In the recent years, there has been a shift in clinical practice toward earlier identification of Alzheimer’s disease (AD) inspired by wider public awareness, larger professional effort, patients’ initiatives and growing scientific evidence. DSM V manual and new research guidelines for AD consider AD as a continuum ranging from preclinical, stage, over prodromal stage characterized by minor cognitive changes referred to as mild cognitive impairment (MCI) to dementia syndrome characterized by major cognitive deficit and impaired functioning. Memory and executive functions impairments are recognized as early symptoms of the disease [1] and new challenging neuropsychological tests are thus being sought to improve early diagnosis. Apart from objective performance in neuropsychological tests, subjective symptoms of memory complaints and slowing of thinking (which may refer to executive dysfunction) are in the center of clinical and research interest.

Memory complaints are a frequent symptom in older adults [2,3] and may reflect both healthy and pathological ageing since they are not always associated with objective impairment in cognitive tests [4]. Furthermore, many studies with non-demented subjects report associations between memory complaints and depression/anxiety [5-10]. The relationship between depression/anxiety, memory complaints and risk of dementia is unclear. Follow-up studies with non-demented subjects report higher rate of conversion to dementia as well as higher incidence of depression/anxiety in subjects with memory complaints compared to non-complainers [2,11-14]. However, this area is even more complex while also depression alone was considered as a risk factor of dementia in some studies; possibly via toxic effects of increased glucocorticoid levels on the hippocampus [15-17]. There is also no consistency in terminology and assessment of subjective memory impairment [18], with some studies using simple yes/no question while other studies using complex questionnaires. Furthermore, studies on subjective memory impairment often do not distinguish cognitively healthy subjects from those with MCI. Many elderly also report problems in other cognitive domains like language, visuo-constructive functions, clumsiness/apraxia, attention, slowing of thinking, or agnosia, despite normal neuropsychological examination. Therefore a new concept was proposed for research of this entity. The concept of subjective cognitive decline (SCD) comprises subjective change in time in any of the cognitive domains in cognitively healthy individuals [19].

Besides the previously mentioned cognitive domains, we reported earlier that allocentric (i.e., with respect to external cues without regard to subject’s own position) spatial navigation (SN) impairment correlates with right hippocampal volume [20], with hippocampal type of memory impairment [21] in MCI as well as with ApoE 4 status [22], putting SN impairment among possible early markers of AD. Spatial disorientation or even being lost is also a well-known and stressful feature reported by caregivers of individuals with dementia due to Alzheimer disease (AD). This suggests potential usefulness of the assessment of SN difficulties on a subjective level, an area not represented in most memory complaints questionnaires.

Data on the relationship between subjective SN complaints (SSNC) and objective cognitive and SN performance in non-demented subjects are limited. Little is also known about the role of depression/anxiety in subjective perception of SN skills. We have developed 15 item SSNC questionnaire assessing navigation performance in one’s daily life and we analyzed the correlation of SSNC with real SN performance in the human analogue of Morris water maze. We hypothesized that according to SCD, depression and anxiety also play a critical role in SSNC reporting.

We investigated 1) whether subjects with MCI report more SSNC than subjects with SCD and 2) whether SSNC reporting and real SN performance depends on depression or/and anxiety.

Subjects

The institutional ethics committee of Motol Hospital approved the study and all participants provided a written informed consent. All procedures comply with the ethical rules for human experimentation that are stated in the Declaration of Helsinki from 1997. Total of 186 individuals with memory complaints were referred to Memory clinic by general practitioners, families or contact sites of Czech Alzheimer Society in 2011-2014. Subjects with a history of neurological or psychiatric disease, psychiatric medication, abnormal neurological examination including gait or movement difficulties, were excluded. All underwent standard neurological and laboratory evaluations, structural magnetic resonance (MR) imaging, clinical scaling and complex neuropsychological testing as well as SN testing. Participants meeting DSM IV -TR criteria for dementia (n=28) were excluded. For the final analyses we also excluded subjects (n=35) which were not able or not willing to undergo any part of the protocol (SN test, neuropsychological examination or fill in the questionnaires). Therefore, the final sample included 123 participants, 71 with MCI and 52 subjects with SCD without detectable objective cognitive impairment.

