Skip to main content
NCBI home page
Human Brain Mapping logoLink to Human Brain Mapping
. 2018 Sep 11;39(12):4987–4997. doi: 10.1002/hbm.24339

Disrupted interactions among the hippocampal, dorsal attention, and central‐executive networks in amnestic mild cognitive impairment

Ganesh B Chand 1,✉, Ihab Hajjar 1,2,†, Deqiang Qiu 3,4,†
PMCID: PMC6866643  PMID: 30272830

Abstract

Neuroimaging investigations consistently demonstrate that the neural processes involve complex interactions between the large‐scale networks. Among those networks, the dorsal attention network (DAN) and the central‐executive network (CEN) have been previously shown to exhibit anti‐correlated activity with the default‐mode network (DMN) in cognitively normal people. In amnestic mild cognitive impairment (MCI) and Alzheimer's disease, the hippocampal network (HCN)—a key memory processing system—and its interactions with other networks have gathered central interest. The current study aims to evaluate the patterns of functional architectures of the HCN with the three networks—DAN, CEN, and DMN—in amnestic MCI and normal controls (NC) to test the hypothesis that the interactions of HCN with other networks alter in MCI. We recorded the resting state functional MRI, assessed patterns of functional architectures between the four networks using dynamical causal modeling, and compared between NC and MCI. Our analysis showed that the DAN modulates the activity between the HCN and the DMN in both MCI and NC. We further uncovered that the DAN modulates the activity between the HCN and the CEN in NC, but such modulation is impaired in MCI. We found an association between impaired modulation and Montreal cognitive assessment (R = 0.349). Overall, our findings provide important insight in understanding the neuroimaging signature of amnestic MCI and/or Alzheimer's disease.

Keywords: neuroimaging, network interactions, functional MRI, dynamical causal modeling

1. INTRODUCTION

The human brain is intrinsically organized into functionally specialized regions that construct several large‐scale networks, including the dorsal attention network (DAN), central‐executive network (CEN), default‐mode network (DMN), and hippocampal network (HCN) (Chand & Dhamala, 2017; Fox et al., 2005; Power et al., 2011; Raichle, 2015; Vincent, Kahn, Snyder, Raichle, & Buckner, 2008). The DAN—consisting of bilateral frontal eye fields, superior parietal lobules, and middle temporal cortices—is associated with externally directed cognition such as eye movement, hand‐eye coordination, and shift of spatial attention (Brissenden, Levin, Osher, Halko, & Somers, 2016; Gao & Lin, 2012; Vincent et al., 2008). The CEN—comprising of bilateral posterior parietal cortices and dorsolateral prefrontal cortices—is responsible in goal‐directed higher order cognition such as attention, working memory, and decision‐making (Bressler & Menon, 2010; Chand & Dhamala, 2016a, 2017; Miller & Cohen, 2001; Zanto, Rubens, Thangavel, & Gazzaley, 2011). The DMN—consisting of bilateral ventromedial prefrontal cortices and posterior cingulate cortices—and the hippocampal network (HCN)—consisting of bilateral hippocampus—are collectively associated during passive mental states linked to internally directed cognition, including recollection of the past and thinking about the future (Huijbers, Pennartz, Cabeza, & Daselaar, 2011; Liang et al., 2014; Vincent et al., 2008). Former studies suggest that the DAN and the CEN exhibit anti‐correlated activity with the DMN (Rohr et al., 2016; Vincent et al., 2008; Wang et al., 2015). Investigations consistently demonstrate that the HCN is a key memory processing system in the human brain and is widely reported to have impaired activity in amnestic mild cognitive impairment (MCI) and/or Alzheimer's disease (Allen et al., 2007; Dubois et al., 2016; Zhang et al., 2011). The patterns of interactions of the HCN with the three key networks—DAN, CEN, and DMN—are yet unknown.

Emerging evidence consistently suggests that cognition involves a set of large‐scale networks and their complex interactions (Biswal, Haughton, & Hyde, 1995; Buckner, Andrews‐Hanna, & Schacter, 2008; Chand & Dhamala, 2016b; Chand, Wu, Hajjar, & Qiu, 2017b; Chand, Wu, Qiu, & Hajjar, 2017c; Deco, Jirsa, & McIntosh, 2011; Fox et al., 2014; Friston, Moran, & Seth, 2013; Menon, 2011; Stephan et al., 2010). Among those networks, the DAN and the CEN have formerly been reported to exhibit anti‐correlated network activity with the DMN (Rohr et al., 2016; Vincent et al., 2008; Wang et al., 2015). The CEN, especially the dorsolateral prefrontal cortex node, has rich connections with several areas in the brain such as visual, somatosensory, and auditory areas, and has been argued to communicate with the occipital, parietal, and temporal cortices in several cognitive functions, such as integrating information from the external environment with the stored internal representation during executive decision‐making processes (Chand & Dhamala, 2017; Chand, Lamichhane, & Dhamala, 2016; Miller & Cohen, 2001; Petrides & Pandya, 1999). However, the functional role of CEN has conflicting reports with the progression of disease (Diener et al., 2012; Li et al., 2015). The activity of the DMN and the HCN has been repeatedly shown to decline in the MCI and/or Alzheimer's disease (Allen et al., 2007; Greicius & Kimmel, 2012; Greicius, Srivastava, Reiss, & Menon, 2004). Moreover, the studies in healthy elderly population show that the HCN is functionally connected with the frontal, parietal, occipital, and temporal lobes and it exhibits the decreased functional connections, especially with the frontal/prefrontal cortex in Alzheimer's disease (Allen et al., 2007). The DAN is involved in top‐down orienting processes and shows elevated activity after presentation of cues indicating where, when, or to what the individuals should direct their attention (Fox, Corbetta, Synder, Vincent, & Raichle, 2006). Recent investigations in MCI/Alzheimer's disease have revealed that the DAN is functionally impaired greater than its ventral subsystem (ventral attention network)—a key system for bottom‐up processes—suggesting that the MCI/Alzheimer's disease patients might have deficits in top‐down control processing than bottom‐up processing (Bokde et al., 2010; Li et al., 2015, 2012).

