Impaired Memory Awareness and Loss Integration in Self-Referential Network Across the Progression of Alzheimer’s Disease Spectrum

Abstract

Background:

Anosognosia, or unawareness of memory deficits, is a common manifestation of Alzheimer’s disease (AD), but greatly variable in subjective cognitive decline (SCD) and amnestic mild cognitive impairment (aMCI) subjects. Self-referential network (SRN) is responsible for self-referential processing and considered to be related to AD progression.

Objective:

Our aim is to explore connectivity changes of SRN and its interaction with memory-related network and primary sensorimotor network (SMN) in the AD spectrum.

Methods:

About 444 Alzheimer’s Disease Neuroimaging Initiative subjects (86 cognitively normal [CN]; 156 SCD; 146 aMCI; 56 AD) were enrolled in our study. The independent component analysis (ICA) method was used to extract the SRN, SMN, and memory-related network from all subjects. The alteration of functional connectivity (FC) within SRN and its connectivity with memory-related network/SMN were compared among four groups and further correlation analysis between altered FC and memory awareness index as well as episodic memory score were performed.

Results:

Compared with CN group, individuals with SCD exhibited hyperconnectivity within SRN, while aMCI and AD patients showed hypoconnectivity. Furthermore, aMCI patients and AD patients both showed the interruption of the FC between the SRN and memory-related network compared to CN group. Pearson correlation analysis showed that disruptive FC within SRN and its interaction with memory-related network were related to memory awareness index and episodic memory scores.

Conclusion:

In conclusion, impaired memory awareness and episodic memory in the AD spectrum are correlated to the disconnection within SRN and its interaction with memory-related network.

Keywords

Alzheimer’s disease amnestic mild cognitive impairment anosognosia memory awareness index memory network network interaction self-referential network subjective cognitive decline

INTRODUCTION

Lack of awareness of cognitive decline, or anosognosia, is a common symptom of people with Alzheimer’s disease (AD). AD patients with anosognosia specifically refer to unawareness of the illness or deficits in daily life [1, 2]. Anosognosia is also correlated with the severity of AD [3, 4]. Thus, it is essential to pay attention to the clinical characteristic in AD patients. Amnestic mild cognitive impairment (aMCI) is the intermediate stage between normal age-related cognitive decline and dementia [5]. Previous studies have demonstrated that aMCI, characterized by memory deficits, has a higher risk of developing into AD dementia [6]. Subjective cognitive decline (SCD), as a preclinical stage of aMCI, is characterized by subjective complaints of memory deterioration without objective cognitive deficits [7, 8]. Although some studies suggested that SCD may merely be a feature of healthy cognitive aging [9]. Increasing evidence indicated that the progression of AD may start from an important first sign of subjective memory complaints to aMCI, and eventually develop into AD dementia [10 –12].

Previous studies have not reached consensus on the alteration of cognitive awareness in MCI patients. They have a highly variable understanding of memory dysfunction, ranging from heightened cognitive awareness and obvious attention to completely unawareness of their cognitive deficits [3, 4]. Moreover, studies found that anosognosia is an independent predictor of the conversion from aMCI to AD dementia in aMCI patients [13]. A large cross-sectional and longitudinal study recently published found that cognitive awareness increased 1.6 years before the diagnosis of MCI and reached to anosognosia 3.2 years before the diagnosis of dementia [4]. This trend is consistent with a previous view [14, 15] that the cognitive awareness varies on a continuum, manifested as normal cognitive awareness in the beginning, followed by a stage of heightened cognitive awareness without cognitive decline, and eventually unawareness. Hence, exploring the alteration of cognitive awareness and its neuropathology in preclinical and prodromal stage of AD is essential for a better understanding the progression of AD.

Self-referential processing is the core of oneself and is critical for elaborating experiential feelings of self. It benefits from the process of self-reference encoding, which links the information involving in memory with the database of the self [16]. Previous studies evinced that anosognosia or a lack of self-awareness was correlated to the impairment of self-referential process [2, 17]. Self-referential network (SRN), extracted from resting state using independent component analysis (ICA) method in functional magnetic resonance imaging (fMRI), is closely associated with self-referential processing [18 –20]. Anatomically, the SRN was mainly located in the midline structures, including the ventromedial prefrontal cortex (VMPFC), the medial orbital prefrontal cortex (MOPFC), the anterior cingulate cortex (ACC), and the gyrus rectus. Neuroimaging studies also verified that anosognosia in AD patients was correlated with the hypoconnectivity and hypoperfusion in regions of SRN [21, 22]. For example, a study found that anosognosia in AD patients was related to dysfunction within midline structures as well as the disconnection of these regions with the medial temporal lobe [21]. But the neural correlates of lack of self-reference was not fully understood and even inconsistent in SCD and aMCI subjects. Some research reported hypoconnectivity in midline structures in aMCI and SCD subjects [23, 24], while another study found hyperconnectivity of these regions in aMCI patients [19]. Therefore, the exploration of alteration of memory awareness and related core networks are crucial to reveal the neuropathology of AD.

It is a remarkable fact that there is the overlap of anatomical location between the SRN and the default mode network (DMN). DMN is related to many regions associated with the posterior cingulate cortex, precuneus, medial prefrontal cortex, superior frontal gyrus, inferior parietal lobule, lateral temporal cortex, hippocampus, and cerebellum, and involves in a great many different mental activities including autobiographical memory, self-referential processing, maintaining awareness, and environmental monitoring in resting state [25 –27]. An increasing number of studies revealed that the DMN is not unitary but consist of heterogenous subsystems, implying that the SRN may be a subnetwork of the DMN [28, 29]. However, the SRN and the DMN identified by the ICA method have different BOLD signal spatial patterns and electrophysiological signals [20, 30] suggested that the two networks are substantially different. Anyway, it is essential to investigate the changes of the SRN and its underlying mechanisms in AD spectrum.

The formation of episodic memory relies on a specific medial prefrontal cortex - hippocampus circuit that integrates information of object, place, and time to construct episodic memory [31]. A meta-analysis pointed out that the regions in SRN in addition to involving in the self-reference processing, are also indispensable in integrating different function, such as the linkage of higher cognitive function and primary/secondary sensorimotor function [32 –34]. For example, the self-related information transferred from exteroceptive sensorimotor modalities is processed in these regions to generate the self-reference, and eventually be integrated into the autobiographical memory in the context of self [17, 35]. Thus, the interaction of self-reference processing and memory encoding is vital in the formation of episodic memory. In addition, the primary sensorimotor network (SMN) plays an important role in transforming a single sensory model into complex self-referential processing by receiving and transmitting self-related information to relevant brain regions [36]. Therefore, the exploration of the altered functional connectivity (FC) of inter-networks related to self-reference processing, sensorimotor processing and advanced memory processing was essential in AD spectrum.

In short, our purpose was to explore the characteristics of memory awareness and the FC within SRN as well as its interaction with memory-related network and SMN in the AD spectrum. ICA, a data-driven processing method identifying brain networks with unique spatio-temporal patterns, was used to separate SRN, SMN, and memory-related network in subjects. Our primary hypotheses were as follows: Firstly, different stages of the AD continuum have characteristic alterations of memory awareness. Secondly, there exists FC alterations within SRN as well as its interaction with memory-related network across the AD continuum. In addition, the FC alteration within SRN and inter-networks interaction may be associated with memory awareness index and episodic memory score.

