An amyloid sensitive composite model to determine subtle cognitive differences in preclinical Alzheimer’s disease

Background: Recently, the focus of Alzheimer's disease (AD) research has shifted from the clinical and symptomatic stage to the preclinical and asymptomatic stage. However, studies on the distinctive neuropsychological features of preclinical AD with adjusted measurement errors are insufficient. The present study aimed to investigate the distinctive cognitive features of preclinical AD and develop a cognitive composite model that can sensitively distinguish between the amyloid positive (Aβ+) and amyloid negative (Aβ-) status in cognitively normal (CN) elderly participants. Methods: A total of 423 CN elderly participants with amyloid positron emission tomography (PET) images were recruited. Using the multiple-indicator multiple-cause (MIMIC) model, sensitive cognitive domains to the Aβ+ group were found. Then, several cognitive tests were selected to create the cognitive composite model as a result of the multivariate analysis of covariance (MANCOVA). Results: The MIMIC model revealed that the domains of episodic memory (verbal and visual) and executive functions were significantly different between Aβ- and Aβ+ ( p <0.05) participants. According to MANCOVA, the Seoul Verbal Learning Test-Elderly’s version (SVLT-E) and the Rey-Osterrieth Complex Figure Test (RCFT) both distinguished between the Aβ+ group and the Aβ- group on the delayed recall tests in terms of episodic memory ( p <0.1). Furthermore, on tests of executive functions, the Aβ+ group performed worse than the Aβ- group in color reading of the Korean-Color Word Stroop test (K-CWST) and the animal naming of the Controlled Oral Word Association Test (COWAT) ( p <0.1). Consequently, the Preclinical Amyloid Sensitive Composite (PASC) model was comprised of SVLT-E delayed recall, RCFT delayed recall, K-CWST color reading, COWAT-animal, and the Korean Mini-Mental State Examination (K-MMSE) according to the results from our study and previous studies. Conclusions: In the present study, we developed our own composite PASC model with the distinctive cognitive profiles of Aβ+ CN individuals. Hence, this composite model can eventually contribute

and asymptomatic stages of AD [1]. Consequently, approximately 20 to 30% of cognitively normal (CN) elderly population is expected to be preclinical AD with amyloid positivity [2][3][4]. These CN individuals with elevated amyloidosis are considered to be more vulnerable to AD progression. This can be demonstrated by the subtle cognitive difference between CN individuals with amyloid positivity and those without Aβ biomarkers in the late preclinical stage [5]. Furthermore, several reports have claimed that approximately 25% of preclinical AD converts to prodromal AD or AD dementia in approximately 3 years [6,7].
Due to the high cost and health risks of Aβ PET, it is usually challenging to obtain a large number of participants with PET data. Thus, it becomes pragmatic and essential for clinicians to predict who might be at high risk of having an Aβ + biomarker with the expectation that future treatment may target Aβ. Given that several prevention trials are currently being conducted in preclinical AD, we need to investigate the distinctive neuropsychological features of preclinical AD, which may help clinicians predict preclinical AD by reducing screen failures and monitoring the therapeutic efficacy of prevention.
Despite attempts to investigate the distinctive neuropsychological features of preclinical AD, the results were inconsistent among studies to date. This may be because previous studies did not consider the effects of measurement errors, although there is always a possibility for the presence of measurement errors in psychometrics [8]. However, by using the principle of factor analysis, the multiple-indicator multiple-cause (MIMIC) model can control for these measurement errors to estimate the latent values. That is, MIMIC may empower a composite model to sensitively detect subtle cognitive differences between Aβ + and Aβ-in CN elderlies. However, to our knowledge, no preclinical AD studies have developed a cognitive composite model using this factor structure method for identifying preclinical AD in the elderly population.
In the present study, we aimed to determine if there are any distinct cognitive domains and cognitive measures of preclinical AD using the MIMIC model, which may yield lower background noise in measurements. We also developed and validated the Preclinical Amyloid Sensitive Composite (PASC), a composite model that precisely distinguishes Aβ + and Aβ-in CN individuals. Considering that subtle deficits in episodic memory and executive functions appear to be critical in preclinical AD due to their strong association with AD progression [9], we hypothesized that memory and executive functions would be the cognitive domains with a significant difference between Aβ + CN and Aβ-CN individuals.

Study participants
A total of 423 CN participants were recruited from September 2015 to December 2018 at the Samsung Medical Center in Seoul, South Korea. All the participants met the following criteria to be qualified as CN: (a) the Korean Mini-Mental State Examination (K-MMSE) ≥ 24 or above − 1.5 standard deviation (SD) from the age-, sex-, and education-adjusted norms if the education period was less than 9 years; (b) above − 1 SD from the age-, sex-, and education-adjusted norms on the delayed recall of the Seoul Verbal Learning Test-Elderly's version (SVLT-E); (c) above − 2 SD from the age-, sex-, and education-adjusted norms on the Korean version of the Boston Naming Test (K-BNT), the Rey-Osterrieth Complex Figure Test  All PET images were reviewed by trained nuclear medicine physicians who were blinded to patient information and dichotomized as Aβ-positive or negative using visual reads. The visual assessments for 18 F-florbetaben PET, 18 F-flutemetamol PET, and 18 F-florbetapir PET were performed with the scoring system that was used in the previous studies [10][11][12][13][14].