Recent studies indicate that functional MRI‐based brain activity is linked with the spatial spreading of Alzheimer's disease‐related brain abnormalities (Jones et al., 2016; Mutlu et al., 2017). The gray matter atrophy—possibly reflecting tau pathology—spreads across regions via specific functionally connected pathways and the hubs such as the HCN and the posterior regions of DMN and CEN are especially vulnerable to amyloid‐beta deposition and glucose hypometabolism (Mutlu et al., 2017). Thus, studying the specific brain networks and/or their interactions is a novel approach that may shed light on our knowledge of MCI and/or Alzheimer's disease pathology. However, a large body of literature in MCI/Alzheimer's disease research have mainly reported the altered activity in individual network such as the reduced activity in the DMN (Brier, Thomas, & Ances, 2014; Greicius & Kimmel, 2012; Greicius et al., 2004), DAN (Li et al., 2012), HCN (Allen et al., 2007), and the increased activity in the prefrontal/frontal cortex of the CEN (Franzmeier, Duering, Weiner, Dichgans, & Ewers, 2017). Investigations have repeatedly highlighted that the HCN is a key memory processing system in the brain and have also indicated that this memory system is declined in amnestic MCI and/or Alzheimer's disease (Allen et al., 2007; Dubois et al., 2016; Zhang et al., 2011). However, a unified model that examines the patterns of interactions of these four key networks consisting of MCI and/or Alzheimer's disease hubs are largely unreported to date. Evaluating the patterns of interactions between these four key networks in elderly subjects with normal cognition and with those of amnestic MCI could help test the unified hypothesis of network alternations in amnestic MCI and thus may provide novel insights in understanding the pathophysiology of amnestic MCI and/or Alzheimer's disease.

In the current study, we aim to probe the patterns of interactions of the HCN with the DAN, CEN, and DMN in elderly people with normal cognition and with those of amnestic MCI. We recorded the resting state functional MRI data, assessed causal interactions between HCN, DAN, CEN, and DMN using dynamical causal modeling methodology, and compared between the normal control group and the amnestic MCI group. We hypothesized that the interactions of the HCN with other networks decline in the MCI group compared with the healthy control group. We further hypothesized that the impaired interactions correlate with the lower cognitive performance on the neuropsychological test.

2. MATERIALS AND METHODS

2.1. Study sample

Institutional Review Board of Emory University reviewed and approved the present study protocol. Our study sample consisted of 29 MCI and 50 cognitively normal control human subjects. A written informed consent was provided by each subject before MRI acquisition. MCI was assessed based on the Diagnostic and Statistical Manual of Mental Disorders‐V (DSM‐V) criteria. MRI data were acquired from MCI subjects who had age ≥50 years, Montreal Cognitive Assessment (MoCA) <26, clinical dementia rating scale and memory sum of boxes (CDR) = 0.5, functional assessment questionnaire (FAQ) <9, abnormal logical memory subscale (delayed paragraph recall) from the Wechsler memory scale‐revised (25 maximum score): <11 for 16 or more years of education, <9 for 8–15 years of education, and <6 for <7 years of education. MRI data were also acquired from cognitively normal control subjects who had age ≥50 years, CDR = 0, FAQ ≥9, and normal logical memory subscale: ≥11 for 16 or more years of education, ≥9 for 8–15 years of education, and ≥6 for <7 years of education. The subject's exclusion criteria were the history of stroke in the past 3 years, ineligible to perform the MRI due to metal implants or cardiac pacemaker, inability to perform the cognitive test (e.g., communication limitations from language or other factors), reports regarding the diagnosis of dementia, and other neurological or psychiatric illness—Parkinson's disease, multiple sclerosis, epilepsy, and schizophrenia—that might affect cognition. The age, sex, and education variables were not statistically different, but the MoCA score and race (black) were statistically different between MCI group and normal control group (Table 1).

Table 1.

Mean (standard deviation) and statistical comparison between cognitively normal control (NC) and mild cognitive impairment (MCI) individuals regarding their age, education, sex, Montreal cognitive assessment (MoCA) score, and race

Characteristics NC MCI p‐value
N 50 29
Age, year 63.6 (8.1) 67.1 (8.2) 0.07
Education, year 16.3 (2.6) 16.1 (2.8) 0.76
Sex, female 35 (70.0%) 16 (55.2%) 0.18
MoCA score 26.8 (2.5) 20.7 (3.4) 10−11
Race
Black 16 (32.0%) 16 (55.2%) 0.04
White 33 (66.0%) 13 (44.8%) 0.07
Hispanic 1 (2.0%) 0 (0.0%)
Open in a new tab

2.2. Image acquisition

MRI images were recorded using Siemens 3 T Trio scanner at Center for Systems Imaging, Emory University, Atlanta. Anatomical 3D images were acquired in sagittal using a T1‐weighted magnetization prepared rapid gradient echo sequence with the repetition time (TR) = 2,300 ms, echo time (TE) = 2.89 ms, inversion time (TI) = 800 ms, flip angle (FA) = 8°, resolution = 256 × 256 matrix, slices = 176, and thickness = 1 mm. Resting state blood oxygenation level dependent (BOLD)‐fMRI images were acquired axially using an echo‐planar imaging sequence with the TR = 2,500 ms, TE = 27 ms, FA = 90°, field of view (FOV) = 22 cm, resolution = 74 × 74 matrix, number of slices = 48, thickness = 3 mm, and bandwidth = 2,598 Hz/Pixel. The subjects were instructed to hold still, keep the eyes open, and not to think anything during resting state BOLD‐fMRI recording.

2.3. Image analysis

Image analysis consists of the following main steps.

