Advanced Circadian Timing and Sleep Fragmentation Differentially Impact on Memory Complaint Subtype in Subjective Cognitive Decline

Abstract

Background:

Increased sleep fragmentation and advanced circadian timing are hallmark phenotypes associated with increased age-related cognitive decline. Subjective cognitive decline (SCD) is considered a prodromal stage of neurodegeneration and dementia; however, little is known about how sleep and circadian timing impact on memory complaints in SCD.

Objective:

To determine how sleep and circadian timing impact on memory complaint subtypes in older adults with SCD.

Methods:

Twenty-five older adults with SCD (mean age = 69.97, SD = 5.33) completed the Memory Functioning Questionnaire to characterize their memory complaints. They also underwent neuropsychological assessment, and completed 1 week of at-home monitoring of sleep with actigraphy and sleep diaries. This was followed by a two-night laboratory visit with overnight polysomnography and a dim light melatonin onset assessment to measure circadian timing.

Results:

Advanced circadian timing was associated with greater memory complaints, specifically poorer memory of past events (r = –0.688, p = 0.002), greater perceived decline over time (r = –0.568, p = 0.022), and increased reliance on mnemonic tools (r = –0.657, p = 0.004). Increased sleep fragmentation was associated with reduced self-reported memory decline (r = 0.529, p = 0.014), and reduced concern about everyday forgetfulness (r = 0.435, p = 0.038).

Conclusion:

Advanced circadian timing was associated with a number of subjective memory complaints and symptoms. By contrast, sleep fragmentation was linked to lowered perceptions of cognitive decline, and less concern about memory failures. As circadian disruption is apparent in both MCI and Alzheimer’s disease, and plays a key role in cognitive function, our findings further support a circadian intervention as a potential therapeutic tool for cognitive decline.

Keywords

Circadian rhythms cognition dementia risk dim light melatonin onset mild cognitive impairment sleep disturbance subjective memory impairment

INTRODUCTION

Approximately 47 million people are living with dementia worldwide [1], which is projected to increase to 131.5 million by 2050. With no cure on the immediate horizon, there is a critical need to identify modifiable targets for improving cognitive health in later life, ideally implemented at early, preclinical phases to delay progression to dementia [2]. Given the abundance of evidence describing the importance of sleep and optimal circadian timing for late life cognitive health, these factors are of increasing scientific and clinical interest for early intervention [3 –5]. However, in order to develop efficacious interventions targeting sleep, it is first important to understand the role that sleep and circadian timing each play in the early stages of cognitive dysfunction.

Changes in the timing and structure of sleep occur across the lifespan. Increased sleep fragmentation and reductions in slow wave sleep (SWS) represent the hallmark signs of age-related changes in sleep [5]. Moreover, these changes are exacerbated in mild cognitive impairment (MCI) and Alzheimer’s disease (AD) (see [4] for a review). Importantly, sleep and cognition do not simply show parallel age-related declines, but instead sleep is likely to play a causal role in cognition, cognitive decline, and dementia progression [6 –9]. Prospective studies have shown that reduced subjective sleep quality, increased sleep fragmentation, and increased sleep disordered breathing are associated with greater cognitive decline, and increased incidence of dementia [10 –13]. From a physiological perspective, poorer sleep quality has been linked to increased amyloid-β (Aβ) accumulation in the brain, and reduced Aβ in the cerebrospinal fluid [14, 15]. While prefrontal atrophy and Aβ are both associated with memory dysfunction, some evidence suggests that this association is not direct, but instead mediated by (slow wave) sleep [16, 17].

