Cognitive reserve (CR) is a concept describing the brain’s ability to adapt to damage or aging. CR is characterized by the capability of engaging neural networks and ensuring adaptive reorganization of brain functions in response to damage or stress [1]. The concept of CR is widely used when describing aging, as well as neurodegenerative and vascular disorders of the brain.

There are two oppositely directed factors playing a key role in the CR organization: education and the effects of stress, which are often replaced by a more measurable trait, stress hormone cortisol levels. Let us consider carefully these three characteristics.

The studies describing the effects of person’s education on his/her body are divided into two large groups: the effects of social benefits associated with pursuing higher education (HE) and alterations of brain’s structural and functional organization occurring under the influence of education. The period of cognitive training associated with pursuing HE can be considered as longer, than that associated with pursuing secondary education (SE). HE is often finished at a relatively young age characterized by high brain plasticity, so there is an opportunity to optimally shape brain networks. The larger accumulated CR also manifesting itself in attitude to stress caused by cognitive stress represents one of the major benefits for individuals with HE.

The course of stress is different in men and women, so it is feasible to consider it separately. This paper is focused on the responses associated with cognitive stress in women with HE and SE suffering from chronic cerebral ischemia (CCI). Women with HE have potentially improved stress regulation mechanisms. This can be due to the dynamic changes in cortisol levels [2]. In women with HE, there is possibility of more intense activation of the prefrontal cortex (PFC) playing an important role in executive functions and emotional regulation [3]. HE is likely to result in stronger PFC cortex activation during stress, thereby contributing to better cognitive control and emotional regulation. This is especially true for interconnections between PFC and cyngulate gyrus. Furthermore, HE is associated with enlargement of the hippocampus playing a role in regulation of memory and responses to stress, possibly ensuring a protective effect of education on hippocampal integrity [4].

One can see various patterns of activation of the amygdala involved in developing responses to fear and threat. HE can result in the less prominent and better controlled activation of the amygdala during stress [5].

Women with lower socioeconomic status that is often associated with little education have elevated basal cortisol levels, even when there seem to be no severe stressors [6]. Reduced hippocampal volume under conditions of chronic stress can negatively affect its functioning, which results in deterioration of memory and impaired stress regulation [7], as well as to reduction of CR.

In individuals with inadequate education, the increased amygdala reactivity and chronic stress can sensitize the amygdala, thereby contributing to enhanced fear response. Furthermore, PFC is not always able to regulate responses of the amygdala, which results in the less effective emotional regulation [8].

Studies of the neuronal correlates of cognitive load in women with different levels of education are still fairly limited, and the results of different studies are not always consistent. Higher cognitive load require larger cerebral energy resources, which results in the increased neuronal activity. This is suggested by the functional magnetic resonance imaging (fMRI) studies, where the increase in BOLD signal (Blood Oxygen Level Dependent) is associated with activation in parietal lobes and PFC. The increased activation in parietal lobes is associated with spatial processing, attention, and working memory, and their activity increases with increasing cognitive problems [9].

METHODS

The study involved 137 women aged 50–85 years suffering from CCI without diabetes. The average duration of CCI was 10.1±0.7 years. CCI patients were divided into two groups: with SE and HE. The main etiological causes of CCI are as follows: atherosclerosis, hypertension (including hypertensive heart disease), venous insufficiency, diabetic angiopathy, vasculitis of various etiology, etc. Inclusion criteria: initial manifestations and subcompensated CCI; no need for permanent care from others in patients' daily life [10]. Non-inclusion criteria: dementia severity score 1 or more (Clinical Dementia Rating) [11]; history of acute cerebrovascular accident, traumatic brain injury, severe heart or renal failure, uncompensated thyroid dysfunction. The diagnosis of CCI was further verified by duplex scan and follow-up MRI. All the patients were righthanded. The patients’ body height and body weight were measured to determine body mass index (BMI). BMI = body weight/body height². Two age characteristics were taken into account: patient’s age at the time of recording and age at first referral to a medical institution due to CCI symptoms. The first referral was usually associated with memory and concentration problems accompanying arterial hypertension, as well as with cerebrovascular accidents. The second trait was age at the time of experimental assessment. It was assumed that patients with different levels of education reached this stage at different age. Both groups of women were mainly knowledge workers, while women with SE were mainly accountants or nursing staff.

Resting state fMRI

The subjects (25 CCI patients) underwent Т2* weighted resting state fMRI of the brain in order to record BOLD signal in the Magnetom Verio magnetic resonance imaging scanner (Siemens, Germany) with the magnetic field strength of 3,0 Tesla. The subjects were offered to follow the instructions: to relax as much as possible, lay still with the eyes closed (to avoid stimulation of visual sensory system), not to think about anything in particular. MRI data were processed using the SPM12 software in the MATLAB computing environment. The CONN-18b application on the SPM-12 platform was used to assess connectivity. Connectivity was assessed in various neural networks of the brain. We compared connectivity in two groups of CCI patients with different levels of education. Significance of differences in these groups was assessed based on the standardized regression coefficient adjusted for multiple comparisons (FDR, False Discovery Rate) in CONN-18b.

