Coloma Andrews LC; Zihl J

Research Article

Austin J Psychiatry Behav Sci. 2014;1(3): 1015.

Cognitive Reserve in Major Depression–Associations with Cognitive Status, Age, Education, Personality, and Depression Severity

Coloma Andrews LC¹* and Zihl J^2,3

¹Institute for Stroke and Dementia Research, Ludwig Maximilian University, Germany

²Max Planck Institute of Psychiatry, Munich University, Germany

³Department of Psychology, Ludwig Maximilian University, Germany

*Corresponding author: Coloma Andrews LC. Institute for Stroke and Dementia Research, Medical Center, Ludwig Maximilian University, Marchioninistr, 15, 81377 Muenchen, Germany

Received: March 31, 2014; Accepted: April 01, 2014; Published: April 03, 2014

Abstract

Cognitive reserve (CR) is understood as a latent potential underlying the flexible adaptation to mental challenges. By optimizing cognitive performance, it can be used to cope with high task demands. This study examines CR in 40 inpatients with unipolar depression compared to 24 healthy control subjects. The size of CR was assessed by calculating the maximal performance improvement over retesting with a digit symbol substitution task. Furthermore, the relation between CR, cognitive status, age, education, and the personality traits Openness for experience and Neuroticism was explored. CR did not differ significantly between controls and the whole group of depressed patients. However, patients who displayed cognitive deficits in one–time neuropsychological testing (50%) showed a lower CR than controls while patients without deficits showed a marginal higher CR. In patients, CR was positively associated with attention, short term memory, and openness for experience. In controls, CR was relatively independent from cognitive status but showed a negative association with age. Our results support the idea that subgroups of patients can be differentiated through cognitive status as well as CR. Furthermore, the marginal higher CR in patients without cognitive impairment suggests that CR acts as a buffer against the development of cognitive deficits in depression.

Keywords: Cognitive Reserve; Cognitive Plasticity; Depression; Cognition; Personality; Testing–the–limits.

Abbreviations

CR: Cognitive Reserve; Ttl: Testing–The–Limits; DSST: Digit Symbol Substitution Test; STM: Short Term Memory; WM: Working Memory; MADR–S: Montgomery Asberg Depression Rating Scale; MMSE: Mini–Mental State Examination; SSRI: Selective Serotonin Reuptake Inhibitors; NASSA: Noradrenergic and Specific Serotonergic Antidepressant; NARI: Selective Noradrenaline Reuptake Inhibitors; D: Depressed Patients without Cognitive Deficits; Ddef: Depressed Patients with Cognitive Deficits.

Introduction

Background

Cognitive deficits in depression are well documented. Impairments have been found in many cognitive domains, including attention [1–3], memory [4,5], and executive function [6,7]. However, it has been doubted whether the standard one–time testing reflects the “latent competence” of a subject. This competence is understood as cognitive reserve (CR), which is activated when required to cope for functional consequences of brain pathology [8]. Therefore, people with high CR have a better ability to compensate for pathologies like Alzheimer’s or Parkinson’s disease than people with low CR [9–12]. In healthy individuals CR has been suggested as a protective factor against cognitive decline in normal ageing [13]. CR is also sought to enable individuals at any age to cope with increased task demands by optimizing cognitive performance, possibly by using neural networks high in efficiency or capacity [8]. To measure CR, proxy variables are often used which are hypothesized to reflect cognitive functioning, like education or intelligence (for a discussion, see [14]). Other authors have favored a more dynamic measure of CR by assessing the potential to adapt performance to a challenging cognitive task (e.g. [15]). In the testing–the–limits (TtL) procedure one’s maximal performance improvement due to training or practice is assessed [16–18]. Within this procedure, CR differs from cognitive status as it reflects the latent competence one has available when high performance is needed.

The first objective: The first objective of the present study is to examine CR in depressed patients with a TtL procedure using a simple retest design. As a CR measure, we used the individual performance improvement in the Digit Symbol Substitution Test (DSST, revised Wechsler Intelligence Scale, german version; Aster [19]), which is a simple but multi faced measure incorporating multiple cognitive abilities [20,21]. Since depression is associated with a wide range of structural and functional neuronal abnormalities (e.g. Hickie and Rogers [22,6], not only detriments in cognitive status but also in CR may be found in depression [23,24]. By using performance gain after memory training as a CR measure, Calero and Galiano [25] found no difference in CR between older subjects with and without high scores in a self–rating depression scale. However, generalization of those results is restricted, since subjects with an elevated score might not have been clinical depressed according to DSM–IV criteria [26]. The present study therefore compares clinically diagnosed depressed patients with healthy controls. As depression is not necessarily associated with cognitive impairments in one–time–testing [27,3], we additionally compared CR in patients with and without clinically relevant cognitive deficits.