Neuropsychology

Comprehensive neuropsychological battery was used to assess all cognitive domains including: 1) memory, measured by Auditory Verbal Learning Test [23], Brief Visuospatial Memory Test-Revised [24] and Enhanced cued recall test Test in Czech validated version [25,26]; 2) attention/processing speed, measured with the Digit Span Backwards [27] and Trail Making Test A [28]; 3) frontal/executive functions, measured with the Trail Making Test B [28] and Controlled Oral Word Association [29]; 4) language, measured with the Boston Naming Test [30]; and 5) visuospatial functions measured with the Rey- Osterreith Complex Figure [24]– copy condition. All participants with MCI met clinical criteria for MCI [31] including cognitive complaints reported by patient or caregiver, evidence of cognitive dysfunction on neuropsychological testing, generally intact activities of daily living, and absence of dementia. Cognitive impairment was established when they scored more than 1.5 standard deviations below the mean of ageand education-adjusted norms on any neuropsychological tests.

Subjective Spatial Navigation Complaints (SSNC) assessment (Figure 1)

Figure 1: Subjective spatial navigation complaints questionnaire.

Since the area of SN research is not standardized, candidate questions were chosen and combined in order to be able to be administered in a novel questionnaire. This multiple choice 15 item SSNC questionnaire assessing SN performance in one’s daily life was developed at Memory Clinic, Motol University Hospital in Prague using content validity criteria. Five experienced clinicians assessing older adults at a memory clinic and an expert on spatial navigation research were involved in development and testing of the questionnaire. It contains questions on subjectively perceived difficulties during spatial orientation, questions on impact of navigational difficulties on everyday life activities, and questions on the impact of certain subjective concerns on everyday life performance. The instrument had a high degree of internal consistency (Cronbach’s alpha=0.894). SSNC was administered together with Geriatric Depression Scale (GDS) [32] and Beck Anxiety Inventory (BAI) [33].

Spatial navigation testing

SN was tested within 2 months period from all other examinations. The real-space navigation setting is called the Blue Velvet Arena located in the Laboratory of Spatial Cognition, a joint workplace of the Department of Neurology, 2nd Faculty of Medicine, Charles University in Prague, and Institute of Physiology, Academy of Sciences of the Czech Republic, Prague. The design of the Blue Velvet Arena and the real-space testing procedure were described in detail elsewhere [34,35]. Briefly, it is a fully-enclosed cylindrical arena 2.8 meters in diameter surrounded by a 2.9 meter high dark blue velvet curtain and it is designed to test two subtypes of navigation; allocentric (world-centered) navigation, which is considered hippocampus-dependent, and where salient distal cues (landmarks) are used for navigation irrespective of an individual’s position [36]. Egocentric (self-centered) navigation is considered parietal cortex-dependent, and relies on an individual’s position and the start location [37]. The egocentric subtask involved using only the start position to locate the goal with no distal orientation cues displayed. The allocentric subtask involved using only two distal orientation cues at the perimeter for navigation to the goal as the start position was unrelated to the goal position. Each subtask involved eight trials. The relative positions of the goal, start position, and both orientation cues were identical across all trials. The performance was measured as the distance error between the subject’s final position and the actual goal location (in centimeters). There was no time limit to find the goal, mainly to reduce bias by differences in cognitive, sensory, and physical functioning. All subjects were tested for both allocentric and egocentric navigation.

Statistical analysis

All analyses were conducted by using IBM SPSS for Windows version 20.0. First, all variables were examined for normality of distribution. Except for SSNC questionnaire composite score, no variable presented significant deviation from normal score distribution. Therefore for SSNC analysis we used non-parametric tests (Mann- Whitney U test and Kruskal-Wallis one-way analysis of variance with post hoc Holm-Bonferroni correction for multiple comparisons. For analysis of other variables we used parametric tests (an independent samples t-test and one way analysis of variance with post hoc Tukey’s test). The χ2 test evaluated differences in proportions (gender). To assess for the influence of depression and anxiety on SSNC scores irrespective of age, gender, education, the real SN performance and cognitive status, we used an ordered logistic regression. Specifically, the SSNC global score was entered into ordered logistic regression model as the outcome variable with the GDS and BAI scores being the main independent variables.

