You are viewing a javascript disabled version of the site. Please enable Javascript for this site to function properly.
Go to headerGo to navigationGo to searchGo to contentsGo to footer
In content section. Select this link to jump to navigation

Is the Discrimination of Subjective Cognitive Decline from Cognitively Healthy Adulthood and Mild Cognitive Impairment Possible? A Pilot Study Utilizing the R4Alz Battery

Abstract

Background:

The early diagnosis of neurocognitive disorders before the symptoms’ onset is the ultimate goal of the scientific community. REMEDES for Alzheimer (R4Alz) is a battery, designed for assessing cognitive control abilities in people with minor and major neurocognitive disorders.

Objective:

To investigate whether the R4Alz battery’s tasks differentiate subjective cognitive decline (SCD) from cognitively healthy adults (CHA) and mild cognitive impairment (MCI).

Methods:

The R4Alz battery was administered to 175 Greek adults, categorized in five groups a) healthy young adults (HYA; n = 42), b) healthy middle-aged adults (HMaA; n = 33), c) healthy older adults (HOA; n = 14), d) community-dwelling older adults with SCD (n = 34), and e) people with MCI (n = 52).

Results:

Between the seven R4Alz subtasks, four showcased the best results for differentiating HOA from SCD: the working memory updating (WMCUT-S3), the inhibition and switching subtask (ICT/RST-S1&S2), the failure sets (FS) of the ICT/RST-S1&S2, and the cognitive flexibility subtask (ICT/RST-S3). The total score of the four R4Alz subtasks (R4AlzTot4) leads to an excellent discrimination among SCD and healthy adulthood, and to fare discrimination among SCD and MCI.

Conclusion:

The R4Alz battery is a novel approach regarding the neuropsychological assessment of people with SCD, since it can very well assist toward discriminating SCD from HOA. The R4Alz is able to measure decline of specific cognitive control abilities - namely of working memory updating, and complex executive functions - which seem to be the neuropsychological substrate of cognitive complaints in community dwelling adults of advancing age.

INTRODUCTION

Nowadays, the percentage of people with neurodegenerative diseases is rapidly rising worldwide [1]. There is scientific evidence indicating that cognitive abilities start to deteriorate several years before the clinical detection of dementia [2]. Older adults who are cognitively healthy (CHA) do not present complaints regarding their cognitive abilities apart from commonly appearing difficulties, such as lag time for names or actions and lapses in concentration, which are usually correlated with affective disorders and general mental health [3, 4]. On the other hand, people with subjective cognitive decline (SCD) are people who usually worry about their cognitive abilities and express prospective memory problems, such as forgetting appointments in the near future, or having episodic memory complains, such as not recalling details of recent events [5]. These cognitive complaints are associated with high rates of cognitive deterioration in daily life and a family history of major neurocognitive disorders that force people with SCD to seek medical attention [6]. Besides the fact that people with SCD present complaints and great worry regarding their general cognitive function, they do not appear to have any measurable decline according to common neuropsychological assessment [7]. In contrast, people with mild cognitive impairment (MCI) or minor neurocognitive disorders (MND) address more cognitive difficulties than expected for their age, which are obvious from the decreased neuropsychological performance (even though significant daily life problems are not usually presented) [8]. Finally, people with major neurocognitive disorders present a major decline in cognitive abilities, causing difficulties in Activities of Daily Life (ADL), with Alzheimer’s disease (AD) being the most common cause of dementia in people over 65 years old [9]. The pathology of AD is extended at least a decade-long before the relative onset of the symptomatology which will set the diagnosis [2].

What we currently know is that adults with SCD and MCI comprise two heterogeneous groups who may or may not be at risk of developing neurocognitive disorders and especially AD. Specifically, there are studies supporting the idea that people with SCD are in great risk to develop MCI or AD [10, 11]. However, there is still a controversy; several other studies consider that SCD is usually associated with mood disorders, personality factors, or chronic health problems, and it presents no association with neurocognitive disorders [12–14]. However, according to the recently diagnostic criteria proposed by SCD-I Working Group, the feelings of worse memory performance should not be associated with the presence of depressive symptoms [7]. On the other hand, MCI is considered to be the preclinical stage of neurocognitive disorders since a great percentage of people with MCI, especially with AD pathology, convert to dementia over the next years [15].

Currently, besides the fact that several neuropsychological tests were developed to capture as soon as possible cognitive decline related to AD, they are usually affected either from the means of the test administration (e.g., tablets or PCs, that older adults are not familiar to their use) [16], or they are education dependent [17]. Moreover, it seems that they are not able to differentiate between cognitively healthy lifespan development and SCD [18]. Therefore, there is an urgent need to develop and implement new cognitive tools that are able to differentially diagnose the two ‘categories’.

The R4Alz battery was designed in order to diagnose cognitive control and attentional control deficits in people with mild and major neurocognitive diseases via subtasks that assess working memory components, attention control abilities, inhibition and switching, and cognitive flexibility [19]. In this study, we tried to examine, at a preliminary level, if the R4Alz battery could successfully detect the earliest signs of the disease at a pre-symptomatic stage, mainly in community-dwelling older adults with SCD, and differentiate them from the spectrum of CHA. Given the knowledge that MCI is considered a potential next step of SCD, at a secondary level, the ability of the R4Alz to differentiate SCD from MCI is examined as well.

The purpose and the hypotheses of the study

The main purpose of the present study was to investigate the potential of discrimination with R4Alz battery of cognitively healthy young adults (HYA), cognitively healthy middle-aged adults (HMaA), cognitively healthy older adults (HOA), as well as MCI patients from people with SCD.

To achieve the purpose, this study aimed 1) to examine whether the performance on every subtask of the battery is affected by gender and educational level, as well as 2) to examine each subtask’s ability to discriminate SCD from the aforementioned cognitive conditions.

Therefore, the study’s hypotheses are formulating as follow:

Hypothesis 1: The battery’s subtasks would not be affected by demographic variables such as gender or education.

Hypothesis 2: The R4Alz total score would adequately differentiate older adults with SCD from the three groups of CHA (young-middle-aged-older), as well as from people with MCI. Specifically, older adults with SCD would be found to present a lower level of cognitive control (at least regarding more complex abilities), as compared to the groups of CHA. On the other hand, they would be found to display a higher level of cognitive control, as compared to MCI patients.

METHODS

Design

Five different diagnostic groups have participated in the study: 1) community-dwelling HYA, 2) community dwelling HMaA, 3) community dwelling HOA, 4) community dwelling adults who fulfilled the criteria for the diagnosis of SCD [7], and 5) people diagnosed with MCI [20]. At this point, it should be mentioned that for the needs of the main aim of the present study only the results regarding the SCD group with the other four (the three groups of the CHA and MCI) will be analyzed, presented, and interpreted since another study is already being performed regarding the differences among the MCI group and the three healthy adult groups. In order to explore which is the best cutoff score that discriminates older adults with SCD from the three cognitively healthy adult groups, two different formulas will be used. The first one will comprise the overall tasks’ score, while the second will include only the tasks with significant differential potential among SCD and HOA groups, which are the most cognitively similar group, compared with the younger groups and the pathological one.

Standard protocol approval and participant’s consents

All the participants were orally and in written informed for the purpose of the study and had the opportunity to ask questions. They were also informed that their data would be confidentially collected in an electronic database. The participants gave written informed consent at the time of their visit, agreeing that their participation was voluntary and that they could withdraw at any time, without giving a reason and without cost. Due to the specific type of the current research, demographic data such as age, gender, or occupation were selected. Since these are considered personal data, the European Union law that exists since 28 May 2018 was applied. According to the law, the use of sensitive personal data is allowed only due to research reasons. Therefore, the participants were informed accordingly and they also agreed that their personal data could be deleted from the web-database after a written request. The study’s protocol was approved by the Scientific and Ethics Committee of the Greek Association of Alzheimer Disease and Related Disorders, and followed the principles outlined in the Helsinki Declaration.

Participants

The total study sample consisted of adults from the broad area of Thessaloniki. In order to carry out the study and attract cognitively healthy people, an invitation brochure was designed. The invitation was 1) notified to the School of Psychology and the School of Electrical and Computer Engineering of the Aristotle University of Thessaloniki, 2) posted to the social networks of the Greek Association of Alzheimer Disease in Thessaloniki (Alzheimer Hellas), and 3) there was an invitation via the newsletter of the Alzheimer Hellas. Therefore, the subsample of the CHA (young and older) consisted of volunteers from the university campus and volunteers from the community. The sub-samples of people with SCD and MCI consisted of visitors of the Day Care Centre “Saint Helen” of the Alzheimer Hellas (DCCAH), in Thessaloniki, during the period of February 2019 to June 2019. The participants with SCD were people who visited the DCCAH for a yearly neuropsychological and psychological check-up routine. As far as the MCI group is concerned, they were participants of the cognitive rehabilitation programs at the DCCAH. Both groups voluntarily accepted to participate in the study, after a kind invitation.

The study sample included 175 people. The participants were categorized in five groups: 1) HYA, (n = 42, 14 men and 28 women, age range: 21 to 39 years, M = 26.00, SD = 5.78, education range: 13 to 23 years, M = 16.61, SD = 1.80), 2) HMaA, (n = 33, 11 men and 22 women, age range: 42 to 57 years, M = 49.57, SD = 4.25, education range: 6 to 22 years, M = 15.12, SD = 3.95), 3) HOA, (n = 14, 6 men and 8 women, age range: 60 to 76 years, M = 68.07, SD = 5.73, education range: 12 to 24 years, M = 16.42, SD = 2.70), 4) people with SCD, (n = 34, 6 men and 28 women, age range: 54 to 86 years, M = 69.14, SD = 6.24, education range: 6 to 21 years, M = 12.97, SD = 3.83), and 5) people with MCI (n = 52), 11 men and 41 women, age range: 58 to 84 years, M = 72.19, SD = 6.55, education range: 6 to 23 years, M = 11.88, SD = 4.16).

