabeza, Anderson, Locantore, & McIntosh, 2002; Grady, McIntosh, & Craik, 2005), and coordination of this activity will likely rely on white matter connections through the genu of the corpus callosum. To what extent divergent results are related to methodological differences or to task differences in visuospatial versus verbal mnemonic networks will require further investigation.

The TBSS technique did not identify large regions where mnemonic function correlated significantly with MD. Although our group and others have reported significant associations between MD and age using TBSS (Barrick et al., 2010; Bendlin et al., 2010), few studies have applied MD-TBSS measures to cognitive analyses. Using regions of interest or voxel based analyses DTI studies have identified significant correlations between mnemonic function and MD (Charlton et al., 2006; Charlton et al., 2010; Kantarci et al., 2011). Despite this, few studies report significant correlations between MD and mnemonic function using TBSS or tractography (Bosch et al., 2012; Metzler-Baddeley et al., 2011). In a tractography investigation of white matter associated with episodic memory performance, Metzler-Baddeley et al. identified significant correlations in the fornix with FA but not with MD; no significant correlations were observed in the uncinate or parahippocampal portion of the cingulate (Metzler-Baddeley et al., 2011). Similarly in a TBSS analysis of older adults, Bosch et al. (2012) found no significant correlations between MD and memory function.

For the WM and LTM tasks, both involving verbal stimuli, there are notable differences in hemispheric laterality. For WM measures we observed associations with white matter in the left, typically language dominant, hemisphere (87/98 right-handed participants). In contrast, for LTM, bilateral white matter microstructure was associated with performance. This may reflect differences between the associative networks important for LTM compared with the arrangement of cognitive operations that support WM. For example, functional neuroimaging results have proposed that LTM relies on both hemispheres for encoding (left prefrontal cortex) and retrieval (right prefrontal cortex) processes, regardless of the modality of the task (Habib, Nyberg, & Tulving, 2003). In addition, theories relating to episodic memory stress the associative links with multiple cortical regions and the core memory structures (Moscovitch et al., 2005; Nadel, Samsonovich, Ryan, Moscovitch, 2000; Nadel & Moscovitch, 1997; McClelland, McNaughton, O’Reilly, 1995; Squire, Cohen, & Nadel, 1984). The results of the current study mirror these notions, LTM being based on distributed associative links to bilateral cortical regions. For WM modality specific effects are more evident, with verbal WM seeming to rely on white matter microstructure in specific left hemisphere structures. This finding illustrates how studies linking variations in white matter tract integrity attributable to pathophysiological changes to specific mnemonic functions can provide corroborative evidence for differently organized functional architecture of the brain.

When age was included as a covariate in the TBSS analysis, no significant associations were observed between white matter microstructure and either WM or LTM. The loss of associations when age was included as a covariate suggests that age is a critical mediator in the observed correlations. These results, not independent of age-effects, differ from previous analyses on this sample. However they are in keeping with the aging literature as a whole, where age, white matter microstructure, and cognitive functions share a substantial proportion of variance (Madden et al., 2012). Such disparities may reflect the impact of methodological differences and stringent multiple comparison correction applied within TBSS analysis, particularly when controlling for robust covariates.

Another issue that has been raised by recent studies is whether integrity of white matter microstructure reflects common whole-brain, possibly lifelong individual differences or differences in specific tracts affected by environmental and genetic factors (Lövdén et al., 2013; Penke et al., 2010). We suggest that microstructure both across the whole brain and within individual tract will be associated with cognitive function but that different mechanisms may be involved. Previous studies have demonstrated strong correlations between childhood intelligence and late-life white matter, suggesting that individual difference in white matter may be fairly stable across the life span and have a lifelong effect on cognition (Deary et al., 2006). In addition, certain white matter tracts such as those passing through the centrum semiovale are more prone to age-related damage, resulting from occlusion of small perforating arterioles as in cerebral small vessel disease (Pantoni, 2010). Therefore one would expect certain tract to be more impacted by age-related vascular risk factors, at least initially, although over time downstream effects may cause damage in more remote brain regions. In normal aging, white matter damage is less severe than in disconnection syndromes and is likely to result in a disruption in network efficiency rather than a total loss of communication between regions. It is also important to note that all of the white matter pathways described in this study may also support other cognitive abilities that were not measured in this analysis. The pattern of gray matter regions required for any task and their interactions may influence the extent to which damage to any single pathway will disrupt function.

