Genetic Risk for Alzheimer’s Disease Alters Perceived Executive Dysfunction in Cognitively Healthy Middle-Aged and Older Adults

INTRODUCTION

Aging is accompanied by a myriad of changes related to physical, mental, and cognitive health. From a cognitive perspective, a modest degree of objective cognitive decline across older adulthood is well- documented and considered typical during the aging process [1, 2]. However, some of these individuals will exhibit severe cognitive decline that progresses to dementia. Prior to this progression, a subset of these elders will perceive changes in their cognition that are not clinically meaningful on neuropsychological tests. These perceived changes are referred to as subjective cognitive complaints (SCC) [3, 4].

Although SCC are prevalent and are not unique to any one etiology underlying cognitive decline, mounting evidence suggests they may specifically portend a diagnosis of Alzheimer’s disease (AD) [5–7]. Specifically, individuals who endorse these complaints are 4.5 times more likely than others to later convert to either mild cognitive impairment (MCI) [8], a transitional stage between normal cognitive aging and dementia, or to dementia, within a seven-year timeframe [9, 10]. Over four years, a recent meta-analysis of longitudinal SCC studies demonstrated the 4-year rate of conversion to MCI was 27% while rate of conversion to dementia was 14% [11]. SCC endorsement is also correlated with biomarkers indicative of impending AD, including hippocampal atrophy [12], elevated amyloid-β deposition (i.e., amyloid burden) [13–21], and altered brain activity [17, 22].

Although the SCC construct is still somewhat controversial due to heterogeneity in the literature, comprehensive reviews and meta-analyses support a significant relationship between SCC and objective cognitive functioning; other factors such as age, sex, and depression add further to the effect sizes [23, 24]. Recent large sample studies of elders with objectively intact cognition have furthermore shown even stronger effects linking SCC to poorer cognition, both globally and in specific domains such as episodic memory, semantic memory/language, and processing speed [19, 26], as well as greater cognitive decline over time in these domains and in visuospatial processing and working memory [25]. These effects held when controlling for depression and demographic factors. However, it is also important to acknowledge that metacognition is required for SCC to accurately reflect cognitive functioning [27]. Notably, anosognosia, the impairment in awareness of deficits in cognitive functioning, is a common feature of AD, particularly in the early stage [28–31], including in approximately 50% of those diagnosed with MCI, a prodromal phase of AD [32, 33]. Other studies report that elders with MCI [34] and also cognitively intact elders have hypernosognosia [35], or hyperawareness of their true deficits, but these findings have not yet successfully replicated [36]. Regardless, anosognosia, measured as the discrepancy between objectively and subjectively assessed cognition, may present an important opportunity for early detection of AD-related cognitive decline. Even the newest, promising pharmaceutical treatments for AD, which act by clearing amyloid deposition, are far more effective in the early stage [37], making any adjuvant predictors incredibly valuable.

The Apolipoprotein E (APOE) ɛ4 allele is a key non-modifiable risk factor for late-onset AD, the most common form of AD. In fact, ɛ4 is the most significant risk factor other than age [38], with carriers of at least one ɛ4 allele having three to four times increased likelihood of developing AD and carriers of both ɛ4 alleles having approximately 12 times increased likelihood [39, 40]. In contrast to other AD biomarkers, APOE ɛ4 is quickly, non-invasively, and inexpensively measured [38]. Furthermore, it is associated with lower cognitive scores and greater amyloid burden after accounting for other AD risks such as sex, education, and lifestyle factors [41, 42]. Importantly, SCC has been shown to be indicative of elevated amyloid burden in younger elders, in particular, in ɛ4-carriers [43]. Moreover, APOE ɛ4 is a primary predictor of future objective cognitive decline in SCC endorsers [44, 45]. Cognitively healthy ɛ4 carriers exhibit compensatory brain activity [46, 47] and accelerated brain atrophy rates [48, 49], as well as a significantly faster rate of cognitive decline overall [10, 51]. While the neuropathology of AD (e.g., amyloid and tau burden) is a primary focus in the current literature, these markers only approach clinical significance, on average, six years before diagnosis of AD [10]. Given that APOE ɛ4 is static and indicative of amyloid burden in those who endorse SCC, the combination of SCC and ɛ4 may provide a valuable proxy for early assessment of AD risk.