MATERIAL AND METHODS

Participants

All subjects were obtained within the Alzheimer’s Disease Neuroimaging Initiative (ADNI; recruitment phase 2/3, ClinicalTrials.gov ID: NCT02854033). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging, positron emission tomography, other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of aMCI and early AD. For up-to-date information, see http://www.adni-info.org. The ADNI recruited participants ranging from 55 to 94 years from multicenter study conducted at 59 locations in North America. Data for the present study came from ADNI2 and ADNI3, and we downloaded the data from the website (http://adni.loni.usc.edu) on October 15, 2020. Then, participants in the baseline period with demographic information, Everyday Cognition scale (ECog), and fMRI scanner were selected in the downloaded data. Finally, a total of 444 individuals were available for the analysis, including 86 individuals with CN, 156 individuals with SCD, 146 individuals with aMCI, and 56 individuals with AD. The inclusion and exclusion criteria of CN, SCD, aMCI, and AD, as well as detailed evaluation criteria, can be obtained in the regular procedure manual on the ADNI website (http://adni.loni.usc.edu/data-samples/access-data/). We also added the inclusion and exclusion criteria in the Supplementary Material.

Self-awareness of memory

As previously used in other ADNI studies [3, 4], the memory awareness index was evaluated by ECog. The ECog is a 39-item specifically used to assess the subjective cognitive decline and daily function of the elderly [37, 38]. The version of ECog of study partner and participant is consist of the same questions which compare participants’ cognitive level with 10 years ago. Consistent with previous studies, we used the memory function subtest of the ECog to evaluate the memory awareness level. The discrepancy score between the participant self-rated memory (EcogPtMem) scores and study partner reporting memory scores (EcogSPMem) in ‘8 items of daily memory’ subtest is memory awareness index. Using the discrepancy scores, a positive score indicates the overestimation of their memory function or a high level of awareness, which means that these people believe that their memory level is higher than the level evaluated by their partner. On the contrary, a negative score indicates low memory awareness, which means that these people think that their memory functions were worse than estimation of their partners. The information study partner reported have been proved to be highly accurate [9, 39].

Cognitive assessments

Global cognitive abilities were obtained by the Mini-Mental State Examination (MMSE) test in the present study. Geriatric Depression Scale (GDS) was dedicated to screening for depression in the elderly, and GDS score > 5 points is considered to have depression performance [40]. Episodic memory was evaluated by the composite score for memory (ADNI-Mem). ADNI-Mem was calculated by combining the specific items of the Rey Auditory Verbal Learning Test (RAVLT), the Alzheimer Disease Assessment Scale-Cognitive (ADAS-Cog), Logical Memory, and MMSE data. The RAVLT includes immediate and delayed recall of a 15-item list of irrelevant words. The ADAS-Cog includes a word list learning task and a word recognition task. For logical memory, the immediate and delayed recall of a brief fact-laden passage are included. The MMSE includes the item of recall of 3 words after interference. Detailed information can be obtained on ADNI website (https://ida.loni.usc.edu/pages/access/studyData.jsp).

MRI acquisition

All MRI scans were acquired on a 3.0T scanner applying unified scanning protocols obtaining from different manufacturers, including Siemens (Munich, Germany), General Electric (Cleveland, OH) and Philips (Best in the Netherlands). More detailed information can be obtained from http://adni.loni.usc.edu/wp-content/uploads/2010/05/ADNI2_MRI_Training_Manual_FINAL.pdf and http://adni.loni.usc.edu/wp-content/uploads/2017/07/ADNI3-MRI-protocols.pdf. In short, a rapid gradient echo sequence prepared by 3DT1 weighted magnetization (isotropic voxel resolution of 1 mm, repetition time (TR = 2,300 ms) was used to record structural MRI. FMRI were recorded with a 3.4 mm isotropic voxel resolution (TR / echo time / flip angle = 3,000 / 30/90°) using a 3D echo planar imaging (EPI) sequence.

Image preprocessing

Statistical parametric mapping (SPM8; http://www.fil.ion.ucl.ac.uk/spm) and the Data Processing Assistant for resting-state fMRI (DPARSF, http://restfmri.net/forum/DPARSF) was used to perform all imaging preprocessing steps. Date from 24 subjects were excluded due to their translation or rotation exceeded 3 mm and 3°, including 4 CN subjects, 5 SCD subjects, 8 aMCI subjects, and 7 AD subjects.

Structural MRI preprocessing

First, the DICOM format images were converted to NIFTI format, and the MPRAGE images were segmented to obtain gray matter, white matter, and cerebrospinal fluid images; the segmented images were standardized to the Montreal Neurological Institute (Montreal Neurological Institute, MNI) standard space, voxel 1.5 mm×1.5 mm×1.5 mm; a smoothing kernel with a full width at half maximum (FWHM) of 8 mm×8 mm×8 mm was used for spatial smoothing to obtain gray matter maps.

Resting-state fMRI preprocessing

The first 10 volumes of each functional time series were discarded before slice time and realignment. The fMRI images were then spatially normalized to the MNI echo-planar imaging template and resampled to a default setting (3×3×3 mm³ voxels). Then, a Gaussian kernel of 8×8×8 mm FWHM was performed to decrease high spatial frequency noise [41]. Though the recruitment of ADNI participants was carried on different MRI scanners in multicenter pattern, previous study found no significant differences in signal to noise ratio of resting state fMRI between scanners [42].

Independent component analysis and identification

ICA is a blind source signal separation method that has been widely used to identify and quantify related activity distribution patterns or spatial networks. All pre-processed date was performing by the informix algorithm in the Group ICA of fMRI Toolbox (GIFT) software (http://icatb.sourceforge.net). GIFT software broke the data into 20 spatially separated components since there’s no standard number of components [43, 44]. Then, principal component analysis (PCA) was used to reduce the dimension of each data set. A single ICA was performed on each subject and a time course and space maps of the single subject were reconstructed from the original data. The intensity values in each graph were converted to a Z values to show the voxel associated with a particular independent component (IC). For each IC, a specific coherent waveform of brain activity corresponds to different time courses, and the principle of associated spatial map was driven by the intensity of this pattern of brain activity penetrating the voxel [27].

After ICs separation, all ICs were spatially sorted in GIFT toolbox matching with standard templates deriving from previous investigations. Based on spatial overlap of ICs with these templates, the most appropriate components for each subject are chose. The standard SRN template was consisted of VMPFC, MOPFC, ACC, and the gyrus rectus [18 , 30]. The standard memory-related network template contained bilateral retrosplenial cortex, posterior cingulate, precuneus, hippocampus, and entorhinal cortex [45]. The standard SMN template contained bilateral primary somatosensory and motor cortices, secondary somatosensory cortex, and the supplementary motor area [46]. These templates were derived from previous research and combined with spatial correlation as selection criterion.

In this study, we used the WFU PickAtlas Tool (The Functional MRI Laboratory, Wake Forest University School of Medicine, Winston-Salem, North Carolina; http://www.ansir.wfubmc.edu) to create these templates.

Although these components in ICA are spatially independent, there is a significant temporal correlation between them. The temporal correlation between ICs reflects the FC between the resting state networks. Thus, in order to get the interaction between the SRN-IC (self-referential processing) and memory-related network-IC (higher-order memory processing), SMN-IC (primary sensorimotor processing), we calculated the temporal correlation coefficients of the functional networks as a correlation among the three networks [47]. The correlation coefficient was measured as the FC strength of inter-networks.

Statistical analysis

The Statistical Package for the Social Sciences (SPSS) software version 22.0 (IBM, Armonk, New York, NY, USA) was used for all statistical analysis. The Kruskal-Wallis test and the chi-square test were performed to compare the differences of demographic and neuropsychological characteristics among four groups. The Mann-Whiney U Tests was used for post hoc analysis with a p-value < 0.05.

