Elsevier

Psychoneuroendocrinology

Volume 38, Issue 8, August 2013, Pages 1388-1396
Psychoneuroendocrinology

When we test, do we stress? Impact of the testing environment on cortisol secretion and memory performance in older adults

https://doi.org/10.1016/j.psyneuen.2012.12.004Get rights and content

Abstract

Context

The majority of studies find that older adults have worse memory performance than young adults. However, contextual features in the testing environment may be perceived as stressful by older adults, increasing their stress hormone levels. Given the evidence that older adults are highly sensitive to the effects of stress hormones (cortisol) on memory performance, it is postulated that a stressful testing environment in older adults can lead to an acute stress response and to memory impairments.

Objective

The current study compared salivary cortisol levels and memory performance in young and older adults tested in environments manipulated to be stressful (unfavourable condition) or not stressful (favourable condition) for each age group.

Methods

28 young adults and 32 older adults were tested in two testing conditions: (1) a condition favouring young adults (constructed to be less stressful for young adults), and (2) a condition favouring older adults (constructed to be less stressful for older adults). The main outcome measure was salivary cortisol levels. Additionally, immediate and delayed memory performances were assessed during each condition.

Results

In older adults only, we found significantly high cortisol levels and low memory performance in the condition favouring young adults. In contrast, cortisol levels were lower and memory performance was better when older adults were tested in conditions favouring them. There was no effect of testing condition in young adults.

Conclusions

The results demonstrate that older adults’ memory performance is highly sensitive to the testing environment. These findings have important implications for both research and clinical settings in which older adults are tested for memory performance.

Introduction

Research suggests that older adults perform worse than young adults on declarative memory tasks (Zacks et al., 1999, Balota et al., 2000). While these deficits are presumed to be due to age-related cognitive decline, contextual features of certain testing environments might be inherently stressful for older adults and artificially result in observed cognitive impairments. In particular, the testing environment, or procedures used therein, may induce an acute stress response in older individuals, leading to acute memory impairments.

The stress hormone cortisol has an important impact on cognition as it crosses the blood–brain barrier and binds to receptors in the hippocampus, amygdala and frontal lobes, each of which are involved in cognitive processes. These regions also regulate cortisol levels by contextualizing relevant environmental stimuli in acutely stressful situations that can impair memory for both young (de Quervain et al., 2000, Lupien et al., 2002a) and older adults (Lupien et al., 1998, Lee et al., 2007). However, older adults’ memory performance is particularly vulnerable to the impact of cortisol when compared to young adults (Seeman et al., 1997, Lee et al., 2007). Indeed, stress-induction paradigms that manipulate cortisol levels (pharmacologically or psychologically) increase forgetting in older adults compared to young adults (Lupien et al., 1997, Lupien et al., 2002b).

Psychosocial researchers have identified four situational determinants that elicit a physiological stress response, namely; novelty, unpredictability, low sense of control and threat to one's ego (Mason, 1968, Dickerson and Kemeny, 2004). In laboratory stress-induction paradigms, such as in the Trier social stress test (TSST) (Kirschbaum et al., 1993), these situational determinants are implemented in order to trigger activity of the hypothalamic–pituitary–adrenal (HPA) axis and consequent surges in cortisol.

In some studies on cognitive ageing that compare memory performance of young and older adults, these situational determinants of the physiological stress response can be imbalanced and unfavourably be more distressing for older adults compared to young (Lupien et al., 2007). Various study design factors dissociate the level to which these situational determinants may be differentially experienced by young and older adults: (1) testing location, (2) age of the research assistant, (3) time of testing, (4) type of memory task and (5) task instructions.

First, many studies on memory performance in young and older adults take place on a university campus, which is familiar to young adults who are often attending as students at the time of the study. However, older adults are faced with the challenge of navigating to this unfamiliar destination. This involves greater novelty and unpredictability for older adults, leading to a more unfavourable (or stressful) testing condition upon arrival. This increased novelty in older adults could trigger a greater stress response in this population. Second, research assistants conducting the studies on ageing and memory are generally young graduate students. Consequently, young participants are tested by their peers, whereas much younger individuals test older adults. This can modify the quality of the relationship between participant and research assistant and/or lead to greater threat to the ego in the older adults and consequently to an increase in cortisol levels.