2.3.1. Image preprocessing

Images were preprocessed using SPM12 (Wellcome Trust Centre for Neuroimaging, London, United Kingdom; http://www.fil.ion.ucl.ac.uk/spm/software/spm12). Preprocessing steps were slice time correction, motion correction, co‐registration to an individual anatomical image, normalization to Montreal Neurological Institute (MNI) template, and spatial smoothing of normalized images using 6 mm isotropic Gaussian kernel. Independent component analysis was then carried out on those preprocessed data.

2.3.2. Regions of interest (ROIs) and independent component analysis (ICA)

The coordinates from previous studies (Chand & Dhamala, 2016a; De Rover et al., 2011; Sridharan, Levitin, & Menon, 2008; Vincent et al., 2008) were used in the MarsBaR package (http://marsbar.sourceforge.net) to create the mask for each network. We defined spherical regions with 6 mm radius based on MNI coordinates centered at the key nodes as shown in Table 2.

Table 2.

Networks, region of interests (ROIs), and MNI coordinates

Network ROI MNI coordinate (x, y, z) mm
Dorsal attention network Middle temporal cortex (R/L) ±45, −69, −2
Frontal eye field (R/L) ±25, −8, 50
Superior parietal lobule (R/L) ±27, −52, 57
Central‐executive network Dorsolateral prefrontal cortex (R/L) ±45, 16, 45
Posterior parietal cortex (R/L) ±54, −50, 50
Hippocampal network Hippocampus (R/L) ±26, −14, −12
Default‐mode network Ventromedial prefrontal cortex (R/L) ±2, 36, −10
Posterior cingulate cortex (R/L) ±7, −43, 33
Open in a new tab

We used the four networks in the Group ICA of fMRI Toolbox (GIFT; http://mialab.mrn.org/software/gift) and computed ICA component of each network. Spatially constrained ICA is a useful technique for the selective network(s) based analyses (Lin, Liu, Zheng, Liang, & Calhoun, 2010). This approach has been formerly reported to provide better representative signal than choosing an average or first eigen‐variate of the template (Chand, Wu, Hajjar, & Qiu, 2017a; Craddock, James, Holtzheimer, Hu, & Mayberg, 2012; Goulden et al., 2014; Shirer, Ryali, Rykhlevskaia, Menon, & Greicius, 2012; Smith et al., 2011). We subsequently carried out a dynamical causal modeling methodology on the ICA‐components.

2.3.3. Dynamical causal modeling (DCM) analysis

DCM analysis (Friston, Harrison, & Penny, 2003) provides the statistical measure of directed functional connectivity between brain areas or networks. DCM is based on Bayesian model selection approach and compares the user‐defined models with the measured data (Stephan et al., 2010) and has recently been implemented in resting state fMRI (Daunizeau, Stephan, & Friston, 2012; Friston, Kahan, Biswal, & Razi, 2014). In model construction, DCM models were designed with full intrinsic connections between the four networks and the modulations were taken to define the models. For example, models 1–3 represent nonlinear modulations from the DMN on the reciprocal connections between the DAN, CEN, and HCN as shown in Figure 3. Similarly, models 4–6 show the modulations from the DAN, models 7–9 indicate the modulations from the CEN, and models 10–12 show the modulations from the HCN to other reciprocal connections between other networks, respectively. DCM analysis was carried out by using SPM12 (Wellcome Trust Centre for Neuroimaging, London, UK; http://www.fil.ion.ucl.ac.uk/spm/software/spm12). Random effect Bayesian model selection method accounts better for heterogeneity of model structure across subjects (Stephan et al., 2010). We used this method to determine an optimal model from the possible models. This method provides results in terms of the posterior model probability—how likely a specific model generated the data of randomly selected subject—and the posterior exceedance probability—how one model is more likely than any other model.

Figure 3.

Figure 3

Open in a new tab

DCM model representation: (M1–M3) specify the modulations by the default‐mode network (DMN) on reciprocal connections between the dorsal attention network (DAN) and the central‐executive network (CEN) (model 1), the DAN and the hippocampal network (HCN) (model 2), the CEN and the HCN (model 3); (M4–M6) specify the modulations by the DAN on the reciprocal connections between the DMN and the CEN (model 4), the DMN and the HCN (model 5), the CEN and the HCN (model 6); (M7–M9) specify the modulations by the CEN on the reciprocal connections between the DMN and the DAN (model 7), the DMN and the HCN (model 8), the DAN and the HCN (model 9); and (M10–M12) specify the modulations by the HCN on the reciprocal connections between the DMN and the DAN (model 10), the DMN and the CEN (model 11), the DAN and the CEN (model 12)

2.4. Cognitive function assessment

Montreal cognitive assessment (MoCA) (Nasreddine et al., 2005) was assessed for each subject. MoCA consists of a set of sub‐tests: short‐term memory recall (5 points), visuospatial (4 points), trail‐making test B (1 point), phonemic fluency (1 point), verbal abstraction (2 points), sustained attention (1 point), serial subtraction (3 points), digit forward (1 point), digit backward (1 point), language test (3 points), repetition and fluency (2 points), and orientation to time and place (6 points) tests. One point is added to the subject's score if his/her formal education is 12 years or less. MoCA is administered out of 30‐point scale in 10 min. MoCA has been previously suggested as a sensitive screening tool for MCI than other existing tools (Dong et al., 2010; Hachinski et al., 2006; Nasreddine et al., 2005; Rossetti, Lacritz, Cullum, & Weiner, 2011).

2.5. Statistical analysis

Normal control group and MCI group were compared using the Mann–Whitney U test regarding the subject's age, education, and MoCA score. The discrete variables that is, race and sex were compared between the groups using the chi‐square test. The comparison of model probabilities within and between groups were assessed using Mann–Whitney U test [adjusted for demographics (age, sex, education, and race) and multiple comparisons—False discovery rate (FDR)]. Pearson's correlation analysis was performed to assess the association between connectivity strength of network modulations with MoCA in the MCI group only because all subjects in the control group had normal MoCA. A p‐value of less than .05 was considered a statistically significant. MATLAB software (Natick, MA; https://www.mathworks.com) was used for data analysis.

