Resting metabolic connectivity in Alzheimer’s disease
Silvia Morbelli
0
1
Dario Arnaldi
0
1
Selene Capitanio
0
1
Agnese Picco
0
1
Ambra Buschiazzo
0
1
Flavio Nobili
0
1
0
S. Morbelli (&) D. Arnaldi S. Capitanio A. Picco A. Buschiazzo F. Nobili Clinical Neurophysiology, Department of Neurosciences, Ophthalmology and Genetics
, IRCCS AOU San Martino-IST,
University of Genoa
, Largo R. Benzi 10, 16132 Genoa,
Italy
1
S. Morbelli S. Capitanio A. Buschiazzo Nuclear Medicine Unit, Department of Health Sciences
, IRCCS AOU San Martino-IST,
University of Genoa
, Largo R. Benzi 10, 16132 Genoa,
Italy
Metabolic connectivity analysis of resting 18F-FDG PET is based on the assumption that brain regions whose metabolism is significantly correlated at rest are functionally associated and that the strength of the association is proportional to the magnitude of the correlation coefficient. Therefore, this method could be used to evaluate connectivity networks independently on the basis of performance in specific tasks. Published studies have provided evidence that metabolic connectivity substantially overlaps underlying anatomical pathways and depends on the location of the analyzed regions, but is not influenced by their size. The present review focuses on the methods and meaning of resting inter-regional correlation analysis of cerebral metabolic rate of glucose consumption in Alzheimer's disease. Accordingly, we describe the evolution of analytical tools from the correlation with a single region of interest to a voxel-based statistical parametric mappingbased approach. We also discuss the pathophysiological implications of metabolic connectivity studies both for Alzheimer-related disconnection syndrome and for default-mode network impairment and compensation mechanisms.
-
Alzheimers disease (AD) is the most prevalent dementing
disorder worldwide and is pathologically defined by the
presence of amyloid aggregations (neuritic plaques, NP)
and tau pathology (neurofibrillary tangles, NFT) [1, 2].
Neuropathological data in AD reveal that NFT are
prominent in the medial temporal lobe (MTL) early in the disease
and then progress outward [3], while NP have a broader
cortical distribution that includes but is not especially
prominent in the MTL [4]. Moreover, recent pathological
studies have led to the hypothesis that NFT predominate in
pyramidal neurons that form corticocortical connections
between and within the cerebral hemispheres, whereas NP
are prevalent at the end of these tracts and in their collateral
branches [5, 6]. Consequently, AD pathophysiology could
be the result not only of damage to one or more neuronal
systems, but also of a disruption of the brains connectivity
due to abnormal interactions between neuronal systems [6].
Consistent with this pathology, magnetic resonance
imaging (MRI) in early stage AD demonstrates MTL
atrophy [7, 8] while functional imaging with 18F-FDG PET
reveals hypometabolism with a characteristic
parietotemporal and posterior cingulate pattern [9].
The concept of a corticocortical disconnection syndrome
in AD was initially investigated using
blood-oxygen-leveldependent (BOLD) fMRI, which can identify functional
networks related to the joint activation of brain areas
involved in different cognitive functions [10]. Several
fMRI studies have shown that patterns of activation are
changed in AD patients during the performance of certain
tasks [10, 11]. Similarly, PET activation studies with H215O
have also been analyzed in terms of functional connectivity
and have demonstrated reduced functional interactions
implying an anterior-posterior disconnection [12]. Since
these approaches focus on the networks that directly
underlie specific task performance, they do not evaluate the
more general concept of impaired connectivity in AD, nor
do they directly address impairment of the default-mode
network (DMN). This network, which includes the medial
prefrontal cortex, posterior cingulate, and inferior parietal
lobule, is thought to be mainly active in resting conditions
when individuals are engaged in internally focused tasks
including autobiographical memory retrieval, visualizing
the future, and conceiving the perspectives of others [13].
Several lines of evidence suggest that in the preclinical
stage of AD, beta amyloid deposition is present mainly in
DMN regions [14] and that impairment of the DMN may
represent a sensitive and specific biomarker of incipient
AD. Among the methods proposed for studying the DMN
in AD, resting-state fMRI reflects spontaneous neuronal
activity and/or the endogenous neurophysiological
processes of the human brain in the resting state (see Liu et al.
for a review [15]), while H215O PET studies allow analysis
of covariance of rest-specific regional cerebral blood flow
patterns [16]. Recent combined resting-state fMRI and
amyloid PET studies have provided, in vivo, further
evidence of the pathophysiological meaning of impaired
connectivity in AD. Indeed, two elegant studies, by
Drzezga et al. [14] and Sheline et al. [17] (...truncated)