Tract-based spatial statistics is not without limitations. First, it does not provide an assessment of gray matter diffusion changes with aging. This is because the skeleton on which statistical analysis is performed provides a map of white matter structural integrity in large white matter pathways and does not include cortical or deep gray structures. Second, it is possible in TBSS analyses that FA may be reduced to such an extent in lesioned tissue that potential areas of interest may be excluded from statistical analysis due to thresholding of the mean FA skeleton (at FA >0.2). In this study, we did not experience this problem, and the skeleton included all large white matter structures. Third, because regions of tissue damage may be located in different anatomical positions in different individuals, it may be that the 1 voxel wide white matter skeleton does not pass through lesion regions in all subjects at the same spatial locations. However, TBSS reduces the effect of lesion location by mapping the maximum local FA perpendicular to the skeleton to individual skeleton voxels. FA values on the skeleton then represent the greatest proximal white matter structural integrity through lesioned regions rather than the greatest effect of the lesion on local structural integrity.

The techniques used in this study allow investigation of changes in structural integrity of white matter across the whole brain and currently provide the most reliable spatial normalization of diffusion data into standard space. A subject-specific template image was determined before image normalization to standard space, a step that provides greater normalization accuracy for datasets that include diffusion images of younger or older subjects. However TBSS analysis may be further improved by incorporating fiber crossing information (Jbabdi, Behrens, & Smith, 2010). Jbabdi et al. (2010) consider the possibility of presence of two fiber populations at each image voxel and determine relative structural integrities along their principal directions. Although this technique potentially provides more accurate coregistration of white matter structures within the white matter skeleton and enables improved interpretation of significant findings, a higher angular resolution diffusion-weighted image acquisition than the 12-direction diffusion-weighted images acquired in this study is required. In our study the acquired number of directions and image resolution (2.5 × 2.5 × 2.8 mm3) are limited compared with current state of the art acquisition, attributable to data collection beginning in 2002. Furthermore, methods that attempt to control for free water partial volume effects within image voxels may increase the power of DTI analyses; however, these techniques are not yet readily available, and their application to clinically acquired datasets is in the early stages (Metzler-Baddeley, O’Sullivan, Bells, Pasternak, & Jones, 2012; Sasson, Doniger, Pasternak, & Assaf, 2010). Because a correction for atrophy was not applied to this analysis, these results should be interpreted with caution, and further studies removing this confound are needed to confirm the overall differential pattern. An additional limitation related to the data is that the sample described here includes only middle-aged and older adults; a similar analysis performed on a sample of young adults from a range of abilities may help to determine whether these white matter pathways are associated with memory systems only after middle age or across the life span.

In summary, we used TBSS to identify white matter tracts associated with verbal mnemonic functions in normal aging. Damage to white matter pathways may either have a widespread effect on multiple functions by causing disruption to multiple distributed networks or effect a specific function through damage limited to a restricted network. Higher-order cognitive functions such as WM and LTM that rely on widespread cortico-cortical and cortico-subcortical networks may be particularly susceptible to white matter damage leading to disruption of networks and cortical disconnection.

Abe, O., Aoki, S., Hayashi, N., Yamada, H., Kunimatsu, A., Mori, H.. . .Ohtomo, K. (2002). Normal aging in the central nervous system: Quantitative MR diffusion-tensor analysis. Neurobiology of Aging, 23, 433–441. doi:10.1016/S0197-4580(01)00318-9

Audoin, B., Guye, M., Reuter, F., Au Duong, M. V., Confort-Gouny, S., Malikova, I.. . .Ranjeva, J.-P. (2007). Structure of WM bundles constituting the working memory system in early multiple sclerosis: A quantitative DTI tractography study. NeuroImage, 36, 1324–1330. doi:10.1016/j.neuroimage.2007.04.038

Baddeley, A., Eysenck, M. W., & Anderson, M. C. (2009). Memory. Hove, UK: Psychology Press Ltd.