The second objective: The second objective of the present study is to explore a number of variables associated with CR in patients and healthy subjects: age, education, cognitive status, and the personality factors Openness and Neuroticism. As healthy aging is associated with a mild decline in cognitive functions [28], higher age may also be negatively associated with CR [16,18,29]. Since past research has repeatedly shown that the DSST is age sensitive (for a meta–analysis, s. Hoyer [30]), this may be especially prone in a DSST–based CR measure. A positive association with education and cognition has been suggested by Richards and Deary [31] and Scarmeas and Stern [32] as it may help to establish efficient cognitive skills. However, positive associations between general cognitive abilities and CR have only been reported in some (e.g. Singer [29]) but not all [15,21] studies using performance improvement as a CR measure. Past research has shown that the “Big Five” [33] personality factor Openness to experience has a direct positive effect on intelligence through environmental enrichment [34] and is positively related to cognitive abilities like memory and executive functions [35– 37]. Neuroticism has frequently reported to be negatively related to cognition (e.g. Chapman [38]). Therefore, Openness (+) and Neuroticism (−) may also be important for the development of CR. However, the relationship between performance improvement and personality has only been studied by few studies, yielding different results [39–41]. Nevertheless, one can assume that a low Neuroticism on the one hand and good cognitive abilities, a high education and a high Openness on the other hand not only favors the development of CR by the accumulation of efficient strategies but are also beneficial in mobilizing CR to improve performance in a current test situation.

In depression, the mobilization of CR may also be affected by the severity of depression as the severity of symptoms can influence the extent of cognitive impairment [42].

The following hypotheses: The following hypotheses have been pre specified in the present study:

Single DSST trials and CR are positively associated with cognitive status
Single DSST trials and CR are positively associated with years of education.
Single DSST trials and CR are negatively associated with age.
Single DSST trials and CR are positively associated with Openness and negatively associated with Neuroticism.
Single DSST trials and CR are negatively associated with severity of depression in patients.
No pre specified hypotheses have been made for group comparisons.

Methods

Study design

Depressed patients and healthy control subjects were tested to assess the individual size of CR, cognitive status, and personality characteristics. Patients were tested on two separate days to avoid fatigue effects. The first day included a routinely administered test battery to comprehensively examine performance in attention, memory, and executive function. The second day implied further study specific measures. Testing in control subjects was executed on a single day including only measures used for the quantitative analysis.

Subjects

Depressed in–patients admitted to the Max Planck Institute of Psychiatry in Munich were originally included when meeting the following inclusion criteria: first episode of unipolar major depression or recurrent depression. Diagnosis of depression was made by the treating psychiatrist according to DSM–IV criteria [26]. Exclusion criteria were other primary psychiatric diagnoses than depression, depression with psychotic symptoms, electroconvulsive therapy within the last three month, present or past neurological illness, present or past substance abuse, un medicated hypo– or hypertonia, diabetes and thyroid dysfunction.

Healthy control subjects without a history of psychiatric or neurological illness were recruited through notices at the Ludwig Maximilian University in Munich and through a control–sample which previously participated in a non–cognitive study at the Max Planck Institute of Psychiatry.

All patients and controls gave written informed consent according to the latest version of the Declaration of Helsinki.

Procedure and Material

Cognitive reserve

CR was measured by a TtL procedure using a retest paradigm with ten consecutive trials of the Digit Symbol Substitution Test (DSST, revised Wechsler Intelligence Scale, german version; Aster [19]) without providing feedback or strategies to improve performance. This simple paradigm was chosen to maximize compliance in spite of the expected reduction of motivation and drive in patients. In each trial, a coding key is presented on the top of the sheet assigning the numbers 1–9 to corresponding symbols. Subjects are asked to use the coding key to note the corresponding symbols under blank fields below a series of digits. The same test was administered ten times in a row with a one minute break in between. To avoid ceiling effects, processing time was limited to 90 sec. The number of correctly written symbols per test served as the outcome measure.