Within network

The most fit components on behalf of SRN were extracted from all subjects. In order to obtain the pattern of SRN in each group, we gathered each group separately to do random-effect analysis using one-sample t-tests. The thresholds were set at p < 0.05, determined by false discovery rate (FDR) correction. One-way ANOVA was performed across four groups to obtain FC difference of SRN between AD spectrum patients and normal people after controlling for the effects of age, gender, education level, GDS scores, and grey matter volume (TFCE-FWE corrected, p < 0.01). Post hoc comparisons using the two-sample T-test to identify detailed changes in the patterns of SRN between groups after controlling for age, gender, education level, GDS scores, and grey matter volume as covariates (TFCE-FWE corrected, p < 0.01). Although we have excluded subjects with severe depression (GDS > 5 scores), memory awareness is closely related to emotions. Thus, GDS scores is also controlled as a covariate. The voxel-based correlation analysis was exerted to assess the relationship between altered FC within SRN and the memory awareness index/episodic memory score between groups (Bonferroni-corrected, p < 0.05, correction for age, gender, education, and GDS scores).

Inter-networks

Pearson’s correlation analysis was used to compute the correlation coefficient of 3 ICs of interest. The correlation coefficient of inter-networks FC (SRN and memory-related network, SRN and SMN) were compared among four groups by one-way ANOVA, and the post-hoc test was obtained by Bonferroni correction (p < 0.05). The Pearson correlation analysis was used to calculate the association between correlation values of inter-networks FC and episodic memory score/memory awareness index (Bonferroni-corrected, p < 0.05, correction for age, gender, education, and GDS scores).

RESULTS

Demographic and neurocognitive characteristics

Table 1 summarized the demographic and clinical characteristics of 444 participants in four groups, including 86 CN (mean age 70.77±6.02), 156 SCD (mean age 71.49±6.37), 146 aMCI (mean age 71.84±7.56), and 56 AD (mean age 74.09±7.66) individuals. Compared to CN group, AD group were older, higher proportion of the male, higher GDS score, lower MMSE score, lower episodic memory score, higher EcogPtMem score, higher EcogSPMem score and lower memory awareness index; aMCI group showed significantly higher proportion of male, higher GDS score, lower MMSE score, lower episodic memory score, higher EcogPtMem score, and higher EcogSPMem score; SCD group exhibited significantly higher GDS score, higher EcogPtMem score and higher memory awareness index. Compared to SCD group, AD group were older, higher proportion of male, higher GDS score, lower education level, lower MMSE score, lower episodic memory score, higher EcogPtMem score, higher EcogSPMem score, and lower memory awareness index; aMCI group had significantly higher proportion of male, higher GDS score, lower MMSE score, lower episodic memory score, higher EcogPtMem score, higher EcogSPMem score and lower memory awareness index. Compared to aMCI group, the AD group exhibited significantly lower MMSE, lower episodic memory score, higher EcogSPMem score, and lower memory awareness index.

Table 1

Demographic and neurocognitive characteristics of CN and patients with SCD, aMCI, and AD

	CN (n = 86)	SCD (n = 156)	aMCI (n = 146)	AD (n = 56)	p
Age (y)	70.77±6.02^a	71.49±6.37^a	71.84±7.56	74.09±7.66	0.011
Gender (female, %)	55 (64) ^a,b	95 (61) ^a,b	63 (43)	24 (43)	0.001
Education (y)	16.71±2.11	16.81±2.42^a	16.29±2.48	15.64±2.42	0.006
MMSE	29.10±1.12^a,b	29.17±1.02^a,b	27.70±1.96^a	22.50±2.30	< 0.001
GDS	0.34±0.68^a,b,c	0.92±1.22^a,b	1.87±11.65	1.86±1.43	< 0.001
EcogPtMem	1.40±0.31^a,b,c	1.76±0.51^a,b	2.30±0.72	2.37±0.76	< 0.001
EcogSPMem	1.21±0.25^a,b	1.35±0.43^a,b	2.26±0.83^a	3.34±0.65	< 0.001
Memory awareness index	0.18±0.37^a,c	0.42±0.52^a,b	0.05±0.89^a	–0.98±1.13	< 0.001
Episodic memory score	1.12±0.50^a,b	0.93±0.56^a,b	0.31±0.53^a	–0.92±0.51	< 0.001

Numbers are given as means±standard deviation unless stated otherwise. The baseline demographic and clinical characteristics of four comparisons were evaluated by Kruskal-Wallis test and chi-square test, and post-hoc test was obtained by Mann-Whiney U Test. CN, cognitively normal; SCD, subjective cognitive decline; aMCI, amnestic mild cognitive impairment; AD, Alzheimer’s disease; MMSE, Mini-Mental State Examination; GDS, Geriatric Depression Scale; EcogPtMem, participant self-rated memory scores; EcogSPMem, study partner reported memory scores. ^aThere was statistical significance in comparison to AD after post hoc analysis, p < 0.05. ^bThere was statistical significance in comparison to aMCI after post hoc analysis, p < 0.05. ^cThere was statistical significance in comparison to SCD after post hoc analysis, p < 0.05.

Table 2

The differences in functional connectivity among AD spectrum patients and normal people

Regions	Hemi	Peak MNI coordinate			F/t	voxel
		X	Y	Z
ANOVA
MOPFC/ACC/VMPFC/gyrus rectus	B	6	51	–6	302.44	887
CN < SCD
MOPFC	B	–3	51	–6	–19.23	76
CN > SCD
gyrus rectus	L	–12	18	–12	11.28	21
ACC	B	6	21	27	4.91	48
CN < aMCI
MOPFC	L	–3	54	–3	–17.73	26
CN > aMCI
MOPFC	B	6	48	–6	24.95	172
CN < AD
MOPFC	R	12	51	–3	–16.20	93
CN > AD
MOPFC	B	–6	48	–6	12.51	103
ACC	B	–6	36	27	7.43	334
SCD > aMCI
MOPFC	B	6	48	–6	24.95	225
SCD < AD
VMPFC	R	6	57	0	–16.42	38
SCD > AD
VMPFC	B	–3	48	–6	17.00	139
ACC	B	–6	36	27	8.29	476
aMCI < AD
MOPFC	R	6	48	–6	–24.84	131
aMCI > AD
ACC	B	–3	51	–6	15.63	458

The differences in functional connectivity among AD spectrum patients and normal people. The x, y, z coordinates are the primary peak locations in the MNI space. Cluster size > 200 voxels in ANOVA analysis, TFCE-FWE corrected, p < 0.01; Cluster size > 20 voxels in two-sample T-test, TFCE-FWE corrected, p < 0.01. CN, cognitively normal; SCD, subjective cognitive decline; aMCI, amnestic mild cognitive impairment; AD, Alzheimer’s disease; MOPFC, the medial orbital prefrontal cortex; VMPFC, the ventromedial prefrontal cortex; ACC, anterior cingulate cortex; Hemi, hemisphere; L, left hemisphere; R, right hemisphere; B, bilateral hemisphere.