Barrick, T. R., Charlton, R. A., Clark, C. A., & Markus, H. S. (2010). White matter structural decline in normal ageing; a prospective longitudinal study using tract based spatial statistics. NeuroImage, 51, 565–577. doi:10.1016/j.neuroimage.2010.02.033

Barrick, T. R., & Clark, C. A. (2004). Singularities in diffusion tensor fields and their relevance in white matter fiber tractography. Neuroimage, 22, 481–491. doi:10.1016/j.neuroimage.2004.02.001

Basser, P. J., Mattiello, J., & LeBihan, D. (1994). Estimation of the effective self-diffusion tensor from the NMR spin echo. Journal of Magnetic Resonance Series B, 103, 247–254. doi:10.1006/jmrb.1994.1037

Bendlin, B. B., Fitzgerald, M. E., Ries, M. L., Xu, G., Kastman, E. K., Thiel, B. W.. . .Johnson, S. C. (2010). White matter in aging and cognition: A cross-sectional study of microstructure in adults aged eighteen to eighty-three. Developmental Neuropsychology, 35, 257–277. doi:10.1080/87565641003696775

Bennett, I. J., Madden, D. J., Vaidya, C. J., Howard, J., & Howard, D. V. (in press). White matter integrity correlates of implicit sequence learning in healthy aging. Neurobiology of Aging.

Bhagat, Y. A., & Beaulieu, C. (2004). Diffusion anisotropy in subcortical white matter and cortical gray matter: Changes with aging and the role of CSF-suppression. Journal of Magnetic Resonance Imaging, 20, 216–227. doi:10.1002/jmri.20102

Bosch, B., Arenaza-Urquijo, E. M., Rami, L., Sala-Llonch, R., Junque, C., Sole-Padulles, C.. . .Bartres-Faz, D. (2012). Multiple DTI index analysis in normal aging, amnestic MCI and AD. Relationship with neuropsychological performance. Neurobiology of Aging, 33, 61–74. doi:10.1016/j.neurobiolaging.2010.02.004

Bucur, B., Madden, D. J., Spaniol, J., Provenzale, J. M., Cabeza, R., White, L. E., & Huettel, S. A. (2008). Age-related slowing of memory retrieval: Contributions of perceptual speed and cerebral white matter integrity. Neurobiology of Aging, 29, 1070–1079. doi:10.1016/j.neurobiolaging.2007.02.008

Burzynska, A. Z., Preuschhof, C., Backman, L., Nyberg, L., Li, S. C., Lindenberger, U., & Heekeren, H. R. (2010). Age-related differences in white matter microstructure: Region-specific patterns of diffusivity. NeuroImage, 49, 2104–2112. doi:10.1016/j.neuroimage.2009.09.041

Cabeza, R., Anderson, N. D., Locantore, J. K., & McIntosh, A. R. (2002). Aging Gracefully: Compensatory Brain Activity in High-Performing Older Adults. Neuroimage, 17, 1394–1402. doi:10.1006/nimg.2002.1280

Catani, M., Howard, R. J., Pajevic, S., & Jones, D. K. (2002). Virtual in vivo interactive dissection of white matter fasciculi in the human brain. Neuroimage, 17, 77–94. doi:10.1006/nimg.2002.1136

Catani, M., & Mesulam, M. (2008). The arcuate fasciculus and the disconnection theme in language and aphasia: History and current state. Cortex: A Journal Devoted to the Study of the Nervous System and Behavior, 44, 953–961. doi:10.1016/j.cortex.2008.04.002

Chabriat, H., Pappata, S., Poupon, C., Clark, C. A., Vahedi, K., Poupon, F.. . .Bousser, M.-G. (1999). Clinical severity in CADASIL related to ultrastructural damage in white matter: In Vivo study with diffusion tensor MRI. Stroke, 30, 2637–2643. doi:10.1161/01.STR.30.12.2637

Charlton, R. A., Barrick, T. R., Lawes, I. N. C., Markus, H. S., & Morris, R. G. (2010). White matter pathways associated with working memory in normal aging. Cortex: A Journal Devoted to the Study of the Nervous System and Behavior, 46, 474–489. doi:10.1016/j.cortex.2009.07.005

Charlton, R. A., Barrick, T. R., Markus, H. S., & Morris, R. G. (2010). The relationship between episodic long-term memory and white matter integrity in normal aging. Neuropsychologia, 48, 114–122. doi:10.1016/j.neuropsychologia.2009.08.018