2.6. Relationship between network interaction and cognitive function

To explore the relationship between network interaction and cognitive function, we investigated the association of impaired network connectivity strength with the global cognitive testing as assessed by the MoCA score in MCI group (since normal control group had normal scores).

3. RESULTS

We used the default‐mode, dorsal attention, central‐executive, and hippocampal networks masks in Group ICA of fMRI Toolbox (GIFT; http://mialab.mrn.org/software/gift) and calculated the network specific ICA component. Figure 1 shows t‐maps of (A) DMN, (B) DAN, (C) CEN, and (D) HCN overlaid on mean BOLD images in the normal control group. Figure 2 displays the t‐maps in MCI group: (A) DMN, (B) DAN, (C) CEN, and (D) HCN overlaid on mean BOLD images.

Figure 1.

Figure 1

Open in a new tab

The t‐maps of (a) default‐mode network (DMN), (b) dorsal attention network (DAN), (c) central‐executive network (CEN), and (d) hippocampal network (HCN) from constrained‐ICA overlaid on mean BOLD images in cognitively normal control group (L: Left and R: Right)

Figure 2.

Figure 2

Open in a new tab

The t‐maps of (a) default‐mode network (DMN), (b) dorsal attention network (DAN), (c) central‐executive network (CEN), and (d) hippocampal network (HCN) from constrained‐ICA overlaid on mean BOLD images in mild cognitive impairment group (L: Left and R: Right)

To compare the normal control and MCI groups, we constructed the 12 possible network models and compared statistically. Figure 3 illustrates the 12 possible DCM models. Panels (M1–M3) specify the modulations by DMN on reciprocal connections between DAN and CEN (model 1), between DAN and HCN (model 2), between CEN and HCN (model 3); panels (M4–M6) specify the modulations by DAN on reciprocal connections between DMN and CEN (model 4), between DMN and HCN (model 5), between CEN and HCN (model 6); panels (M7–M9) specify the modulations by CEN on reciprocal connections between DMN and DAN (model 7), between DMN and HCN (model 8), between DAN and HCN (model 9); and panels (M10–M12) specify the modulations by HCN on reciprocal connections between DMN and DAN (model 10), between DMN and CEN (model 11), between DAN and CEN (model 12).

We carried out the DCM analysis using random effects‐based Bayesian model selection. The first column of Figure 4 shows the expected posterior model probability and the exceedance probability in the normal control group. DCM analysis revealed that the model 5 (the modulations by DAN on reciprocal connections between DMN and HCN) and model 6 (the modulations by DAN on reciprocal connections between CEN and HCN) have higher probability values than other models in the normal control group. The second column of Figure 4 displays the expected posterior model probability and the exceedance probability in the MCI group. DCM analysis demonstrated that the model 5 (the modulations by DAN on reciprocal connections between DMN and HCN) has a higher probability than other models; however, model 6 (the modulations by DAN on reciprocal connections between CEN and HCN) has no longer higher probability value than other models in MCI group.

Figure 4.

Figure 4

Open in a new tab

Random effects results for the normal controls (NC) group and the mild cognitive impairment (MCI) group in terms of expected and exceedance probabilities: The first column demonstrates that the probabilities of model 5 and of model 6 are higher than other models in the NC group, and the second column demonstrates that the probability of model 5 is higher like the NC group, but of model 6 is decreased in the MCI

The probabilities of models were further compared within the normal control group and MCI group using the Mann–Whitney U test (Figure 5). In the normal control group (Figure 5a), our analysis demonstrated that model 5 and model 6 had significantly higher probability than that of other models (p < .05; adjusted for demographics and FDR), but there was no statistical difference between model 5 and model 6 (p > .05; adjusted for demographics and FDR). In MCI group (Figure 5b), we revealed that model 5 had significantly higher probability than that of model 6 (p < .05; adjusted for demographics and FDR) and of other models (p < .05; adjusted for demographics and FDR). To assess the alterations in the winning model probabilities (model 5 and model 6) of the normal control group with the disease progression into MCI, we further compared two groups within model 5 and model 6 (Figure 6). We uncovered that the probability of model 5 had no significant difference between normal control and MCI (p > .05; adjusted for demographics and FDR), but the probability of model 6 had significantly lower value in MCI group than that of normal control group (p < .05; adjusted for demographics and FDR).

Figure 5.

Figure 5

Open in a new tab

Comparison between models within the normal control (NC) and mild cognitive impairment (MCI) groups: (a) both model 5 and model 6 had statistically higher probability than other models in NC, but (b) only model 5 had statistically higher probability than other models in MCI. (Note: * represents statistically significant and n.s. represents statistically not significant)

Figure 6.

Figure 6

Open in a new tab

Alterations in the winning model probabilities (model 5 and model 6) of normal control (NC) with mild cognitive impairment (MCI): The probability of model 5 had no significant (n.s.) difference between NC and MCI, but the probability of model 6 had significantly lower value in MCI than that of NC (*: p < .05)

3.1. Relationship between network interaction and cognitive function

We investigated the association between the impaired brain connectivity measures and the behavioral performance on MoCA in MCI group (since normal control group had normal scores). Lower connectivity strengths derived from DCM model parameter estimation was associated with lower performance on cognitive testing as reflected by the MOCA scores (R = 0.349). These results are presented in Figure 7.

Figure 7.