Notably absent in the literature are studies of SCC outside of the domain of memory. Although SCC is not limited to memory functions, memory has been the focus of SCC research due to the ubiquity of episodic memory impairment in AD [52, 53]. However, there is accumulating evidence that very early changes also occur in non-amnestic domains, such as in executive functioning [19, 54], including in cognitively intact elders who have family history of AD [55] or who carry ɛ4 [46, 47]. Executive functioning is an umbrella term for multiple interacting cognitive processes that underlie goal-directed behaviors [56], including working memory (updating), task switching, and countermanding dominant or prepotent responses (i.e., inhibitory control) [57, 58]. These processes, predominantly mediated by the prefrontal cortex [59, 60], play an important role in the successful completion of activities of daily living, a fundamental component of the AD diagnosis. Importantly, Grober et al. [61] found tests of executive functioning to have comparable predictive validity to tests of memory function, particularly several years before an MCI or AD diagnosis, which may coincide with the onset of neural dysfunction (i.e., compensation) [62, 63] and when individuals begin to endorse SCC. Some evidence suggests that, amongst those with SCC, baseline executive functioning and language are more closely associated with progression to non-amnestic MCI, while verbal memory is more associated with amnestic MCI [64].

Despite an absence of objective memory impairment, individuals with SCC in memory have evidenced poorer performance, in some cases comparable to those diagnosed with MCI, on non-amnestic cognitive tasks that tap frontal executive networks including, sustained and divided attention [65], verbal fluency [66–68], visual working memory [69], inhibition and interference [67, 68]. The link between memory complaints and executive functioning also corresponds to greater AD neuropathology, such as amyloid burden [70, 71] and future dementia [72]. Cognitively normal individuals with elevated amyloid burden have also been shown to endorse more SCC specific to language and executive functioning than those with low amyloid burden [68]. As such, subjective executive functioning (SEF) is a gap in the literature that needs study as a predictor of risk for AD.

The current study examined the relationship between subjective and objective executive functioning (i.e., awareness of executive functioning), in cognitively intact middle aged and older adults who were stratified by risk for AD using the APOE ɛ4 allele. We hypothesized that the relationship between subjective and objective executive functioning would be moderated by APOE ɛ4, such that individuals at a greater risk for AD (APOE ɛ4 carriers; ɛ4+) would have reduced awareness of their executive functioning relative to non-carriers (ɛ4-).

MATERIALS AND METHODS

RESULTS

Moderation model

We hypothesized that the relationship between objective executive functioning and SEF would be moderated by APOE ɛ4, such that ɛ4+ older adults would be less aware of their actual executive functioning than ɛ4- participants. This theoretical model is depicted in Fig. 1.

OPEN IN VIEWER

Fig. 1

Theoretical model depicting APOE ɛ4 (W) moderating the relationship between objective executive functioning (X, predictor) and subjective executive functioning (Y, outcome).

The results of the moderation analysis are presented in Table 3. After accounting for the contributions of age, depression (GDS) and anxiety (BAI), objective executive function performance (EF_Factor) was significantly associated with subjective executive function (EXECDYS), and APOE ɛ4 moderated this relationship. Specifically, those with poorer objectively assessed executive function had greater perceived executive dysfunction, but this was clarified by an interaction with ɛ4. In ɛ4-, those with poorer objective executive functioning also had greater perceived executive dysfunction, while in ɛ4+, those with poorer objective executive functioning had less perceived executive dysfunction. Thus, ɛ4+ were less accurate in assessing their executive functioning than ɛ4-. These differing patterns are shown in Fig. 2.

OPEN IN VIEWER

Table 3

Moderation model

	Model summary
	R	R2	MSE	F(6,47)	p
EXECDYS	0.58	0.34	39.12	3.94	0.003
	Model parameters
	Coeff.	se	t	p	LLCI	ULCI
Constant	21.557	7.692	2.803	0.007	6.083	37.031
EF_Factor	–2.878	1.233	–2.334	0.024	–5.359	–0.397
ɛ4	–1.793	1.973	–0.986	0.368	–5.762	2.176
Interaction	3.873	1.832	2.114	0.04	0.187	7.559
Covariates
Age	0.114	0.116	0.99	0.327	–0.118	0.347
GDS	1.291	0.628	2.058	0.045	0.029	2.554
BAI	0.589	0.323	1.822	0.075	–0.061	1.24
Highest order unconditional interaction	Δ R 2	F(1,47)	p
EF_Factor X ɛ4	0.063	4.468	0.04
Conditional effects	Effect	se	t	p	LLCI	ULCI
ɛ4-	–2.878	1.233	–2.334	0.024	–5.359	–0.397
ɛ4+	0.995	1.53	0.65	0.519	–2.083	4.072