Charlton, R. A., Barrick, T. R., McIntyre, D. J. O., Shen, Y., O’Sullivan, M., Howe, F. A.. . .Markus, H. S. (2006). White matter damage on diffusion tensor imaging correlates with age related cognitive decline. Neurology, 66, 217–222. doi:10.1212/01.wnl.0000194256.15247.83

Charlton, R. A., Landau, S., Schiavone, F., Barrick, T. R., Clark, C. A., Markus, H. S., & Morris, R. G. (2008). A structural equation modeling investigation of age related variance in executive function and DTI measured white matter damage. Neurobiology of Aging, 29, 1547–1555. doi:10.1016/j.neurobiolaging.2007.03.017

Damoiseaux, J. S., Smith, S. M., Witter, M. P., Sanz-Arigita, E. J., Barkhof, F., Scheltens, P.. . .Rombouts, S. A. R. B. (2009). White matter tract integrity in aging and Alzheimer’s disease. Human Brain Mapping, 30, 1051–1059. doi:10.1002/hbm.20563

Daselaar, S. M., Veltman, D. J., Rombouts, S. A. R. B., Raaijmakers, J. G. W., & Jonker, C. (2003). Neuroanatomical correlates of episodic encoding and retrieval in young and elderly subjects. Brain: A Journal of Neurology, 126, 43–56. doi:10.1093/brain/awg005

Davis, S. W., Dennis, N. A., Buchler, N. G., White, L. E., Madden, D. J., & Cabeza, R. (2009). Assessing the effects of age on long white matter tracts using diffusion tensor tractography. NeuroImage, 46, 530–541. doi:10.1016/j.neuroimage.2009.01.068

Deary, I. J. P., Bastin, M. E. D., Pattie, A. B., Clayden, J. D. M., Whalley, L. J. M., Starr, J. M. M., & Wardlaw, J. M. (2006). White matter integrity and cognition in childhood and old age. Neurology, 66, 505–512. doi:10.1212/01.wnl.0000199954.81900.e2

Dineen, R. A., Vilisaar, J., Hlinka, J., Bradshaw, C. M., Morgan, P. S., Constantinescu, C. S., & Auer, D. P. (2009). Disconnection as a mechanism for cognitive dysfunction in multiple sclerosis. Brain: A Journal of Neurology, 132, 239–249. doi:10.1093/brain/awn275

Furutani, K., Harada, M., Minato, M., Morita, N., & Nishitani, H. (2005). Regional changes of fractional anisotropy with normal aging using Statistical Parametric Mapping (SPM). Journal of Medical Investigation, 52, 186–190. doi:10.2152/jmi.52.186

Grady, C. L., McIntosh, A. R., & Craik, F. I. (2003). Age-related differences in the functional connectivity of the hippocampus during memory encoding. Hippocampus, 13, 572–586. doi:10.1002/hipo.10114

Grady, C. L., McIntosh, A. R., & Craik, F. I. M. (2005). Task-related activity in prefrontal cortex and its relation to recognition memory performance in young and old adults. Neuropsychologia, 43, 1466–1481. doi:10.1016/j.neuropsychologia.2004.12.016

Gunning-Dixon, F. M., & Raz, N. (2000). The cognitive correlates of white matter abnormalities in normal aging: A quantitative review. Neuropsychology, 14, 224–232. doi:10.1037/0894-4105.14.2.224

Gunning-Dixon, F. M., & Raz, N. (2003). Neuroanatomical correlates of selected executive functions in middle-aged and older adults: A prospective MRI study. Neuropsychologia, 41, 1929–1941. doi:10.1016/S0028-3932(03)00129-5

Habib, R., Nyberg, L., & Tulving, E. (2003). Hemispheric asymmetries of memory: The HERA model revisited. Trends in Cognitive Sciences, 7, 241–245. doi:10.1016/S1364-6613(03)00110-4

Henson, R. N., & Gagnepain, P. (2010). Predictive, interactive multiple memory systems. Hippocampus, 20, 1315–1326. doi:10.1002/hipo.20857

Jbabdi, S., Behrens, T. E., & Smith, S. M. (2010). Crossing fibres in tract-based spatial statistics. NeuroImage, 49, 249–256. doi:10.1016/j.neuroimage.2009.08.039