Figure 7

Open in a new tab

Correlation analysis in model 6: (a) the modulation connectivity of dorsal attention network (DAN) over the connection from central‐executive network (CEN) to hippocampal network (HCN) versus the Montreal cognitive assessment (MoCA) score (Pearson's correlation: R = 0.101), and (b) the modulation connectivity of DAN over the connection from HCN to CEN versus the MoCA (R = 0.349)

4. DISCUSSION

Dynamic coordinated interactions between the large‐scale networks are crucial in cognitive health (Bressler & Menon, 2010; Chand & Dhamala, 2016a; Menon, 2015; Uddin, 2015). Here we evaluated the patterns of coordinated interactions among the HCN, DAN, CEN, and DMN in elderly people with normal cognition and with those of MCI. In the normal control group, we found that the DAN modulated the activity between the HCN and the DMN, which was in accord with the previously reported top‐down control role of the DAN (Fox et al., 2006; Li et al., 2012). In addition, the DAN modulated the activity between the HCN and the CEN, suggesting the control role of DAN in effective coordination between the HCN and the CEN in normal cognitive functions. In the MCI group, the DAN modulated the activity between the HCN and the DMN like the control group; however, the DAN was not able to modulate the activity between the HCN and the CEN. Our findings that the impaired DAN activity is in line with former findings that the dorsal pathways are disrupted in MCI and/or Alzheimer's disease pathology (Bokde et al., 2010; Horwitz et al., 1995). We further assessed an association between the impaired connectivity and the cognitive function reflected by global neuropsychological assessment suggesting a possible brain‐behavior relationship.

Normal cognitive functions are thought to be explained by the large‐scale networks and their dynamic interactions (Buckner et al., 2008; Chand, Wu, Qiu, et al., 2017c; Deco et al., 2011; Fox et al., 2014; Friston et al., 2013; Menon, 2011). Both the DAN and the CEN have been formerly shown to anti‐correlate with the DMN (Rohr et al., 2016; Vincent et al., 2008; Wang et al., 2015). Functional role of the DAN has been mainly demonstrated in top‐down control processes (Fox et al., 2006) during externally oriented cognitive actions, including eye movement, hand‐eye coordination, and shift of spatial attention (Brissenden et al., 2016; Gao & Lin, 2012; Vincent et al., 2008). The DMN and the HCN functions are suggested during passive mental states linked to internally directed cognition, including memory recall and thinking about the future (Huijbers et al., 2011; Vincent et al., 2008). The HCN is functionally connected with the frontal, parietal, occipital, and temporal cortices (Allen et al., 2007). The CEN, especially its the dorsolateral prefrontal cortex node, is structurally connected with the occipital, parietal and temporal cortices, and therefore this network is well situated to support a broad range of cognitive processes, such as integrating information from the external environment with the stored internal representation during executive decision‐making processes (Chand & Dhamala, 2017; Chand et al., 2016; Miller & Cohen, 2001; Petrides & Pandya, 1999). Our findings and former reports might indicate collectively that the modulating effects of the DAN with the HCN and CEN are crucial in supporting normal cognitive functions.

In MCI and/or Alzheimer's disease, emerging reports demonstrate that the impaired patterns of interactions between the large‐scale networks (Chand, Wu, Qiu, et al., 2017c; Greicius & Kimmel, 2012; Wang et al., 2015). A large body of functional MRI and/or PET analyses have consistently illustrated that the DAN—a key system for top‐down processes—is impaired but not the ventral subsystem—a key system for bottom‐up processes—in the MCI and/or Alzheimer's disease pathology (Bokde et al., 2008, 2010; Li et al., 2015, 2012; McIntosh et al., 1994). Functional connections of the HCN are declined with the frontal/prefrontal cortex in Alzheimer's disease (Allen et al., 2007). Amnestic MCI and/or Alzheimer's disease investigations show the declined activity of the HCN—a key memory processing system in the brain—and of the DMN (Allen et al., 2007; Dubois et al., 2016; Greicius & Kimmel, 2012; Greicius et al., 2004). However, the functional role of the CEN activity has remained unclear with the progression of disease (Diener et al., 2012; Li et al., 2015). Our results are broadly in line with recent findings that support a link between functional MRI‐based brain activity and the spatial spreading of Alzheimer's disease‐related pathology (Jones et al., 2016; Mutlu et al., 2017). Those studies suggest that the gray matter atrophy might reflect tau pathology and the atrophy spreads across regions via specific functionally connected pathways. The key hubs of the MCI and/or Alzheimer's disease especially the HCN and the posterior regions of DMN and CEN are suggested to be more vulnerable to amyloid‐beta deposition and glucose hypometabolism (Mutlu et al., 2017). Our results that unaltered modulating features of the DAN over the DMN and HCN in both normal control and MCI groups possibly imply that the interactions between the DMN and the HCN and/or control of the DAN is less affected, at least in the preclinical MCI stage. Our findings that the disrupted modulating features of DAN over the CEN and HCN in the MCI group might also possibly be due to the impaired reciprocal connections between the CEN and the HCN. As Alzheimer's disease pathology triggers a cascading failure of targeted functional networks (Jones et al., 2016), it is worth evaluating in the future that whether there is a progressive alteration in the DAN modulation over the DMN and HCN in a group of Alzheimer's disease patients.

Although we leveraged a novel unified network approach focusing on the HCN, DAN, CEN, and DMN, and found the disrupted functional architectures in amnestic MCI, it is worth to note that our study sample is not relatively high. While the control group and MCI group are age‐matched, the two groups are different in the race (albeit weak). To account for those differences, we leveraged a generalized linear regression of demographics (age, education, sex, and race) and believe that those demographic differences have influenced less in our findings. A recent study (Zhou et al., 2018) in healthy adolescent and young adults investigate interaction patterns among the DMN, DAN, and salience network, and suggest that the salience network resides at the apex of the hierarchy; however, the HCN and CEN were not considered in that analysis. Investigating such hierarchy in the MCI and Alzheimer's disease would be an interesting direction for the future. The main challenge of including the salience network in our current the four networks model would be that the number of possible DCM models increases significantly and the accuracy of DCM model comparison decreases (Stephan et al., 2010).