The Pearson’s chi square analysis was used for investigating the gender differences between groups. The analysis showed that the three cognitively healthy groups did not differ in gender (p = 0.790). According to one-way ANOVA analysis the three groups did also not differ in education (p = 0.210). Regarding education, the Scheffe post hoc comparisons showed that HYA did not significantly differ from HMaA, I–J = 0.23, p > 0.05, and from HOA, I–J = 0.14, p > 0.05, while HMaA did not differ from HOA, I–J = –0.09, p > 0.05. The three groups of HOA and adults with SCD and MCI, did not significantly differ in gender (p = 0.152), and age (p = 0.151), but they differed in education (p = 0.006). Regarding age, the Scheffe post hoc comparisons showed that HOA did not differ from people with SCD, I–J = –0.08, p > 0.05, and MCI, I–J = –0.31, p > 0.05, and people with SCD did not differ from people with MCI, I–J = –0.23, p > 0.05. As far as education is concerned, the Scheffe post hoc comparisons showed that HOA did not significantly differ in education from SCD, I–J = 0.61, p > 0.05; however, they differed from MCI, I–J = 0.78, p < 0.05. People with MCI had lower education than HOA while the SCD group did also not differ from MCI group, I–J = 0.17, p > 0.05. Study sample characteristics are presented in Table 1.

Table 1

Demographic characteristics of the participants of the study (n = 175)

Diagnosis
CharacteristicsHYAHMaAHOASCDMCIp*p**
(n = 42)(n = 33)(n = 14)(n = 34)(n = 52)
Age M (SD)26.00 (5.78)49.37 (4.21)67.20 (6.47)69.14 (6.24)72.19 (6.55)>0.05
Gender (Male/Female)14M/28W11M/22W6M/8F6M/28F11M/41F>0.05>0.05
Educational level (low/middle/high/very high)0/0/36/62/6/18/70/1/12/15/13/14/211/22/16/3>0.05<0.05
MMSE M (SD)29.45 (.88)29.15 (1,21)29.06 (.92)27.53 (2.25)<0.05
FUCAS M (SD)42.10 (.45)42.14 (.53)42.80 (.92)44.02 (3.65)<0.05

HYA, healthy young adults; HMaA, healthy middle-aged adults; HOA, healthy older adults; SCD, older adults with subjective cognitive decline; MCI, people with mild cognitive impairment; MMSE, Mini-Mental State Examination; FUCAS, Functional Cognitive Assessment Scale, p*, significant difference between the three groups of healthy adults; p**, significant difference between the three groups of older adults.

The three cognitively healthy control groups came from the general population. The exclusion criteria were: 1) history of psychiatric illness or affective disorder (Major Depression-General Anxiety Disorder), 2) substance abuse or alcoholism, 3) history of traumatic brain injury, 4) brain tumor, encephalitis, epilepsy history, Parkinson’s disease, stroke history, and other neurological disorders such as hydrocephalus, 5) cancer in the last 5 years, myocardial infarction in the last 6 months, or pacemaker, 6) thyroid or diabetes, 7) drug treatment with opioids, B12, folate, or thyroid, 8) sensory deficits, 9) drug treatment with opioids or medication for B12 vitamin deficiency, and 10) absence of subjective cognitive complaints. The above criteria were selected because the authors wanted to ensure to the larger possible extent that the adults control subjects do not have factors that can cause cognitive problems. All participants who were over 50 years of age completed an extended neuropsychological assessment, including a specific battery which is presented below, in order to determine precisely their cognitive status.

The group of SCD comprised people underwent an extended neuropsychological assessment as part of a yearly routine check-up. The assessment is used in order to support the diagnosis of SCD. The battery included depression scales (Geriatric Depression Scale and Beck Depression Inventory) [21–23], anxiety inventories such as the Short Anxiety Screening Test [24, 25] and the Beck Anxiety Inventory [26] for the exclusion of affective disorders, and the Neuropsychiatric Inventory [27, 28], for the exclusion of neuropsychiatric symptoms. The Mini-Mental State Examination (MMSE) [29, 30] was also used, as the most well-known and commonly used short and reliable screening tool for the assessment of general cognitive performance. The Functional Cognitive Assessment [31] was used in order to assess their ability to organize and execute six different Activities of Daily Living (ADL: telephone communication, shopping, orientation in place, taking of medication, personal hygiene, and clothing) directly from the subject. Furthermore, tests for the assessment of general cognitive and functional abilities, memory, language, executive function, and attention were used as well. The entirety of the neuropsychological tests included in the battery is presented in detail in Tsolaki et al., 2017 [32].

People with SCD were older adults who expressed worries regarding their cognitive difficulties, mostly in naming, prospective memory, and attention abilities. The inclusion criteria for people with SCD, were based on the diagnostic criteria proposed by SCD-I Working Group [7]: 1) feelings of worse memory performance, not associated with the presence of depressive symptoms, 2) absence of objective cognitive deficits, according to the neuropsychological tests, and 3) stage 2 of the disease according to Global Deterioration Scale [33]. As far as the exclusion criteria are concerned, these were the same with the HOA group apart from the absence of subjective cognitive complaints.

As far as people with MCI are concerned, they are people who participated in the cognitive training programs of the DCCAH. For their diagnosis, the DSM-5 criteria for mild neurocognitive disorders were followed [20]. The diagnosis was also supported by neurological examination, neuropsychological and also neuropsychiatric assessment, neuroimaging (computed tomography or magnetic resonance imaging) and blood tests, by a consensus of specialized health professionals, considered experts in neurocognitive disorders. The inclusion criteria were: 1) diagnosis of Minor Neurocognitive Disorders according to DSM-5, 2) MMSE total score≥26, 3) stage 3 of the disease according to Global Deterioration Scale, and 4) 1.5 standard deviation (SD) below the normal mean according to age and education, in at least one cognitive domain according to the utilized neuropsychological tests. The battery of the neuropsychological tests which was used is presented above. Regarding the exclusion criteria, it were the same as the other groups of our study.

Tools

The R4Alz battery is a system that does not require the examinee to operate a tablet, PC, or other electronic tools for the implementation of the clinical assessment, rather, it utilizes the REMEDES system, which measures reflexes using visual and auditory triggers (https://remedes.eu/). The system comprises physical, three-dimensional devices, the so-called REMEDES pads. Every R4Alz battery setup includes seven (7) REMEDES devices/pads that can be activated providing visual or auditory stimuli and can be accompanied by a figure/image, depending on the task. For the needs of the R4Alz battery, specific figures were printed and placed in front of every pad. The battery’s instructions are both verbal and non-verbal (sketches). According to the task’s instructions the examinee is asked either to deactivate the REMEDES pads or to count them or both and to react verbally to specific instructions. Via this system, the proposed battery becomes ecologically valid and alleviates the issues that generate assessment faults for older adults and people with cognitive diseases.

The R4Alz battery was designed to evaluate cognitive control abilities. It comprises three main tasks. The first task assesses Working Memory Capacity & Updating (WMCUT). It includes 1) a working memory (WM) component subtask of short-term store (WMCUT-S1), 2) a WM component subtask of Central Executive (processing) (WMCUT-S2), and 3) a WM updating subtask & WM component of episodic buffer (WMCUT-S3), which include 6 conditions. The second task of the R4Alz battery assesses attention control (ACT). It consists of 9 subtasks: 1) a visual selective attention subtask, 2) an auditory sustained attention subtask, 3) a divided attention subtask (3 levels of difficulty), 4) a divided auditory attention subtask, 5) a divided visual-auditory attention subtask and, 6) a divided visual-auditory attention and visual - motion attention subtasks (2 levels of difficulty). The last task of the R4Alz battery assesses Inhibitory Control & Task/Rule Switching (ICT/RST). The ICT/RST comprises three subtasks (ICT/RST-1, ICT/RST-2, and ICT/RST-3). The ICT/RST-1 subtask assesses inhibition and includes four conditions, the ICT/RST-2 subtask assesses inhibition and task/rule switching, and includes two conditions. Within the second condition of ICT/RST-S1&S2 which assesses inhibition and switching of attention, an additional score is being calculated (apart from the total score of ICT/RST-S2) that measures the number of the switching sets that had at least one failed step. The ICT/RST-3 subtask includes four conditions and assesses cognitive flexibility, and specifically a combination of inhibitory control and task/rule switching. The full design and setup methodology, as well as the detailed description of the tasks are presented in Poptsi et al. [19]. More information regarding the R4Alz battery with some examples with videos are also presented in https://r4a.issel.ee.auth.gr/the-r4alz-battery/.

Statistical analysis

For data analysis, the IBM SPSS Statistics for Windows (version 23.0; IBM Corp, Armonk, New York) was used [34]. The variables representing demographic characteristics of the sample were gender, and educational level. Regarding education the sample was categorized in 4 groups: a) 1–6 years, b) 7–12, c) 13–18, d) >19 years of education.

The analyses carried out included Multivariate Analysis of Variance (MANOVA) and subsequent ANOVAs, by using as independent variables the demographic characteristics (gender and education) and group, and as dependent variables the total scores on the seven R4Alz subtasks. Levene’s test was used to assess the equality of variances, and Box’s M Test for the assessment of the equivalence of covariance matrices. Partial eta-squared (ηp2) was used for the estimation of the effect size. Bonferroni correction was adopted for mean comparisons and multiple testing: an alpha level equal to 0.007 (0.05/7) was adopted.