Cognitive status, personality and severity of depression

Table (1) shows the cognitive tests used in the test battery routinely administered at the Max Planck Institute of Psychiatry in Munich. To compare patients with and without cognitive deficits, performance in every test was rated. A clinical relevant deficit was defined as having a test score lower than 1.5 SD below the test specific norm in one or more measures. To examine variables associated with CR, we limited the analysis to three cognitive tests as valid measures of short term⁄ working memory, selective attention, and problem solving.

CR_i = 1+(1+x₁ ⁄ x_maxgroup )*[(x_i - x₁ ) ⁄ x_maxgroup ]

Selective visual attention: Selective visual attention was measured by the d2 test of attention [43]: Participants were required to cross out the letter d with two dashes out of 14 lines of letters p and d with one to four dashes arranged above and below each letter. The number of correctly crossed out letters minus the number of errors served as the attention score.

Verbal short term and working memory: Verbal short term and working memory was measured by the subtest digit span forward and backward of the revised Wechsler Memory Scale (german version; Härting [44]). Subjects were asked to repeat strings of digits in increasing length either in the same (short term memory; STM) or the reversed order (Working Memory; WM).

Problem solving : Problem solving was measured by the subtest matrix reasoning (matrices) of the Wechsler Adult Intelligence Scale– III (german version; Aster [19]): Subjects were instructed to complete 26 geometric patterns of increasing difficulty by choosing the correctout of five inserts.

Personality: Personality was assessed by a computerized version of the NEO–Fife–Factor–Inventory (NEO–FFI; german version; Borkenau and Ostendorf [46]). Subjects were instructed to indicate on a five–point Likert scale how each of 60 statements suits their personality. The test–score for Neuroticism reflecting nervousness and anxiety and Openness to experience reflecting curiosity and creativity were used for analyses.

Severity of depression: Severity of depression in patients was measured by the Montgomery Asberg Depression Rating Scale (MADR–S; Montgomery and Asberg [47]). The MADR–S is an external rating instrument consisting of ten items representing symptoms of major depressive disorder. The rating was based on a structured interview [48] (Table 1).

Table 1: Cognitive tests used in the routinely administered neuropsychological test battery.



  
    Table 1. Cognitive tests used in the routinely    administered neuropsychological test battery.
  
  
    Cognitive domain
    Standardized cognitive test
  
  
    Attention
    Processing speed
    Zahlenverbindungstest    A & B
      comparable    to trail making test [72]
  
  
    Selective attention
    d2 test    of attention
      [43] 
  
  
    Alertness
    Testbatterie zur    Aufmerksamkeitsprüfung (TAP)
      computerized reaction-time test [73]
  
  
    Memory
    Verbal memory
    Wechsler    Memory Scale (German version) logical memory
      [44] 
  
  
    Verbal short term/working memory
    Wechsler    Memory Scale (German version) digit span forward and backward
      [44] 
  
  
    Verbal learning and recall
    Verbaler    Lern- und Merkfähigkeitstest
      comparable    to California Verbal Learning Test [74]
       
      If the    age of the patient exceeded 65 years, word list learning, recall and    recognition of the consortium to establish a registry for Alzheimer's disease    (CERAD) test battery was rated instead [45].
  
  
    Executive functions
    Problem solving
    Wechsler Intelligence Scale    (German version) matrices
      [19] 
  
  
    Cognitive flexibility
    Computerized    version of Wisconsin Card Sorting Test [75]
      Regensburger    Wortflüssigkeits-Test (word    fluency test; Aschenbrenner [76])



Table 1: Cognitive tests used in the routinely administered neuropsychological test battery.

Statistical Analysis

As a measure of CR, we used the following formula:

Improvement from the first DSST (x₁) to each of the following nine trials (x_i) is calculated. The formula accounts for possible ceiling effects, which can be expected due to psychomotor restrictions, by compensating for the baseline (x₁) and the maximum score occurring in the respective group (x_{max group} of patients⁄controls). The maximal score of the resulting nine scores is used as the CR measure as it reflects the individual maximal performance improvement.