In summary, we examined the patterns of interactions among the HCN, DAN, CEN, and DMN in elderly people with MCI and compared with a normal control group. We revealed that the DAN modulates the activity between the HCN and the DMN in both MCI and normal control groups; however, the DAN could not modulate the reciprocal connections between the HCN and the CEN in MCI. Those findings help us better understand the alternations in network patterns in MCI/Alzheimer's disease pathology.

CONFLICT OF INTEREST

The authors declare that they have no conflicts of interest to declare.

ACKNOWLEDGMENTS

The authors thank BSHARP team members and the volunteers for their participation. National Institutes of Health grants RF1AG051633 and R01AG042127 supported to the author IH. National Institutes of Health grants AG25688, AG42127, AG49752, AG51633 supported to the author DQ.

Chand GB, Hajjar I, Qiu D. Disrupted interactions among the hippocampal, dorsal attention, and central‐executive networks in amnestic mild cognitive impairment. Hum Brain Mapp. 2018;39:4987–4997. 10.1002/hbm.24339

Funding information National Institutes of Health, Grant/Award Numbers: AG51633, AG49752, AG42127, AG25688; National Institutes of Health, Grant/ Award Numbers: R01AG042127, RF1AG051633

REFERENCES

  1. Allen, G. , Barnard, H. , McColl, R. , Hester, A. L. , Fields, J. A. , Weiner, M. F. , … Cullum, C. M. (2007). Reduced hippocampal functional connectivity in Alzheimer disease. Archives of Neurology, 64, 1482–1487. [DOI] [PubMed] [Google Scholar]
  2. Biswal, B. , Haughton, V. M. , & Hyde, J. (1995). Functional connectivity in the motor cortex resting human brain using echo‐planar MRI. Magnetic Resonance in Medicine, 34, 537–541. [DOI] [PubMed] [Google Scholar]
  3. Bokde, A. L. , Lopez‐Bayo, P. , Born, C. , Dong, W. , Meindl, T. , Leinsinger, G. , … Hampel, H. (2008). Functional abnormalities of the visual processing system in subjects with mild cognitive impairment: An fMRI study. Psychiatry Research, 163, 248–259. [DOI] [PubMed] [Google Scholar]
  4. Bokde, A. L. , Lopez‐Bayo, P. , Born, C. , Ewers, M. , Meindl, T. , Teipel, S. J. , … Hampel, H. (2010). Alzheimer disease: Functional abnormalities in the dorsal visual pathway. Radiology, 254, 219–226. [DOI] [PubMed] [Google Scholar]
  5. Bressler, S. L. , & Menon, V. (2010). Large‐scale brain networks in cognition: Emerging methods and principles. Trends in Cognitive Sciences, 14, 277–290. [DOI] [PubMed] [Google Scholar]
  6. Brier, M. R. , Thomas, J. B. , & Ances, B. M. (2014). Network dysfunction in Alzheimer's disease: Refining the disconnection hypothesis. Brain Connectivity, 4, 299–311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Brissenden, J. A. , Levin, E. J. , Osher, D. E. , Halko, M. A. , & Somers, D. C. (2016). Functional evidence for a cerebellar node of the dorsal attention network. The Journal of Neuroscience, 36, 6083–6096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Buckner, R. L. , Andrews‐Hanna, J. R. , & Schacter, D. L. (2008). The brain's default network: Anatomy, function, and relevance to disease. Annals of the New York Academy of Sciences, 1124, 1–38. [DOI] [PubMed] [Google Scholar]
  9. Chand, G. B. , & Dhamala, M. (2016a). Interactions among the brain default‐mode, salience, and central‐executive networks during perceptual decision‐making of moving dots. Brain Connectivity, 6, 249–254. [DOI] [PubMed] [Google Scholar]
  10. Chand, G. B. , & Dhamala, M. (2016b). The salience network dynamics in perceptual decision‐making. NeuroImage, 134, 85–93. [DOI] [PubMed] [Google Scholar]
  11. Chand, G. B. , & Dhamala, M. (2017). Interactions between the anterior cingulate‐insula network and the fronto‐parietal network during perceptual decision‐making. NeuroImage, 152, 381–389. [DOI] [PubMed] [Google Scholar]
  12. Chand, G. B. , Lamichhane, B. , & Dhamala, M. (2016). Face or house image perception: Beta and Gamma bands of oscillations in brain networks carry out decision‐making. Brain Connectivity, 6(8), 621–631. [DOI] [PubMed] [Google Scholar]
  13. Chand, G. B. , Wu, J. , Hajjar, I. , & Qiu, D. (2017a). Interactions of insula subdivisions‐based networks with default‐mode and central‐executive networks in mild cognitive impairment. Frontiers in Aging Neuroscience, 9, 367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Chand, G. B. , Wu, J. , Hajjar, I. , & Qiu, D. (2017b). Interactions of the salience network and its subsystems with the default‐mode and the central‐executive networks in normal aging and mild cognitive impairment. Brain Connectivity, 7, 401–412. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Chand, G. B. , Wu, J. , Qiu, D. , & Hajjar, I. (2017c). Racial differences in insular connectivity and thickness and related cognitive impairment in hypertension. Frontiers in Aging Neuroscience, 9, 177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Craddock, R. C. , James, G. A. , Holtzheimer, P. E., 3rd , Hu, X. P. , & Mayberg, H. S. (2012). A whole brain fMRI atlas generated via spatially constrained spectral clustering. Human Brain Mapping, 33, 1914–1928. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Daunizeau, J. , Stephan, K. E. , & Friston, K. J. (2012). Stochastic dynamic causal modelling of fMRI data: Should we care about neural noise? NeuroImage, 62, 464–481. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. De Rover, M. , Pironti, V. A. , McCabe, J. A. , Acosta‐Cabronero, J. , Arana, F. S. , Morein‐Zamir, S. , … Sahakian, B. J. (2011). Hippocampal dysfunction in patients with mild cognitive impairment: A functional neuroimaging study of a visuospatial paired associates learning task. Neuropsychologia, 49, 2060–2070. [DOI] [PubMed] [Google Scholar]