Neuropsychological assessments
The second edition of the Seoul Neuropsychological Screening Battery (SNSB-II) was administered to all the participants to assess their cognitive functions [15]. The SNSB-II was standardized on 1,067 CN The RCFT involved copying, immediate recall, 20-minute delayed recall, and recognition tests.
Similarly, the SVLT-E was composed of immediate recall trials, delayed recall, and recognition tests. In addition to the tests mentioned above, the K-MMSE was also used for the global mental state assessments of the participants [16].

Statistical analyses
Demographic characteristics were compared between the Aβ + and Aβ-groups using the independent sample t-test if the variables were continuous and the chi-square test if the variables were categorical.
Confirmatory factor analysis (CFA) was yielded to validate the structure of the five cognitive domains.
CFA is one of the multiple forms of structural equation modeling (SEM), which confirms whether a prespecified factor structure fits the data well [8]. We validated the CFA model for the neuropsychological test battery to control for measurement errors. The tests included in each cognitive domain were the same as those described earlier, and the language domain consisted of a single test score. The subtests of the SVLT-E and the RCFT in the memory domain were measured respectively using the same method. Therefore, it was considered acceptable to add an error covariance between the residual variances associated with SVLT-E immediate and delayed recalls and RCFT immediate and delayed recalls. Since our factor structure included both reflective and causal indicators, we used the MIMIC model to compare the latent means in the cognitive domains between the Aβ + and Aβ-groups ( Fig. 1).
Multivariate analysis of covariance (MANCOVA) was performed to see if any neuropsychological tests showed a significant difference between the two groups. Since the measurement errors were not treated in the MANCOVA, we deliberately set the cutoff for significance to be less conservative in order to increase the power and reduce the risk of type II errors. Thus, the tests with p-value < 0.1 were selected to be included in the composite model. Lastly, the MIMIC model was repeated to identify whether these tests were sensitive to differences between Aβ + and Aβ-in CN elderlies as a composite.
Raw scores were used in each test in all statistical analyses. The K-TMT-E part B was log-transformed for accuracy of the estimate due to its large range (0-300) and non-normality. Multiple imputation and full information maximum likelihood estimations were used to treat missing values.
IBM SPSS (version 25.0, SPSS Statistics/IBM Corp, Armonk NY, USA) was used for demographic analysis and MANCOVA. For comparisons of latent means between the groups, maximum likelihood estimation was analyzed by Mplus (version 8.0) [17]. However, due to the violation of normality, biascorrected bootstraps were performed together.

Demographic characteristics of participants
The demographic and neuropsychological characteristics of the study participants are presented in Table 1. The overall mean age of the participants was 69.9 years. Among the 423 participants, 75 were Aβ+ (17.7%). The frequency of APOE ε4 carriers was 24.6%.
The Aβ + group was significantly older than the Aβ-group (71.5 ± 6.8 years vs. 69.5 ± 8.4 years, p < 0.05). The Aβ + group also displayed a higher percentage of APOE ε4 carriers compared with the Aβgroup (53.3% vs. 16.4%, p < 0.001). However, the two groups did not significantly differ in education level and proportion of female participants. The result revealed that the difference between the Aβ + and Aβ-groups in attention, visuospatial function, and language function were not significant, but the latent means in the Aβ + group were significantly lower than the Aβ-group in the three domains of verbal memory, visual memory, and executive functions ( Table 2).  Aβ-groups when sex, education, and age were controlled. The result of the MANCOVA is shown in Table 3.
Few neuropsychological subtests showed meaningful differences between the groups. Primarily, we set the level of significance at 0.

Preclinical Amyloid Sensitive Composite (PASC)
Based on the MANCOVA results and clinical significance, the following 5 tests were finally selected: the SVLT-E delayed recall; the RCFT delayed recall; the K-CWST color reading; the COWAT animal; and the K-MMSE. The K-MMSE was added for examining global cognition. The PASC model presented a good fit with the data (x 2 = 4.757, p = .933; RMSEA = .000; CFI = 1.000; TLI = 1.001; SRMR = .014). All factor loadings in the model were significant between .56 and .73 (Fig. 2).