Since most of the data is lacking normal distribution, nonparametric tests are used for statistical analyzing. For analyses between two groups, Mann–Whitney U tests are computed. Differences between three groups are calculated with Kruskal– Wallis–Tests using post–hoc Mann–Whitney U tests with Bonferroni adjusted p values are calculated to detect the location of differences. To examine performance differences in subsequent testing, Wilcoxon signed–rank tests are calculated. To analyze relations between variables, one–tailed Kendall tau–b rank correlation coefficients are computed. To control for variables such as age, years of education, and depression severity, regression residuals of forced entry multiple linear regression analyses are used for further testing. For correlations with CR, the upper (positive relation) or lower (negative relation) levels of bias–corrected and accelerated 90% bootstrap confidence intervals are reported. Bootstrap resampling was set at 2000.

By default, a–levels were set to .05. Bonferroni adjusted p values (multiplied by 4) are reported when calculations imply cognitive status (attention, STM, WM, problem solving).

Results

Subjects

Depressed patients: Fifty–four depressed in–patients were originally included in the study. Fourteen patients were excluded from analysis due to meeting exclusion criteria at discharge (n = 9), measurement errors (n = 1) or prolonged latency between test sessions (n = 4). The final sample included 40 patients (21 females, 19 males; M age = 44.9, SD = 13.5, range 18–70; M years of education = 12.98, SD = 3.9, range 9–21) either diagnosed with a first depressive episode (moderate: n = 4; severe: n = 9) or recurrent depression (moderate: n = 10; severe: n = 17). All but three patients received medication by the time of testing (table (2)). Severity of depression was assessed by the Montgomery Asberg Depression Rating Scale (MADR–S; Montgomery and Asberg [47]) with a score of 0 indicating absence of symptoms and a score of 60 indicating severe depression. The mean MADR–S sum score was 19.8 (SD = 10.5, range = 3–45). One subject reaching a score of 3 in the MADR–S was not excluded since he was not considered as remitted both by himself and the treating psychiatrist. All patients completed the Mini–Mental State Examination (MMSE; Folstein [49]) as a dementia screening instrument. The mean score was 29.6 (SD = 0.6, range 28–30) indicating normal general cognition in all patients (Table 2).

Table 2: Medication in depressed patients (N = 37).



  
    Table 2. Medication in depressed patients (N = 37).
     
  
  
    Type of medication
    n 
  
  
    Sedative Antidepressants (Mirtazapine, Trazodon, tricyclic    antidepressants amitriptylin-type)
    12
  
  
    Nonsedative Antidepressants (tricyclic antidepressants imipramin and    desipramin type, SSRI, NASSA, NARI, monoamine oxidase inhibitors)
    28
  
  
    Benzodiazepine
    7
  
  
    Neuroleptics
    8
  
  
    Lamotrigin
    14
  
  
    Note. 3 Patients refused medication by the time of    testing.



Table 2:  Medication in depressed patients (N = 37).

Healthy control subjects: Twenty–four age, gender, and education–matched controls with no history of psychiatric diagnoses were included (13 females, 11 males; M age = 43.7, SD = 16.1, range 20–70; M years of education = 13.1, SD = 4.1, range 8–21). None of the controls reached a score higher than 14 points in the Beck– Depression–Inventory–II (BDI–II; german version; Hautzinger [50]) indicating the absence of mild or more severe depression. None of the controls reported subjective cognitive deficits. The mean score of the MMSE was 29.7 (SD = 0.5, range 29–30).

Performance in the digit symbol substitution test

Figure (1) displays the results of the ten DSST trials in depressed patients and control subjects. In DSST1, patients solved on average 51.35 symbols (SD = 14.74) and increased their performance on average by 1.63 SD of baseline performance (24.08⁄14.74 = 1.63 SD) to a mean individual maximum of 75.43 symbols (SD = 11.34). Control subjects showed a mean initial performance of 56.63 symbols (SD = 15.08) in DSST1 and improved their performance on average by 1.55 SD (23.38⁄15.08 = 1.55 SD) to a maximum of 80.00 symbols (SD = 18.45). Wilcox on signed–rank tests revealed a significant improvement from initial to maximal performance in both, patients (z = 5.51, p < .001) and controls (z = 4.29, p < .001) (Figure 1).

Figure 1 :Mean performance in the ten retest trials of the Digit Symbol Substitution Test (DSST) in depressed patients (N = 40) and control subjects (N = 24).



Figure 1 :Mean performance in the ten retest trials of the Digit Symbol Substitution Test (DSST) in depressed patients (N = 40) and control subjects (N = 24).

