Prior work has shown that DMN functionality is disrupted in AD23, 24
as well as in cognitively normal individuals at increased risk for developing AD (i.e., individuals having at least one APOE ε4 allele or participants with amyloid plaques using Pittsburgh Compound B (PiB))31-33
. In the present work, we demonstrate reduced functional connectivity within specific regions of the DMN (PCC-MTL) in asymptomatic individuals whose parents have late-onset AD. This effect was detectable in non-carriers of the APOE ε4 allele. Observed decreases in functional connectivity were not attributable to structural MTL atrophy. In addition, other networks were not affected (i.e., motor, visual, and auditory). Disruption of PCC-MTL functional connectivity may be an early pathophysiological marker of AD reflecting unknown genetic factors.
Previous studies using neuronal tract tracing techniques have described anatomical connections between the PCC and MTL in non-human primates 34, 35
. Diffusin tensor imaging studies of humans have identified structural connectivity (the cingulum bundle) between the two regions 36, 37
. Recent work using rs-fcMRI has demonstrated PCC-MTL functional connections in humans 23
that might be relevant for episodic memory 38
. The present result, that is, selective reduction in the PCC-MTL connectivity in asymptomatic individuals, is consistent with a model in which pathology (neurofibrillary tangles) initially appears within transentorhinal regions and subsequently spreads within the limbic system 39
. Furthermore, prior work has shown that symptomatic AD patients have reduced functional connectivity not only between the PCC and MTL, but also between the PCC and other regions of the DMN 23, 24, 40
. Hence, the available data suggest that reduced functional connectivity in AD progresses in an orderly manner, the earliest changes occurring between the PCC-MTL, with subsequent spread to other DMN regions. Longitudinal studies tracking the effects of AD progression on functional networks and identifying pathological basis of functional network disruption are therefore needed.
An age-related increase in amyloid deposition has been previously observed in individuals with a family history of AD but not in indiviudals without a family history of AD 26
. Similarly, the present work found an age-dependent decrease in the PCC-RSC functional connectivity only within individuals with a family history of AD but not in those without. Prior work has related functional abnormalities in the PCC and RSC to the increase of amyloid burden 41
. Thus, it is possible that the decreased functional connectivity may be related to amyloid deposition in an age-dependent fashion in individuals with a family history of AD. Future work assessing the relationships between family history of AD, amyloid deposition, and resting state networks is needed.
We observed that the effect of family history of late-onset AD was significant in APOE ε4 non-carriers but not in APOE ε4 carriers. The biological mechanisms underlying this differential influence of APOE genotype on the expression of the family history remain unclear. It is possible that the negative results seen in APOE ε4 carriers are a consequence of insufficient statistical power; the smallest group of participants were those who did not have a family history but had at least one APOE ε4 allele (N=35). Further studies are required to determine whether gene-gene interactions impact DMN integrity.
Previous investigations of APOE ε4 effects on the DMN typically studied individuals in relatively narrow age ranges33, 42-45
. It has been reported that, relative to APOE ε4 non-carriers, functional connectivity in carriers is increased in young adulthood (mean age 28 years)42
, unchanged in middle adulthood (mean age 46 years)44
(mean age 59 and 63 years respectively) in middle age, and decreased in old age (median age 78 years)43
. These discrepancies are difficult to reconcile and may reflect differences in the ages of the study samples and experimental strategies (i.e., seed-based vs. independent component analysis). Our study cohort was 68 years old on average, ranging from 45 to 91 years. The possibility of a non-linear relationship between age and APOE ε4 related changes in DMN functional connectivity was assessed but not observed (data not shown). Further work investigating a lifespan cohort that includes sufficient sample sizes in each age range is needed to reveal the trajectory of APOE ε4 related changes in functional connectivity.
We found no evidence that the decreased PCC-MTL functional connectivity is attributable to MTL atrophy. This observation is consistent with previous studies reporting that reduced DMN functional connectivity associated with risk factors for AD in general is not attributable to atrophy 31-43
. Collectively, these findings suggest that rs-fcMRI may be more sensitive than structural MRI for the detection of group differences at preclinical stages of AD.
The present work has several limitations. It is possible that some participants whose parents had neurodegenerative dementias other than AD were included, as inclusion was based on reports of the offspring (i.e., the participants) and parental diagnoses generally were not confirmed by neuropathological examination. Moreover, it is also possible that some of our participants with no family history of AD have parents who will eventually develop AD. Inclusion of such participants would have reduced the sensitivity of our analyses. In addition, our study is cross-sectional in design. Additional longitudinal studies are required to determine if decreased functional connectivity remains restricted to PCC-MTL, or if additional DMN regions are disrupted. The present results show that rs-fcMRI measures are affected by unknown genetic factors embodied in a family history of late-onset AD. Future research is warranted to determine the extent to which specific genetic factors, such as novel AD susceptibility genes identified by recent genome-wide association studies 46
, account for the observed changes.