  19. Deco, G. , Jirsa, V. K. , & McIntosh, A. R. (2011). Emerging concepts for the dynamical organization of resting‐state activity in the brain. Nature Reviews. Neuroscience, 12, 43–56. [DOI] [PubMed] [Google Scholar]
  20. Diener, C. , Kuehner, C. , Brusniak, W. , Ubl, B. , Wessa, M. , & Flor, H. (2012). A meta‐analysis of neurofunctional imaging studies of emotion and cognition in major depression. NeuroImage, 61, 677–685. [DOI] [PubMed] [Google Scholar]
  21. Dong, Y. , Sharma, V. K. , Chan, B. P. , Venketasubramanian, N. , Teoh, H. L. , Seet, R. C. , … Chen, C. (2010). The Montreal cognitive assessment (MoCA) is superior to the mini‐mental state examination (MMSE) for the detection of vascular cognitive impairment after acute stroke. Journal of the Neurological Sciences, 299, 15–18. [DOI] [PubMed] [Google Scholar]
  22. Dubois, B. , Hampel, H. , Feldman, H. H. , Scheltens, P. , Aisen, P. , Andrieu, S. , … Jack, C. R., Jr. (2016). Preclinical Alzheimer's disease: Definition, natural history, and diagnostic criteria. Alzheimers Dement, 12, 292–323. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Fox, M. D. , Buckner, R. L. , Liu, H. , Chakravarty, M. M. , Lozano, A. M. , & Pascual‐Leone, A. (2014). Resting‐state networks link invasive and noninvasive brain stimulation across diverse psychiatric and neurological diseases. Proceedings of the National Academy of Sciences of the United States of America, 111, E4367–E4375. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Fox, M. D. , Corbetta, M. , Synder, A. Z. , Vincent, J. L. , & Raichle, M. E. (2006). Spontaneous neuronal activity distinguishes human dorsal and ventral attention systems. Proceedings of the National Academy of Sciences of the United States of America, 103, 10046–10051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Fox, M. D. , Snyder, A. Z. , Vincent, J. L. , Corbetta, M. , Van Essen, D. C. , & Raichle, M. E. (2005). The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proceedings of the National Academy of Sciences of the United States of America, 102, 9673–9678. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Franzmeier, N. , Duering, M. , Weiner, M. W. , Dichgans, M. , & Ewers, M. (2017). Left frontal cortex connectivity underlies cognitive reserve in prodromal Alzheimer disease. Neurology, 88, 1054–1061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Friston, K. J. , Harrison, L. , & Penny, W. (2003). Dynamic causal modelling. NeuroImage, 19, 1273–1302. [DOI] [PubMed] [Google Scholar]
  28. Friston, K. J. , Kahan, J. , Biswal, B. , & Razi, A. (2014). A DCM for resting state fMRI. NeuroImage, 94, 396–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Friston, K. , Moran, R. , & Seth, A. K. (2013). Analysing connectivity with granger causality and dynamic causal modelling. Current Opinion in Neurobiology, 23, 172–178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Gao, W. , & Lin, W. (2012). Frontal parietal control network regulates the anti‐correlated default and dorsal attention networks. Human Brain Mapping, 33, 192–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Goulden, N. , Khusnulina, A. , Davis, N. J. , Bracewell, R. M. , Bokde, A. L. , McNulty, J. P. , & Mullins, P. G. (2014). The salience network is responsible for switching between the default mode network and the central executive network: Replication from DCM. NeuroImage, 99, 180–190. [DOI] [PubMed] [Google Scholar]
  32. Greicius, M. D. , & Kimmel, D. L. (2012). Neuroimaging insights into network‐based neurodegeneration. Current Opinion in Neurology, 25, 727–734. [DOI] [PubMed] [Google Scholar]
  33. Greicius, M. D. , Srivastava, G. , Reiss, A. L. , & Menon, V. (2004). Default‐mode network activity distinguishes Alzheimer's disease from healthy aging: Evidence from functional MRI. Proceedings of the National Academy of Sciences of the United States of America, 101, 4637–4642. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Hachinski, V. , Iadecola, C. , Petersen, R. C. , Breteler, M. M. , Nyenhuis, D. L. , … Leblanc, G. G. (2006). National Institute of Neurological Disorders and Stroke‐Canadian stroke network vascular cognitive impairment harmonization standards. Stroke, 37, 2220–2241. [DOI] [PubMed] [Google Scholar]
  35. Horwitz, B. , McIntosh, A. R. , Haxby, J. V. , Furey, M. , Salerno, J. A. , Schapiro, M. B. , … Grady, C. L. (1995). Network analysis of PET‐mapped visual pathways in Alzheimer type dementia. Neuroreport, 6(17), 2287–2292. [DOI] [PubMed] [Google Scholar]
  36. Huijbers, W. , Pennartz, C. M. , Cabeza, R. , & Daselaar, S. M. (2011). The hippocampus is coupled with the default network during memory retrieval but not during memory encoding. PLoS One, 6, e17463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Jones, D. T. , Knopman, D. S. , Gunter, J. L. , Graff‐Radford, J. , Vemuri, P. , Boeve, B. F. , … Alzheimer's Disease Neuroimaging Imaging . (2016). Cascading network failure across the Alzheimer's disease spectrum. Brain, 139, 547–562. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Li, H. J. , Hou, X. H. , Liu, H. H. , Yue, C. L. , He, Y. , & Zuo, X. N. (2015). Toward systems neuroscience in mild cognitive impairment and Alzheimer's disease: A meta‐analysis of 75 fMRI studies. Human Brain Mapping, 36, 1217–1232. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Li, R. , Wu, X. , Fleisher, A. S. , Reiman, E. M. , Chen, K. , & Yao, L. (2012). Attention‐related networks in Alzheimer's disease: A resting functional MRI study. Human Brain Mapping, 33, 1076–1088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Liang, P. , Li, Z. , Deshpande, G. , Wang, Z. , Hu, X. , & Li, K. (2014). Altered causal connectivity of resting state brain networks in amnesic MCI. PLoS One, 9, e88476. