Prospective Memory and Default Mode Network Functional Connectivity in Normal and Pathological Aging

Abstract

Background:

Prospective memory (PM), the ability to execute a previously formed intention given the proper circumstance, has been proven to be vulnerable to Alzheimer’s disease. Previous studies have indicated the involvement of the frontoparietal networks; however, it is proposed that PM may also be associated with other neural substrates that support stimulus-dependent spontaneous cognition.

Objective:

The present study aimed to examine the hypothesis that PM deficit in Alzheimer’s disease is related to altered functional connectivity (FC) within the default mode network (DMN).

Methods:

Thirty-four patients with very mild or mild dementia (17 with Alzheimer’s disease and 17 with subcortical ischemic vascular disease) and 22 cognitively-normal participants aged above 60 received a computerized PM task and resting-state functional magnetic resonance imaging study. Seed-based functional connectivity analysis was performed at group level within the DMN.

Results:

We found that the dementia groups showed worse PM performance and altered FC within the DMN as compared to the normal aging individuals. The FC between the medial prefrontal cortices and precuneus/posterior cingulate cortex was significantly correlated with PM in normal aging, while the FC between the right precuneus and bilateral inferior parietal lobules was correlated with PM in patients with Alzheimer’s disease.

Conclusion:

These findings support a potential role for the DMN in PM, and corroborate that PM deficit in Alzheimer’s disease was associated with altered FC within the posterior hubs of the DMN, with spatial patterning different from normal aging.

Keywords

Alzheimer’s disease cerebral small vessel disease default mode network memory for intention prospective memory

INTRODUCTION

Prospective memory (PM) is the realization of an intended action in the future. It requires disengagement from external stimuli and reflection upon internal thoughts that have been previously formulated. PM is a functionally relevant psychological construct [1] and is susceptible to Alzheimer’s disease (AD) [2 –5].

Several theoretical accounts of PM have emphasized the need of resource-consuming strategic moni-toring, or cognitive control [6 –8]. Meanwhile, a spontaneous retrieval process that resembles a “pop-up into mind” experience is also considered an essential mechanism [8]. Findings from neuroimaging studies affirm this dual process view by demonstrating that both top-down and bottom-up processes supported by the dorsal and ventral frontoparietal networks, respectively, were closely associated with PM function [9, 10].

We argued that the default mode network (DMN), which is associated with self-referential as well as future-oriented thoughts, may also play a part in successful PM, and that disruption of the DMN may be partly responsible for PM deficit in AD. Buckner and Carroll [11] proposed that there would be a core neural network responsible for self-projection, presumably in the DMN, that allows humans to live not only here and now but also to travel between time and space. The mental processes involving self-projection that allows us to shift the immediate perception to an alternative situation include remembering a past event, taking others’ perspectives, mental navigation, and prospection. Prospection is the preconstruction of a future scene, and this ability to mentally simulate an upcoming event is important to PM [12]. Neuroimaging studies also showed that mental imagery is associated with activities in or with the DMN [13, 14].

In addition, the DMN has been linked with spontaneous cognition or the off-task “mind-wandering” state through its widespread networks connected by hubs [15]. PM requires a similar cognitive maneuver that directs the attention away from the concurrent, ongoing activities toward the previously formed intention inwards. Previous studies also reported brain activation within the DMN during spontaneous retrieval of PM intention, including the ventral parietal cortex and precuneus [10, 16]. The incipient converging evidence indicated a potential role of the DMN in PM processing.

Recently, Kvavilashvili and colleagues [17] contended that PM can be viewed as a type of “stimulus-dependent spontaneous cognition”, or mind-wandering, that is subserved by the DMN. They hypothesized that the focal PM deficits often seen in AD may result from the early amyloidosis typically distributed in the DMN that overlaps with the areas supporting PM. More specifically, focal PM requires similar cognitive processes for the ongoing task and PM cue detection, which could be managed relatively spontaneously, or automatically. On the contrary, nonfocal PM requires a different cognitive operation to detect the PM cue, and is thought to rely on strategic monitoring and controlled processes [8]. Due to the disruption of the neural network that supports stimulus-dependent spontaneous cognition, individuals with AD thereby become less likely to fulfill a focal PM task upon encountering an action cue that would otherwise trigger the spontaneous retrieval of the intended action.

Although previous studies sporadically reported associations between PM and activities in discrete brain areas within the territory of the DMN in normal aging, to the best of our knowledge, none has explicitly examined the role of DMN functional connectivity (FC) in PM. Since altered FC within the DMN has been reported in pathological aging such as AD and vascular dementia [18, 19], we hypothesized that there would be a relationship between PM that requires spontaneous processing and FC within the DMN in individuals with these two pathological aging conditions.