Kantarci, K., Senjem, M. L., Avula, R., Zhang, B., Samikoglu, A. R., Weigand, S. D.. . .Jack, C. R.Jr. (2011). Diffusion tensor imaging and cognitive function in older adults with no dementia. Neurology, 77, 26–34. doi:10.1212/WNL.0b013e31822313dc

Karlsgodt, K. H., van Erp, T. G. M., Poldrack, R. A., Bearden, C. E., Nuechterlein, K. H., & Cannon, T. D. (2008). Diffusion tensor imaging of the superior longitudinal fasciculus and working memory in recent-onset schizophrenia. Biological Psychiatry, 63, 512–518. doi:10.1016/j.biopsych.2007.06.017

Kennedy, K. M., & Raz, N. (2009a). Aging white matter and cognition: Differential effects of regional variations in diffusion properties on memory, executive functions, and speed. Neuropsychologia, 47, 916–927. doi:10.1016/j.neuropsychologia.2009.01.001

Kennedy, K. M., & Raz, N. (2009b). Pattern of normal age-related regional differences in white matter microstructure is modified by vascular risk. Brain Research, 1297, 41–56. doi:10.1016/j.brainres.2009.08.058

Kircher, T., Weis, S., Leube, D., Freymann, K., Erb, M., Jessen, F.. . .Krach, S. (2008). Anterior hippocampus orchestrates successful encoding and retrieval of non-relational memory: An event-related fMRI study. European Archives of Psychiatry and Clinical Neuroscience, 258, 363–372. doi:10.1007/s00406-008-0805-z

Lövdén, M., Laukka, E. J., Rieckmann, A., Kalpouzos, G., Li, T. Q., Jonsson, T., . . .Bäckman, L. (2013). The dimensionality of between-person differences in white matter microstructure in old age. Human Brain Mapping, 34, 1386–1398. doi:10.1002/hbm.21518

Madden, D. J., Bennett, I. J., Burzynska, A., Potter, G. G., Chen, N. K., & Song, A. W. (2012). Diffusion tensor imaging of cerebral white matter integrity in cognitive aging. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease, 1822, 386–400. doi:10.1016/j.bbadis.2011.08.003

McClelland, J. L., McNaughton, B. L., & O’Reilly, R. C. (1995). Why there are complementary learning systems in the hippocampus and neocortex: insights from the successes and failures of connectionist models of learning and memory. Psychological Review, 102, 419–457. doi:10.1037/0033-295X.102.3.419

McDonald, R. J., Devan, B. D., & Hong, N. S. (2004). Multiple memory systems: The power of interactions. Neurobiology of Learning and Memory, 82, 333–346. doi:10.1016/j.nlm.2004.05.009

Mencl, W. E., Pugh, K. R., Shaywitz, S. E., Shaywitz, B. A., Fulbright, R. K., Constable, R. T.. . .Gore, J. C. (2000). Network analysis of brain activations in working memory: Behavior and age relationships. Microscopy Research & Technique, 51, 64–74. doi:10.1002/1097-0029(20001001)51:1<64::AID-JEMT7>3.0.CO;2-D

Metzler-Baddeley, C., Jones, D. K., Belaroussi, B., Aggleton, J. P., & O’Sullivan, M. J. (2011). Frontotemporal connections in episodic memory and aging: A diffusion MRI tractography study. The Journal of Neuroscience, 31, 13236–13245. doi:10.1523/JNEUROSCI.2317-11.2011

Metzler-Baddeley, C., O’Sullivan, M. J., Bells, S., Pasternak, O., & Jones, D. K. (2012). How and how not to correct for CSF-contamination in diffusion MRI. NeuroImage, 59, 1394–1403. doi:10.1016/j.neuroimage.2011.08.043

Moscovitch, M., Rosenbaum, R. S., Gilboa, A., Addis, D. R., Westmacott, R., Grady, C.. . .Nadel, L. (2005). Functional neuroanatomy of remote episodic, semantic and spatial memory: A unified account based on multiple trace theory. Journal of Anatomy, 207, 35–66. doi:10.1111/j.1469-7580.2005.00421.x