Third, it is reported that studies on memory performance tend to schedule participants’ appointments in the afternoon (May et al., 1993). Yet, young adults perform significantly better when tested in the afternoon (between 1300 h and 1700 h), while older adults perform significantly better when tested in the morning (between 0800 h and 1100 h) (May et al., 1993, Intons-Peterson et al., 1998, Intons-Peterson et al., 1999, May and Hasher, 1998, Winocur and Hasher, 2004). Studies generally report that when older adults are tested in the afternoon, their performance is significantly worse than young adults, whereas when they are tested in the morning, their performance does not differ (Borella et al., 2011). When the aforementioned studies report that they test participants in the morning, the time at which cognitive testing starts and ends tends to vary across studies; starting time may be anywhere between 0800 h and 0900 h and may end between 1000 h and 1200 h. Therefore, the exact time window of optimal testing time is presently difficult to determine, yet the advantage of morning testing time is found when cognitive testing ends as late as noon (Borella et al., 2011). Episodic memory, working memory and in particular tasks requiring inhibitory processes are influenced by the time of testing (May and Hasher, 1998, West et al., 2002, Yang et al., 2007). By contrast, tasks assessing semantic knowledge including verbal fluency and vocabulary are less sensitive and remain consistent throughout the day (May and Hasher, 1998, Brown et al., 1999, Martin et al., 2008). Taken together, the evidence suggests that older adults are at a disadvantage compared to young adults when tested during their non-optimal testing time.

Fourth, participants are typically instructed to memorize items (word lists; Chantome et al., 1999 or paragraphs; Foster et al., 1999) for memory tasks. Learning and memorizing are routine tasks for university students, but not for older adults. Memorizing may thus be perceived as novel and unpredictable for older adults, resulting in a greater physiological stress response in this population. Finally, it has been demonstrated that task instructions may also impact older adults’ memory performance. Instructions tend to include several reminders that the task at hand aims to measure memory capacities, where participants are expected to memorize in order to be tested later (Rahhal et al., 2001). Studies have demonstrated that when instructions for memory tasks are modified to decrease the emphasis on the memory component of the task, previously observed differences between older and young adults are abolished (Hasher et al., 1999, Rahhal et al., 2001).

Although many researchers in the field of ageing and memory are aware of these issues and – for example – schedule testing times according to individual time preferences or include familiarization phases in the testing protocol, we have previously suggested that in some studies assessing the relationship between age and memory performance, older adults are commonly tested in unfavourable conditions compared to young individuals, which may lead to stress-induced memory impairments in older adults (Lupien et al., 2007). In these studies, the cognitive impairments observed in older adults may be interpreted as a consequence of ageing, when some proportion of the memory impairment could be due to the inadvertent product of a stressful testing environment. While the argument may be made that older adults display declines in memory function due to age-related elevated basal cortisol levels, it has been demonstrated that healthy older adults do not show higher cortisol levels compared to young adults under basal non-stressful conditions in their home environment (Waltman et al., 1991, Lupien et al., 1994). Thus, elevations in cortisol may be confined to the testing environment, leading to impairment in memory performance among older adults.

The primary goal of this study was to examine the effect of the testing environment on cortisol secretion in young and older adults. The secondary goal was to assess whether changes in cortisol associated with different testing environments influence memory performance in both age groups. Participants were tested in experimental settings that were created to be favourable (i.e. leading to a low stress response) or unfavourable (i.e. leading to a high stress response) for each respective age group (i.e. young adults versus older adults). To ensure that the unfavourable testing environment created a stress response in each age group, home-based basal cortisol levels were measured for comparisons. Home-based sampling further enabled this study to replicate previous findings that older adults do not secrete higher basal stress hormone levels compared to their younger counterparts. In addition, due to the potential influence of circadian acrophase on arousal and cognition (May and Hasher, 1998, West et al., 2002, Winocur and Hasher, 2004, Martin et al., 2008), awakening time of both age groups were compared, and the time elapsed between awakening and testing time was calculated for each testing environment. It was hypothesized that age-appropriate unfavourable testing conditions would increase cortisol secretion and decrease memory performance in both age groups but that older adults would be more sensitive than young adults to these environmental manipulations.