MATERIALS AND METHODS

Participants

The study initially recruited 31 normal aging (NA) adults, 22 patients diagnosed with probable AD, and 20 patients with subcortical ischemic vascular disease (SIVD). Among them, 26 NA, 20 AD, and 20 SIVD completed both cognitive tests and resting-state functional magnetic resonance imaging (rs-fMRI) examinations. The inclusion criteria were age above 60; fulfill the diagnosis of AD according to the National Institute on Aging-Alzheimer’s Association Criteria [20] or SIVD according to Erkinjuntti et al.’s [21] criteria; and rated as mild or very mild on the Clinical Dementia Rating (CDR) [22] (i.e., CDR = 0.5–1). Participants in the NA group had to show no evidence of cognitive deficits, and had no profound structural abnormalities on a conventional MRI. The exclusion criteria were stroke event within the past two weeks; signs of hemorrhages, cortical watershed infarction, and specific causes of white matter lesions; hydrocephalus; delirium; abnormal serology results that may affect cognitive function; and severe visual or hearing impairment. The study was completed in accordance with Helsinki Declaration and approved by a local research ethics committee (REC106-48). All participants signed an informed written consent document before entering the study.

Neuropsychological tests

The participants received the Taiwanese Mini-Mental State Examination [23, 24] and a computerized PM task. The PM task was designed with the software E-Prime [25] in the format of the McDaniel-Einstein paradigm [26], with modification in Chinese. The participants were asked to engage in a semantic categorization task as the ongoing task, in which they had to decide if an exemplar presented on the bottom right of the computer screen (e.g., “”, which means “scooter”) belongs to the category indicated on the upper left (“”, which means “transportation”) by pressing the designated keys. For the PM trials, the participants had to form the intention that they would press an additional key (i.e., the PM action) when encountering a Chinese character that is encircled on the outside, such as “” (i.e., the PM cue), while performing the concurrent semantic categorization task. There were 13 PM cues randomly embedded in the 78 trials of the semantic categorization task. In the present study, we used the hit rate (%) as an index of PM function.

Neuroimaging examination

MR experiments were performed on a 3T MRI system (Discovery MR750; GE Medical System, Milwaukee, WI, USA) with an eight-channel phased-array head coil. The sequence parameters for T1-weighted imaging were inversion time (TI) of 450 ms, flip angle of 12°, field-of-view (FOV) of 240 mm, matrix size of 240×240, slice thickness of 1 mm, and 160 slices. For rs-fMRI, the sequence parameters of the echo planar imaging (EPI) were repetition time (TR) of 2,500 ms, echo time (TE) of 30 ms, flip angle of 90°, FOV of 200 mm, matrix size of 64×64, SL of 3 mm, and 47 slices. Participants were asked to “rest but not fall asleep during the examination”.

The rs-fMRI data preprocessing procedures were performed by the software Data Processing Assistant for Resting-State fMRI (DPARSF) [27], including slice timing, realignment, co-registration, spatial normalization, nuisance covariate regression (white matter signals, cerebrospinal fluid signals, and head motion parameters) and band-pass filtering. The images were then resampled into 3-mm cubic voxels. Quality control was performed, in which participants with T1, EPI or results of normalization rated less than three on a five-point Likert scale (5: Very good; 4: Good; 3: Fair; 2: Poor; 1: Very poor) were excluded. Head motion was controlled by removing data with mean framewise displacement (FD) larger than 0.2 mm and incorporating the mean FD as covariates in the statistical analyses.

Four regions-of-interest (ROIs) with 8 mm radius were used for seed-based FC analysis, including the posterior cingulate cortex (PCC), ventromedial prefrontal cortex (vmPFC), and the left and the right inferior parietal lobules (LIPL and RIPL, respectively). Table 1 shows the coordinates based on Elton and Gao [28]. The time series were extracted from the four ROIs and correlated with the time series of all other voxels in the brain. The Pearson correlation coefficients at each voxel was then converted to z values using Fisher’s r-to-z transformation, generating the FC maps for each participant.

Table 1

Montreal Neurological Institute (MNI) coordinates of the seeds

Seeds	MNI space			BA
	X	y	z
PCC	1	–55	17	30
vmPFC	0	51	–7	11
LIPL	–47	–71	29	39
RIPL	50	–64	27	39

PCC, posterior cingulate cortex; vmPFC, ventromedial prefrontal cortex; LIPL, left posterior inferior parietal lobule; RIPL, right posterior inferior parietal lobule; BA, Brodmann area.