We also examined whether the association between BAI and SSNC was modified by GDS by creating a categorical variable to reflect combinations of anxiety and depression using the median for BAI (≤10) and GDS (≤ 4) as follows: not anxious/not depressed (0; reference), anxious/not depressed (1), not anxious/depressed (2) and anxious/ depressed (3). We controlled for the same covariates as in the main regression model.

There were no differences in GDS and BAI scores between MCI and SCD groups. As expected, the MCI group performed worse in the real SN testing in both egocentric (p<0.001) and allocentric (p<0.001) navigation tasks compared to SCD group. However, MCI subjects did not report more SSNC than SCD subjects (p=0.916) (Table 1 for the characteristics of the MCI and SCD groups).

Table 1: Characteristics of SCD and MCI groups. Note: ¹Mean distance from the correct position of the goal in cm; MMSE: Minimental state evaluation; GDS: Geriatric depression scale; BAI: Beck anxiety inventory); Normal ranges: MMSE 27-30, GDS 0-4, BAI 0-10, BVA ego 0-29, BVA allo 0-39, SSNC - not set.

When we performed ordered logistic regression including all the subjects using SSNC global score as dependent variable and GDS and BAI score as an independent variable, anxiety significantly increased the likelihood of reporting more spatial navigation complaints controlling for cognitive status, real SN performance, and depression (Estimate=0.077, p<0.001). In the same model, the association between depression and the likelihood of reporting spatial navigation complaints was not statistically significant (Estimate=-0.069, p=0.311). When age, gender, and education were also controlled (Appendix 1), greater anxiety was still associated with an increased likelihood of reporting spatial navigation complaints (Estimate=0.105, p=0.003; Appendix 1). Depression showed the opposite trend, whereby greater depression was related to lower likelihood of reporting spatial navigation complaints when all study covariates and anxiety were controlled, although this association was not statistically significant (Estimate=-0.196, p=0.079).

In order to learn more about the effect of depression and anxiety on subjective perception of spatial navigation deficit we divided the sample regardless of cognitive status based on median BAI (≤ 10) and GDS (≤ 4) scores into 4 groups – not anxious/not depressed (n=44), anxious/ not depressed (n=18), not anxious/depressed (n=13) and anxious/ depressed (n=34). These 4 groups characterized by GDS and BAI scores did not differ in age, MMSE score or the real SN performance. The main characteristics of these 4 groups are displayed in Table 2. Results with “anxiety/depression” combinations as the main independent variable are shown in Appendix 2. We found that those in the anxious/ not depressed group were significantly more likely to report spatial navigation complaints than those in the not anxious/not depressed group (Estimate=2.011, p=0.012). In addition, we observed greater likelihood of spatial navigation complaints in the anxious/depressed group, but this result did not reach our preset threshold for statistical significance (Estimate=1.039, p=0.078).

Table 2: Groups according to depression and anxiety (median GDS ≤ 4, median BAI ≤ 10). Note: **p<0.01, *p<0.05 (compared to normal group); †p<0.05 (compared to depressive group); ¹Mean distance from the correct position of the goal in cm; MMSE (Minimental state evaluation); Normal ranges: MMSE 27-30, GDS 0-4, BAI 0-10, BVA ego 0-29, BVA allo 0-39, SSNC - not set.

We found that anxiety rather than depression influenced subjective perception of SN abilities in non-demented elderly, regardless of age, sex, education, real-space SN performance and objective cognitive abilities. We also found that the association between anxiety and SSNC did not vary as a function of depression, but rather that the association between anxiety and SSNC showed a relatively consistent magnitude across levels of depression. Finally, we found that SSNC scores were not reliably related to age and gender of participants. We also found that higher education was associated with greater SSNC, but this relationship did not reach statistical significance (p=0.068). It may be that more educated individuals are more aware of spatial navigation problems, possible as a result of greater daily geographical mobility.