Due to the fact that the group of the healthy older adult participants was small (smaller than the other four groups), the initial alpha level, equal to 0.05, regarding the comparisons with this group, was kept. The Scheffe test (plus Bonferroni correction) was used for post hoc multiple comparisons. The Scheffe post hoc test was chosen since it tests all possible comparisons, it is robust in relation to non-normality, and it provides maximum protection against type I error [35].

In order to formulate a total score of the tasks of the R4Alz battery, the individual scores were normalized and were changed to negative (the higher the score, the worst). Furthermore, two separate total scores were also created, one incorporating all subtasks (R4AlzTot7) and a second utilizing the R4Alz subtasks that seemed to better differentiate the HOA from the SCD groups (R4AlzTot4).

Receiver operating characteristic curve (ROC curve) analysis was also used for assessing the predictive value of the R4Alz tasks to discriminate the three healthy adult control groups and MCI group (HYA, HMaA, HOA, and MCI) from community dwelling adults with SCD. The cutoff points were determined by maximizing the Youden index [36]. The area under the curve (AUC) of the ROC curve was used in order to quantify the R4AlzTot7 and R4AlzTot4 discriminant potential in fair, good, perfect, or excellent according to the relative literature (AUC values from 1.0 are perfect, 0.9–0.99 is excellent, 0.8–0.89 is good, 0.7–0.79 is fair and 0.51–0.69 is a poor test [37, 38].

RESULTS

Before the examination of the R4Alz battery discriminant potential, we examined the effects of individual-demographic factors (gender and educational level) on the battery’s tasks’ scores.

At first, MANOVA was applied to the data of the R4Alz sub scores with individual-demographic factors and group as independent variables and performance on the seven subtasks (WMCUT-S1, WMCUT-S2, WMCUT-S3, ACT, ICT/RST-S1&S2, ICT/RST-S3, ICT/RST-FS) as the dependent variables. Diagnostic group effect was found significant, F(4,175) = 3.84, p < 0.001, while a gender effect was also noticed F(1,175) = 3.65, p < 0.05. A significant effect of education was not found.

Working memory capacity and updating task (WMCUT)

Specifically, the WMCUT-S1 (S1 - WM component of short-term store) was found to be significantly affected by gender (since in this ability males had better performance than women) F(1,175) = 13.91, p < 0.001, n2 = 0.09, and group, F(4,175) = 14.41, p < 0.001, n2 = 0.29. Scheffe post hoc comparisons showed that WMCUT-S1 discriminates SCD from HYA, I–J = –1.16, p < 0.001, and also SCD from HMaA, I–J = –1.09, p < 0.001.

As far as the WMCUT-Subtask 2 [S2 - WM component of Central Executive (processing)] is concerned, ANOVA showed that only group significantly affected performance, F(4, 175) = 6.933, p < 0.001, η2 = 0.16. Scheffe post hoc comparisons showed that WMCUT-S2 discriminates SCD from YHA, I–J = –1.22, p < 0.001, and SCD from HMaA, I–J = 1.05, p = 0.001.

The WMCUT-Subtask 3 (S3 - WM updating & WM component of the episodic buffer) was also affected by group, F(4, 175) = 7.421, p < 0.001, η2 = 0.17. Scheffe post hoc comparisons showed that WMCUT-S3, discriminates SCD from HYA, I–J = 2.42, p < 0.001, from MAaA, I–J = –2.33, p = 0.001, and from HOA, I–J = –2,13, p < 0.05.

Means and standard deviations regarding all groups at the WMCUT subtasks are presented in Fig. 1. As evident the performance at all WMCUT subtasks worsens when age or cognitive deficits increase.

Fig. 1

Means (M) and standard deviations (SD) for the three Working Memory Updating Subtasks (WMCUT-S1, WMCUT-S2, WMCUT-S3), of cognitively healthy young adults (HYA), healthy middle-aged adults (HMaA), cognitively healthy older adults (HOA), adults with subjective cognitive decline (SCD), and diagnosed with mild cognitive impairment (MCI). WMCUT-S1, working memory component of short-term store subtask; WMCUT-S2, working memory component of central executive (processing); WMCUT-S3, updating of working memory

Means (M) and standard deviations (SD) for the three Working Memory Updating Subtasks (WMCUT-S1, WMCUT-S2, WMCUT-S3), of cognitively healthy young adults (HYA), healthy middle-aged adults (HMaA), cognitively healthy older adults (HOA), adults with subjective cognitive decline (SCD), and diagnosed with mild cognitive impairment (MCI). WMCUT-S1, working memory component of short-term store subtask; WMCUT-S2, working memory component of central executive (processing); WMCUT-S3, updating of working memory

Attention control task (ACT)

The ACT was found to be affected by group, F(4, 175) = 24.59, p < 0.001, η2 = 0.41. With regards to the discriminative ability, this subtask was found to be able to differentiate people with SCD from HYA, I–J = 10.02, p < 0.001, as well as from HMaA, I–J = 8.05, p < 0.001, and SCD from MCI, I–J = –6.37, p < 0.001.

Means and standard deviations regarding all groups at the ACT are presented in Fig. 2. As evident the performance at ACT worsens when age or cognitive deficits increase.

Fig. 2

Means (M) and standard deviations (SD) for the attention control task (ACT), of cognitively healthy young adults (HYA), healthy middle-aged adults (HMaA), cognitively healthy older adults (HOA), adults with subjective cognitive decline (SCD), and diagnosed with mild cognitive impairment (MCI)

Means (M) and standard deviations (SD) for the attention control task (ACT), of cognitively healthy young adults (HYA), healthy middle-aged adults (HMaA), cognitively healthy older adults (HOA), adults with subjective cognitive decline (SCD), and diagnosed with mild cognitive impairment (MCI)

Inhibitory control and task/rule switching task (ICT/RST)

In relation to the ICT/RST-S1 & S2 (the total score of inhibition subtask plus the total score of the task/rule switching subtask), a significant effect of group, F(4, 175) = 11.98, p < 0.001, η2 = 0.25, was also noticed. According to Scheffe post hoc comparisons, the ICT/RST-S1 & S2 was found to discriminate SCD from all the other groups: from HYA, I–J = 4.33, p < 0.001, from HMaA, I–J = 3.59, p < 0.007, from HOA, I–J = 3.55, p < 0.05, and from MCI, I–J = –4.80, p < 0.01.

The ICT/RST-FS (the total number of failed sets on the inhibition and the task/rule switching subtask), was also found to be affected by group, F(4, 175) = 15.25, p < 0.001, η2 = 0.30. Regarding its discrimination ability, the ICT/RST-FS was found to be able to differentiate SCD from all the other groups: from HYA, I–J = 1.54, p < 0.001, from HMaA, I–J = 1.16, p < 0.007, as well as from HOA, I–J = 1.26, p < 0.05, and from MCI, I–J = –1.57, p < 0.001.

Finally, regarding the ICT/RST-S3 (cognitive flexibility subtask, which comprise a combination of inhibition and task/rule switching functions) was also found to be affected by group, F(4, 175) = 13.53, p < 0.001, η2 = 0.28. Regarding its discrimination ability, the ICT/RST-FS was found to be able to differentiate SCD from HYA, I–J = 7.95, p < 0.001, from HMaA, I–J = 6.05, p < 0.001, and from HOA, I–J = 5.14, p = 0.004.

Means and standard deviations regarding all groups at the ICT/RST subtasks are presented in Fig. 3. Similarly to the WMCUT and ACT tasks, ICT/RST performance worsens when age or cognitive deficits increase.

Fig. 3

Means (M) and standard deviations (SD) for the Inhibitory Control and task/rule Switching subtasks (ICT/RST-S1&S2, ICT/RST-S3, ICT/RST-FS), of cognitively healthy young adults (HYA), healthy middle-aged adults (HMaA), cognitively healthy older adults (HOA), adults with subjective cognitive decline (SCD), and diagnosed with mild cognitive impairment (MCI). ICT/RST/FS, Cognitive Flexibility Subtask, failure of sets; ICT/RST-S1&2, Inhibition Subtask plus total score of the task/rule Switching subtask; ICT/RST-S3, Cognitive Flexibility Subtask, a combination of Inhibition and task/rule Switching

Means (M) and standard deviations (SD) for the Inhibitory Control and task/rule Switching subtasks (ICT/RST-S1&S2, ICT/RST-S3, ICT/RST-FS), of cognitively healthy young adults (HYA), healthy middle-aged adults (HMaA), cognitively healthy older adults (HOA), adults with subjective cognitive decline (SCD), and diagnosed with mild cognitive impairment (MCI). ICT/RST/FS, Cognitive Flexibility Subtask, failure of sets; ICT/RST-S1&2, Inhibition Subtask plus total score of the task/rule Switching subtask; ICT/RST-S3, Cognitive Flexibility Subtask, a combination of Inhibition and task/rule Switching

Discriminant potential of the R4Alz battery

As the discriminant potential of the R4Alz battery is concerned, it was decided to provide two different total scores: 1) a total score by including all subtasks of the R4Alz battery (R4AlzTot7), and 2) a second one comprising only the total score of the subtasks in which SCD and HOA were found to differ (R4AlzTot4). The concept behind R4AlzTot4 was based on the idea that besides the fact that people with SCD actually differ in many tasks with the cognitively healthy young and middle-aged adults, the damage which is probably correlated with the SCD, is mainly focused at the subtasks that seem to better differentiate the SCD from the HOA, which is the most similar group in terms of age with the adults of advancing age who present subjective complaints.