Nadel, L., & Moscovitch, M. (1997). Memory consolidation, retrograde amnesia and the hippocampal complex. Current Opinion in Neurobiology, 7, 217–227. doi:10.1016/S0959-4388(97)80010-4

Nadel, L., Samsonovich, A., Ryan, L., & Moscovitch, M. (2000). Multiple trace theory of human memory: Computational, neuroimaging, and neuropsychological results. Hippocampus, 10, 352–368. doi:10.1002/1098-1063(2000)10:4<352::AID-HIPO2>3.0.CO;2-D

Narayanan, N. S., Prabhakaran, V., Bunge, S. A., Christoff, K., Fine, E. M., & Gabrieli, J. D. (2005). The role of the prefrontal cortex in the maintenance of verbal working memory: An event-related fMRI analysis. Neuropsychology, 19, 223–232. doi:10.1037/0894-4105.19.2.223

Naveh-Benjamin, M., Moscovitch, M., & Roediger, H. L. Perspectives on human memory and cognitive aging: Essays in honour of Fergus I. M. Craik (2002). Philadelphia, PA: Psychology Press.

O’Sullivan, M., Jones, D. K., Summers, P. E., Morris, R. G., Williams, S. C. R., & Markus, H. S. (2001). Evidence for cortical “disconnection” as a mechanism of age-related cognitive decline. Neurology, 57, 632–638. doi:10.1212/WNL.57.4.632

O’Sullivan, M., Morris, R. G., Huckstep, B., Jones, D. K., Williams, S. C. R., & Markus, H. S. (2004). Diffusion tensor MRI correlates with executive dysfunction in patients with ischaemic leukoaraiosis. Journal of Neurology, Neurosurgery & Psychiatry, 75, 441–447. doi:10.1136/jnnp.2003.014910

Owen, A. M., McMillan, K. M., Laird, A. R., & Bullmore, E. T. (2005). N-Back working memory paradigm: A meta-analysis of normative functional neuroimaging studies. Human Brain Mapping, 25, 46–59. doi:10.1002/hbm.20131

Pantoni, L. (2010). Cerebral small vessel disease: From pathogenesis and clinical characteristics to therapeutic challenges. The Lancet Neurology, 9, 689–701. doi:10.1016/S1474-4422(10)70104-6

Penke, L., Maniega, S. M., Murray, C., Gow, A. J., Valdés Hernández, M. C., Clayden, J. D.. . .Deary, I. J. (2010). A general factor of brain white matter integrity predicts information processing speed in healthy older people. The Journal of Neuroscience, 30, 7569–7574. doi:10.1523/JNEUROSCI.1553-10.2010

Perry, M. E., McDonald, C. R., Hagler, J., Gharapetian, L., Kuperman, J. M., Koyama, A. K.. . .McEvoy, L. K. (2009). White matter tracts associated with set-shifting in healthy aging. Neuropsychologia, 47, 2835–2842. doi:10.1016/j.neuropsychologia.2009.06.008

Persson, J., Nyberg, L., Lind, J., Larsson, A., Nilsson, L. G., Ingvar, M., & Buckner, R. L. (2006). Structure-function correlates of cognitive decline in aging. Cerebral Cortex, 16, 907–915. doi:10.1093/cercor/bhj036

Pierpaoli, C., & Basser, P. J. (1996). Toward a quantitative assessment of diffusion anisotropy. Magnetic Resonance in Medicine, 66, 893–906. doi:10.1002/mrm.1910360612

Prins, N. D., van Dijk, E. J., den, H. T., Vermeer, S. E., Jolles, J., Koudstaal, P. J.. . .Breteler, M. M. B. (2005). Cerebral small-vessel disease and decline in information processing speed, executive function and memory. Brain: A Journal of Neurology, 128, 2034–2041. doi:10.1093/brain/awh553

Ranganath, C., Johnson, M. K., & D’Esposito, M. (2003). Prefrontal activity associated with working memory and episodic long-term memory. Neuropsychologia, 41, 378–389. doi:10.1016/S0028-3932(02)00169-0

Reuter-Lorenz, P. A., & Sylvester, C. C. (2005). The cognitive neuroscience of working memory and aging. In R.Cabeza, L.Nyberg, & D.Park (Eds.), Cognitive neuroscience of aging (pp. 186–217). New York, NY: Oxford University Press.