Statistical analysis

Behavioral data were analyzed with the software IBM SPSS Statistics 25. Demographic data were examined by analysis of variance (ANOVA) and Chi-square tests, where appropriate. Normality of residuals was checked by the normal quantile–quantile plot of the dependent variables. Welch’s ANOVA tests were used for data violating the assumption of homogeneity. Post hoc tests for significant results were performed using Bonferroni or Games-Howell tests, where appropriate. Analyses of covariance (ANCOVA) with age, education, and sex as covariates were used to examine PM performances between groups, followed by Bonferroni correction for the post hoc tests.

rs-fMRI data were analyzed using a toolbox for Data Processing & Analysis for Brain Imaging (DPABI) [29]. One-sample t tests were first used to examine areas showing significant FC with the pre-determined seeds (Table 1) for visual inspection, and the results can be seen in Supplementary Figure 1. The differences on the FC maps between the SIVD and NA, and between AD and NA were compared using the DPABI, with age, sex, education, and mean FD as covariates. Subsequently, partial correlation analyses were performed to evaluate the associations between PM performance and the FC values generated by the seed ROIs in each group, which were constrained to a predetermined DMN mask [30, 31], with age, sex, education, and mean FD as covariates. Familywise error rate was controlled by Gaussian Random Field (GRF) Theory Correction with voxel p value set at 0.001 and cluster p value set at 0.05.

RESULTS

Overall, 56 participants (22 NA, 17 AD, and 17 SIVD) with mean age of 69.68 (SD = 12.30) completed the MRI survey with acceptable image quality. Table 2 shows the demographic data in each group. All participants were right-handed. There was a significant difference between groups on age (W = 34.81, p = 0.006), in which the NA group was younger than the AD group (p < 0.05). There was no significant difference between groups on education (W = 34.53, p = 0.052) or gender (χ² = 0.77, p = 0.68). Nonetheless, age and education were included as covariates in the subsequent analyses to parcel out potential confounding effects.

Table 2

Basic information of the participants and performance on the computerized prospective memory task

	SIVD (n = 17)		AD (n = 17)		NA (n = 22)		ANOVA/ANCOVA^§
	Mean	SD	Mean	SD	Mean	SD	W^†/χ²^‡/F^§	p
Age (y)	69.59	8.26	75.35	6.55	65.36	12.30	34.81^†	0.006^b
Education (y)	6.94	3.77	7.88	4.69	10.41	4.81	34.53^†	0.052^a
Sex (% of male)	53%		53%		41%		0.77^‡	0.679
MMSE	22.35	4.43	22.53	4.24	28.14	1.93	12.2^§	< 0.001^a,b
PM hit rate (%)	0.31	0.26	0.21	0.25	0.70	0.30	10.13^§	< 0.001^a,b

SIVD, subcortical ischemic vascular disease; AD, Alzheimer’s disease; NA, normal aging; ANOVA, analysis of variance; ANCOVA, analysis of covariance. ^†Welch’s ANOVA; ^‡Chi-square test; ^§Age and education as covariates; MMSE, Mini-Mental State Examination; PM, prospective memory; ^aSignificant difference (p < 0.05) between SIVD and NA groups; ^bSignificant difference (p < 0.05) between AD and NA groups.

Behavioral results

Table 2 shows that there was a significant group difference on PM performance (F = 10.13, p < 0.001, $η_{p}^{2}$ = 0.28). Post hoc analyses showed worse PM performance in patients with SIVD (p = 0.003) and AD (p = 0.001) as compared to the NA group. There was no significant difference on PM performance between the SIVD and AD groups.

Neuroimaging results

When comparing the AD and NA groups, there were several clusters showing significant group difference of FC with different seed ROIs (Table 3). The AD group showed significantly lower FC between PCC and the left orbitofrontal cortex, left precuneus, left anterior cingulate cortex, and anterior prefrontal cortex (all p < 0.001). There was also significant difference of FC between the LIPL and the left posterior cingulate cortex, and between the RIPL and the left retrosplenial cortex, left posterior cingulate cortex, and left precuneus (all p < 0.001). When compared with the NA group, the SIVD group showed significantly higher FC only between LIPL and a cluster over the medial orbital prefrontal area (p < 0.001; Table 3).

Table 3

Functional connectivity showing significant difference from the normal aging group

		Significant clusters^*
Groups	Seeds	Brain region	MNI space			Peak intensity (t)	Cluster size (mm³)
			x	y	z
AD	PCC	Left orbitofrontal	–3	54	–18	–3.38	216
		Left precuneus	–15	–57	12	–3.38	216
		Left precuneus	0	–51	60	–3.37	216
		Left anterior cingulate	–3	42	6	–3.41	162
		Anterior prefrontal	–15	60	12	–3.45	162
	LIPL	Left posterior cingulate	–3	–42	9	–3.36	189
	RIPL	Left retrosplenial	–12	–48	18	–3.39	297
		Left posterior cingulate	–6	–42	12	–3.43	243
		Left precuneus	0	–54	60	–3.37	162
SIVD	LIPL	Medial orbital prefrontal	–3	27	–12	4.37	297

^*Voxel p < 0.001; cluster p < 0.05. MNI, Montreal Neurological Institute; SIVD, subcortical ischemic vascular disease; AD, Alzheimer’s disease; LIPL, left posterior inferior parietal lobule; PCC, posterior cingulate cortex; RIPL, right posterior inferior parietal lobule.