The MAGNETOM Verio magnetic resonance imaging unit (Siemens, Germany) had the magnetic field strength of 3,0 Tesla. Functional scans were acquired in the resting state using the Т2* weighted EPI sequence: TR = 1500 ms, TE = 30 ms, flip angle 70°, slice thickness 2 mm, FOV = 190 mm, FoV phase 100.0%.

Cognitive tests

Patients were assessed using the previously reported correction test, verbal fluency test, Luria’s verbal working memory test, МоСA test [12]. Furthermore, we recorded blood pressure, calculated pulse pressure (difference between systolic and diastolic blood pressure) and heart rate.

Biochemical tests

The patients’ salivary cortisol levels were determined using the Abbott i2000 ARCHITECT immunochemiluminescence analyzer (Abbott Laboratories, Illinois, USA) using the reagent kit of the same brand.

Saliva samples were collected in accordance with the previously reported protocol [13]. The patients did not drink alcohol for a week, tea or coffee for 1 h before saliva collection, rinsed their mouth with water 10 min before this. Saliva collection was accomplished through spitting into a test tube with the volume of at least 1.5 mL. Saliva samples contaminated with blood were excluded from the study. For that the enzyme immunoassay kit was used to determine contamination of saliva with blood [13].

Statistical processing

Analysis of the data obtained was performed using the Statistica-12 software package (Dell, USA). The Kolmogorov–Smirnov test was used to test the distribution for normality. We calculated mean values, standard errors, conducted one-way analysis of variance and correlation analysis. To analyze neural networks, Student’s t-test was calculated, and adjustment for multiple comparisons was applied – FDR (False Discovery Rate).

Assessment procedure

Saliva was first collected at least 2 h after the meal; initially, we performed background recording of slow electrical activity (for no more than 5 min), then, with the 1–2 min break, the patients performed three cognitive tests: Kirchner’ correction test (3 min), verbal fluency test (phonemic variant) (4–5 min), Luria’s verbal working memory test (5–7 min). Saliva was collected again immediately after performing tests, within 1–2 min. The МоСа test was performed after the end of the experiments and saliva collection.

RESULTS

In CCI patients, contingency between the level of education and age characteristics is significantly manifested in women. Table tab. 1 provides demographic data showing the role of education in the dynamics of abnormal vascular aging (CCI).

There is a clearly visible effect of HE on the women’s age characteristics: women with HE seek medical care for CCI later (tab. 1). Initial presentation to the medical institution of individuals with SE took place about 7 years earlier compared to patients with HE. Since there are literature data on the impact of continued education on stress, let us consider the effects of education on the stress mechanisms, specifically on cortisol, the stress hormone (tab. 2).

In women with HE, baseline cortisol levels were about 14 nmol/L lower, and responsiveness to cognitive load was about 3.5–5 times higher (tab. 2). There are no significant differences in cortisol values after cognitive load in two groups. This largely explains different cortisol shift values reported for both groups. It seems that there is some maximum possible cortisol level associated with cognitive load, and the difference in cortisol shift values is explained by different baseline cortisol levels reported for both groups. We have revealed a significant correlation between baseline cortisol levels and relative cortisol reactivity: r = 0.41; n = 88; p = 0.00008.

The cortisol level shift associated with cognitive load is related to the age of first referral to the medical institution due to the emergence of CCI symptoms in patients with HE (fig. 1).

The greater the cortisol reactivity, the later the patient seeks medical help due to the emergence of CCI symptoms. It should be noted that patients with almost the same disease severity were selected in the hospital for assessment. Individuals with HE had lower baseline cortisol levels, than women with SE (tab. 1), while the patients’ age had no significant effect on cortisol levels (p = 0.14). Under cognitive load, cortisol levels of individuals with HE and SE were almost the same: on average 63.0 ± 3.4 nmol/L (n = 83).

What are the differences in organization of neural networks based on fMRI data in patients with different levels of education? Let us consider the differences in indicators of connectivities connecting brain regions in individuals with HE and SE (fig. 2).

Predominance of positive connectivity differences showing that contingency of some brain structures in patients with HE is higher, than in patients with SE, can be seen (fig. 2). These differences relate mainly to the temporal regions of both hemispheres and hippocampus, as well as to some regions of the cerebellum and cerebellar vermis.

Let us consider statistical characteristics of the difference in connectivities significant taking into account multiple comparisons and different in both groups of patients (tab. 3). The analysis of these connectivities shows, which connections predominate in patients with HE and SE.