The R4Alz battery comprises several tasks and subtasks with a different type of rating. The WMCUT subtasks have positive rating; specifically, the total score of the WMCUT-S1 and the total score of WMCUT-S2, is the number of the correct answers, while the total score of WMCUT-S3 is the sum of the correct number of answers minus the errors. The ACT has negative rating since the total score is calculated as the sum of the total errors. The three subtasks of the ICT/RST also have negative rating; the scores of ICT/RST-S1 and ICT/RST-S2 are calculated together as a total score, while the ICT/RST-S3 subtask has a separate score. The ICT/RST-FS is a separate score from the ICT/RST-S1 and ICT/RST-S2 tasks, which also has a negative rating, since it measures the number of the total sets that the examinee failed to correctly complete. For the needs of the study, two total scores were created, one from all seven subtasks (R4AlzTot7) and one from the four subtasks that differentiate SCD from HOA (R4AlzTot4).

In order to formulate a total score, the next procedure was followed:

  • A) The tasks that were rated with a positive score were inverted, so as for all ratings to be negative in nature (i.e., to always measure errors).

  • B) All ratings were normalized to the [0, 1] range, using the real maximum and minimum scores of all groups as reference and not the tasks’ theoretical ones. This decision was taken, in order to correctly normalize the rates to the specific sample, since the theoretical maximum in some subtasks was quite large. Next, the normalization formulas for a variable x are presented:

  • x normalization with inversion: xn=(x)-x(x)-(x)xn=(x)-x(x)-(x)

  • x normalization: xn=x-(x)(x)-(x)

The normalization of each subtask’s scores is presented in Table 2.

Table 2

Normalization of R4Alz sub-tasks’ scores

NameRangeNormalized nameNormalization formula
WMCUT-S1 xWMS1[3, 7] Positive scoringxWMS1n 7-xWMS14
WMCUT-S2 xWMS2[2, 7] Positive scoringxWMS2n 7-xWMS15
WMCUT-S3 xWMS3[0, 14] Positive scoringxWMS3n 14-xWMS114
ACT xACT[0, 40] Negative scoringxACTn xACT40
ICT/RST-FS xIRFS[0, 6] Negative scoringxIRFSn xIRFS40
ICT/RST-S1&S2 xIR1&2[0, 23] Negative scoringxIR1&2n xIR1&223
ICT/RST-S3 xIR3[0, 35] Negative scoringxIR3n xIR335

WMCUT-S1, Working Memory component of short-term store subtask; WMCUT-S2, Working Memory component of Central Executive (processing); WMCUT-S3, Updating of Working Memory; ACT, Attention Control; ICT/RST/FS, Cognitive Flexibility Subtask, failure of sets; ICT/RST-S1&2, Inhibition Subtask plus total score of the task/rule Switching subtask; ICT/RST-S3, Cognitive Flexibility Subtask, a combination of Inhibition and task/rule Switching.

The final score was formulated as the Power Mean (or Hölder Mean)1 of the individual scores, denoted as x-(m)=(1ni=1nxim)1m , where m = 2 and n is the number of subtasks. Thus, the R4AlzTot7 and R4AlzTot4 scores have a typical range of [0, 1] and are formulated as such:

R4AlzTot7=20×17×(xWMS1n2+xWMS2n2+xWMS3n2+xACTn2+xIRFSn2+xIR1&2n2+xIR3n2)
R4AlzTot4=20×14×(xWMS3n2+xIRFSn2+xIR1&2n2+xIR3n2)

Each R4Alz subtask’s score was raised to the square power, so as to be stricter with scores that indicated a higher number of errors. For creating a more “human-friendly” cutoff score (i.e., not in the range of 0–1 but in the range of 0–10), we multiplied both total scores with a coefficient of 20. Of course, the multiplication of the total score with a coefficient does not alter the statistical analyses’ results. A point worth mentioning, concerning the “typical” range notion, is that if a future user has a score that is outside of the utilized ranges (e.g., a score over 35 in ICT/RST-S3), this does not affect the discrimination potential, as their total score will be increased, thus it will be more distant from the calculated cutoff, further strengthening the assumption of them being SCD or MCI (and not HOA).

ROC curves were computed to evaluate the diagnostic sensitivity and specificity of the R4AlzTot7 and R4AlzTot4. The AUC of the ROC curves showed that the discriminant potential of both scores that discriminate the SCD group from the healthy control ones and the MCI group, were ranged from excellent to fare. The results of the ROC curves in relation to the diagnostic groups are as follow and summarized in Table 3.

Table 3

Diagnostic R4Alz total score and subtask’s score classification between HYA, HMaA, HOA, SCD, and MCI

GroupsTasksCutoffAUCSensitivity %Specificity %95% CIp
SCD versus HYAR4AlzTot75.180.95688.287.80.917–0.996<0.001
R4AlzTot42.380.99110097.60.974–1.000<0.001
SCD versus HMaAR4AlzTot75.170.94188.287.10.888–0.994<0.001
R4AlzTot42.420.98110093.50.952–1.000<0.001
SCD versus HOAR4AlzTot77.210.78458.81000.656–0.9110.002
R4AlzTot42.930.96085.392.90.911–1.000<0.001
SCD versus MCIR4AlzTot79.100.78972.379.40.688–0.890<0.001
R4AlzTot45.820.75876.673.50.651–0.866<0.001

HYA, healthy young adults; HMaA, healthy middle-aged adults; HOA, healthy older adults; SCD, older adults with subjective cognitive decline; MCI, people with mild cognitive impairment; AUC, area under the curve; CI, confidence interval

Furthermore, in order to easily classify each participant according to their results in the subtasks, an electronic tool was created, capable of calculating both total scores, as well as indicating the group the participant seems to belong: https://r4alz.bss.design/.

SCD versus HYA

Regarding the ability of R4AlzTot7 score, to discriminate SCD from HYA, a cutoff of 5.18 resulted in excellent sensitivity (88.2%) and specificity (87.8%) (AUC = 0.95 95% CI = 0.9–0.9, p < 0.001), while a cutoff of 2.38 of the R4AlzTot4 resulted in excellent sensitivity (100%) and specificity (97.6%) (AUC = 0.99 95% CI = 0.9–1.00, p < 0.001) (Fig. 4).

Fig. 4

ROC curve analysis of the R4AlzTot7 versus the R4AlzTot4 that seem to discriminate between subjective cognitive decline (SCD) and cognitively healthy young (HYA) adults.

ROC curve analysis of the R4AlzTot7 versus the R4AlzTot4 that seem to discriminate between subjective cognitive decline (SCD) and cognitively healthy young (HYA) adults.

SCD versus HMaA

As far as the ability of R4AlzTot7, to discriminate SCD from HMaA, a cutoff of 5.17 resulted in excellent sensitivity (88.2%) and specificity (87.1%) (AUC = 0.941 95% CI = 0.8–0.0.9, p < 0.001), while a cutoff of 2.42 of the R4AlzTot4 resulted in excellent and higher sensitivity (100%) and specificity (93.3%) (AUC = 0.98 95% CI = 0.9–1.00, p < 0.001) (Fig. 5).

Fig. 5

ROC curve analysis of the R4AlzTot7 versus the R4AlzTot4 that seems to discriminate between subjective cognitive decline (SCD) and healthy middle-aged (HMaA) adults.

ROC curve analysis of the R4AlzTot7 versus the R4AlzTot4 that seems to discriminate between subjective cognitive decline (SCD) and healthy middle-aged (HMaA) adults.

SCD versus HOA

Regarding the ability of R4AlzTot7, to discriminate SCD from HOA, a cutoff of 7.21 resulted in fair sensitivity (58.80%) and excellent specificity (100%) (AUC = 0.78 95% CI = 0.6–0.9, p = 0.002), while a cutoff of 2.93 of the R4AlzTot4 resulted in both excellent sensitivity (85.3%) and specificity (92.9%) (AUC = 0.96 95% CI = 0.9–1.00, p < 0.001) (Fig. 6).

Fig. 6

ROC curve analysis of the R4AlzTot7 versus the R4AlzTot4 that seems to discriminate between subjective cognitive decline (SCD) and cognitively healthy older (HOA) adults.

ROC curve analysis of the R4AlzTot7 versus the R4AlzTot4 that seems to discriminate between subjective cognitive decline (SCD) and cognitively healthy older (HOA) adults.

SCD versus MCI

With regards to the ability of R4AlzTot7, to discriminate SCD from MCI, a cutoff of 9.10 resulted in fair sensitivity (72.3%) and specificity (79.4%) (AUC = 0.78 95% CI = 0.6–0.8, p < 0.001), while a cutoff of 5.82 of the R4AlzTot4 resulted also in fair sensitivity (76.6%) and specificity (73.5%) (AUC = 0.75 95% CI = 0.6–0.8, p < 0.001) (Fig. 7).

Fig. 7

ROC curve analysis of the R4AlzTot7 versus the R4AlzTot4 that seems to discriminate between subjective cognitive decline (SCD) and mild cognitive impairment (MCI) adults.

ROC curve analysis of the R4AlzTot7 versus the R4AlzTot4 that seems to discriminate between subjective cognitive decline (SCD) and mild cognitive impairment (MCI) adults.