Subsequently, we conducted correlational analyses between seed-based FC within the DMN and PM performance in each group. Partial correlational analyses controlling for age, sex, education, and mean FD in the NA group showed that PM performance was significantly correlated with FC between the PCC and a cluster over the left superior medial PFC (peak intensity r = 0.78, p < 0.001, Fig. 1A), and between vmPFC and a cluster over the left precuneus (peak intensity r = 0.82, p < 0.001, Fig. 1B). In the AD group, PM performance was significantly correlated with FC between the LIPL and a cluster in the right precuneus (peak intensity r = 0.90, p < 0.001, Fig. 2A), and between the RIPL and a cluster in the right precuneus (peak intensity r = 0.84, p < 0.001, Fig. 2B). In the SIVD group, there was no significant correlation between PM and FC within the DMN.

Fig. 1

Significant correlations between prospective memory and seeds-to-voxel functional connectivity (FC) in the normal aging (NA) group. (A) FC between the posterior cingulate cortex (PCC) and left superior medial frontal gyrus (threshold r > 0.68, p < 0.001; peak voxel x = 3, y = 42, z = 30; cluster size: 216 mm³). (B) FC between the ventromedial prefrontal cortex (vmPFC) and left precuneus (threshold r > 0.68, p < 0.001; peak voxel x = –6, y = –42, z = 51; cluster size: 270 mm³).

Fig. 2

Significant correlations between prospective memory (PM) performances and seeds-to-voxel functional connectivity (FC) in the Alzheimer’s disease (AD) group. The FC of both the left inferior parietal lobule (LIPL) and the right inferior parietal lobule (RIPL) with the right precuneus was found to be correlated with PM performance (threshold r > 0.77, p < 0.001): (A) FC between the LIPL and right precuneus (peak voxel x = 9, y = –60, z = 42; cluster size: 243 mm³); (B) FC between the RIPL and right precuneus (peak voxel x = 3, y = –63, z = 39; cluster size: 162 mm³).

DISCUSSION

The present study attempted to examine the hypothesis that PM would be related to FC within the DMN, which is associated with self-generated and future-oriented thoughts. We found that in the NA group, PM was related to FC along the anterior-posterior axis of the DMN, between the precuneus/PCC and vmPFC. The results also revealed altered FC within the DMN in pathological aging, in which the AD group demonstrated a greater number of FC changes than the SIVD group when compared with the NA group. Furthermore, the FC correlates of PM in AD showed a different pattern from NA, which lied posteriorly between the right precuneus and bilateral IPLs.

Buckner and Carrol [11] contended that the DMN may be the common neural substrates underlying cognitive processes involving projecting oneself into an alternative scenario, including remembering the past and simulating future events. These mental works are essential to planning and recalling an intention that has to be performed. Kvavilashvili et al. [17] further proposed that PM deficits often seen in patients with AD may arise from pathological involvement of the DMN. They explained that PM can be viewed as a kind of internally-generated cognition, which falls into the same scope as the function of DMN in terms of spontaneous cognition and future thinking. Empirical studies to support this view have been limited and circumstantial. A recent magnetoencephalography study [32] showed that theta oscillations, which were considered pivotal to working memory and memory integration, were associated with PM performance not only over lateral brain areas but also overlapping areas of the DMN. Studies of structural connectivity also showed that PM was associated with microstructural integrity in cingulum [33, 34] and superior longitudinal fasciculus [35]; these white matter bundles were found to be correlated with FC of the DMN [36]. To the best of our knowledge, this is the first study to directly examine the relationship between PM and FC in the DMN. Together with the accumulating evidence, the present study provides further empirical corroboration to the theoretical account that DMN is important to PM.

Our results in both normal and pathological aging implied an important role for a key hub of the DMN, the precuneus. The precuneus was critically involved in episodic retrieval [37] but was also found to be important to the maintenance of PM intention in task-related fMRI studies [10 , 38–40]. The adjacent PCC played a part in attention orientation, especially when shifting the thoughts inward, such as encoding and retrieving a PM intention [9]. Cona et al. [9] proposed that the PCC may work with the parietal areas to direct attention during PM processes. This view is corroborated by our finding of a significant relationship between PM and the FC between the PCC and bilateral IPLs. On the other hand, vmPFC was considered to be important for outward attention that deals with external stimuli [9, 41]. Deactivation of similar areas in the medial rostral PFC was found to be induced by PM activity in healthy young adults [38]. Note that the vmPFC examined in the current study lies along the cortical midline (x = 0, y = 51, z = –7), which locates differently from the anterior prefrontal cortex (BA10) examined in previous studies which emphasized the role of frontoparietal networks in PM (e.g., [10]).