Relations between CR and variables of interest

Table (3) shows the one–tailed Kendall–Tau–b rank correlations in depressed patients. As highlighted in the table, CR was positively related to attention (τ_b = .26, p = .036) and STM (τ_b = .35, p = .004). However, only the relation between attention and CR remained significant when controlling for age, years of education, and severity of depression (τ_b = .42, p = .02). Furthermore, rank correlations revealed a positive relation between CR and Openness to experience (τ_b = .20, p = .036) and a negative relation between CR and severity of depression (τ_b = .19, p = .039). The latter remained even when controlling for age and education (τ_b = .36, p = .014). No further significant correlation was found for cognitive status (WM, problem solving), age, education and personality (Neuroticism).

Table 3: Kendall Tau-b rank correlations (one-tailed) with DSST trials 1-10 and cognitive reserve in patients (N = 40).



  
    Depressed    patients
     
    1
     
    2
     
    3
     
    4
     
    5
     
    6
     
    7
     
    8
     
    9
     
    10
     
    CR
    CI level
  
  
    Attention (d2)
    t_b
     
    .51^**
     
    .45^**
     
    .45^**
     
    .46^**
     
    .48^**
     
    .42^**
     
    .45^**
     
    .47^**
     
    .46^**
     
    .48^**
     
    .26^*
     
    .41
  
  
    p
     
    <.01 
     
    <.01 
     
    <.01 
     
    <.01 
     
    <.01 
     
    <.01 
     
    <.01 
     
    <.01 
     
    <.01
     
    <.01
     
    .04
     
     
  
  
    STM (digit span forward)
    t_b
     
    .17
     
    .18
     
    .24
     
    .24
     
    .29^*
     
    .28^*
     
    .33^*
     
    .30^*
     
    .34^**
     
    .40^**
     
    .35^** 
     
    .52
  
  
    p
     
    .28
     
    .23
     
    .09
     
    .09
     
    .03
     
    .03
     
    .01
     
    .02
     
    <.01
     
    <.01
     
    <.01
     
     
  
  
    WM (digit span backward)
    t_b
     
    .31^*
     
    .24
     
    .28^*
     
    .34^**
     
    .33^*
     
    .33^*
     
    .33^*
     
    .26
     
    .26
     
    .27^*
     
    .15
     
    .32
  
  
    p
     
    .01
     
    .08
     
    .04
     
    <.01
     
    .01
     
    .01
     
    .01
     
    .06
     
    .06
     
    .04
     
    .39
     
     
  
  
    Problem solving (matrices)
    t_b
     
    .27^*
     
    .24
     
    .22
     
    .32^*
     
    .27^*
     
    .29^*
     
    .28^*
     
    .27^*
     
    .32^*
     
    .32^*
     
    .16
     
    .35
  
  
    p
     
    .04
     
    .07
     
    .10
     
    .01
     
    .04
     
    .03
     
    .03
     
    .04
     
    .01
     
    .01
     
    .30
     
     
  
  
    Age
    t_b
    -
    .46^**
    -
    .31^**
    -
    .29^**
    -
    .30^**
    -
    .24^*
    -
    .23^*
    -
    .20^*
    -
    .21^*
    -
    .22^*
    -
    .20^*
     
    .05 
    -
    .14
  
  
    p
     
    < .01
     
    <.01
     
    <.01
     
    <.01
     
    .01
     
    .02
     
    .04
     
    .03
     
    .02
     
    .04
     
    .67
     
     
  
  
    Years of education
    t_b
     
    .05
     
    .05
     
    .07
     
    .16
     
    .15
     
    .11
     
    .16
     
    .13
     
    .13
     
    .17
     
    .12 
     
    .36
  
  
    p
     
    .35
     
    .35
     
    .27
     
    .10
     
    .10
     
    .19
     
    .09
     
    .13
     
    .15
     
    .07
     
    .15
     
     
  
  
    MADR-S (severity of    depression)
    t_b
    -
    .18
    -
    .23^*
    -
    .26^*
    -
    .24^*
    -
    .29^**
    -
    .29^**
    -
    .32^**
    -
    .33^**
    -
    .33^**
    -
    .33^**
    - 
    .19^* 
    -
    .36
  
  
    p
     
    .06
     
    .02
     
    .01
     
    .02
     
    <.01
     
    <.01
     
    <.01
     
    <.01
     
    <.01
     
    <.01
     
    .04
     
     
  