DISCUSSION

The main purpose of the present study was to investigate the discriminant potential of the R4Alz between people with SCD, healthy adult population (young, middle-aged, and older), and people with MCI. On that ground, our first concern was to explore whether the battery is affected by demographic characteristics such as gender and educational level. The demographic effects on the performance of R4Alz battery are as follows:

Gender and educational level effects

The great challenge that neuropsychology faces these days is the development of new sensitive tools that will not be affected by demographic characteristics, such as gender and particularly education. In our study, our results showed that all the R4Alz subtasks are free of education and gender effects, except for the WM component of short-term store, which seems to be affected by gender.

As far as the gender effects on cognitive performance is concerned, there is controversy, with some studies indicating that gender does not usually affect the performance on cognitive tests [17, 39], while others support the idea that gender differences in favor of males is a common finding and specifically in cognitive areas of verbal and visual spatial abilities [40–42]. These differences that are noticed among the genders in visual-spatial abilities reflect differences on the strategies that the two genders use [43]. Males seem to have better visual working memory, since they tend to use more effective holistic strategies in order to complete a task, compared with females who seem to select fewer effective strategies and are better in verbal working memory tasks [43]. In our study, a gender effect was noticed only in the WM subtask of short-term store. Consistent with the literature, according to the mean scores, males in our sample showed better performance in WMCUT-S1 than women, indicating that males are better in spatial working memory abilities than females (males: M = 5.82, SD = 0.98, females: M = 5.26, SD = 1.03). The specific subtask was designed in order to assess the storage of spatial (visual) information over a limited period. It requires from the examinee to keep and recall the correct sequence of series of two to seven lightened pads. Taking into account that the more pads the examinee has to remember, the more effort is required, and the examinees are usually forced to use strategies in order to find a way to give a correct answer, which is a possible reason for the age-related effect, since the literature has shown that males select more effective strategies than females.

The most important result regarding the effects of demographics is related to the (non)effect of education on the scores of the R4Alz subtasks. Education usually seems to have an important impact on the scores of neuropsychological tests [44]. High level of education, as well known, is correlated with cognition of older adults in different ways and is considered to be a protective factor for the cognitive loss according to the “cognitive reserve theory” [45]. This means that people with high educational level and therefore high cognitive reserve possibly present a greater degree of pathology; they continue to perform within normal limits on cognitive tests [45, 46], a fact that can usually lead to misdiagnosis. The R4Alz battery was found to be independent from the effects of education, possibly because it was designed to require fluid and not crystallized intelligence for the task implementation. Fluid intelligence refers to the ability to reason and think flexibly and is considered as independent of learning and education [47].

R4Alz differential ability

The main aim of the present study was to assess the R4Alz battery’s ability to differentiate SCD from cognitively healthy adulthood (young to older adults) and MCI. Even though a few studies indicate that some cognitive deficits such as in semantic verbal fluency are detectable in people with SCD [48], to our knowledge, no specific cutoff scores for differentiating SCD from cognitively CHA exist till now. On the contrary, the criteria for the diagnosis are based on the criteria of the SCD-I Working Group and on specific self-administered memory complaint questionnaires that are usually used [46, 49, 50]. Therefore, the definition of SCD by specific and objective clinical criteria still is the ultimate goal. Motivated by this, we examined which R4Alz subtasks better discriminate SCD from cognitively healthy young, middle-aged, and older adults and people with MCI. According to the results the SCD group differs from HYA and HMaA in all subtasks, while three subtasks (but four scores, since for the ICT/RCT-S1&2 subtask two scores were created) were found to discriminate SCD from HOA and these comprise 1) the WM updating, 2) the ICT/RCT-S1&2, and ICT/RST-FS, and 3) the ICT/RST-S3. On the other hand, the ACT subtask and all the ICT/RCT scores were found to better discriminate the SCD group from the MCI one.

WMCUT subtasks differential abilities

The WMCUT task comprised three sub-tasks: a WM component of short-term store, a WM component of central executive, and a WM updating subtask enriched by a component of episodic buffer.

WMCUT-S1 and WMCUT-S2 require from the examinees to remember and deactivate in the correct (forward) or in the diverse (backward) order a sequence of two to seven activated pads, therefore, they measure visual-spatial WM. These subtasks seem to differentiate the group of SCD from HYA and HMaA. As known, the spatial working memory is sensitive to age effects during lifespan, with healthy young adults to score better than older adults [51, 52]. In our sample, people with SCD are older adults over 60 years old; therefore, in consistency with the literature, the older adults have diminished performance in WM tasks, especially in storage and processing spatial ones. However, since age is not an explanatory process for brain and cognitive changes during lifespan, other factors such as “health issues” is possible to be engaged in cognitive changes [53]. The SCD group, besides the fact that they did not have serious health problems, they mostly had elevations in various diseases and vascular risk factors. Therefore, taking into account the vascular hypothesis of cognitive aging, there is a continuum of vascular degeneration which results in different degrees of cognitive decline [53]. Conclusively it makes sense that people with SCD can be differentiated by a subtask of working memory from healthy younger or middle-aged adults.

Regarding the differential ability of SCD from HOA, it seems that WMCUT-S1 and WMCUT-S2 are not able to differentiate these groups. A possible reason is the “age-relation theory”; however, the level of complexity of these subtasks is also a potential explanation. In order to investigate the age relation theory in our sample, besides the fact that is not at the study’s aims, an extra analysis regarding the differential ability between all the healthy groups was performed. Regarding the WMCUT-S1, the analysis showed that HOA differ from HYA, I–J = –0.98, p = 0.003 and HOA from HMaA I–J = –0.91, p = 0.010 fact that further strengthens the age-relation theory. As far as WMCUT-S2 is concerned, according to the extra analysis, a differentiation between the three groups of CHA was not noticed. Even though a relation seems to exist among processing and age [54], the specific subtask is proved to not be age related. Of course, the high complexity of the subtask is a potential differentiation reason. According to the theory, working memory has a limited capacity compared to long-term memory. People can usually store approximately three to four items in their working memory for a short period of time [55], whereas for more items specific number of combinations or verbalization may be activated. Since we utilize up to 7 pads, apart from spatial WM storage and processing, a combination of verbal and visual strategies is required as for the examinee to keep in memory the correct or the diverse longer sequence. Therefore, people with SCD seem to be capable enough to utilize these strategies as well as healthy older adults do, leading to poor differentiation between these groups.

Finally, both WMCUT-S1 and WMCUT-S2 do not differentiate SCD from MCI. People with MCI usually present difficulties in spatial working memory and spatial digit span tasks compared to healthy controls [56], while there is no evidence in the literature regarding differences among SCD and MCI in the above abilities. In our results, if we observe the SCD and MCI scores’ means of working memory, a difference is evident in favor of people with SCD. Nevertheless, a relatively high standard deviation leads to poor differentiation capacity.

Regarding the WMCUT-S3, the results showed that it can differentiate between SCD and all the groups of CHA. The third subtask of WM is called WM updating enriched by a component of episodic buffer abilities, comprising a combination of long-term memory, short term store, processing, and perception. This subtask includes conditions with increasing level of difficulty, where 1) the first two mostly require short term store and updating abilities, 2) the third requires the previous ones but also long-term memory, 3) the fourth condition requires additional processing, and 4) the last two conditions demand an additional high level of memory load. Therefore, the examinee needs to temporarily store the new information, to manipulate the other stimuli, and finally to appropriately modify the new information. Secondarily, inhibition and cognitive flexibility are also activated during the last two conditions. Our results indicate that WMCUT-S3 can differentiate SCD from healthy young and healthy middle-aged adults. As aforementioned, WM abilities are diminished by age, specifically in tests of updating of WM such as n-back tests [57]. According to the literature, the age-related decline in WM is due to updating and inhibition functions [58], which are required in WMCUT-S3 as well [59]. As far the discriminant ability of HOA from SCD is concerned, WMCUT-S3 can also differentiate these groups. To our knowledge, until today there are no studies to support the existence of difficulties of SCD in updating of WM and episodic buffer. The WMCUT-S3, as described above, comprises different levels of difficulties and assesses several working memory abilities, while others are also activated during the subtask’s implementation. Therefore, it is more complex and more sensitive in comparison to the other WM subtasks of the battery, making it capable of detecting even the subtlest cognitive decline. Therefore, our results support the idea that people with SCD is a group with cognitive deficits that differs from HOA.

Finally, the WMCUT-S3 seems that cannot differentiate people with SCD from MCI. Besides the fact that people with MCI appear to present difficulties in updating tasks [60] and to have lower performance compared to HOA, it seems that this specific subtask is more complex than a single test of updating, since it engages additional cognitive functions, such as episodic buffer. Episodic buffer as a temporary storage system with a limited capacity, integrating information from a variety of sources into episodes [61]. As the results show, people with SCD didn’t manage to successfully cope with these complex subtasks that demand concurrent updating, enhanced by episodic buffer. The result was lower scores, which were more similar to MCI scores than HOA ones. Therefore, this subtask could not distinguish these two groups.

ACT differential ability

The attention control task was designed in order to assess different aspects of attention control abilities and comprises conditions of 1) visual selective attention, 2) auditory sustained attention, 3) divided visual and auditory attention, and 4) a combination of divided visual-auditory and motion attention, spanning in different levels of difficulty subtasks. Besides the fact that the specific subtask was designed in order to assess attentional control abilities, other abilities are also required, such as shifting of attention and inhibition, since multiple similar stimuli are present. Attention abilities such as selective attention and sustained attention seem to change during lifespan since there is an almost linear decline in older adults [62]. Divided attention is also affected by age as studies of the effect of age in dual task and cognitive performance shows [63]. In our study, ACT seems to differentiate SCD from HYA, HMaA, and from MCI, while a significant difference between SCD and HOA was not noticed and this is the most important result. Again, if we pay attention to the task’s means, a difference is evident in favor of SCD but a fairly large standard deviation leads to poor discrimination potential. Therefore, according to our results, people with SCD seem to maintain a good level of attention control abilities, in contrast to other abilities that seem to drop early in the course of neurocognitive diseases.