Different patterns of FC correlates for PM between the AD and NA groups were clearly delineated in the current study. In the NA group, PM was related to the FC between the anterior and the posterior hubs, which includes the vmPFC and was associated with self-referential thoughts [14]. On the contrary, in the AD group, the PM performance was correlated with the FC of the posterior DMN, which includes the PCC and inferior parietal lobes, and was associated with both episodic retrieval and future thinking [14]. The results indicate that performance difference on PM in normal aging may be related to individual difference within the prefrontal areas. In contrast, early-stage AD may compromise PM function through the pathological changes emanating from the PCC, which is closely connected with the hippocampus [42]. Regarding the SIVD group, we did not find any significant correlation between PM and FC in the DMN. The PM deficit observed in the SIVD group may arise from pathological involvement in other brain areas or networks outside of the DMN, such as the frontoparietal network [43].

Past research has shown disruption of the DMN in several pathological conditions. A meta-analysis [44] indicated a robust change of connectivity within the DMN in patients with AD. Our results that PM was associated with reduced FC between the posterior hubs of the DMN is consistent to the Spontaneous Retrieval Deficit Hypothesis [17], which predicts that the risk-groups of AD would show deficits on stimulus-dependent spontaneous retrieval processes that rely on brain areas involved in spontaneous cognition, such as the PCC. That is, the tendency of individuals with incipient AD to be triggered by a PM cue and to bring back a task-unrelated (unrelated to the ongoing task) thoughts (i.e., the PM intention) would be reduced due to the disruption of the DMN which presumably supports the “stimulus-dependent spontaneous retrieval process”. A few studies also revealed decreased structural connectivity and maladaptive change of FC in the DMN among patients with SIVD [19, 45]. The current study found a greater number of FC changes within the DMN in AD as compared to SIVD, and the less prominent alteration in the latter group was in line with the fact that SIVD generally exhibits greater subcortical changes than cortical ones [46, 47]. Overall, the current study confirmed that there was altered FC between the hubs within the DMN in patients with AD and SIVD. The results further implied a cognitive impact of altered FC in the DMN on the ability to remember to carry out future actions in AD.

Nonetheless, future studies may provide further clarification and insight by comparing focal and nonfocal PM regarding their associations with the DMN. The Spontaneous Retrieval Deficit Hypothesis asserts that individuals with AD would show a deficit in cognitive functions that rely on spontaneous retrieval. It predicts that a performance deficit would be seen in AD on focal PM tasks, which depend upon stimulus-dependent spontaneous retrieval and which would be associated with the DMN. The present study adopted a Chinese semantic categorization task as the ongoing task, and a particular perceptual feature of the Chinese character (a closed shape) as its PM cue. On one hand, the grapheme-based property of Chinese language [48] allows for automatic processing in this task. On the other hand, a nonfocal feature cannot be excluded as the PM cue detection does not fully depend on semantic processing. As suggested by the Dynamic Multiprocess Framework [49], the spontaneous retrieval and monitoring processes may interact in a dynamic way to support PM. Therefore, our results provide partial evidence for the Spontaneous Retrieval Deficit Hypothesis by showing that the performance on a PM task that partly relied on spontaneous retrieval was related to altered FC within the DMN in individuals with AD. Nevertheless, it will require additional studies that tease apart the two types of event-based PM to provide further evidence.

The results of an association between PM and DMN FC in normal aging is intriguing. We did not include a healthy young adult group to compare the PM performances between different age groups. However, the age effect had long been a paradox in PM literature. Unlike many other cognitive functions that show age-associated deterioration in later life, PM research has not been able to find consistent evidence of age effect. This may be a result of different test settings, methodology, or task types [50]. For example, there is little evidence of age-associated change in focal PM tasks [51, 52], which again corroborates the statement that focal PM relies strongly on automatic, spontaneous processes. This also demonstrates a clear distinction from individuals with AD who showed focal PM deficit [53, 54], which could be attributed to DMN disruption, instead of normal aging processes, as addressed in the Spontaneous Retrieval Deficit hypothesis [17, 55].