EXECDYS, Frontal Assessment Battery executive dysfunction subscore (subjective executive functioning); EF_Factor, factor score of objective executive functioning battery; ɛ4, APOE ɛ4; Coeff, unstandardized coefficient; MSE, mean square error; LLCI, lower-level confidence interval; ULCI, upper level confidence interval; GDS, Geriatric Depression Scale total; BAI, Beck Anxiety Inventory total.

OPEN IN VIEWER

Fig. 2

Moderation analysis [simple-slopes analysis via PROCESS; 92] showing the interaction of objectively measured executive functioning (EF_Factor) and APOE ɛ4 group, predicting subjective executive functioning (EXECDYS). Age, depression, and anxiety were covaried. APOE ɛ4 carriers (ɛ4+) were less accurate in identifying executive concerns than ɛ4 non-carriers (ɛ4-). Specifically, in ɛ4-, perceived executive dysfunction was greater in those with poorer objective executive functioning, while in ɛ4+, the opposite pattern was evident; perceived executive dysfunction was lower in those with poorer objective executive functioning.

DISCUSSION

Mounting evidence suggests that SCC might foreshadow an AD diagnosis [5–7, 10] and its related accumulating neuropathology [13–21, 94]. Moreover, there has been little study of perceived dysfunction in non-amnestic cognitive domains, despite the demonstrated importance of domains such as executive functioning in the earliest signs of impending dementia [70, 95]. Even less is known about the role of APOE ɛ4, a primary risk factor for AD [38], in non-amnestic SCC. Thus, this study examined the relationship between subjective and objective executive functioning (i.e., awareness) in healthy, cognitively intact, middle-aged and older adults, as a function of APOE ɛ4. The results demonstrated that SEF was significantly correlated with objectively measured executive function performance in cognitively intact older adults, and, as predicted, ɛ4 carriers exhibited significantly less accurate awareness of their executive functioning performance than ɛ4 non-carriers. A post-hoc speculative analysis also reinforced that these results are specific to executive functioning, rather than cognition in general or memory and suggest that executive functioning may be more sensitive to very early dysfunction in awareness. These results therefore support the use of subjective executive dysfunction along with APOE ɛ4 as possible sensitive, early markers of impending cognitive decline. Specifically, poor awareness of cognitive functioning, otherwise referred to as anosognosia, in cognitively intact ɛ4 carriers may specifically portend the development of AD. Given the promising new amyloid clearing pharmaceutical treatments for AD are primarily effective in the early stage [37], anosognosia in executive function, particularly in ɛ4 carriers may be a valuable early adjuvant to predicting AD risk.

Our SEF findings add to the growing evidence in the memory domain showing a significant relationship between subjective and objective cognitive performance. The largest memory SCC studies indeed also show poorer cognitive functioning and greater decline over time in multiple domains outside of memory in those who endorse SCC [19, 26], particularly in complex cognitive functions that rely on frontal executive networks [65–69]. Thus, our findings extend the literature beyond the memory domain into executive functioning, and further reinforcing the possible importance of executive function in the earliest signs of impending dementia [70, 95].

Our results do not directly address AD neuropathology. However, our findings are consistent with studies showing greater amyloid burden and elevated risk of cognitive decline in cognitively intact ɛ4 carriers who endorse SCC [43–45]. Indeed, since cognitively normal individuals with significant amyloid burden have been shown to endorse SCC specifically in the executive domains [68], our results reinforce the potential value of assessing subjective executive dysfunction as an early risk for future dementia. Moreover, while altered memory awareness has been shown in AD risk via amyloid burden, our results suggest that these alterations are comparably detectable simply via APOE ɛ4, a cost-effective, non-invasive method of examining AD risk. Additional research is needed to determine whether ɛ4 alongside subjective executive dysfunction might be able to detect risk for future cognitive decline at an earlier stage than amyloid burden. Indeed, our findings further suggest that ɛ4 status might be crucial to a clear assessment of the relationship between SCC and AD risk because only ɛ4 carriers had poor awareness of their functioning. It is also noteworthy that despite their elevated risk, only about half of all ɛ4 carriers go on to develop AD [96]. Yet, since those who do develop AD exhibit an accelerated rate of decline [10, 42], an early index of the degree of risk and need for early intervention could help to focus early assessment efforts on those most at risk, rather than attempting to indiscriminately evaluate all ɛ4 carriers. Thus, early impairment of awareness in ɛ4 carriers could possibly serve as a precursor sign for early assessment of amyloid and tau burden.

