*3.2. rs-fMRI Studies Associated with DMN in Late-Life Depression (LLD)*

The fact that DMN is related to processes mostly employed during rest, such as self-generated thought, has gained significant attention, especially with studies related to depression [89]. DMN activity is considered to be negatively correlated to the ECN activity because reducing the brain's perspective processes seems necessary to focus on the imminent task [84]. In this line, failure to reduce DMN activity has been suggested as a sign of an inability to quiet or inhibit internal mentation or emotional processing [90]. Although not the focus of this review, the relative increases in DMN connectivity during tasks has been consistently reported in various task-based fMRI studies in individuals with depression [91,92]. Several rs-fMRI studies have also reported a relative increase in DMN connectivity [93].

In addition, the difference of connectivity pattern between the anterior and the posterior DMNs in individuals with LLD has been reported. Decreased FC in the posterior DMN have been reported in individuals with LLD compared to healthy controls by rs-fMRI studies using ReHo and ALFF [94–96]. Opposing elevation of FC in the anterior DMN was also observed in LLD patients [96]. This difference of FC between the anterior and the posterior DMNs, increased in the anterior DMN and decreased in the posterior DMN, has been also reported in rs-fMRI studies in younger adults with depressed moods compared to healthy controls [77,97]. Although the results in the elderly were not presented, several studies reported that the anterior and the posterior DMNs were associated with different depressive symptoms, rumination, and autobiographical memory, respectively [98]. Moreover, this difference persisted after 12 weeks of antidepressant treatment in young subjects who recovered from MDD [99].

Seed-based analysis that used seed regions of the PCC and precuneus reported interesting results. Unmedicated LLD patients presented with decreased PCC connectivity with increased connectivity in the anterior DMN at baseline. This decrease in connectivity was partly restored after 12 weeks of treatment with paroxetine [100], suggesting that connectivity between the anterior and posterior DMN regions reflects treatment effects. Seed-based analysis presented that the dissociation between the posterior DMN and ECN was also reported in LLD individuals with current depression compared to the healthy control group [92]. During the restoration of this dissociation after antidepressant treatment, it was also reported that the connectivity between PCC and MFG decreased at baseline, then the FC from PCC to the bilateral medial frontal gyrus increased after 12 weeks of antidepressant treatment in LLD patients [101]. In the seed-based analysis using mPFC as a seed region, the dissociation between the anterior and posterior DMNs in depression has been consistently reported in young adults [102,103]. Van Tol et al. (2014) reported increased connectivity between the mPFC and left anterior insula, indicating increased connectivity between the anterior DMN and the SN [103]. We presented key findings associated with LLD in Table 1 and characteristics of main rs-fMRI studies in Supplementary Table S1.



Abbreviations: DMN, Default mode network.

*3.3. rs-fMRI Studies Associated with DMN in Alzheimer's Disease (AD) and Mild Cognitive Impairment (MCI)*

The DMN has garnered considerable attention in rs-fMRI studies of neurodegenerative diseases, and the findings have been rather consistent. Early rs-fMRI studies focused on the hippocampus [104], because amyloidosis and tau pathology initially appear in the hippocampus [105–107], and hippocampal volume loss during the progression of AD is directly associated with cognitive decline in longitudinal studies [108,109]. Various seedbased analyses have reported that less hippocampal FC was found in a broad spectrum of cortical and subcortical regions in AD patients than in healthy individuals [110–112]. This altered hippocampal FC has been replicated in more recent seed-based analyses [113–117].

Various rs-fMRI studies that used ICA, fALFF, and ALFF to assess broader networks have also reported consistent results. While there are some inconsistent results in the exact regions reported as being affected by decreased connectivity, there are common significant regions that are nodes of the DMN in AD, including the precuneus and PCC [118–132]. Decreased connectivity within the DMN is often accompanied by increased connectivity in the frontoparietal network and SN [133]. In addition to these well-established results of the entire DMN, further studies addressed the dissociation between subdivisions of the DMN (anterior and posterior), suggested by ICA studies in AD patients [120]. There are both results with connectivity reductions mainly in the posterior DMN [134], but with altered connectivity to the anterior DMN [135,136]. In the dissociation between subdivisions of the DMN, an interesting result was reported in longitudinal studies. Findings from patients with early-onset Alzheimer's disease revealed an increase in the anterior DMN and decreased posterior DMN connectivity [120].

Analysis based on graph theory to assess the alteration of brain networks in AD has shown impressive results. The degree of centrality and clustering coefficients represent the density of a network that is reduced in AD patients [137–140], and networks in AD had longer distances than healthy controls with the loss of edges [141–143]. These studies also reported a negative correlation between small-worldness that reflected a balance between local processing and global integration in the human brain and disease severity [138–140]. Similar to overall network changes, small-worldness has been consistently reported in AD patients, asymptomatic apolipoprotein Apo ε4 mutation carriers, and the aging elderly [140,144]. However, inconsistent results have reported an increase in the clustering coefficient in AD compared to healthy subjects [128,143].

Alteration of DMN connectivity is associated with a genetic mutation in AD. In particular, autosomal-dominant mutation carriers (PSEN1, PSEN2, or APP), who were young and asymptomatic, presented with altered DMN connectivity [145–147]. Regarding the Apo ε4 allele, various studies have reported diminished DMN connectivity in carriers of at least one Apo ε4 allele in all age ranges [144,146–151]. These results suggest some potential for the use of DMN connectivity for early identification of AD in young adults who carry relevant genetic mutations. Moreover, rs-fMRI studies have also reported DMN connectivity changes before the amyloidosis detected by Pittsburgh compound B [152–154], which can support the potential of DMN connectivity as an early marker of AD.

The clinical implication of DMN connectivity has been investigated in various areas. Altered DMN connectivity was correlated with the extent of cognitive decline in middle-aged and elderly Apo ε4 allele carriers [155–157]. This association has been shown consistently in AD or MCI patients related to global cognition and episodic memory performance [127,158–161]. With consistent results of altered DMN connectivity in rs-fMRI studies, attention has been focused on how these alterations can be counteracted by treatment [78]. Studies on donepezil's effect on the resting-state networks in AD have found that the application of donepezil leads to an increase in previously reduced connectivity with no differences in study groups at baseline [162,163].

Additionally, altered connectivity between the anterior and posterior DMNs is associated with aging and age-related cognitive decline [147,164]. This dissociation in DMN subdivision has also been shown in the cognitively normal elderly who presented with abnormal cerebrospinal fluid amyloid or tau proteins [165], or cerebral amyloidosis detected by PET [166]. These results are congruent with the idea that AD patients have a long preclinical period with functional alterations before the onset of disease symptoms. For the network connectivity changes in the progression of AD, longitudinal studies reported

decreased connectivity between the precuneus and ECN [167], different local aging patterns in the FC between the left hippocampus and the PCC [168], and decreased global connectivity associated with the striatum [169]. Based on the suggested potential of DMN to provide biomarkers, several rs-fMRI studies have addressed early detection, classification, and prediction in AD and MCI. These studies have shown relatively high performances: ICA [161,170–173], seed-based analysis [174], and graph theory [175]. We presented key findings associated with AD and MCI in Table 2 and characteristics of main rs-fMRI studies in Supplementary Tables S2 and S3.

**Table 2.** Summary of key findings of rs-fMRI studies associated with the default mode network (DMN) in Alzheimer's disease (AD) and mild cognitive impairment (MCI) patients included in the review.


Abbreviations: DMN, default mode network.

#### **4. The Executive Control Network (ECN)**