The current findings did not contradict the robust evidence that PM relies on the anterior prefrontal cortex or the frontoparietal network [9]. For example, Massa et al.’s [56] PET study on MCI due to AD revealed that PM performance was positively associated with metabolic levels in the right-sided rostral prefrontal cortex and its interhemispheric FC. In Cona et al.’s review [9], the PCC was hypothesized to interact with the frontoparietal areas to control the locus of attention. A post hoc analysis of our data revealed an anticorrelation between vmPFC and superior medial PFC (t = –0.55, p < 0.001; cluster size: 162 mm³). This is in line with the conjecture that the anterior prefrontal cortex may serve as the attention gateway [57]. Future studies may set out to delineate the potentially complex relationships between multiple neural networks subserving PM, which may help reconcile the traditional and novel views of neural substrates of PM.

The present study included patients with very mild and mild dementia. Hence, we could not exclude the possibility that the results reported here were due to more advanced amyloidosis and tauopathy that affect the precuneus and PCC. We conducted post hoc analysis on the AD subgroup with CDR 0.5 (N = 10), and found that PM performance was significantly correlated with FC between the PCC and a cluster over the right angular gyrus after controlling for age, sex, education, and mean FD, and after GRF correction (peak intensity r = 0.95, p < 0.001; peak voxel x = 48, y = –51, z = 33; cluster size: 108 mm³), indicating that similar phenomenon of PM correlating to FC within the posterior DMN could be observed in the very early stage of AD. However, the sample size of this tentative analysis was too small for a conclusive remark. Future studies may further explore if our finding would be replicated in a larger sample of very mild AD or MCI due to AD.

The present study has several other limitations. Despite abiding by specific diagnostic criteria, incorporating amyloid imaging in future studies may strengthen the findings with a more solid neuropathological basis. Moreover, future studies examining the interplay and causal dynamics within and between the brain networks using faster rs-fMRI data may help better understand the neural underpinning of PM. Another limitation is that we only included event-based PM. However, whether the findings of this work would apply to time-based PM, or merely reflect the event-based one, awaits future study to examine. Nonetheless, the findings still provide valuable information and extend the relatively new insight upon the underlying neural mechanism of PM.

Footnotes

ACKNOWLEDGMENTS

We thank the participants for making this study possible. We also thank Mr. Jon Renzella for English editing and proofreading.

This work was supported by the Ministry of Science and Technology in Taiwan [MOST 108-2410-H-194-045; MOST 109-2410-H-194 -053 -MY2]; Taichung Tzu Chi Hospital Grant [TTCRD 106-12]; and National Health Research Institute [NHRI-110-BN-PP-06].

Authors’ disclosures available online ().

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

References

Woods

, Weinborn

, Velnoweth

, Rooney

, Bucks

(2012) Memory for intentions is uniquely associated with instrumental activities of daily living in healthy older adults. J Int Neuropsychol Soc 18, 134–138.

Hsu

, Huang

, Tu

, Hua

(2015) Prospective memory in subjective cognitive decline: A preliminary study on the role of early cognitive marker in dementia. Alzheimer Dis Assoc Disord 29, 229–235.

Costa

, Fadda

, Perri

, Brisindi

, Lombardi

, Caltagirone

, Carlesimo

(2015) Sensitivity of a time-based prospective memory procedure in the assessment of amnestic mild cognitive impairment. J Alzheimers Dis 44, 63–67.

Duchek

, Balota

, Cortese

(2006) Prospective memory and apolipoprotein E in healthy aging and early stage Alzheimer’s disease. Neuropsychology 20, 633–644.

Jones

, Livner

, Backman

(2006) Patterns of prospective and retrospective memory impairment in preclinical Alzheimer’s disease. Neuropsychology 20, 144–152.

Smith

, Bayen

(2004) A multinomial model of event-based prospective memory. J Exp Psychol Learn Mem Cogn 30, 756–777.

Strickland

, Loft

, Remington

, Heathcote

(2018) Racing to remember: A theory of decision control in event-based prospective memory. Psychol Rev 125, 851–887.

McDaniel

, Einstein

(2000) Strategic and automatic processes in prospective memory retrieval: A multiprocess framework. Appl Cogn Psychol 14, S127–S144.

Cona

, Scarpazza

, Sartori

, Moscovitch

, Bisiacchi

(2015) Neural bases of prospective memory: A meta-analysis and the “Attention to Delayed Intention” (AtoDI) model. Neurosci Biobehav Rev 52, 21–37.

10.

McDaniel

, Lamontagne

, Beck

, Scullin

, Braver

(2013) Dissociable neural routes to successful prospective memory. Psychol Sci 24, 1791–1800.

11.

Buckner

, Carroll

(2007) Self-projection and the brain. Trends Cogn Sci 11, 49–57.

12.

Cheng

, Tu

, Huang

, Hsu

(2021) Effects of mental imagery on prospective memory: A process analysis in individuals with amnestic mild cognitive impairment. Gerontology 67, 718–728.

13.