The comparisons of intra-network FC in SRN among CN, SCD, aMCI, and AD group

We obtained a pattern of SRN consisted of VMPFC, MOPFC, ACC, and the gyrus rectus among four groups (Fig. 1). One-way ANOVA analysis showed significantly altered FC in regions of VMPFC, MOPFC, ACC, and the gyrus rectus among the four groups (Fig. 2a, TFCE-FWE corrected, p < 0.01, cluster size > 200 voxels). Post hoc analysis of the FC values extracted from SRN (Fig. 2b) showed that compared to CN group, individuals with SCD exhibited hyperconnectivity (p_corrected < 0.05), individuals with AD exhibited hypoconnectivity (p_corrected < 0.05); Compared to SCD group, individuals with aMCI exhibited hypoconnectivity (p_corrected < 0.05), individuals with AD exhibited hypoconnectivity (p_corrected < 0.05). The peak connectivity was observed in SCD group in the curve resembling the “inverse U-shaped curve”.

Fig. 1

The spatial expression of self-referential network in CN, SCD, aMCI, and AD groups. The pattern of self-referential network consists of MOPFC, VMPFC, ACC, and the gyrus rectus among four groups. The thresholds were set at a corrected p < 0.05, determined by FDR. CN, cognitively normal; SCD, subjective cognitive decline; aMCI, amnestic mild cognitive impairment; AD, Alzheimer’s disease; MOPFC, the medial orbital prefrontal cortex; VMPFC, the ventromedial prefrontal cortex; ACC, anterior cingulate cortex. L, left hemisphere; R, right hemisphere.

Fig. 2

a) One-way ANOVA analysis showed significantly altered brain regions within SRN among AD spectrum patients in comparison to CN group (TFCE-FWE corrected, p < 0.01). Age, sex, education level, GDS scores and gray matter volume were controlled for covariates. b) The FC values within SRN were extracted from four groups for comparisons. Further post hoc analysis showed that compared to CN group, individuals with SCD exhibited hyperconnectivity in SRN, individuals with aMCI and AD exhibited hypoconnectivity in SRN (Bonferroni corrected, p < 0.05); Age, sex, education level, GDS scores and gray matter volume were controlled for covariates. *Bonferroni corrected, p < 0.05; CN, cognitively normal; SCD, subjective cognitive decline; aMCI, amnestic mild cognitive impairment; AD, Alzheimer’s disease; MOPFC, the medial orbital prefrontal cortex; VMPFC, the ventromedial prefrontal cortex; ACC, anterior cingulate cortex; CN, cognitively normal; SCD, subjective cognitive decline; aMCI, amnestic mild cognitive impairment; AD, Alzheimer’s disease; L, left hemisphere; R, right hemisphere.

In order to obtain the discrepant brain regions, the two-sample T test was performed for post hoc comparison in any two groups (Fig. 3, TFCE-FWE corrected, cluster size > 20, p < 0.01). Compared to the CN, SCD group showed significantly increased FC in the bilateral MOPFC, and decreased FC in the left gyrus rectus; aMCI group showed significantly increased FC in the upper part of left MOPFC, and decreased FC in the right MOPFC and the lower part of left MOPFC; AD group showed significantly increased FC in the upper part of right MOPFC, decreased FC in the left MOPFC, the lower part of right MOPFC and the bilateral ACC. Compared to the SCD, aMCI group showed significantly decreased FC in the bilateral MOPFC; AD group showed significantly increased FC in the upper part of right VMPFC, and decreased FC in the left VMPFC, the lower part of right VMPFC, and bilateral ACC. Compared to aMCI group, AD group showed significantly increased FC in the right MOPFC, and decreased FC in the bilateral ACC. Age, gender, education level, GDS scores and gray matter volume covariates were removed in all results.

Fig. 3

Compared to the CN group, SCD group showed significantly increased FC in the bilateral MOPFC, and decreased FC in the left gyrus rectus; aMCI group showed significantly increased FC in the upper part of left MOPFC, and decreased FC in the right MOPFC and the lower part of left MOPFC; AD group showed significantly increased FC in the upper part of right MOPFC, decreased FC in the left MOPFC, the lower part of right MOPFC and the bilateral ACC. Compared to the SCD, aMCI group showed significantly decreased FC in the bilateral MOPFC; AD group showed significantly increased FC in the upper part of right VMPFC, and decreased FC in the left VMPFC, the lower part of right VMPFC, and bilateral ACC. Compared to aMCI group, AD group showed significantly increased FC in the right MOPFC, and decreased FC in the bilateral ACC. Age, gender, education level, GDS scores and gray matter volume covariates were removed in all results (TFCE-FWE corrected, p < 0.01). L, left hemisphere; R, right hemisphere.

The imaging results were visualized using BrainNet Viewer (http://www.nitrc.org/projects/bnv/) [48]. The component 6, 15, and 16 were best-fit with SRN, SMN, and memory-related network after template matching, which are shown in Supplementary Figure 1.

The comparisons of inter-networks FC in CN, SCD, aMCI, and AD groups

The correlation coefficients of the SRN and memory-related network as well as SMN were calculated to reflect the interaction of inter-networks connectivity. After controlling age, sex, education level, and GDS scores, one-way ANOVA analysis showed that there was a significant group difference for the magnitude of correlations between SRN and memory-related network (Fig. 5a, p < 0.05, corrected). Post hoc comparisons showed that aMCI and AD groups exhibited significant differences than CN and SCD groups (Bonferroni-corrected, p < 0.05). The correlation coefficient between SRN and memory-related network also showed a trend of gradually diminishing across the AD spectrum. In detail, the SRN and the memory-related network were positively correlated in the CN and SCD groups. When progressed to the aMCI stage, the correlation coefficient reduced to a negative value, and continued to diminish in AD group.

Fig. 5

a) There was significant difference in inter-network FC among CN, SCD, aMCI, and AD groups using one-way ANOVA (p < 0.001). Further post hoc analysis showed that aMCI and AD group exhibited significant difference than CN and SCD group (Bonferroni corrected, p < 0.05); The SRN and the memory-related network (MN) were positively correlated in the CN and SCD groups and negatively correlated in the aMCI and AD groups; ^* Bonferroni corrected, p < 0.05; Error bar: 95%confidence intervals (CIs). b) In the AD spectrum patients, a positive correlation between the inter-network FC and episodic memory score (r = 0.232, p < 0.001); c) In the AD spectrum patients, a positive correlation between the inter-network FC and memory awareness index (r = 0.254, p < 0.001). Age, sex, education, and GDS scores were controlled for all these results; The statistical threshold was set at p < 0.05 with Bonferroni correction for all these results. CN, cognitively normal; SCD, subjective cognitive decline; aMCI, amnestic mild cognitive impairment; AD, Alzheimer’s disease.

The correlation coefficient of time courses between SRN and SMN was negative. Moreover, one-way ANOVA analysis showed no significant group effect between the SRN and the SMN among the AD spectrum (Bonferroni-corrected, p = 0.377).

Behavioral significance of the disrupted FC

Pearson correlation analysis was performed between altered FC within SRN and neuropsychological scales (Fig. 4, Bonferroni corrected, p < 0.05). In the AD spectrum, the analysis showed that the altered FC within SRN was positively correlated with memory awareness index (r = 0.190, p < 0.001) and episodic memory score (r = 0.301, p < 0.001). In the groups that contained SCD and AD, the analysis showed significant correlation between the altered FC in ACC and memory awareness index/episodic memory score (r = 0.316, p < 0.001; r = 0.407, p < 0.001); the analysis showed significant correlation between the altered FC in VMPFC and memory awareness index/episodic memory score (r = 0.466, p < 0.001; r = 0.613, p < 0.001); In the groups that contained aMCI and AD, the analysis showed significant correlation between the altered FC in MOPFC and memory awareness index/episodic memory score (r = –0.368, p < 0.001; r = –0.598, p < 0.001); In the groups that contained SCD and aMCI, the analysis showed significant correlation between the altered FC in MOPFC and memory awareness index/episodic memory score (r = 0.187, p = 0.001; r = 0.402, p < 0.001). Besides, the correlation analysis between the magnitude of correlations between SRN and memory-related network and episodic memory score showed positively correlation (r = 0.254, p < 0.001). The correlation analysis between the magnitude of correlations between SRN and memory-related network and memory awareness index showed positively correlation (r = 0.232, p < 0.001). Age, gender, and education, and GDS scores were used as nuisance variables in the analysis.