Discussion
We investigated the distinctive neuropsychological features of Aβ + CN elderlies in a carefully phenotyped, CN cohort that underwent detailed neuropsychological tests, MRI, and amyloid PET scans with the standardized protocols. Accordingly, there were several significant neuropsychological findings in this study. First, the MIMIC model found a difference in the latent mean between the Aβ + and Aβ-groups in the domains of verbal memory, visual memory, and executive functions.
Furthermore, MANCOVA showed that the Aβ + group performed worse in the RCFT delayed recall, SVLT-E delayed recall, COWAT animal, and K-CWST color reading within the three cognitive domains.
Lastly, based on our results and clinical significance, we developed the PASC with the RCFT delayed recall, SVLT-E delayed recall, COWAT animal, K-CWST color reading, and K-MMSE that were found to be critical for amyloid deposition and global cognition. Hence, the PASC can be used to develop a composite score for detecting Aβ positivity in CN individuals, which can eventually help with early intervention of AD.
The demographic profile of our participants was extremely close to that of the previously reported Asian society profile. The CN Aβ + percentage in Asian countries is known to be lower than that in western countries. The percentage of CN amyloid positivity in the Asian population ranged between 18% and 25% according to the Korean Brain Aging Study for the Elderly Diagnosis and Prediction of Alzheimer's disease (KBASE) and J-ADNI [18,19]. On the other hand, the western population, represented by ADNI, was reported to range approximately from 25 to 45% Aβ positivity rate [20,21].
Our study exhibited approximately 18% Aβ positivity in the 423 CN individuals, which was in line with that in the Asian population. The discrepancy between our results and that of the western society may be explained by the differences in the frequency of APOE ε4 and the age of the study participants. Our cohort seemed to have a lower percentage of APOE ε4 (23%) than that reported by ADNI (27%) [22]. Moreover, the younger age of our cohort (mean, 69.9 years) compared to that of the ADNI CN individuals (mean, 75.8 years) may have affected the lower rate of amyloid positivity [22].
Regardless of these disparities, the APOE ε4 rate and age of our cohort were still at comparable levels to J-ADNI's APOE ε4 rate (24%) and CN individuals' ages (mean, 67.9) [18].
Our major finding was that the Aβ + CN individuals presented a lower performance in verbal memory, visual memory, and executive functions compared to Aβ-CN, which was generally consistent with the findings of previous meta-analyses. In terms of memory, there has been a consensus that episodic memory has a strong association with Aβ burden [23][24][25]. Unlike episodic memory, the results regarding executive functions in the previous studies are not entirely consistent. A recent metaanalysis suggested a significant difference in executive function [24], while two others showed either a small effect size or a weak association with Aβ burden [23,25]. This may be because the previous studies did not consider the effects of measurement errors that could impact the individual test scores. However, applying a factor analysis with the latent variables, we controlled for the measurement errors from each test score for more precise measurement of the corresponding cognitive function.
In the present study, we developed the PASC to predict amyloid positivity using the SVLT-E delayed recall, the RCFT delayed recall, the COWAT animal, and the CWST color reading from the three cognitive domains, and the K-MMSE. We included the K-MMSE in the PASC because the Mini-Mental State Exam (MMSE) is a practical neuropsychological test to examine individual cognitive function holistically [26]. The PASC seems similar to the Preclinical Alzheimer Cognitive Composite (PACC) [27].
However, the PACC was not developed for differentiation of preclinical AD, but rather has been used in prevention trials for preclinical AD [1,27].
There are a few reasons why we tried to observe cognitive differences in CN individuals with a unidimensional outcome. First, there is a need to create a novel and reliable measure to holistically assess cognitive domains specific to preclinical AD. AD pathology progression involves the deterioration of multiple cognitive domains instead of a single cognitive function. Currently, the MMSE [26] and the Clinical Dementia Rating (CDR) [28] are commonly used to assess individual cognitive function holistically. However, they often display ceiling effects in CN individuals [29,30]. Therefore, they are not sensitive measures for CN individuals. Furthermore, the ratings of the CDR primarily rely on clinicians' judgments following patient and caregiver interviews. Thus, bias is rarely avoidable.
Another advantage of obtaining a composite model is that a single factor model allows for a more precise differentiation of distinct groups. Compared to multi-outcomes, a primary outcome usually yields lower background noise in the measurement, which derives a lower risk of Type-I error [8,31].
Therefore, a primary outcome has better reliability and sensitivity in terms of detecting subtle cognitive differences.
The strength of our study is the large sample size of the CN cohort who underwent amyloid PET. In spite of this strength, there are a few limitations to our study, as well. First, our composite model was not cross-validated. Future studies are needed to cross-validate this composite model in different patient groups and data sets. Second, the participants went through different types of PET ligands.
The variety of the tracers may have affected the visual reads of amyloid deposition. However, this limitation can be somewhat alleviated by the high correlations among the different ligands [32,33].
Moreover, we did not explore the clinical effects of the PASC. Future studies with clinical impacts of the PASC on other biomarkers like tau or cortical atrophy may be recommended. Lastly, our study was a single-center study with the cohort recruited from a hospital. Since we did not recruit the participants from the general population, the generalizability of our study may be challenging.

Conclusions
Our study created the PASC that is a sensitive cognitive composite model for Aβ + in CN elderly individuals, since we also investigated some distinctive cognitive features of Aβ + in CN elderly individuals. The PASC, which employed significant tests in episodic memory and executive functions, along with the global cognitive measure of the K-MMSE, can be used to develop a sensitive composite  The CFA model of the PASC Figure 3 The MIMIC model of the PASC for latent mean comparison between Aβ+ and Aβ-