Zhang

, Zhang

, Wang

, Li

, Lin

, Chang

, Huang

, Liu

(2018) Intrinsic neural linkage between primary visual area and default mode network in human brain: Evidence from visual mental imagery. Neuroscience 379, 13–21.

14.

, Yuan

, Lei

(2016) Activation and connectivity within the default mode network contribute independently to future-oriented thought. Sci Rep 6, 21001.

15.

Raichle

(2015) The brain’s default mode network. Annu Rev Neurosci 38, 433–447.

16.

Beck

, Ruge

, Walser

, Goschke

(2014) The functional neuroanatomy of spontaneous retrieval and strategic monitoring of delayed intentions. Neuropsychologia 52, 37–50.

17.

Kvavilashvili

, Niedźwieńska

, Gilbert

, Markostamou

(2020) Deficits in spontaneous cognition as an early marker of Alzheimer’s disease. Trends Cogn Sci 24, 285–301.

18.

Kim

, Cha

, Lee

, Shin

, Jung

, Kim

, Choe

, Lee

, Kim

, Lee

, Na

, Seo

(2016) Distinctive resting state network disruptions among Alzheimer’s disease, subcortical vascular dementia, and mixed dementia patients. J Alzheimers Dis 50, 709–718.

19.

Liu

, Chen

, Cheng

, Luo

, Lv

, Fang

, Gong

, Jiang

(2019) Altered functional connectivity in patients with subcortical ischemic vascular disease: A resting-state fMRI study. Brain Res 1715, 126–133.

20.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

, Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dementia 7, 263–269.

21.

Erkinjuntti

, Inzitari

, Pantoni

, Wallin

, Scheltens

, Rockwood

, Roman

, Chui

, Desmond

(2000) Research criteria for subcortical vascular dementia in clinical trials. J Neural Transm Suppl 59, 23–30.

22.

Morris

(1993) The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology 43, 2412–2414.

23.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

24.

Shyu

, Yip

(2001) Factor structure and explanatory variables of the Mini-Mental State Examination (MMSE) for elderly persons in Taiwan. J Formos Med Assoc 100, 676–683.

25.

Schneider

, Eschman

, Zuccolotto

(2002) E-Prime Reference Guide, Psychology Software Tools, Pittsburgh, PA.

26.

Einstein

, McDaniel

(1990) Normal aging and prospective memory. J Exp Psychol Learn Mem Cogn 16, 717–726.

27.

Yan

, Zang

(2010) DPARSF: A MATLAB toolbox for “Pipeline” data analysis of resting-state fMRI. Front Syst Neurosci 4, 13.

28.

Elton

, Gao

(2015) Task-positive functional connectivity of the default mode network transcends task domain. J Cogn Neurosci 27, 2369–2381.

29.

Yan

, Wang

, Zuo

, Zang

(2016) DPABI: Data processing & analysis for (resting-state) brain imaging. Neuroinformatics 14, 339–351.

30.

Brown

, Schmitt

, Smith

, Gold

(2019) Distinct patterns of default mode and executive control network circuitry contribute to present and future executive function in older adults. Neuroimage 195, 320–332.

31.

Brown

, Jiang

, Smith

, Gold

(2018) Age and Alzheimer’s pathology disrupt default mode network functioning via alterations in white matter microstructure but not hyperintensities. Cortex 104, 58–74.

32.

Cona

, Chiossi

, Di Tomasso

, Pellegrino

, Piccione

, Bisiacchi

, Arcara

(2020) Theta and alpha oscillations as signatures of internal and external attention to delayed intentions: A magnetoencephalography (MEG) study. Neuroimage 205, 116295.

33.

Hoare

, Westgarth-Taylor

, Fouche

, Spottiswoode

, Paul

, Thomas

, Stein

, Joska

(2012) A diffusion tensor imaging and neuropsychological study of prospective memory impairment in South African HIV positive individuals. Metab Brain Dis 27, 289–297.

34.

McCauley

, Wilde

, Bigler

, Chu

, Yallampalli

, Oni

, Wu

, Ramos

, Pedroza

, Vasquez

, Hunter

, Levin

(2011) Diffusion tensor imaging of incentive effects in prospective memory after pediatric traumatic brain injury.. J Neurotrauma 28, 503–516.

35.

Hsu

, Huang

, Deng

, Tu

(2019) Microstructural correlates and laterality effect of prospective memory in non-demented adults with memory complaints. Dement Geriatr Cogn Disord 47, 375–384.

36.

van den Heuvel

, Mandl

, Luigjes

, Hulshoff Pol

(2008) Microstructural organization of the cingulum tract and the level of default mode functional connectivity. J Neurosci 28, 10844–10851.

37.

Krause

, Schmidt

, Mottaghy

, Taylor

, Halsband

, Herzog

, Tellmann

, Müller-Gärtner

(1999) Episodic retrieval activates the precuneus irrespective of the imagery content of word pair associates. A PET study. Brain 122(Pt 2), 255–263.