Fig. 4

Significant associations between altered FC within SRN and neurocognitive scales including memory awareness and episodic memory (Bonferroni corrected, p < 0.05). Age, sex, education, and GDS scores were controlled for all these results. CN, cognitively normal; SCD, subjective cognitive decline; aMCI, amnestic mild cognitive impairment; AD, Alzheimer’s disease; MOPFC, the medial orbital prefrontal cortex; VMPFC, the ventromedial prefrontal cortex; ACC, anterior cingulate cortex; CN, cognitively normal; SCD, subjective cognitive decline; aMCI, amnestic mild cognitive impairment; AD, Alzheimer’s disease; L, left hemisphere; R, right hemisphere.

DISCUSSION

As an aspecific manifestation of AD disease, impaired memory awareness has not been paid attention by clinicians and caregivers. In recent years, more and more studies have shown that impaired memory awareness was not only an important manifestation of AD disease, but also closely related to the severity of the disease. In this study, we delineated alterations of memory awareness, intra-network connectivity in SRN as well as inter-network connectivity between SRN and memory-related network/SMN in the AD spectrum. Consistent with incipient hypotheses, our study revealed the following findings. Firstly, alteration of memory awareness was observed across AD spectrum, manifested as heightened memory awareness in SCD group but low memory awareness even anosognosia in aMCI and AD groups. Secondly, SRN showed damaged in the AD spectrum, mainly manifested as compensatory hyperconnectivity in SCD stage, a continuous decline in aMCI stage, as well as a disordered functional compensation and decline coexisting in AD stage. Thirdly, the interaction of SRN with memory-related network diminished as the cognitive impairment loads increased and broke down in aMCI and AD individuals. The correlation analysis demonstrated that the altered FC within SRN and correlation coefficient of inter-networks were related to impaired memory awareness as well as episodic memory level. All the above results indicated that the disruption of SRN and its loss interaction with memory-related network may play an important role in the progression of AD spectrum, which was also in line with the proposal that the AD may be a disconnection syndrome caused by the disruption of the long-range transcortical temporo-frontal circuit [49, 50].

Self-awareness of memory in AD spectrum

Our results found that the self-awareness of memory was distorted across AD spectrum. Compared with CN group, individuals with SCD presented heightened level of memory awareness, while individuals with aMCI and AD presented low awareness of memory levels, which was consistent with previous study [51]. Although self-reported scores of memory deficits were higher than CN group in aMCI and AD groups, their subjective evaluation of memory deficits were still lower than their study partner’ reports. The trend of memory awareness levels in our study was also corresponding with a proposed model [4 , 52] that the cognitive awareness progresses from normal cognition, to higher awareness, and finally to anosognosia across the progression of AD spectrum. Thus, memory awareness level is not in line with self-reported complaints, and low memory awareness level, rather than complaints, may play a crucial role in monitoring the progression from SCD to aMCI/AD [53 –55]. However, this study is a cross-sectional study, and longitudinal study is expected to validate this cognitive awareness model in the future.

The FC of intra-network in SRN among AD spectrum

Our results demonstrated that the AD spectrum patients showed a widespread alteration of FC in SRN. In addition, the alteration was significantly correlated with the memory awareness index and episodic memory score. Moreover, the curve of the FC within SRN was consistent with the curve of “inverse U-shaped curve” in the progression of AD [56].

The VMPFC and the MOPFC have very extensively functional and anatomical connections, thus playing a very key role in the diverse cognitive activity of the human brain. A great number of the resting state and task state neuroimaging studies had verified that the VMPFC and the MOPFC participate in processing self-referential stimulus [19 , 57–59]. Both of them receive self-related information from primary sensory modes and process the information to form cognitive self-awareness [35]. In detail, the MOPFC is associated with the evaluation of self-referential information and the VMPFC is mainly involved in the representation of self-referential information. Besides, the VMPFC, coordinated with the MOPFC, refers to processing and judging emotional information related to oneself. In fact, the MOPFC and the VMPFC all anatomically belong to the MPFC, which is not limited to performing self-reference tasks as well as involves in cognitive function and other mental activities, such as planning, attention, social and emotional behaviors [60]. It explained why the FC alteration of SRN was both correlated with memory awareness index and episodic memory score in our study. Besides, the ACC integrates various information and links them with strategies to exert high-level control over decision-making and behaviors, and therefore contributes to cognitive monitoring and behavioral control [61]. The gyrus rectus, the extension of the anterior cingulate on the frontal lobe, plays a synergistic role in cognitive self-awareness together with the ACC [62]. The alteration of FC in the ACC and the gyrus rectus are also consistent with previous studies [19, 63].

In our study, the hyperconnectivity of SRN in SCD group can be a compensatory mechanism to recruit abundant neurons to maintain normal brain network function, which was manifested as heightened memory awareness and normal cognitive level. However, this compensation is limited and not incessant. With the progression of the AD spectrum (i.e., aMCI stage), plenty of neurons were exhausted and the FC was gradually declining, manifested as low memory awareness and cognitive impairment. Specifically, our results indicate that patients with SCD have significantly increased FC in bilateral MOPFC than CN/aMCI group and decreased FC in left gyrus rectus and bilateral ACC than CN group. The positive correlation between the high activity of MOPFC and memory awareness index indicates it is a potential mechanism for self-awareness, while low connectivity between ACC and gyrus rectus in SCD group proved that the occurrence of FC disorder precedes the appearance of clinical symptom [64]. In the comparisons of other groups, in the gyrus rectus, part of upper MOPFC and lower part of VMPFC showed increasing FC in aMCI group and AD group compared to SCD and CN group, and the hyperconnectivity in AD group was more prominent even than aMCI group. The hyperconnectivity in these regions were in line with different proposed models that the increased engagement of neurons is trying to withstand the progression of cognitive impairment in aMCI and AD disease [65]. Although the brain is working hard to activate the remaining neurons to compensate for the decline in cognitive level, this compensation effect is too weak to restore the impaired memory awareness and cognitive function.

In brief, our findings revealed the pattern of altered FC in SRN across the progression of AD spectrum, which was resembling a ‘inverse U-shaped curve’ reporting in previous studies. The correlation analysis between altered of FC in SRN and memory awareness index ae well as episodic memory score verified the critical role the SRN playing in self-referential processing and cognitive function. Moreover, the peak connectivity observed in SCD subjects and then declining in aMCI patients indicated that the decreasing FC in the SRN may have hints for the conversion from SCD stage to aMCI stage. However, subgroup analyses of SCD subjects are expected to explore this statement in future.

The inter-networks connectivity analysis of SRN and memory-related network, SRN and SMN among AD spectrum

In addition to FC changes within SRN, our finding also found that the FC correlation between SRN and memory-related network was gradually decline with the progression of AD spectrum and negative correlation was observed in aMCI and AD groups. The inter-networks FC in AD spectrum was positively correlated with memory awareness index and episodic memory score. Moreover, no significant difference was observed in the FC correlation between SRN and SMN among AD spectrum.