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Lin, Q. H. , Liu, J. , Zheng, Y. R. , Liang, H. , & Calhoun, V. D. (2010). Semiblind spatial ICA of fMRI using spatial constraints. Human Brain Mapping, 31, 1076–1088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. McIntosh, A. R. , Grady, C. L. , Ungerleider, L. G. , Haxby, J. V. , Rapoport, S. I. , & Horwitz, B. (1994). Network analysis of cortical visual pathways mapped with PET. The Journal of Neuroscience, 14(2), 655–666. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Menon, V. (2011). Large‐scale brain networks and psychopathology: A unifying triple network model. Trends in Cognitive Sciences, 15, 483–506. [DOI] [PubMed] [Google Scholar]
  44. Menon, V. (2015). Salience network. Brain Mapping: An Enclopedic Reference, 2, 597–611. [Google Scholar]
  45. Miller, E. K. , & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual Review of Neuroscience (Palo Alto, CA), 24, 167–202. [DOI] [PubMed] [Google Scholar]
  46. Mutlu, J. , Landeau, B. , Gaubert, M. , de La Sayette, V. , Desgranges, B. , & Chetelat, G. (2017). Distinct influence of specific versus global connectivity on the different Alzheimer's disease biomarkers. Brain, 140, 3317–3328. [DOI] [PubMed] [Google Scholar]
  47. Nasreddine, Z. S. , Phillips, N. A. , Bedirian, V. , Charbonneanu, S. , Whitehead, V. , Collin, I. , … Chertkow, H. (2005). The Montreal cognitive Assement, MoCA: A brief screening tool for mild cognitive impairment. Journal of the American Geriatrics Society, 53, 695–699. [DOI] [PubMed] [Google Scholar]
  48. Petrides, M. , & Pandya, D. N. (1999). Dorsolateral prefrontal cortex: Comparative cytoarchitectonic analysis in the human and the macaque brain and corticocortical connection patterns. The European Journal of Neuroscience, 11, 1011–1036. [DOI] [PubMed] [Google Scholar]
  49. Power, J. D. , Cohen, A. L. , Nelson, S. M. , Wig, G. S. , Barnes, K. A. , Church, J. A. , … Petersen, S. E. (2011). Functional network organization of the human brain. Neuron, 72, 665–678. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Raichle, M. E. (2015). The restless brain: How intrinsic activity organizes brain function. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 370, 20140172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Rohr, C. S. , Vinette, S. A. , Parsons, K. A. , Cho, I. Y. , Dimond, D. , Benischek, A. , … Bray, S. (2016). Functional connectivity of the dorsal attention network predicts selective attention in 4‐7 year‐old girls. Cerebral Cortex, 27(9), 4350–4360. [DOI] [PubMed] [Google Scholar]
  52. Rossetti, H. C. , Lacritz, L. H. , Cullum, C. M. , & Weiner, M. F. (2011). Normative data for the Montreal cognitive assessment (MoCA) in a population‐based sample. Neurology, 77, 1272–1275. [DOI] [PubMed] [Google Scholar]
  53. Shirer, W. R. , Ryali, S. , Rykhlevskaia, E. , Menon, V. , & Greicius, M. D. (2012). Decoding subject‐driven cognitive states with whole‐brain connectivity patterns. Cerebral Cortex, 22, 158–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Smith, S. M. , Miller, K. L. , Salimi‐Khorshidi, G. , Webster, M. , Beckmann, C. F. , Nichols, T. E. , … Woolrich, M. W. (2011). Network modelling methods for FMRI. NeuroImage, 54, 875–891. [DOI] [PubMed] [Google Scholar]
  55. Sridharan, D. , Levitin, D. J. , & Menon, V. (2008). A critical role for the right fronto‐insular cortex in switching between central‐executive and default‐mode networks. Proceedings of the National Academy of Sciences of the United States of America, 105, 12569–12574. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Stephan, K. E. , Penny, W. D. , Moran, R. J. , den Ouden, H. E. , Daunizeau, J. , & Friston, K. J. (2010). Ten simple rules for dynamic causal modeling. NeuroImage, 49, 3099–3109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Uddin, L. Q. (2015). Salience processing and insular cortical function and dysfunction. Nature Reviews. Neuroscience, 16, 55–61. [DOI] [PubMed] [Google Scholar]
  58. Vincent, J. L. , Kahn, I. , Snyder, A. Z. , Raichle, M. E. , & Buckner, R. L. (2008). Evidence for a frontoparietal control system revealed by intrinsic functional connectivity. Journal of Neurophysiology, 100, 3328–3342. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Wang, P. , Zhou, B. , Yao, H. , Zhan, Y. , Zhang, Z. , Cui, Y. , … Jiang, T. (2015). Aberrant intra‐ and inter‐network connectivity architectures in Alzheimer's disease and mild cognitive impairment. Scientific Reports, 5, 14824. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Zanto, T. P. , Rubens, M. T. , Thangavel, A. , & Gazzaley, A. (2011). Causal role of the prefrontal cortex in top‐down modulation of visual processing and working memory. Nature Neuroscience, 14, 656–661. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Zhang, D. , Wang, Y. , Zhou, L. , Yuan, H. , Shen, D. , & Alzheimer's Disease Neuroimaging Imaging . (2011). Multimodal classification of Alzheimer's disease and mild cognitive impairment. NeuroImage, 55, 856–867. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Zhou, Y. , Friston, K. J. , Zeidman, P. , Chen, J. , Li, S. , & Razi, A. (2018). The hierarchical organization of the default, dorsal attention and salience networks in adolescents and young adults. Cerebral Cortex, 28, 726–737. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Human Brain Mapping are provided here courtesy of Wiley

RESOURCES