Section snippets

Participants

Measures were obtained from a sample of 32 healthy older adults (14 men, 18 women) between the ages of 60 and 75 (M = 67.4, S.D. = 4.64) and 28 young adults (9 men, 19 women) between the ages of 18 and 35 (M = 24.4, S.D. = 4.60). Older adults were recruited through local newspaper advertisements, while young adults were recruited through McGill University's classified advertising (Montreal, Canada). Among older adults, the group mean years of education was 14.40 years (S.D. = 3.16), whereas young adults

Results

Preliminary analyses performed on time of awakening showed that older adults woke up significantly earlier (7:14 ± 0.2 am) than young adults (8:59 ± 0.23 am; t(57) = 28.89; p < .0001). Given that young and older adults did not wake up at similar times, it was important to see whether the time that elapsed between waking time and testing time for each subject was different across groups. The analysis performed in the Favouring-Young (PM testing) condition showed that there was no significant difference

Discussion

Supporting our hypothesis of the impact of stressful testing environments on both cortisol levels and memory performance in older adults, the present study shows that when older adults are tested in an environment that is unfavourable for their age group, they present increased cortisol levels (compared to home-based basal levels) and a steep forgetting rate in memory performance, as shown by the decrease in delayed recall. However, when older adults are tested in an environment that is

Role of funding source

This study was supported by a grant from the Canadian Institutes for Health Research (Grant #77788) and by a Senior Investigator Chair from the Canadian Institutes of Health Research, Institute of Gender and Health to S.J.L. S.S is supported by a Frederick Banting and Charles Best Canada Graduate Doctoral Scholarship Awarded by the Canadian Institutes of Health Research.

Conflict of interest statement

The authors unequivocally state no conflict of interests.

Acknowledgement

The authors would also like to thank Helen Findlay at the Centre for Studies on Human Stress laboratory for conducting the cortisol assays.

References (37)

  • D.J. de Quervain et al.

    Acute cortisone administration impairs retrieval of long-term declarative memory in humans

    Nat. Neurosci.

    (2000)
  • S.S. Dickerson et al.

    Acute stressors and cortisol responses: a theoretical integration and synthesis of laboratory research

    Psychol. Bull.

    (2004)
  • J.K. Foster et al.

    The hippocampus and delayed recall: bigger is not necessarily better?

    Memory

    (1999)
  • H.H. Fung et al.

    Sending memorable messages to the old: age differences in preferences and memory for advertisements

    J. Pers. Soc. Psychol.

    (2003)
  • M. Grundman et al.

    Hippocampal volume is associated with memory but not monmemory cognitive performance in patients with mild cognitive impairment

    J. Mol. Neurosci.

    (2003)
  • L. Hasher et al.

    Timing, instructions, and inhibitory control: some missing factors in the age and memory debate

    Gerontology

    (1999)
  • M.J. Intons-Peterson et al.

    Aging, optimal testing times, and negative priming

    J. Exp. Psychol. Learn.

    (1998)
  • M.J. Intons-Peterson et al.

    Age, testing at preferred or nonpreferred times (testing optimality), and false memory

    J. Exp. Psychol. Learn. Mem. Cogn.

    (1999)
  • Cited by (56)

    • Personality traits affect older adults' memory differently depending on the environmental support provided at encoding

      2022, Personality and Individual Differences
      Citation Excerpt :

      Thus, environmental support can correspond to any experimental situation that promotes memorization. For example, Sindi et al. (2013) showed that when older adults performed a memory task under conditions that were favorable to them (the experimenter was the same age as the participants, instructions did not mention a memory task to avoid stereotype threat, experimentation was conducted in a familiar place and in the morning, which is the preferred time of day of older adults), they were more successful than when they performed the task in more unfavorable conditions. To create conditions of environmental support, we provided older adults with prior task success (Experiment 1) and encouraged effective study strategies using generation versus reading (Experiment 2).

    • Does older adults’ cognition particularly suffer from stress? A systematic review of acute stress effects on cognition in older age

      2022, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      Thus, older adults cannot, and should not, be considered as a homogenous population. As a result, cognitive decline, stereotype threat, testing environment, age differences in stress appraisal and regulation are all potential confounders of stress effects on cognition in the absence of a validated stressor and if not properly controlled for (e.g., Sindi et al., 2013; Zandara et al., 2016; Zuber et al., 2019). In this regard, we would like to underline the importance of having an age-adapted stressor.

    • The Applied Implications of Age-Based Stereotype Threat for Older Adults

      2020, Journal of Applied Research in Memory and Cognition
    View all citing articles on Scopus
    View full text