The implementation of self-referential processing and various cognitive function in regions of SRN benefited by its complex anatomical and functional connections with various functional regions [17]. The MPFC and the medial temporal lobe form the medial prefrontal cortex-hippocampus circuit, which is associated with consolidation of memory [32, 66]. In our results, SCD group showed a little bit low but relatively normal FC between SRN and memory-related network in comparison with CN group, which was in line with maintained cognitive function in clinical feature of SCD subjects. However, the negative FC between the SRN and the memory-related network in aMCI and AD groups reflected the disconnection of inter-networks, which rendered the inability of the SRN to assess and integrate appropriate information received from primary sensory system into memory-related regions [21, 67]. Thus, the breakdown of inter-networks was manifested as cognitive impairment especially memory dysfunction in clinical feature, which was in line with the performances of aMCI and AD groups in our study. Moreover, the network breakdown may also lead to the mismatch retrieval between current and past episodic experiences and fail to integrate current memory into the self-database, which make patients have a false awareness of their memory level [59, 68]. This theoretical assumption, from the perspective of inter-network interaction, further explained the neuroimaging mechanism of low memory awareness/anosognosia in aMCI and AD patients. These findings were also consistent with previous studies that found hypometabolism and reduced intrinsic connectivity in orbitofrontal cortex as well as the disconnection between these regions and the medial temporal lobe in AD patients [21, 69]. A task fMRI study [70] showed that the comparisons of remembered task versus forgotten and self-descriptive task versus non-self-descriptive both referred to the activation of the MOPFC also suggested that the regions participated in the interaction between memory retrieval and self-relevance. Besides, no significant difference in the interaction between the SRN and the SMN across AD spectrum was also consistent with the relative retention of sensory and motor function in AD spectrum patients. All of results revealed that the impairment of memory awareness and episodic memory was due to the disorder of the internal SRN and the interaction with the memory-related network, rather than the interruption of primary sensorimotor mode.

In summary, our results suggested that low memory awareness/anosognosia and cognitive impairment in aMCI and AD patients was involved in the disconnection between self-related and memory-related networks.

Limitations

There are some limitations in our study. First, the fMRI date was acquired from ADNI database, in order to reduce selection bias and include as many as possible samples for analysis, we only deleted some subjects lacking imaging data and basic statistical information. Therefore, there were significant differences between groups in age, gender, and education level. However, in order to avoid the effect of age, gender, and education level on the results, we have removed these factors as covariates in the statistical analysis. Second, the interaction of inter-networks involved in our study are limited. Considering the characteristics of impaired memory awareness and episodic memory in AD spectrum patients, we only conducted inter-networks interaction between SRN and memory-related network, SRN and SMN. However, other networks may play an important role in the pathophysiology of AD spectrum patients. Therefore, detecting the alteration of FC in multiple networks is essential in the future. Finally, the ADNI database is a cross-sectional and a retrospective study, thus the findings in our study need to be further verified by future longitudinal studies.

CONCLUSION

Our research validated the previously proposed hypothesis that AD was a disconnection syndrome, which is manifested as networks broken both within SRN and its interaction with memory-related network referred to episodic memory encoding and retrieval. In addition, the disconnection of self-related and memory-related networks resulted in the failure of processing proper self-awareness and the inability to integrate and encode autobiographical experiences into memory-related regions, and finally leaded to the impairment of memory awareness and episodic memory in AD spectrum. In summary, the alteration of FC within SRN and breakdown between SRN and memory-related network revealed in this study may be used as a predictor of the progression of AD spectrum. And future attention and intervention on memory awareness level in the AD spectrum may play an important role in delaying the progression of AD disease.

Footnotes

ACKNOWLEDGMENTS

This study was supported by the National Natural Science Foundation of China (No. 81701675); the Key Project supported by Medical Science and technology development Foundation, Nanjing Department of Health (No. JQX18005); the Cooperative Research Project of Southeast University-Nanjing Medical University (No. 2018DN0031); the Key Research and Development Plan (Social Development) Project of Jiangsu Province (No. BE2018608); and Innovation and Entrepreneurship Training Program for College Students in Jiangsu Province (No.201810312061X; 201910312035Z). Key scientific research projects of colleges and universities in Henan province (No:18A190003).

Data collection and sharing for this project was in part funded by the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; BristolMyers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.;Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Starkstein

, Chemerinski

, Sabe

, Kuzis

, Petracca

, Teson

, Leiguarda

(1997) Prospective longitudinal study of depression and anosognosia in Alzheimer’s disease. Br J Psychiatry 171, 47–52.

Hallam

, Chan

, Gonzalez Costafreda

, Bhome

, Huntley

(2020) What are the neural correlates of meta-cognition and anosognosia in Alzheimer’s disease? A systematic review. Neurobiol Aging 94, 250–264.

Therriault

, Ng

, Pascoal

, Mathotaarachchi

, Kang

, Struyfs

, Shin

, Benedet

, Walpola

, Nair

, Gauthier

, Rosa-Neto

(2018) Anosognosia predicts default mode network hypometabolism and clinical progression to dementia. Neurology 90, e932–e939.

Hanseeuw

, Scott

, Sikkes

SAM

, Properzi

, Gatchel

, Salmon

, Marshall

, Vannini

, Alzheimer’s Disease Neuroimaging Initiative (2020) Evolution of anosognosia in Alzheimer’s disease and its relationship to amyloid. Ann Neurol 87, 267–280.

Petersen

(2016) Mild cognitive impairment. Continuum (Minneap Minn) 22, 404–418.

Chen

, Xu

, Xue

, Hu

, Ma

, Qi

, Dong

, Lin

, Chen

(2020) Voxelwise meta-analysis of gray matter abnormalities in mild cognitive impairment and subjective cognitive decline using activation likelihood estimation. J Alzheimers Dis 77, 1495–1512.

Jessen

, Amariglio

, Buckley

, van der Flier

, Han

, Molinuevo

, Rabin

, Rentz

, Rodriguez-Gomez

, Saykin

, Sikkes

SAM

, Smart

, Wolfsgruber

, Wagner

(2020) The characterisation of subjective cognitive decline. Lancet Neurol 19, 271–278.

Jessen

, Amariglio

, Buckley

, van der Flier

, Han

, Molinuevo

, Rabin

, Rentz

, Rodriguez-Gomez

, Saykin

, Sikkes

SAM

, Smart

, Wolfsgruber

, Wagner

(2020) The characterisation of subjective cognitive decline. Lancet Neurol 19, 271–278.

van Harten

, Mielke

, Swenson-Dravis

, Hagen

, Edwards

, Roberts

, Geda

, Knopman

, Petersen

(2018) Subjective cognitive decline and risk of MCI: The Mayo Clinic Study of Aging. Neurology 91, e300–e312.

10.

van Oijen

, de Jong

, Hofman

, Koudstaal

, Breteler

(2007) Subjective memory complaints, education, and risk of Alzheimer’s disease. Alzheimers Dement 3, 92–97.

11.

Jessen

, Wiese

, Bachmann

, Eifflaender-Gorfer

, Haller

, Kolsch

, Luck

, Mosch

, van den Bussche

, Wagner

, Wollny

, Zimmermann

, Pentzek

, Riedel-Heller

, Romberg

, Weyerer

, Kaduszkiewicz

, Maier

, Bickel

, German Study on Aging, Cognition and Dementia in Primary Care Patients Study Group (2010) Prediction of dementia by subjective memory impairment: Effects of severity and temporal association with cognitive impairment. Arch Gen Psychiatry 67, 414–422.

12.