In our study, patients with SCD did not have impaired SN performance, which is in agreement with our previously published data where SCD subjects performed similarly to healthy controls [38] and patients with MCI had similar quality and magnitude of SN impairment as patients with mild AD [34,35]. Interestingly, the difference in the real performance in Blue Velvet Arena was not reflected by a difference of SSNC total score between these two groups, MCI subjects did not report generally more difficulties than SCD subjects in the SSNC questionnaire. This corresponds to the findings from cross-sectional studies on non-demented elderly where subjectively perceived memory problems do not usually reflect a true cognitive deficit [4,5,8]. However, some limitation of those studies might be a different sensitivity of the psychometric tools used for the detection of the underlying cognitive deficit in memory complainers [39]. In this regard, studies with SCD subjects suggest that not all cognitive complaints are of equal significance for the risk of future cognitive decline; hence it is important which questions are asked [40]. More research is needed to determine which concrete complaints may be most significant for future cognitive decline and which are purely associated with anxiety. Our cross sectional data do not allow to analyze weather SSNC which is not associated with anxiety would lead to increased risk of dementia, however this is going to be a question of farther research using data from the longitudinal follow up of these patients.

It is also well documented that memory complaints are often associated with depression [5,7,8]. Therefore, it was surprising that we did not observe such an effect of depression on subjective SN complaints. The post hoc analyses revealed that most of the SSNC were anxiety driven. The finding that greater depressive symptoms were associated with lower likelihood of spatial navigation complaints, albeit not significantly, also deserves comment. It may be that more depressed patients live in a more restricted life-space, hence not “testing” their spatial navigation in real life situations as much as individuals with low depressive symptoms. This finding also provides additional evidence for the notion that anxiety rather than depression may be the more reliable indicator of spatial navigation problems.

The relationship between anxiety and memory complaints, or between anxiety and SN complaints, has been studied in few studies. In a study with 283 community-dwelling people in which authors examined correlations of personality variables of conscientiousness and neuroticism in relation to subjective memory in older adults, anxiety together with self-consciousness explained almost one third of the variance in subjective memory complaints while only 4% unique variance was associated with the objective memory measure [9,10].

On a practical level, subjective memory complaints may interfere with activities of daily living and quality of life in a different way. Those with subjective SN complaints, rather than subjects with subjective memory complaints, may be more fixed to their home environment resulting in reduced capacity of independent shopping, socializing or searching for medical services. These are the reasons why SN capacity and its objective grounding should be of a thorough and separate assessment beyond and above of subjective memory complaints. The focus on patients’ history and self-reporting including that of subjective memory complaints and subjective SN complaints combines benefits of easy to perform and potentially efficient screening tool.

Another issue is every day functioning of individuals with anxiety. Regardless of their otherwise preserved memory and navigational abilities, individuals with increased anxiety might be avoiding situations in which they feel insecure and this might have an impact on their quality of life.

There are some limitations. The self-reported questionnaire may have a ceiling effect for MCI subjects, whereby we may have observed more severe complaints in this group otherwise. In addition, decreased awareness of cognitive problems in cognitively impaired individuals (i.e., anosognosia) may have affected the results. We found that MCI patients and participants with SCD did not differ with respect to anxiety, depression, or memory complaints. We also found that patients with MCI performed worse on the objectively measured navigation tasks, suggesting possible lack of awareness of spatial navigation problems in the MCI subgroup. Future research should examine this possibility.

General subjective complaints about SN performance do not distinguish patients with MCI with objective cognitive decline from subjects with SCD reporting memory complaints but no impaired cognition. Depression but mainly anxiety appears to play a role in subjective perception of navigational skills. Our results suggest that screening for anxiety, and not only for depression, may be useful to elucidate the relevance of subjective cognitive complaints in elderly, specifically SSNC. In the clinical setting, specific questions may yield different clinically useful information as it needs not reflect functional impairment, but may be associated with anxiety. This is important in the context that otherwise SSNC may be associated with real SN deficits which are present early in AD and which may interfere with quality of life and everyday functioning. Decision-making on further examination and management of non-demented subjects who frequently report SSNC needs to consider anxiety as an important contributor to patients’ way of self-assessment.

p>This work was funded in part from the European Union Regional Development Fund – Project FNUSA-ICRC (CZ.1.05/1.1.00/02.0123) and by project ICRCERA- Human Bridge (No. 316345), Grant Agency of Charles University in Prague (624012, 546113 and 1108214).