As far as the MCI group is concerned, a clear discrimination from SCD is observed. MCI is a group which usually appears reduced attention abilities, especially divided attention [64]. As previously mentioned, the ACT task is complex and comprises different levels of difficulty, also requiring inhibition abilities. As inhibition and attention deficits are common in people with MCI, it seems that they could not manage to cope with the tasks requirements compared to SCD, resulting in a good differentiation between the two groups.

ICT/RST-S1&S2 and ICT/RST-FS differential ability

The ICT/RST-S1&S2 and ICT/RST-FS scores seem able to differentiate SCD from all the other groups. The total score of ICT/RST-S1&S2 evaluates inhibition & task/rule switching and is a sum of these subtasks’ conditions. Even though the idea that older adults, compared to younger ones, are less capable for overcoming dominant responses and also to ignore distracting information exist since many decades [65], nowadays there is a controversy regarding whether inhibition and task/rule switching is age dependent. According to some studies inhibition and also switching are age-related declining abilities [66, 67], while others support the opposite idea [68, 69]. The ICT/RCT-S1&S2 comprises conditions with different levels of difficulty, that require inhibition of the dominant responses, as well as a great effort for switching from one rule to the other. On the other hand, the ICT/RST-FS is a score that measures the total number of sets that contain at least one error, allowing to define the capacity of the examinee, not only for switching but also for maintaining a set. Even though it was under debate whether inhibition and switching are age-related abilities, according to our study, both abilities seem to change during lifespan. Our study comes to an agreement with others that show a decrease in activation of inhibitory control [70] and switching as well [71]. Therefore, ICT/RST-S1&S2 and ICT/RST-FS in consistency with the age-relation literature, can differentiate the SCD people from the HYA and HMaA.

Regarding SCD and HOA, discrimination ability of the ICT/RCT-S1&S2 and also of the ICT/RCT-FS was noticed. According to our results, the seemingly cognitively healthy SCD group, presents many difficulties compared to HOA both in inhibition and switching. According to the calculated means, people with SCD have a great difficulty to inhibit dominant and automatic responses, in order to successfully complete ICT/RCT-S1&S2. Moreover, an extra difficulty is presented when along with the inhibition of automatic responses, have to switch between the subtasks. Even though the particular subtasks are quite hard, HOA are perfectly capable of completing them with only a few errors, in contrast to people with SCD. Therefore, our results support the idea that the discrimination between SCD and HOA is possible, when the used neuropsychological batteries or tests are quite demanding, discovering the subtle deficits that may be presented in SCD, and not in HOA. According to our knowledge there are no other studies that show differences between SCD and HOA in inhibition and switching, while others showed a difference in inhibition and switching of cognitively HOA compared to cognitively HOA with vascular risk [72]. On the other hand, difficulties in inhibition control and switching are usually observed in neurodegenerative diseases, such as in MCI [73, 74]. As it seems people with MCI have diminished abilities for overcoming dominant responses, ignoring distracting information and rule switching abilities, resulting to reduced capability to complete the subtask with more errors than SCD, leading to good discrimination.

ICT/RST-S3 differential ability

The ICT/RST-S3 subtask was designed to assess cognitive flexibility as the application of inhibitory control plus rule switching in a cognitive task, including three different visual and three different auditory stimuli. Therefore, the participant must 1) act by following a specific given instruction, 2) inhibit all combinations of similar but erroneous actions according to the specific condition, 3) be disengaged from the previously correct stimuli and subsequently be involved in a new condition, and finally 4) shift their behavior to continually changing demands. Again, ICT/RST-S3 contains tasks with progressive difficulty, leading to the activation of additional abilities, including selective attention and processing, as well as updating.

Performance in ICT/RST-S3 was found to significantly differ between SCD and the healthy adult groups. Besides the fact that no studies exist to report cognitive flexibility deficits in people with SCD, cognitive control difficulties are usually presented in older adults as a result of a degree of vascular abnormality. Specifically, there are indications that older adults with risk factors for vascular disease development present lower performance compared to young adults in inhibitory control and inhibition plus switching [72, 75, 76]. Our study comes to contribute by also supporting the idea that cognitive flexibility deficits exist in older adults with SCD, since ICT/RST-S3 is sensitive enough to detect them.

Regarding the ability of the ICT/RST-S3 to differentiate SCD from MCI, it seems that the specific subtask is not sensitive toward this discrimination. Besides the fact that cognitive flexibility and inhibition are impaired in MCI [77], the specific subtask is not able to differentiate this group from SCD, even though a difference in the scores’ means once more existed. Specifically, we noticed that the SCD scores were quite heterogeneous, since some scored significantly lower in comparison to others. The most probable explanation is that the disorganization of cognitive flexibility, as a combination of inhibitory and switching, seems to initiate early at the course of cognitive deficits, starting at SCD, while it seems to be stabilized at MCI. Furthermore, another possible factor behind the inability to differentiate may be SCD group’s heterogeneity, since there are people that reverse to healthy cognitive aging without cognitive complaints (underlying emotional or medical conditions etiology), people who are stable and non-reversible during the years, and some who progress in cognitive impairments and dementia [18]. Therefore, the specific subtask’s increased complexity made it quite difficult, not only for people with MCI but for people with SCD as well.

Cutoff scores of R4Alz battery to discriminate SCD from CHA and MCI

As for the discriminant potential of the R4Alz battery, for the needs of this study we provided two different cutoff scores, one for the total score of the subtasks that were found to better differentiate SCD from HOA (R4AlzTot4), and one for the total score of all subtasks (R4AlzTot7). Therefore, these total scores combined with specific cutoff scores are proposed to be used for the differential diagnosis of SCD from the three groups of CHA (young, middle-aged, and older). Therefore, the cutoff of 2.38 and the cutoff score of 2.42 of the R4AlzTot4 were the best scores that differentiated SCD from HYA and HMaA. Respectively, with an excellent sensitivity and specificity, a cutoff of 2.93, seems to be the best balance for the discrimination of SCD from HOA. On the other hand, the R4AlzTot7 score that we also provide seems to have also an excellent ability for the discrimination of SCD from healthy young and middle-aged adults but a fare discrimination ability regarding the differentiation of SCD from HOA. The results regarding the two different scores make actually sense, since the R4AlzTot7 includes all the battery’s subtasks and not only the subtasks that the analysis showed that better differentiate the two demographically similar groups of SCD and HOA. However, is very important that even this score also leads to a differentiation. Based on the fact that until today there are no tests or other instruments that provide validated cutoff scores for differentiating older adults with subjective complains from older adults without complains [18], it very important that in this study a new tool is proposed that is possible to differentiate these groups. Moreover, if one considers that the present battery is not affected by demographic factors and especially education, it adds an extra strength in this work.

As far as the discriminant ability of the two scores of R4Alz battery to differentiate SCD from MCI, none of them leads to an excellent discrimination but both scores had fare discrimination ability. A possible explanation for this lack of an excellent differentiation is that some people with SCD showed lower performances compared with others. This is possibly related with the SCD sample which is characterized by great heterogeneity [18]. Furthermore, the R4Alz battery as a tool which includes different levels of difficulty and requires several mental abilities for its task implementation, is probably more difficult from other neuropsychological tests or batteries that were designed to assess strictly specific abilities, especially for people with advanced neurodegeneration, such as dementia. Therefore, R4Alz was able to detect even people with SCD who had diminished performance, close to people with MCI. Moreover, people with SCD is possible to share more similarities than differences with people with MCI; however, the existent neuropsychological tests are not able to detect them. To be more precise, there are several studies which propose to consider SCD as a preclinical stage of AD and that there are biological similarities between SCD and MCI in terms of functional connectivity [78], presence of an ApoE ɛ4 allele, and hippocampal volume [79]. Therefore, it is possible that people with SCD have some slight cognitive decline that the until today tools were not able to detect it.

Another topic that is worth mentioning is that the R4Alz battery, because of its different levels of difficulty, requires for its overall administration almost 45 minutes to one hour, according to the participant’s cognitive level. This actually means that the better the examinee’s cognitive function is, the shorter amount time that is needed for its administration. Therefore, in this study we propose a cutoff score from three subtasks (four scores) so as to minimize the required administration time, since these three subtasks approximately last fifteen minutes. We consider that based on this total score, it is possible for SCD to be detected by objective criteria, which could lead to early preventive interventions.

Strengths and limitations and future work

According to the literature, even though people with SCD present normal cognitive performance in common neuropsychological measures, a subtle decline from the prior level on objective cognitive tests appeared [46]. Therefore, the design of new and sensitive neuropsychological tools that will be capable to identify the slight cognitive decline of people with subjective cognitive complaints, many years before the onset and the diagnosis of minor or major cognitive deficits, is always a crucial need. Thus, one of the greatest strengths of the present study was that we determined the most sensitive subtasks of the R4Alz battery, in order to propose cutoff scores which seem to differentiate older adults with SCD from cognitively healthy young, middle-aged, and especially older adults. The fact that the administration of the subtasks that seem sensitive for discrimination of SCD from HOA and the other groups of CHA, last only fifteen minutes, makes R4Alz an appropriate screening test that can be used in every clinical setting. Another strength of the study is that according to the results, R4Alz battery is not affected by educational level. This is very important since we can have more reliable and accurate results, avoiding the misdiagnosis of people with SCD, especially those who are in high educational status (cognitive reserve) [80, 81]. The limitation of this pilot study is the small subsample of cognitively HOA.