Xue

, Yuan

, Yue

, Xu

, Wang

, Wu

, Ji

, Zhou

, Zhao

, Rao

, Yang

, Xiao

, Chen

(2019) Distinct disruptive patterns of default mode subnetwork connectivity across the spectrum of preclinical Alzheimer’s disease. Front Aging Neurosci 11, 307.

13.

Gerretsen

, Chung

, Shah

, Plitman

, Iwata

, Caravaggio

, Nakajima

, Pollock

, Graff-Guerrero

, Alzheimer’s Disease Neuroimaging Initiative (2017) Anosognosia is an independent predictor of conversion from mild cognitive impairment to Alzheimer’s disease and is associated with reduced brain metabolism. J Clin Psychiatry 78, e1187–e1196.

14.

Vannini

, Amariglio

, Hanseeuw

, Johnson

, McLaren

, Chhatwal

, Pascual-Leone

, Rentz

, Sperling

(2017) Memory self-awareness in the preclinical and prodromal stages of Alzheimer’s disease. Neuropsychologia 99, 343–349.

15.

Dalla Barba

, La Corte

, Dubois

(2015) For a cognitive model of subjective memory awareness. J Alzheimers Dis 48 Suppl 1, S57–61.

16.

Kalenzaga

, Clarys

(2013) Self-referential processing in Alzheimer’s disease: Two different ways of processing self-knowledge? J Clin Exp Neuropsychol 35, 455–471.

17.

Northoff

, Heinzel

, de Greck

, Bermpohl

, Dobrowolny

, Panksepp

(2006) Self-referential processing in our brain–a meta-analysis of imaging studies on the self. Neuroimage 31, 440–457.

18.

Liao

, Mantini

, Zhang

, Pan

, Ding

, Gong

, Yang

, Chen

(2010) Evaluating the effective connectivity of resting state networks using conditional Granger causality. Biol Cybern 102, 57–69.

19.

Bai

, Shi

, Yuan

, Wang

, Yue

, Teng

, Wu

, Zhang

, Jia

, Zhang

(2012) Altered self-referential network in resting-state amnestic type mild cognitive impairment. Cortex 48, 604–613.

20.

Mantini

, Perrucci

, Del Gratta

, Romani

, Corbetta

(2007) Electrophysiological signatures of resting state networks in the human brain. Proc Natl Acad Sci U S A 104, 13170–13175.

21.

Perrotin

, Desgranges

, Landeau

, Mezenge

, La Joie

, Egret

, Pelerin

, de la Sayette

, Eustache

, Chetelat

(2015) Anosognosia in Alzheimer disease: Disconnection between memory and self-related brain networks. Ann Neurol 78, 477–486.

22.

Antoine

, Bahri

, Bastin

, Collette

, Phillips

, Balteau

, Genon

, Salmon

(2019) Anosognosia and default mode subnetwork dysfunction in Alzheimer’s disease. Hum Brain Mapp 40, 5330–5340.

23.

Yasuno

, Kazui

, Yamamoto

, Morita

, Kajimoto

, Ihara

, Taguchi

, Matsuoka

, Kosaka

, Tanaka

, Kudo

, Takeda

, Nagatsuka

, Iida

, Kishimoto

(2015) Resting-state synchrony between the retrosplenial cortex and anterior medial cortical structures relates to memory complaints in subjective cognitive impairment. Neurobiol Aging 36, 2145–2152.

24.

Vannini

, Hanseeuw

, Munro

, Amariglio

, Marshall

, Rentz

, Pascual-Leone

, Johnson

, Sperling

(2017) Anosognosia for memory deficits in mild cognitive impairment: Insight into the neural mechanism using functional and molecular imaging. Neuroimage Clin 15, 408–414.

25.

Qin

, Northoff

(2011) How is our self related to midline regions and the default-mode network? Neuroimage 57, 1221–1233.

26.

Eyler

, Elman

, Hatton

, Gough

, Mischel

, Hagler

, Franz

, Docherty

, Fennema-Notestine

, Gillespie

, Gustavson

, Lyons

, Neale

, Panizzon

, Dale

, Kremen

(2019) Resting state abnormalities of the default mode network in mild cognitive impairment: A systematic review and meta-analysis. J Alzheimers Dis 70, 107–120.

27.

Mak

, Minuzzi

, MacQueen

, Hall

, Kennedy

, Milev

(2017) The default mode network in healthy individuals: A systematic review and meta-analysis. Brain Connect 7, 25–33.

28.

Chen

, Chen

, Shen

, Li

, Lu

, Zhu

, Fan

, Yan

(2020) The subsystem mechanism of default mode network underlying rumination: A reproducible neuroimaging study. Neuroimage 221, 117185.

29.

van Buuren

, Gladwin

, Zandbelt

, Kahn

, Vink

(2010) Reduced functional coupling in the default-mode network during self-referential processing. Hum Brain Mapp 31, 1117–1127.

30.

, Chen

, Si

, Li

, Zhang

, Liao

, Jiang

, Yao

, Xu

(2020) Constructing large-scale cortical brain networks from scalp EEG with Bayesian nonnegative matrix factorization. Neural Netw 125, 338–348.

31.

Kurczek

, Wechsler

, Ahuja

, Jensen

, Cohen

, Tranel

, Duff

(2015) Differential contributions of hippocampus and medial prefrontal cortex to self-projection and self-referential processing. Neuropsychologia 73, 116–126.

32.

Chao

, de Souza Silva

, Yang

, Huston

(2020) The medial prefrontal cortex - hippocampus circuit that integrates information of object, place and time to construct episodic memory in rodents: Behavioral, anatomical and neurochemical properties. Neurosci Biobehav Rev 113, 373–407.

33.

Sui

, Humphreys

(2015) The integrative self: How self-reference integrates perception and memory. Trends Cogn Sci 19, 719–728.

34.

Durbin

, Mitchell

, Johnson

(2017) Source memory that encoding was self-referential: The influence of stimulus characteristics. Memory 25, 1191–1200.

35.

Northoff

, Bermpohl

(2004) Cortical midline structures and the self. Trends Cogn Sci 8, 102–107.

36.

Kringelbach

, Rolls

(2004) The functional neuroanatomy of the human orbitofrontal cortex: Evidence from neuroimaging and neuropsychology. Prog Neurobiol 72, 341–372.

37.

Tomaszewski Farias

, Mungas

, Harvey

, Simmons

, Reed

, Decarli

(2011) The measurement of everyday cognition: Development and validation of a short form of the Everyday Cognition scales. Alzheimers Dement 7, 593–601.

38.

Rueda

, Lau

, Saito

, Harvey

, Risacher

, Aisen

, Petersen

, Saykin

, Farias

, Alzheimer’s Disease Neuroimaging Initiative (2015) Self-rated and informant-rated everyday function in comparison to objective markers of Alzheimer’s disease. Alzheimers Dement 11, 1080–1089.

39.

Potter

, Plassman

, Burke

, Kabeto

, Langa

, Llewellyn

, Rogers

, Steffens

(2009) Cognitive performance and informant reports in the diagnosis of cognitive impairment and dementia in African Americans and whites. Alzheimers Dement 5, 445–453.

40.

Shin

, Park

, Lee

, Ko

, Kim

, Han

, Jeong

, Han

(2019) Usefulness of the 15-item geriatric depression scale (GDS-15) for classifying minor and major depressive disorders among community-dwelling elders. J Affect Disord 259, 370–375.

41.

Zhu

, Wang

, Xiao

, Liao

, Zhong

, Wang

, Yao

(2012) Evidence of a dissociation pattern in resting-state default mode network connectivity in first-episode, treatment-naive major depression patients. Biol Psychiatry 71, 611–617.