Sasson, E., Doniger, G. M., Pasternak, O., & Assaf, Y. (2010). Structural correlates of memory performance with diffusion tensor imaging. NeuroImage, 50, 1231–1242. doi:10.1016/j.neuroimage.2009.12.079

Schmahmann, J. D., Pandya, D. N., Wang, R., Dai, G., D’Arceuil, H. E., de Crespigny, A. J.. . .Wedeen, V. J. (2007). Association fibre pathways of the brain: Parallel observations from diffusion spectrum imaging and autoradiography. Brain: A Journal of Neurology, 130, 630–653. doi:10.1093/brain/awl359

Smith, S. M. (2002). Fast robust automated brain extraction. Human Brain Mapping, 17, 143–155. doi:10.1002/hbm.10062

Smith, S. M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T. E., MacKay, C. E.. . .Behrens, T. E. J. (2006). Tract-based spatial statistics: Voxelwise analysis of multi-subject diffusion data. NeuroImage, 31, 1487–1505. doi:10.1016/j.neuroimage.2006.02.024

Smith, S. M., & Nichols, T. E. (2009). Threshold-free cluster enhancement: Addressing problems of smoothing, threshold dependence and localisation in cluster inference. NeuroImage, 44, 83–98. doi:10.1016/j.neuroimage.2008.03.061

Squire, L. R. (1992). Declarative and Nondeclarative Memory: Multiple Brain Systems Supporting Learning and Memory. Journal of Cognitive Neuroscience, 4, 232–243. doi:10.1162/jocn.1992.4.3.232

Squire, L. R., Cohen, N. J., & Nadel, L. (1984). The medial temporal region and memory consolidation: a new hypothesis. In H.Weingartner & E.Parker (Eds.), Memory Consolidation (pp. 185–210). Hillsdale, NJ: Lawrence Erlbaum.

Stadlbauer, A., Salomonowitz, E., Strunk, G., Hammen, T., & Ganslandt, O. (2008). Age-related degradation in the central nervous system: Assessment with diffusion-tensor imaging and quantitative fiber tracking. Radiology, 247, 179–188. doi:10.1148/radiol.2471070707

Tulving, E. (1983). Elements of episodic memory. Oxford, UK: Clarendon Press.

Van Petten, C., Plante, E., Davidson, P. S. R., Kuo, T. Y., Bajuscak, L., & Glisky, E. L. (2004). Memory and executive function in older adults: Relationships with temporal and prefrontal gray matter volumes and white matter hyperintensities. Neuropsychologia, 42, 1313–1335. doi:10.1016/j.neuropsychologia.2004.02.009

Wechsler, D., Wycherley, R. J., Benjamin, L., Callanan, M., Lavender, T., Crawford, J. R., & Mockler, D. (1998). Wechsler Memory Scale-III. Third Edition. London, UK: The Psychological Corporation.

Woods, R. P., Grafton, S. T., Holmes, C. H., Cherry, S. R., & Mazziota, J. C. (1998). Automated image registration: II. Intersubject validation of linear and non-linear models. Journal of Computer Assisted Tomography, 22, 153–165. doi:10.1097/00004728-199801000-00028

Woods, R. P., Grafton, S. T., Holmes, C. J., Cherry, S. R., & Mazziotta, J. C. (1998). Automated image registration: I. General methods and intrasubject, intramodality validation. Journal of Computer Assisted Tomography, 22, 139–152. doi:10.1097/00004728-199801000-00027

Wozniak, J. R., & Lim, K. O. (2006). Advances in white matter imaging: A review of in vivo magnetic resonance methodologies and their applicability to the study of development and aging. Neuroscience and Biobehavioral Reviews, 30, 762–774. doi:10.1016/j.neubiorev.2006.06.003

Zahr, N. M., Rohlfing, T., Pfefferbaum, A., & Sullivan, E. V. (2009). Problem solving, working memory, and motor correlates of association and commissural fiber bundles in normal aging: A quantitative fiber tracking study. NeuroImage, 44, 1050–1062. doi:10.1016/j.neuroimage.2008.09.046

Submitted: April 11, 2012 Revised: February 20, 2013 Accepted: March 4, 2013