Therefore, the future goal is to increase our subsamples and make them balanced in numbers, in order to evaluate the performance in each age range. Furthermore, new appropriately designed subtasks will also be added to the existent battery in order to achieve better differentiation between SCD from HOA and people with MCI. That way we will be able to present separate cutoff scores for each age and for each diagnostic group. As a result, a more early, accurate, and reliable diagnosis will be provided. To conclude with, the R4Alz battery seems to be a novel technological approach which can be used in clinical practice in order to detect people with SCD, and discriminate them from CHA. Therefore, the R4Alz battery is proposed for early diagnosis and for the early design of protective strategies, in order to prevent the cognitive decline.

DISCLOSURE STATEMENT

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/20-0562r1).

REFERENCES

[1] 

Patterson C (2018) 2018 World Alzheimer Report. The State of the Art of Dementia Research: New Frontiers. Alzheimer’s Disease International, London.

[2] 

Huang Y , Mucke L (2012) Alzheimer mechanisms and therapeutic strategies. Cell 148, 1204–1222.

[3] 

Seo E , Kim H , Choi K , Lee K , Choo I (2017) Association of subjective memory complaint and depressive symptoms with objective cognitive functions in prodromal Alzheimer’s disease including pre-mild cognitive impairment. J Affect Disord 217, 24–28.

[4] 

Weaver Cargin J , Collie A , Masters C , Maruff P (2008) The nature of cognitive complaints in healthy older adults with and without objective memory decline. J Clin Exp Neuropsychol 30, 245–257.

[5] 

van Norden A , Fick W , de Laat K , van Uden I , van Oudheusden L , Tendolkar I , Zwiers M , de Leeuw F (2008) Subjective cognitive failures and hippocampal volume in elderly with white matter lesions. Neurology 71, 1152–1159.

[6] 

Ramakers I , Visser P , Bittermann A , Ponds R , van Boxtel M , Verhey F (2009) Characteristics of help-seeking behaviour in subjects with subjective memory complaints at a memory clinic: A case-control study. Int J Geriatr Psychiatry 24, 190–196.

[7] 

Jessen F , Amariglio RE , Van Boxtel M , Breteler M , Ceccaldi M , Chételat G , Dubois B , Dufouil C , Ellis KA , Van Der Flier WM , Glodzik L (2014) A conceptual framework for research on subjective cognitive decline in preclinical Alzheimer’s disease. Alzheimers Dement 10, 844–852.

[8] 

Albert M , DeKosky S , Dickson D , Dubois B , Feldman H , Fox N , Gamst A , Holtzman D , Jagust W , Petersen R , Snyder P , Carrillo M , Thies B , Phelps C (2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 270–279.

[9] 

Uddin MS , Ashraf GM (2018) Introductory chapter: Alzheimer’s disease—The most common cause of dementia. In Advances in Dementia Research, Ashraf GM, ed. IntechOpen. https://www.intechopen.com/books/advances-in-dementia-research/introductory-chapter-alzheimer-s-disease-the-most-common-cause-of-dementia

[10] 

Farias S , Lau K , Harvey D , Denny K , Barba C , Mefford A (2017) Early functional limitations in cognitively normal older adults predict diagnostic conversion to mild cognitive impairment. J Am Geriatr Soc 65, 1152–1158.

[11] 

Mitchell A , Beaumont H , Ferguson D , Yadegarfar M , Stubbs B (2014) Risk of dementia and mild cognitive impairment in older people with subjective memory complaints: Meta-analysis. Acta Psychiatr Scand 130, 439–451.

[12] 

Boone K (2009) Fixed belief in cognitive dysfunction despite normal neuropsychological scores: Neurocognitive hypochondriasis? Clin Neuropsychol 23, 1016–1036.

[13] 

Chu C , Sun I , Begum A , Liu S , Chang C , Chiu W , Chen C , Tang H , Yang C , Lin Y , Chiu C , Stewart R (2017) The association between subjective memory complaint and objective cognitive function in older people with previous major depression. PLoS One 12, e0173027.

[14] 

Minett T , Da Silva R , Ortiz K , Bertolucci P (2007) Subjective memory complaints in an elderly sample: A cross-sectional study. Int J Geriatr Psychiatry 23, 49–54.

[15] 

Prestia A , Caroli A , van der Flier W , Ossenkoppele R , Van Berckel B , Barkhof F , Teunissen C , Wall A , Carter S , Scholl M , Choo I , Nordberg A , Scheltens P , Frisoni G (2013) Prediction of dementia in MCI patients based on core diagnostic markers for Alzheimer disease. Neurology 80, 1048–1056.

[16] 

Iverson G , Brooks B , Ashton V , Johnson L , Gualtieri C (2009) Does familiarity with computers affect computerized neuropsychological test performance? J Clin Exp Neuropsychol 31, 594–604.

[17] 

Poptsi E , Moraitou D , Eleftheriou M , Kounti-Zafeiropoulou F , Papasozomenou C , Agogiatou C , Bakoglidou E , Batsila G , Liapi D , Markou N , Nikolaidou E , Ouzouni F , Soumpourou A , Vasiloglou M , Tsolaki M (2019) Normative data for the montreal cognitive assessment in greek older adults with subjective cognitive decline, mild cognitive impairment and dementia. J Geriatr Psychiatry Neurol 32, 265–274.

[18] 

Jessen F , Amariglio RE , Buckley RF , van der Flier WM , Han Y , Molinuevo JL , Rabin L , Rentz DM , Rodriguez-Gomez O , Saykin AJ , Sikkes SA (2020) The characterisation of subjective cognitive decline. Lancet Neurol 19, 271–278.

[19] 

Poptsi E , Tsardoulias E , Moraitou D , Symeonidis A , Tsolaki M (2019) REMEDES for Alzheimer-R4Alz battery: Design and development of a new tool of cognitive control assessment for the diagnosis of minor and major neurocognitive disorders. J Alzheimers Dis 72, 783–801.

[20] 

American Psychiatric Association (2013) Diagnostic and statisticalmanual ofmental disorders (DSM-5®), American Psychiatric Publishing.

[21] 

Yesavage J , Brink T , Rose T , Lum O , Huang V , Adey M , Leirer V (1982) Development and validation of a geriatric depression screening scale: A preliminary report. J Psychiatr Res 17, 37–49.

[22] 

Fountoulakis K , Tsolaki M , Iacovides A , Yesavage J , O’Hara R , Kazis A , Ierodiakonou C (1999) The validation of the short form of the Geriatric Depression Scale (GDS) in Greece. Aging Clin Exp Res 11, 367–372.

[23] 

Beck AT , Ward C , Mendelson M , Mock J , Erbaugh J (1961) Beck depression inventory (BDI). Arch Gen Psychiatry 46, 561–571.

[24] 

Sinoff G , Ore L , Zlotogorsky D , Tamir A (1999) Short anxiety screening test—a brief instrument for detecting anxiety in the elderly. Int J Geriatr Psychiatry 14, 1062–1071.

[25] 

Grammatikopoulos I , Sinoff G , Alegakis A , Kounalakis D , Antonopoulou M , Lionis C (2010) The Short Anxiety Screening Test in Greek: Translation and validation. Ann Gen Psychiatry 9, 1.

[26] 

Beck A , Epstein N , Brown G , Steer R (1988) An inventory for measuring clinical anxiety: Psychometric properties. J Consult Clin Psychol 56, 893–897.

[27] 

Cummings J , Mega M , Gray K , Rosenberg-Thompson S , Carusi D , Gornbein J (1994) The neuropsychiatric inventory: Comprehensive assessment of psychopathology in dementia. Neurology 44, 2308–2308.

[28] 

Politis A , Mayer L , Passa M , Maillis A , Lyketsos C (2004) Validity and reliability of the newly translated Hellenic Neuropsychiatric Inventory(H-NPI) applied to Greek outpatients with Alzheimer’s disease: A study of disturbing behaviors among referrals to a memory clinic. Int J Geriatr Psychiatry 19, 203–208.

[29] 

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

[30] 

Fountoulakis K , Tsolaki M , Chantzi H , Kazis A (2000) Mini Mental State Examination (MMSE): A validation study in Greece. Am J Alzheimers Dis Other Demen 15, 342–345.

[31] 

Kounti F , Tsolaki M , Kiosseoglou G (2006) Functional cognitive assessment scale (FUCAS): A new scale to assess executive cognitive function in daily life activities in patients with dementia and mild cognitive impairment. Hum Psychopharmacol 21, 305–311.

[32] 

Tsolaki M , Poptsi E , Aggogiatou C , Markou N , Zafeiropoulos S , Kounti F (2017) Computer-based cognitive training versus paper and pencil training: Which is more effective? a randomized controlled trial in people with mild cognitive impairment. JSM Alzheimers Dis Relat Dement 4 1, 1032.

[33] 

Reisberg B , Ferris SH , de Leon MJ , Crook T (1982) The Global Deterioration Scale for assessment of primary degenerative dementia. Am J Psychiatry 139, 1136–1139.

[34] 

IBM SPSS Statistics for Windows, Version 23.0. 2015, IBM Corp Armonk, New York.

[35] 

Armstrong RA , Hilton A (2006) ANOVA tests. Microbiologist 6, 34–36.

[36] 

Youden W (1950) Index for rating diagnostic tests. Cancer 3, 32–35.

[37] 

Hosmer DW Jr , Lemeshow S , Sturdivant RX (2013) Applied logistic regression, John Wiley & Sons, USA.