38.

Gonneaud

, Rauchs

, Groussard

, Landeau

, Mézenge

, de La Sayette

, Eustache

, Desgranges

(2014) How do we process event-based and time-based intentions in the brain? An fMRI study of prospective memory in healthy individuals. Hum Brain Mapp 35, 3066–3082.

39.

Gilbert

, Armbruster

, Panagiotidi

(2012) Similarity between brain activity at encoding and retrieval predicts successful realization of delayed intentions. J Cogn Neurosci 24, 93–105.

40.

Burgess

, Gonen-Yaacovi

, Volle

(2011) Functional neuroimaging studies of prospective memory: What have we learnt so far? Neuropsychologia 49, 2246–2257.

41.

Gilbert

, Spengler

, Simons

, Steele

, Lawrie

, Frith

, Burgess

(2006) Functional specialization within rostral prefrontal cortex (Area 10): A meta-analysis. J Cogn Neurosci 18, 932–948.

42.

Mutlu

, Landeau

, Tomadesso

, de Flores

, Mézenge

, de La Sayette

, Eustache

, Chételat

(2016) Connectivity disruption, atrophy, and hypometabolism within posterior cingulate networks in Alzheimer’s disease. Front Neurosci 10, 582.

43.

Zhuang

, Kuo

, Lin

, Yang

, Tu

, Hsu

(2021) Prospective memory and regional functional connectivity in subcortical ischemic vascular disease. Front Aging Neurosci 13, 686040.

44.

Badhwar

, Tam

, Dansereau

, Orban

, Hoffstaedter

, Bellec

(2017) Resting-state network dysfunction in Alzheimer’s disease: A systematic review and meta-analysis. Alzheimers Dement (Amst) 8, 73–85.

45.

Sang

, Liu

, Wang

, Zhang

, Li

, Qiao

, Li

, Qiu

(2020) Disrupted brain structural connectivity network in subcortical ischemic vascular cognitive impairment with no dementia. Front Aging Neurosci 12, 6.

46.

, Huang

, Hsu

, Yang

, Lin

, Kuo

(2021) Discriminating subcortical ischemic vascular disease and Alzheimer’s disease by diffusion kurtosis imaging in segregated thalamic regions. Hum Brain Mapp 42, 2018–2031.

47.

, Lo

, Huang

, Hsu

, Huang

, Deng

, Lee

(2017) Effectiveness of diffusion tensor imaging in differentiating early-stage subcortical ischemic vascular disease, Alzheimer’s disease and normal ageing. PLoS One 12, e0175143.

48.

Yeh

, Li

, Chen

(1997) The perceptual dimensions underlying the classification of the shapes of Chinese characters. Chinese J Psychol 39, 47–74.

49.

Scullin

, McDaniel

, Shelton

(2013) The dynamic multiprocess framework: Evidence from prospective memory with contextual variability. Cogn Psychol 67, 55–71.

50.

Henry

, MacLeod

, Phillips

, Crawford

(2004) A meta-analytic review of prospective memory and aging. Psychol Aging 19, 27–39.

51.

Kliegel

, Jager

, Phillips

(2008) Adult age differences in event-based prospective memory: A meta-analysis on the role of focal versus nonfocal cues. Psychol Aging 23, 203–208.

52.

Schnitzspahn

, Kvavilashvili

, Altgassen

(2020) Redefining the pattern of age-prospective memory-paradox: New insights on age effects in lab-based, naturalistic, and self-assigned tasks. Psychol Res 84, 1370–1386.

53.

McDaniel

, Shelton

, Breneiser

, Moynan

, Balota

(2011) Focal and nonfocal prospective memory performance in very mild dementia: A signature decline. Neuropsychology 25, 387–396.

54.

Niedźwieńska

, Kvavilashvili

(2018) Reduced mind-wandering in mild cognitive impairment: Testing the spontaneous retrieval deficit hypothesis. Neuropsychology 32, 711–723.

55.

Henry

(2021) Prospective memory impairment in neurological disorders: Implications and management. Nat Rev Neurol 17, 297–307.

56.

Massa

, Grisanti

, Brugnolo

, Doglione

, Orso

, Morbelli

, Bauckneht

, Origone

, Filippi

, Arnaldi

, De Carli

, Pardini

, Pagani

, Nobili

, Girtler

(2020) The role of anterior prefrontal cortex in prospective memory: An exploratory FDG-PET study in early Alzheimer’s disease. Neurobiol Aging 96, 117–127.

57.

Burgess

, Dumontheil

, Gilbert

(2007) The gateway hypothesis of rostral prefrontal cortex (area 10) function. Trends Cogn Sci 11, 290–298.