42.

Neitzel

, Franzmeier

, Rubinski

, Ewers

(2019) Left frontal connectivity attenuates the adverse effect of entorhinal tau pathology on memory. Neurology 93, E347–E357.

43.

, Liu

, Friston

, Shen

, Wang

, Zeng

, Hu

(2013) A treatment-resistant default mode subnetwork in major depression. Biol Psychiatry 74, 48–54.

44.

Savio

, Funger

, Tahmasian

, Rachakonda

, Manoliu

, Sorg

, Grimmer

, Calhoun

, Drzezga

, Riedl

, Yakushev

(2017) Resting-state networks as simultaneously measured with functional MRI and PET. J Nucl Med 58, 1314–1317.

45.

Fjell

, Amlien

, Westlye

, Stenset

, Fladby

, Skinningsrud

, Eilsertsen

, Bjornerud

, Walhovd

(2010) CSF biomarker pathology correlates with a medial temporo-parietal network affected by very mild to moderate Alzheimer’s disease but not a fronto-striatal network affected by healthy aging. Neuroimage 49, 1820–1830.

46.

Biswal

, Yetkin

, Haughton

, Hyde

(1995) Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med 34, 537–541.

47.

Jafri

, Pearlson

, Stevens

, Calhoun

(2008) A method for functional network connectivity among spatially independent resting-state components in schizophrenia. Neuroimage 39, 1666–1681.

48.

Xia

, Wang

, He

(2013) BrainNet Viewer: A network visualization tool for human brain connectomics. PLoS One 8, e68910.

49.

Ott

, Lafleche

, Whelihan

, Buongiorno

, Albert

, Fogel

(1996) Impaired awareness of deficits in Alzheimer disease. Alzheimer Dis Assoc Disord 10, 68–76.

50.

Contreras

, Avena-Koenigsberger

, Risacher

, West

, Tallman

, McDonald

, Farlow

, Apostolova

, Goni

, Dzemidzic

, Wu

, Kessler

, Jeub

, Fortunato

, Saykin

, Sporns

(2019) Resting state network modularity along the prodromal late onset Alzheimer’s disease continuum. Neuroimage Clin 22, 101687.

51.

Lehrner

, Kogler

, Lamm

, Moser

, Klug

, Pusswald

, Dal-Bianco

, Pirker

, Auff

(2015) Awareness of memory deficits in subjective cognitive decline, mild cognitive impairment, Alzheimer’s disease and Parkinson’s disease. Int Psychogeriatr 27, 357–366.

52.

Wilson

, Boyle

, Yu

, Barnes

, Sytsma

, Buchman

, Bennett

, Schneider

(2015) Temporal course and pathologic basis of unawareness of memory loss in dementia. Neurology 85, 984–991.

53.

Cacciamani

, Tandetnik

, Gagliardi

, Bertin

, Habert

, Hampel

, Boukadida

, Revillon

, Epelbaum

, Dubois

, Grp IN-PS (2017) Low cognitive awareness, but not complaint, is a good marker of preclinical Alzheimer’s disease. J Alzheimers Dis 59, 753–762.

54.

Cacciamani

, Sambati

, Houot

, Habert

, Dubois

, Epelbaum

, INSIGHT-PreAD study group (2020) Awareness of cognitive decline trajectories in asymptomatic individuals at risk for AD. Alzheimers Res Ther 12, 129.

55.

Vannini

, Hanseeuw

, Gatchel

, Sikkes

SAM

, Alzate

, Zuluaga

, Moreno

, Mendez

, Baena

, Ospina-Lopera

, Tirado

, Henao

, Acosta-Baena

, Giraldo

, Lopera

, Quiroz

(2020) Trajectory of unawareness of memory decline in individuals with autosomal dominant Alzheimer disease. JAMA Netw Open 3, e2027472.

56.

, Chen

, Su

, Shu

, Gong

, Xie

, Zhou

, Bai

(2018) An inverse U-shaped curve of resting-state networks in individuals at high risk of Alzheimer’s disease. J Clin Psychiatry 79, 17m11583.

57.

D’Argembeau

, Collette

, Van der Linden

, Laureys

, Del Fiore

, Degueldre

, Luxen

, Salmon

(2005) Self-referential reflective activity and its relationship with rest: A PET study. Neuroimage 25, 616–624.

58.

Gusnard

, Akbudak

, Shulman

, Raichle

(2001) Medial prefrontal cortex and self-referential mental activity: Relation to a default mode of brain function. Proc Natl Acad Sci U S A 98, 4259–4264.

59.

Frewen

, Schroeter

, Riva

, Cipresso

, Fairfield

, Padulo

, Kemp

, Palaniyappan

, Owolabi

, Kusi-Mensah

, Polyakova

, Fehertoi

, D’Andrea

, Lowe

, Northoff

(2020) Neuroimaging the consciousness of self: Review, and conceptual-methodological framework. Neurosci Biobehav Rev 112, 164–212.

60.

Riga

, Matos

, Glas

, Smit

, Spijker

, Van den Oever

(2014) Optogenetic dissection of medial prefrontal cortex circuitry. Front Syst Neurosci 8, 230.

61.

Wallis

(2007) Orbitofrontal cortex and its contribution to decision-making. Ann Rev Neurosci 30, 31–56.

62.

Ballmaier

, Toga

, Blanton

, Sowell

, Lavretsky

, Peterson

, Pham

, Kumar

(2004) Anterior cingulate, gyrus rectus, and orbitofrontal abnormalities in elderly depressed patients: An MRI-based parcellation of the prefrontal cortex. Am J Psychiatry 161, 99–108.

63.

, Di

, Eickhoff

, Zhang

, Peng

, Guo

, Sui

(2016) Distinct and common aspects of physical and psychological self-representation in the brain: A meta-analysis of self-bias in facial and self-referential judgements. Neurosci Biobehav Rev 61, 197–207

64.

Hayes

, Tang

, Viviano

, van Rooden

, Ofen

, Damoiseaux

(2017) Subjective memory complaints are associated with brain activation supporting successful memory encoding. Neurobiol Aging 60, 71–80.

65.

Delli Pizzi

, Punzi

, Sensi

, Alzheimer’s Disease Neuroimaging Initiative (2019) Functional signature of conversion of patients with mild cognitive impairment. Neurobiol Aging 74, 21–37.

66.

Subramaniam

, Hinkley

LBN

, Mizuiri

, Kothare

, Cai

, Garrett

, Findlay

, Houde

, Nagarajan

(2019) Beta-band activity in medial prefrontal cortex predicts source memory encoding and retrieval accuracy. Sci Rep 9, 6814.

67.

d’Oleire Uquillas

, Jacobs

HIL

, Schultz

, Hanseeuw

, Buckley

, Sepulcre

, Pascual-Leone

, Donovan

, Johnson

, Sperling

, Vannini

(2020) Functional and pathological correlates of judgments of learning in cognitively unimpaired older adults. Cereb Cortex 30, 1974–1983.

68.

Nogueira

, Abolafia

, Drugowitsch

, Balaguer-Ballester

, Sanchez-Vives

, Moreno-Bote

(2017) Lateral orbitofrontal cortex anticipates choices and integrates prior with current information. Nat Commun 8, 14823.

69.

Morris

, Mograbi

(2013) Anosognosia, autobiographical memory and self knowledge in Alzheimer’s disease. Cortex 49, 1553–1565.

70.

Macrae

, Moran

, Heatherton

, Banfield

, Kelley

(2004) Medial prefrontal activity predicts memory for self. Cereb Cortex 14, 647–654.