  
    Openness for experience
    t_b
     
    .11
     
    .20^*
     
    .18
     
    .17
     
    .18
     
    .17
     
    .17
     
    .19^*
     
    .21^*
     
    .17
     
    .20^* 
     
    .40
  
  
    p
     
    .18
     
    .04
     
    .06
     
    .06
     
    .06
     
    .06
     
    .07
     
    .05
     
    .03
     
    .07
     
    .03
     
     
  
  
    Neuroticism
    t_b
    -
    .02
    -
    .05
    -
    .12
    -
    .08
    -
    .10
    -
    .11
    -
    .09
    -
    .05
    -
    .09
    -
    .10
    -
    .12
    -
    .28
  
  
    p
     
    .42
     
    .32
     
    .14
     
    .24
     
    .18
     
    .17
     
    .21
     
    .32
     
    .21
     
    .18
     
    .14
     
     
  
  
    Note. CR, cognitive reserve; d2, d2 test of attention; CI    level, upper (attention, STM, WM, problem solving, years of education,    Openness for experience) or lower (age, MADR-S, Neuroticism) 90% bootstrap    confidence level for correlations with cognitive reserve; MADR-S, Montgomery    Asberg Depression Rating Scale; * p <    .05; ** p < .01;
      p-values for cognitive status (attention, STM, WM problem solving) are    Bonferroni adjusted.



Table 3:  Kendall Tau-b rank correlations (one-tailed) with DSST trials 1-10 and cognitive reserve in patients (N = 40).

Table (4) shows the one–tailed rank correlations in control subjects. No significant relation between cognitive status, personality and CR was found (all p <.05). However, rank correlations revealed a negative association between age and CR (τ_b = −.26, p = .041). No significant positive association was found between years of education and CR and the upper level of the 90% bootstrap confidence interval indicates a negative association (−.04).

Relations between single DSST trials and variables of interest

Table (3) shows that in patients, attention, problem solving, and WM were all positively related to DSST1. No significant relation between STM and DSST1 was found but correlation coefficients marginally increased from D_SST1 to D_SST10 (z = 1.59, p = .055). In patients, a negative relation was found between age and single DSST trials. However, this relation significantly decreased from D_SST1 to D_SST10 (z = 1.83, p = .034). No significant correlation between severity of depression and first DSST performance was found. To further analyze the association between severity of depression and cognitive status, we calculated partial correlations with attention, STM, WM and problem solving controlling for age and education. Those revealed only one significant negative association with STM (τ_b = .32, p = .024).

Table (4) shows that in control subjects, attention and STM were positively related to DSST 1–10 while age was negatively related to single DSST trials. No further significant associations were found for single DSST trials and cognitive status (WM, problem solving), education or personality (all p >.05) (Tables 3 and 4).

Table 4: Kendall Tau-b rank correlations (one-tailed) with DSST trials 1-10 and cognitive reserve in controls (N = 24).



  
    Control subjects
     
     
    1
     
    2
     
    3
     
    4
     
    5
     
    6
     
    7
     
    8
     
    9
     
    10
     
    CR
    CI level
  
  
    attention (d2)
    t_b
     
    .58^**
     
    .56^**
     
    .60^**
     
    .51^**
     
    .46^**
     
    .51^**
     
    .47^**
     
    .42^**
     
    .41^*
     
    .39^*
     
    .11
     
    .40
  
  
    p
     
    <.01
     
    <.01
     
    <.01
     
    <.01
     
    .01
     
    <.01
     
    <.01
     
    <.01
     
    .01
     
    .02
     
    .92
     
     
  
  
    STM (digit span forward)
    t_b
     
    .35^*
     
    .36^*
     
    .39^*
     
    .37^*
     
    .37^*
     
    .42^*
     
    .41^*
     
    .43^*
     
    .35^*
     
    .40^*
     
    .20
     
    .45
  
  
    p
     
    .04
     
    .04
     
    .02
     
    .04
     
    .03
     
    .01
     
    .02
     
    .01
     
    .04
     
    .02
     
    .38
     
     
  
  
    WM (digit span backward)
    t_b
     
    .14
     
    .18
     
    .14
     
    .19
     
    .22
     
    .21
     
    .20
     
    .24
     
    .20
     
    .28
     
    .20
     
    48
  
  
    p
     
    .77
     
    .50
     
    .74
     
    .46
     
    .32
     
    .37
     
    .42
     
    .26
     
    .40
     
    .16
     
    .39
     
     
  