[38] 

Hanley J , McNeil B (1982) The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 143, 29–36.

[39] 

Magnusdottir B , Haraldsson H , Sigurdsson E (2019) Trail making test, stroop, and verbal fluency: Regression-based norms for the icelandic population. Arch Clin Neuropsychol, doi: 10.1093/arclin/acz049

[40] 

Voyer D , Voyer S , Saint-Aubin J (2016) Sex differences in visual-spatial working memory: A meta-analysis. Psychon Bull Rev 24, 307–334.

[41] 

Elosúa M , Ciudad M , Contreras M (2017) Gender differences in verbal and visuospatial working memory tasks in patients with mild cognitive impairment and Alzheimer disease. Dement Geriatr Cogn Dis Extra 7, 101–108.

[42] 

Zilles D , Lewandowski M , Vieker H , Henseler I , Diekhof E , Melcher T , Keil M , Gruber O (2016) Gender differences in verbal and visuospatial working memory performance and networks. Neuropsychobiology 73, 52–63.

[43] 

Wang L , Carr M (2014) Working memory and strategy use contribute to gender differences in spatial ability. Educ Psychol 49, 261–282.

[44] 

Ostrosky-Solis F , Ramirez M , Ardila A (2004) Effects of culture and education on neuropsychological testing: A preliminary study with indigenous and nonindigenous population. Appl Neuropsychol 11, 186–193.

[45] 

Stern Y (2003) The concept of cognitive reserve: A catalyst for research. J Clin Exp Neuropsychol 25, 589–593.

[46] 

Molinuevo JL , Rabin LA , Amariglio R , Buckley R , Dubois B , Ellis KA , Ewers M , Hampel H , Klöppel S , Rami L , Reisberg B (2017) Implementation of subjective cognitive decline criteria in research studies. Alzheimers Dement 13, 3 296–311.

[47] 

Kievit R , Davis S , Griffiths J , Correia M , Cam-Can , Henson R (2016) A watershed model of individual differences in fluid intelligence. Neuropsychologia 91, 186–198.

[48] 

Nikolai T , Bezdicek O , Markova H , Stepankova H , Michalec J , Kopecek M , Dokoupilova M , Hort J , Vyhnalek M (2017) Semantic verbal fluency impairment is detectable in patients with subjective cognitive decline. Appl Neuropsychol Adult 25, 448–457.

[49] 

Papaliagkas V , Papantoniou G , Tsolaki M , Moraitou D (2017) Self-report instruments of cognitive failures as screening tools for subjective cognitive impairment in older adults. Hell J Nucl Med 20, 58–70.

[50] 

Hong Y , Park K , Kang D , Lee J (2019) Prediction of Alzheimer’s pathological changes in subjective cognitive decline using the self-report questionnaire and neuroimaging biomarkers. Dement Neurocogn Disord 18, 19.

[51] 

Brown L (2016) Spatial-sequential working memory in younger and older adults: Age predicts backward recall performance within both age groups. Front Psychol 7, 1514.

[52] 

Woods D , Wyma J , Herron T , Yund E (2015) An improved spatial span test of visuospatial memory. Memory 24, 1142–1155.

[53] 

Spiro A , Brady C (2011) Integrating health into cognitive aging: Toward a preventive cognitive neuroscience of aging. J Gerontol B Psychol Sci Soc Sci 66, i17–i25.

[54] 

Dai R , Thomas A , Taylor H (2018) Age-related differences in the use of spatial and categorical relationships in a visuo-spatial working memory task. Mem Cognit 46, 809–825.

[55] 

Adam K , Vogel E , Awh E (2017) Clear evidence for item limits in visual working memory. Cogn Psychol 97, 79–97.

[56] 

Pandey P , Thapa K (2017) Spatial span and digit span’s task performance in patients with mild cognitive impairment and matched controls. Indian J Health Wellbeing 8,, 898–901.

[57] 

Bopp K , Verhaeghen P (2020) Aging and n-Back performance: A meta-analysis. J Gerontol B Psychol Sci Soc Sci 75, 229–240.

[58] 

Zuber S , Ihle A , Loaiza V , Schnitzspahn K , Stahl C , Phillips L , Kaller C , Kliegel M (2019) Explaining age differences in working memory: The role of updating, inhibition, and shifting. Psychol Neurosci 12, 191–208.

[59] 

Dumas J , Kutz A , McDonald B , Naylor M , Pfaff A , Saykin A , Newhouse P (2013) Increased working memory-related brain activity in middle-aged women with cognitive complaints. Neurobiol Aging 34, 1145–1147.

[60] 

Fraga F , Mamani G , Johns E , Tavares G , Falk T , Phillips N (2018) Early diagnosis of mild cognitive impairment and Alzheimer’s with event-related potentials and event-related desynchronization in N-back working memory tasks. Comput Methods Programs Biomed 164, 1–13.

[61] 

Baddeley A (2000) The episodic buffer: A new component of working memory? Trends Cogn Sci 4, 417–423.

[62] 

McAvinue L , Habekost T , Johnson K , Kyllingsbæk S , Vangkilde S , Bundesen C , Robertson I (2012) Sustained attention, attentional selectivity, and attentional capacity across the lifespan. Atten Percept Psychophys 74, 1570–1582.

[63] 

Brustio P , Magistro D , Zecca M , Rabaglietti E , Liubicich M (2017) Age-related decrements in dual-task performance: Comparison of different mobility and cognitive tasks. A cross sectional study. PLoS One 12, e0181698.

[64] 

Okonkwo O , Wadley V , Ball K , Vance D , Crowe M (2008) Dissociations in visual attention deficits among persons with mild cognitive impairment. Neuropsychol Dev Cogn B Aging Neuropsychol Cogn 15, 492–505.

[65] 

Hasher L , Zacks RT , Bower GH (1988) The Psychology of Learning and Motivation, Academic Press, New York.

[66] 

Clapp W , Rubens M , Sabharwal J , Gazzaley A (2011) Deficit in switching between functional brain networks underlies the impact of multitasking on working memory in older adults. Proc Natl Acad Sci U S A 108, 7212–7217.

[67] 

Turner G , Spreng R (2012) Executive functions and neurocognitive aging: Dissociable patterns of brain activity. Neurobiol Aging 33, 826.e1–826.e13.

[68] 

Sebastian A , Baldermann C , Feige B , Katzev M , Scheller E , Hellwig B , Lieb K , Weiller C , Tüscher O , Klöppel S (2013) Differential effects of age on subcomponents of response inhibition. Neurobiol Aging 34, 2183–2193.

[69] 

Schuch S (2016) Task inhibition and response inhibition in older vs. younger adults: A diffusion model analysis. Front Psychol 7, 1722.

[70] 

Hu S , Ide J , Chao H , Castagna B , Fischer K , Zhang S , Li C (2018) Structural and functional cerebral bases of diminished inhibitory control during healthy aging. Hum Brain Mapp 39, 5085–5096.

[71] 

Callaghan E , Holland C , Kessler K (2017) Age-related changes in the ability to switch between temporal and spatial attention. Front Aging Neurosci 9, 28.

[72] 

Pantsiou K , Sfakianaki O , Papaliagkas V , Savvoulidou D , Costa V , Papantoniou G , Moraitou D (2018) Inhibitory control, task/rule switching, and cognitive planning in vascular dementia: Are there any differences from vascular aging? Front Aging Neurosci 10, 330.

[73] 

Rabi R , Vasquez B , Alain C , Hasher L , Belleville S , Anderson N (2020) Inhibitory control deficits in individuals with amnestic mild cognitive impairment: A meta-analysis. Neuropsychol Rev 30, 97–125.

[74] 

Gonçalves A , Tarrasconi M , Holz M , Kochhann R , Paz Fonseca R (2019) Cognitive flexibility and inhibition in single-versus multiple-domain mild cognitive impairment: A comparative and discriminative analysis. Psychol Neurosci 12, 209–223.

[75] 

Adornetti I (2016) On the phylogenesis of executive functions and their connection with language evolution. Front Psychol 7, 1426.

[76] 

Lam C , Yiend J , Lee T (2017) Imaging and neuropsychological correlates of white matter lesions in different subtypes of mild cognitive impairment: A systematic review. Neurorehabilitation 41, 189–204.

[77] 

Guarino A , Forte G , Giovannoli J , Casagrande M (2019) Executive functions in the elderly with mild cognitive impairment: A systematic review on motor and cognitive inhibition, conflict control and cognitive flexibility. Aging Ment Health 2, 1–18.

[78] 

López-Sanz D , Bruña R , Garcés P , Martín-Buro M , Walter S , Delgado M , Montenegro M , López Higes R , Marcos A , Maestú F (2017) Functional connectivity disruption in subjective cognitive decline and mild cognitive impairment: A common pattern of alterations. Front Aging Neurosci 9, 109.

[79] 

Caillaud M , Maltezos S , Hudon C , Brambati S , Duchesne S , Lorrain D , Gagnon J , Mellah S , Phillips N , Belleville S (2019) P4-251: Similarities between the cognitive profile of individuals with subjective cognitive and that of persons with mild cognitive impairment: A study from the CIMA-Q cohort. Alzheimers Dement 15, 1376–1376.

[80] 

Stern Y (2012) Cognitive reserve in ageing and Alzheimer’s disease. Lancet Neurol 11, 1006–1012.

[81] 

van Oijen M , de Jong F , Hofman A , Koudstaal P , Breteler M (2007) Subjective memory complaints, education, and risk of Alzheimer’s disease. Alzheimers Dement 3, 92–97.

Notes