Abnormal whole-brain functional connection in amnestic mild cognitive impairment patients

Abstract

Amnestic mild cognitive impairment (aMCI) patients are thought to be particularly vulnerable to convert to clinical AD where functional disconnection is a major feature of the cortical neuropathology. However, the presence and extent of whole-brain connectivity disturbances is largely unknown in aMCI patients. Twenty-six aMCI patients and eighteen matched healthy subjects were evaluated at baseline and at mean 20 months follow up. Temporal correlations between spatially distinct regions were evaluated by using longitudinal resting-state fMRI. Compared to normal aging controls, patterns of abnormal interregional correlations in widely dispersed brain areas were identified in the patients, which also changed with disease progression. These disturbances were found particularly in subcortical regions and frontal cortex. Importantly, significantly decreased negative functional connection may be specifically associated with the development of aMCI patients. This suggests a compensatory mechanism is underway where local processing deficits are offset by recruitment of more dispersed cortical regions. In addition, the presence of this increased connectivity is seen to eventually weaken with disease progression. The results suggest that patterns of whole-brain functional connection may be a useful risk marker for conversion to AD in aMCI patients.

Introduction

Alzheimer's disease (AD) is the major cause of dementia in the elderly population. The increase of the life expectancy and the lack of effective treatments continue to lead to a rapid increase in patient numbers. Individuals with amnestic mild cognitive impairment (aMCI), who are thought to be particularly vulnerable to convert to AD [12], [19], [38]. However, currently no diagnostic tool is available which reliably separates aMCI patients into those at risk of conversion and those not. It is becoming increasingly important to identify such tools so that management and intervention can be initiated as early as possible in the disease course and focused on those who need it. One avenue that may prove productive is to explore the patterns of altered brain function in aMCI patients, which may assist with producing early diagnostic indicators of AD development.

Increasing attention is being turned toward examining functional connections in the brain. fMRI permits the creation of functional connection maps of spatially distinct but temporally correlated brain regions called functional networks. Recent functional connection analyses have shown that certain brain networks are especially active when an individual is at rest, and these have been termed “resting state networks”. Biswal and colleagues were the first to observe that low-frequency (<0.08 Hz) fluctuation (LFF) of bold oxygenation level dependent (BOLD) was highly synchronous within motor cortices in this resting-state [8], and the term ‘default-mode’ network (DMN) came into use in a series of publications [15], [16], [17], [18], [29]. However, resting-state networks are not limited to the DMN described above. Damoiseaux et al. found that resting state LFF were identified in the DMN, sensory and motor networks, visual and auditory cortices, and two additional networks most likely involved in memory and executing functions [10]. Mantini et al. observed six resting state networks including the DMN, dorsal attention network, visual and auditory networks, somato-motor network and a network postulated to reflect self-referential processing [20]. Obviously, temporally synchronized activity across distributed regions of the brain is a key feature of normal brain function even when an individual is not engaged in any specific goal-directed behaviour.

Recently, evidence of impaired resting-state functional connectivity has been reported in AD and MCI patients, including structures such as hippocampus [1], [36] and posterior cingulate cortex [4], [42]. In addition, an unbiased whole-brain resting-state functional connection study (based on interregional correlations of 116 separate regions) suggested that AD patients had decreased positive correlations between the frontal and parietal lobes, but increased positive correlations within the frontal lobe, parietal lobe, and occipital lobe [37]. Importantly, these findings support the contention that the key underlying pathology in AD involves a progressive cortical disconnection syndrome [6], [11]. There is also evidence of similar disturbances in aMCI patients involving the DMN [27], [28], [30], [34] and the executive attention networks [34]. However, a more detailed mapping of the disrupted whole-brain connection patterns awaits completion in aMCI patients. Moreover, longitudinal functional information is still extremely limited, and there remains an urgency to identify putative markers that may have utility in predicting progression from aMCI to AD.

In the present study, using resting-state fMRI, we tested the hypotheses that aMCI patients display increasingly disturbed patterns of functional connection in the brain over time. We explored the properties of whole-brain networks using multivariate spectral analysis of fMRI time-series measured in 116 cortical and subcortical regions in the resting state. This same approach has been widely utilized in previous studies on healthy subjects [31], schizophrenia [21], blind [23], epilepsy [22] and AD patients [37].