Our results are consistent with anosognosia [97, 98], rather than hypernosognosia [34–36], as a potential early indicator of future cognitive decline in high risk older adults who are currently cognitively intact. Yet, Cacciamani et al. [97] reported that although cognitively intact older adults with elevated amyloid burden had anosognosia, there was no difference in cognitive awareness between APOE ɛ4 carriers and noncarriers. Notably, while we used a self-appraisal of awareness, Cacciamani and colleagues operationalized awareness using the discrepancy between self- and informant appraisals, as a number of studies have suggested that informant appraisals may be more sensitive than self-appraisals [99–102]. However, a number of cross-sectional studies have also shown that self-report is as effective as informant reports when there is little actual impairment [103, 104], as is the case with our study. Indeed a review and meta-analysis concluded that self-reports may be better at indexing function prior to the transition to dementia, while informant reports may be better at the more advanced and transitional stage [102]. Thus, we can conclude that the ɛ4 carriers in our sample likely had more impaired awareness than in their sample. Perhaps more importantly, we note that the combination of subjective and objective assessment, as used in our study, might be more sensitive to anosognosia, and therefore to AD risk, than the discrepancy between two purely subjective assessments.

It is important to acknowledge that given the cognitively intact status of the participants and the cross-sectional design of the current study, it is not known which of these participants will ultimately exhibit cognitive decline or convert to MCI or AD, or whether they have elevated amyloid burden. However, the elevated risk for AD in the ɛ4 carriers is well established [39, 106]. Thus, longitudinal data that includes assessment prior to and after onset of symptoms would be helpful toward confirming the value of the assessment of subjective executive dysfunction to early AD prediction given the non-specific nature of SCC [107, 108]. Furthermore, consideration of life course factors, particularly those associated with cognitive reserve, will provide a more comprehensive understanding of variability in the trajectories of those with SCC. Menopause and estrogen are also potentially influential factors that were not considered in the current study. Additionally, we relied on a global subjective executive functioning score rather than perceived dysfunction related to individual processes of executive functioning. Synchronized assessment of both subjective and objective executive subprocesses might be even more illuminating, possibly with differential findings across executive subprocesses. Lastly, this study had a relatively small sample and did not assess the underlying neural mechanisms that might be responsible for subjective concerns or that might correlate earlier and more strongly with subjective dysfunction than objective testing. Understanding the relevant neural mechanisms may also better explain differences in awareness observed in cognitively intact older adults, including underestimation and overestimation of cognitive abilities.

In conclusion, the findings of the current study add to the small but growing body of literature on awareness of cognitive functioning, and specifically in executive functioning, in cognitively normal older adults who have elevated risk for AD. Our findings suggest that combined self-appraisal of executive functioning used alongside objective neurocognitive assessment may be helpful toward early identification of impending cognitive decline. This may be particularly valuable and important to discerning who amongst APOE ɛ4 carriers have substantively elevated risk for AD and thus may be more likely to exhibit future cognitive decline. Moreover, the current findings suggest that the assessment of SEF complements objective neurocognitive assessment of executive functioning and, taken with neuroimaging and other biomarkers, has potential to improve prediction of future cognitive decline while effective intervention is still possible.

CRediT AUTHOR STATEMENT

Sarah A. Evans (Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Validation; Visualization; Writing –original draft; Writing –review & editing); Elizabeth R. Paitel (Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Validation; Visualization; Writing –original draft; Writing –review & editing); Riya Bhasin (Data curation; Investigation; Writing –original draft; Writing –review & editing); Kristy A. Nielson (Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Validation; Visualization; Writing –original draft; Writing –review & editing).