  
    Problem solving (matrices)
    t_b
     
    .10
     
    .12
     
    .07
     
    .19
     
    .13
     
    .18
     
    .18
     
    .16
     
    .16
     
    .13
     
    .02
     
    .28
  
  
    p
     
    .99
     
    .85
     
    .9
     
    .42
     
    .79
     
    .46
     
    .46
     
    .61
     
    .58
     
    .81
     
    .99
     
     
  
  
    Age
    t_b
    -
    .34^*
    -
    .33^*
    -
    .29^*
    -
    .37^**
    -
    .35^**
    -
    .39^**
    -
    .36^**
    -
    .44^**
    -
    .30^*
    -
    .34^*
    -
    .26^* 
    -
    .46
  
  
    p
     
    .01
     
    .01
     
    .02
     
    <.01
     
    <.01
     
    <.01
     
    <.01
     
    <.01
     
    .02
     
    .01
     
    .04
     
     
  
  
    Years of education
    t_b
    -
    .28
    -
    .32
    -
    .26
    -
    .33
    -
    .35
    -
    .23
    -
    .28
    -
    .26
    -
    .33
    -
    .34
    -
    .33
    -
    .04
  
  
    p
     
    .96
     
    .98
     
    .95
     
    .98
     
    .99
     
    .93
     
    .96
     
    .95
     
    .98
     
    .99
     
    .98
     
     
  
  
    Openness for experience
    t_b
     
    .05
     
    .02
     
    .08
     
    .12
     
    .05
     
    .08
     
    .09
     
    .01
     
    .03
    -
    .03
    - 
    .12 
     
    .14
  
  
    p
     
    .37
     
    .44
     
    .30
     
    .22
     
    .39
     
    .30
     
    .27
     
    .48
     
    .41
     
    .58
     
    .80
     
     
  
  
    Neuroticism
    t_b
    -
    .08
    -
    .07
    -
    .09
    -
    .10
    -
    .08
    -
    .09
    -
    .09
    -
    .04
    -
    .07
     
    .01
     
    .004
    -
    .31
  
  
    p
     
    .30
     
    .33
     
    .28
     
    .24
     
    .30
     
    .28
     
    .28
     
    .60
     
    .32
     
    .52
     
    .51
     
     
  
  
    Note. CR, cognitive reserve; d2, d2 test of attention; CI level, upper    (attention, STM, WM, problem solving, years of education, Openness for    experience) or lower (age, Neuroticism) 90% bootstrap confidence level for    correlations with cognitive reserve; * p < .05; ** p < .01; p-values for cognitive status    (attention, STM, WM problem solving) are Bonferroni adjusted.



Table 4: Kendall Tau-b rank correlations (one-tailed) with DSST trials 1-10 and cognitive reserve in controls (N = 24).

Group differences in CR, first DSST and cognitive status

To examine group differences between patients and controls in CR, DSST1 and attention, memory and problem solving abilities, Mann–Whitney U tests were calculated. The results are depicted in table (5). Patients increased their performance on average by 24 symbols from baseline to maximum (SD = 11.34), controls by 23 symbols (SD = 7.79). The calculation of CR scores revealed a slightly lower performance of patients (M_CR controls – M_CR patients = 0.03) when accounting for the higher baseline performance in controls (M_DSST1 patients = 51.35; M_DSST1 controls = 56.63). Mann–Whitney U Tests showed that this small group difference in CR is not significant. No further group differences were found (all p >.05) (Table 5).

Table 5: Descriptive statistics of weight gain by drugs.



  
    Variable
    Control subjects (N = 24)
     
    Depressed patients (N = 40)
    Mann-Whitney Test
  
  
    M
    Mdn 
    SD
     
    M
    Mdn
    SD
     
    U
    p
  
  
    CR
    1.36
    1.33
    0.14
     
    1.33
    1.29
    0.18
     
    406.0
    .31
  
  
    DSST1
    56.63
    60
    15.08
     
    51.35
    52
    14.74
     
    368.0
    .12
  
  
    Attention 

      (d2)
    167.58
    164
    33.09
     
    150.51
    156
    42.45
     
    362.0
    .56
  
  
    STM 

      (digit span forward)
    8.17
    8
    1.99
     
    7.78
    8
    1.92
     
    420.5
    >.99
  
  
    WM 

      (digit span backward)
    7.50
    7
    2.19
     
    6.95
    7
    1.80
     
    438.5
    >.99
  
  
    Problem solving (matrices)
    20.00
    20.5
    3.73
     
    18.15
    20
    4.59
     
    354.5
    .32
  
  
    Note. d2, d2 test of attention; DSST1, Digit Symbol Substitution Test trial    1; CR, cognitive reserve; p-values    for cognitive status (attention, STM, WM problem solving) are Bonferroni    adjusted.