It should be noted that positive connections predominate in individuals with HE (tab. 3). The only negative connectivity linking the Heschl's gyrus with the cerebellar vermis is nonsignificant considering multiple comparisons (р-FDR > 0.5). This connectivity is provided in tab. 3, since the p-FDR value exceeds significant by less than 0.008.

Connectivities between the above regions can be digitized and presented as vectors, where a regression coefficient would be correlated to each connectivity of each patient. These connectivities show significant differences in both groups based on the F-test varying between 17.5 and 27.4 with the significance levels between 0.00035 and 0.00003.

Some connectivities are related to cortisol regulation (tab. 4).

Connectivities that are positive by sign are negatively correlated to baseline cortisol levels and positively correlated with relative cortisol reactivity; negative connectivity (HGrVer10) is negatively correlated to relative cortisol reactivity.

Thus, certain connectivities are significantly different in individuals with different levels of education and are associated with regulation of cortisol levels. The contribution of the PPrpSTGl, COr-PTl, and HGr-Ver10 connectivities to ensuring cognitive reserve is likely to be associated with their connections with the auditory-speech system.

DISCUSSION

Our studies have revealed a later incidence of CCI in women with HE. It is likely that direct impact of education has a significant effect on CR. CR of women with HE combined with such factors, as social status and a more healthy lifestyle, contribute to stress reduction and, therefore, to the later CCI onset. This is also confirmed by lower baseline cortisol levels in patients with HE and increased cortisol responsiveness to cognitive load. High cortisol reactivity in such individuals supports a more effective energy supply of cognitive functions and their successful realization, especially when the mental load is relatively brief. The fact of higher brain connectivity in women with HE and the relationship between these connectivities and cognitive functions seems to be important.

According to our data, women with HE develop chronic cerebral ischemia (CCI) later, which is suggested by their later referral to medical institutions (on average by 7 years) and the older age of the beginning of assessment, given they have the same disease stage corresponding to stage 1 or 2 dyscirculatory encephalopathy. The fact that CCI patients with HE age slower is confirmed by higher connectivity of the brain neural networks, which is generally considered to be the sign of the better functioning and younger brain [14].

Modern literature describes differences in aging of women with HE and SE. These differences relate to biological age and external manifestations of age-related changes. Undoubtedly, several factors, including the level of education, contribute to slowing down the aging processes [15]. HE is often correlated to higher socioeconomic status, which results in better access to the resources affecting health and aging. This includes access to high-quality healthcare, proper diet and healthy habits. Welleducated women can be more prone to maintaining healthy lifestyle, including regular physical exercise, smoking cessation, and body weight control. All of them also know more about the health risks and preventive measures.

HE is associated with the larger CR, i.e. the brain’s ability to cope with damage and maintain normal functioning [1]. This can result in the later onset or slower progression of the agerelated cognitive decline. Women with HE develop CCI at older age and show slower disease progression, which is also true for Alzheimer's disease and the development of other dementia types [16].

The relationship between education and cortisol levels suggesting lower baseline cortisol levels in individuals with HE is also worth attention. The issue of connectivities that are different in individuals with different levels of education and sensitive to cortisol levels suggests that there is a principal relationship between two major CR factors.

Functional connectivity between certain brain regions, such as posterior parietal region (PPr), posterior superior temporal gyrus (pSTG), corticomotor output region (COr), parietal cortex (PTl), and the cerebellar vermis (HGr-Ver10), play an important role in auditory processing, speech and cognitive functions in general. Neuroimaging studies have shown that the characteristics of these connectivities are often different in patients with various disorders of auditory processing, such as verbal and auditory load. These connectivities are also correlated to cortisol levels, especially in response to cognitive load [17–18].

Based on the structural and physiological organization of these brain regions, it can be assumed that the PPrpSTG connectivity being part of the auditory association system contributes to integration of auditory information and understanding of language [19]. The COr-PTl connection is considered to be involved in synthesis of speech and sensorimotor integration essential for speech functions [20]. The HGr-Ver10 connectivity connecting primary auditory cortex with the cerebellar vermis is considered to be basic for auditory-motor regulation, which contributes to implementation of the tasks requiring coordination of auditory stimuli and motor responses under normal and abnormal conditions [21]. Thus, the role of these connections in construction and organization of CR becomes more evident.

CONCLUSION

HE of women is highly likely to ensure a number of extra benefits. This is often associated with the higher material wealth, awareness of healthy lifestyle, lower adherence to harmful habits, and other factors. All the above, along with lower cortisol levels in individuals with HE, contribute to higher CR and longer health preservation. This is also confirmed by higher connectivity values associated with lower cortisol levels in individuals with HE.