Functional implications of hippocampal volume and diffusivity in mild cognitive impairment

Abstract

Hippocampal atrophy has been related to mild cognitive impairment (MCI) and early Alzheimer disease (AD), but the diagnostic significance of cross-sectionally determined hippocampal volumes is still ambiguous. Diffusion-Tensor-Imaging (DTI) in MCI patients revealed an association of microstructural changes in hippocampal areas with verbal memory decline. MRI volumetry and DTI were combined to investigate 18 MCI patients attending a memory clinic, and 18 carefully age- and gender-matched healthy controls. Neuropsychological testing, high resolution T1-weighted volume MRI scans, and DTI scans with regions-of-interest in hippocampal areas were applied. Left hippocampal volume was significantly lower (−11%, P = 0.02) in MCI patients than in control subjects. No significant differences were found for the right hippocampus (−4%). Mean diffusivity (MD) was significantly elevated in MCI patients vs. controls in left (+ 10%, P = 0.002) and right hippocampal areas (+ 13%, P = 0.02). Hippocampal volume and MD values were not significantly correlated. Combining left hippocampal volume and MD measures showed that lower left hippocampal volumes were associated with poor verbal memory performance particularly when co-occurring with high MD values. No comparable associations could be found regarding the right hippocampal formation and with respect to non-verbal memory function. The results demonstrate that microstructural abnormalities as revealed by DTI are very sensitive early indicators of hippocampal dysfunction. The combination of macro- and microstructural parameters in hippocampal areas could be promising in early detection of neurodegenerative processes.

Introduction

There is convincing evidence that degenerative processes of the entorhinal–hippocampal formation occur at very early stages of Alzheimer's Disease (AD) (Braak and Braak, 1991, Braak and Braak, 1997). Atrophy of the medial temporal lobe, particularly the parahippocampal and hippocampal regions, seems associated with memory loss (Laakso et al., 2000a, Pennanen et al., 2004). Thus, reduction in hippocampal volume represents one of the structural hallmarks of incipient AD (Chetelat and Baron, 2003, Kantarci and Jack, 2003, Wolf et al., 2004). From a clinical standpoint, the amnestic type of mild cognitive impairment (MCI) (Petersen et al., 1999, Petersen et al., 2001) is at present the most valid concept indicative of incipient or early AD (Morris et al., 2001, Grundman et al., 2004, Crocco and Loewenstein, 2005). Consistent with this, patients with MCI showed hippocampal size reductions, particularly on the left side, compared to healthy controls (Du et al., 2001, Pennanen et al., 2004, De Toledo-Morrell et al., 2004), and longitudinal studies demonstrated that hippocampal volume reduction predicts transition from MCI to dementia (Laakso et al., 2000b, Kantarci and Jack, 2003). On a functional level, hippocampal atrophy was associated with impaired verbal memory function deficits comprising both encoding and retrieval (Chetelat et al., 2003), particularly delayed verbal recall (Walhovd et al., 2004). Previous studies have also shown a specific correlation between hippocampal neuron loss and verbal memory deficits in patients with left temporal lobe epilepsy (Sass et al., 1995). A meta-analysis of 33 studies on the relationship between memory performance and hippocampal volumes (Van Petten, 2004) yielded, however, an extreme variability and surprisingly little support for the “bigger-is-better” hypothesis in older adults. Thus, there is uncertainty regarding the relationship between hippocampal volume and memory. Reports on age-related volume reductions of hippocampal structures are inconsistent (Mueller et al., 1998, Pruessner et al., 2001, Scahill et al., 2003, Szentkuti et al., 2004); recently, preserved hippocampal volumes across an age range from 20 to 85 years have been suggested by a study involving 128 subjects (Sullivan et al., 2005). Twin studies (Järvenpää et al., 2004, Van Erp et al., 2004) show that genetic factors determine approximately 30–40% of the variance in hippocampal volume. However, among different brain regions, the hippocampus formation seems to be least genetically determined (Sullivan et al., 2001, Järvenpää et al., 2004) and most prone to developmental and environmental influences. Additional risk factors for hippocampal shrinkage could be hypertension and further vascular risk factors (Strassburger et al., 1997, Raz et al., 2005). A study has recently shown reduced hippocampal size (−14%) in 20 male patients with coronary artery disease (CAD) without any cognitive impairment compared to matched controls without CAD (Koschack and Irle, 2005). Other influencing factors discussed in the literature are education and cognitive exercise, traumatic brain injury, endocrine-metabolic factors, and psychiatric disorders.

A meta-analysis of 12 studies on hippocampal volume reduction in patients with major depression has revealed a reduction of both hippocampi (−8% left; −10% right side) in a total of 351 patients with life-time depression and 279 healthy controls after adjustment for age and gender differences (Videbech and Ravnkilde, 2004). Limbic brain structures and particularly the hippocampus formation are assumed to mediate the relationship between depression and dementia (Steffens et al., 2002). In patients with post-traumatic stress disorder (PTSD), an 8% smaller right hippocampal volume relative to control subjects has been found (Bremner et al., 1995). Smaller right hippocampi have also been found in patients with recent-onset PTSD (Wignall et al., 2004), leading to the hypothesis that reduced right hippocampal size could be a risk factor for affective and stress disorders (Gilbertson et al., 2002). The specificity of these findings is nevertheless questionable, as in schizophrenic patients brain volume reductions including right and left hippocampal gray matter have been repeatedly reported (Yamasue et al., 2004, Rupp et al., 2005, Weiss et al., 2005) though the underlying mechanisms are not fully known.

Two studies (Kantarci et al., 2001, Fellgiebel et al., 2004) have demonstrated that the random, undirected motion of water molecules assessed with diffusion-tensor-imaging (DTI) was significantly elevated in hippocampal regions of MCI patients compared to age-matched controls. Supporting the functional relevance of elevated diffusivity in hippocampal areas, DTI parameters were correlated with sensitive measures of memory dysfunction, e.g., delayed verbal recall test (DVRT) performance (Fellgiebel et al., 2004). It is assumed that reduced water diffusion parallel to axonal tracts (Fractional Anisotropy, FA) is indicative of axonal degeneration, and increased water diffusion perpendicular to axonal tracts (Mean Diffusivity, MD) may be associated with changes in water content, disruption and partial break-down of tissue cytoarchitecture (Beaulieu, 2002, Neil et al., 2002), sclerosis (Assaf et al., 2003), or demyelinating processes (Bartzokis, 2004, Song et al., 2002, Song et al., 2003, Sun et al., 2005). Such processes are supposed to occur before neuronal degradation and atrophy are detectable on a macroscopic level. Accordingly, DTI parameters might be more sensitive and quantifiable measures of early degeneration in AD than conventional MRI imaging techniques (Neil et al., 2002, Sundgren et al., 2004, Sun et al., 2005). The diagnostic utility of different MR modalities and their combination in early AD has been explored in a few studies (Kantarci et al., 2001, Kantarci et al., 2002, Kantarci and Jack, 2003). In schizophrenic patients, the co-registration of DTI supplementary to hippocampus volumetry could detect subtle regional alterations (Kalus et al., 2004). However, the combination of MRI-based volumetry and DTI measures in MCI has not been considered yet. The combination of both techniques could help improve the interpretation of hippocampal size in cross-sectional studies and could elucidate functional implications of morphological and ultrastructural changes in early AD.

Therefore, the present study was designed to combine bilateral hippocampus volumetry and DTI measures in hippocampal regions-of-interest for comparison of MCI patients with healthy subjects and to estimate the functional association of both measures with memory performance.