Table 5: Descriptive statistics of weight gain by drugs.

To examine CR in patients with and without clinical relevant cognitive deficits, the whole group of depressed patients was stratified by their performance in a routinely administered test battery. A cognitive deficit was defined by a score of 1.5 SD below the mean of the age specific norm population in at least one cognitive test. Deficits were present in 19 (47.5%, Ddef) vs. 21 (52.5%, D) patients. As figure (2) depicts, with the exception of problem solving, deficits were found in all assessed cognitive domains. D and Ddef did not differ in age and education and there were no significant differences in the distribution of gender, medication or diagnosis at discharge (all p > .05). However, Ddef showed a marginal higher severity of depression (M = 22.68, SD = 11.37, Mdn = 24) than D (M = 17.10, SD = 9.21, Mdn = 16; Mann–Whitney U = 142.00, p = .06) (Figure 2).

Figure 2: Percentage of depressed patients with impaired performance in attention (including d2 test of attention), memory (including short term memory, STM, and working memory, WM), and executive functions (including problem solving).



Figure 2: Percentage of depressed patients with impaired performance in attention (including d2 test of attention), memory (including short term memory, STM, and working memory, WM), and executive functions (including  problem solving).

Figure (3) shows the CR score, performance in the first DSST trail and attention, STM, WM and problem solving in controls, D and Ddef, respectively. In the first DSST trial, Ddef showed a numerically lower performance than both, D and controls. In the course of training Ddef increased their performance on average by 16 symbols (SD = 6.93) from baseline to maximum, D by 31 symbols (SD = 9.78), and controls by 23 symbols (SD = 7.79). Kruskal– Wallis–Tests revealed significant group differences in DSST1 and CR (DSST1: H (2) = 11.57, p = .003; CR: H (2) = 26.45, p < .001). Post– Hoc Mann–Whitney Tests showed that this difference was due to a significant lower performance in Ddef than in both, D and control subjects in DSST1 (Ddef vs. D: Mann–Whitney U = 82.5, p = .002; Ddef vs. controls: Mann–Whitney U = 114.00 p = .005) and in CR (Ddef vs. D: Mann–Whitney U = 24.00, p < .001; Ddef vs. controls: Mann–Whitney U = 72.00, p < .001), respectively. Differences in DSST1 between D and control subjects were not significant (Mann–Whitney U = 250.00, p = .960). Surprisingly, there was a marginal higher level in CR in D than in control subjects (Mann– Whitney U = 170.00, p = .062). We reanalyzed the group differences between D and Ddef using regression residuals controlling for depression severity. This confirmed the above pattern showing that Ddef reached lower scores in DSST1 (Mann–Whitney U = 91.5, p = .003) and in CR (Mann–Whitney U = 44.00, p < .001).

As figure (3) depicts, control subjects and D showed a numerically comparable cognitive status with an almost identical performance in attention. Performance in cognitive status is numerically lower in Ddef than in both, controls and D. However, Kruskal–Wallis–Tests revealed that differences in attention (H (2) = 7.29, p = .105), STM (H (2) = 8.16, p = .068), WM (H (2) = 4.01, p = .537), and problem olving (H (2) = 6.88, p = .129) did not reach significance when controlling for the four statistical tests.

Influence of medication

To identify a possible bias due to the use of sedative medication, we explored the association between medication and CR, attention, memory, and problem solving by comparing cognition in patients using sedative (n = 15) or nonsedative drugs (n = 22). Mann– Whitney U tests using regression residuals of the cognitive tests scores corrected for age, education, and severity of depression revealed no significant differences in CR and other cognitive scores (all p > .05)(Figure 3).

Download PDF

Citation: Coloma Andrews LC, Zihl J. Cognitive Reserve in Major Depression-Associations with Cognitive Status, Age, Education, Personality, and Depression Severity. Austin J Psychiatry Behav Sci. 2014;1(3): 1015. ISSN: 2381-9006.

Instruction for Authors

Submit Your Article