The internal biological clock is integral to ensuring optimal timing and structure of sleep, through the generation of approximately 24 hour “circadian” rhythms which are regulated by the suprachiasmatic nucleus (SCN) in the hypothalamus [18]. The internal body clock can be measured through output signals such as core body temperature (CBT, e.g. the CBT nadir at ∼5 am [19]) and the hormone melatonin (e.g., the rise in melatonin in the evening in the absence of light, termed “dim light melatonin onset” [DLMO]). There are a number of age-related changes that occur to the biological clock and its outputs, including decreased volume and cell number within the SCN, decreased amplitude of CBT, and advanced timing of circadian phase (e.g., CBT nadir and DLMO occurring at an earlier clock time) (see [20] for a review). The advance in circadian timing relative to clock time is the main circadian-related change observed in older adults [19 , 22]. This advance is exacerbated in people with MCI, who show a circadian advance compared to age-matched healthy controls, with both an earlier DLMO and habitual bed time [23]. While there are a number of challenges in accurately measuring circadian rhythms in people with dementia, clear changes in circadian rhythmicity have been reported (e.g., decreased amplitude in plasma melatonin, and disrupted melatonin secretion across the 24-hour day - see [24]). Behaviorally, MCI and AD patients show disturbed rest-activity rhythms, which correlate with greater memory and cognitive disruption, increased risk of conversion, and postmortem neurodegeneration in the SCN [25 –27]. While rest-activity rhythms do not represent a pure measure of the circadian clock, these findings suggest that disrupted circadian rhythmicity is an important risk factor for cognitive decline.

Early detection of risk factors for dementia is necessary to facilitate the most effective intervention strategies [28]. While the MCI period is an optimal phase for secondary prevention strategies [29], it may be necessary to target the cognitive-promoting sleep and circadian system before the onset of any cognitive decline, particularly since it is now well established that neurodegenerative changes leading to AD are occurring in the brain at least 10–20 years before MCI [30]. Subjective cognitive decline (SCD), often referred to as subjective memory decline, or subjective memory impairment, refers to an individual who subjectively perceives worsening of cognitive function in the absence of objectively identifiable cognitive deficits [31]. Individuals with SCD are two times more likely to develop dementia than those without SCD [32], and demonstrate cortical thinning, amyloid burden, and grey matter changes consistent with AD pathology [33 –36]. Individuals reporting SCD therefore represent an important group for early intervention strategies. SCD has been associated with increased daytime sleepiness [37], greater subjective sleep complaints [38 –40], and actigraphic differences in sleep efficiency and fragmentation compared to controls. Objective changes to sleep, however, have been inconclusive: for instance, while some have reported greater SCD was associated with improved sleep efficiency and reduced sleep fragmentation [41], others have reported that participants with SCD had lower sleep efficiency and higher wakefulness than those without SCD [42].

The lack of consistency between sleep and SCD may relate to the way in which SCD is measured. The construct of memory is multidimensional, thus SCD may encompass a wide range of changes or impairments. For instance, qualitative work suggests that complaints can be divided into sub-themes such as retrospective functioning, everyday cognitive difficulties, and changes in insight, function, and identity [43]. Further work into complaint subtype may help elucidate some of the inconsistencies in the SCD literature, including the inconsistent association between sleep and SCD.

To address these gaps, the current study aimed to characterize sleep architecture and circadian timing for the first time in participants with SCD. We further aimed to examine how sleep and circadian characteristics is associated with subtypes of subjective memory complaints. In line with previous work in MCI [23], it was hypothesized that advanced circadian timing would be associated with more pronounced subjective memory complaints. Despite conflicting results in SCD samples, it was also hypothesized that reduced SWS, and increased sleep fragmentation would be associated with greater subjective memory complaints.

METHOD

Participants

Twenty-five older adults aged 60–79 (M = 69.97, SD = 5.33) with SCD were recruited from the community (18 females). SCD was defined as a self-reported decline in memory in the past five years beyond that expected of normal aging. To be eligible, participants must not have had any diagnosis of major or minor neurocognitive disorder, neurological or sleep disorders; current psychiatric diagnoses; prior stroke or head injury; inadequate English fluency for neuropsychological assessment; current use of beta-blockers, ACE inhibitors, or sedative hypnotics; chemotherapy or radiation in the past 12 months; general anesthetics, shift work, or trans meridian travel (>2 time zones) in the past 3 months; and consumption of >14 standard alcoholic drinks per week. In addition, participants were excluded if they exhibited objective cognitive deficits upon neuropsychological testing (see below), or presented with undiagnosed sleep apnea (AHI >20 [44]) or periodic limb movements with associated cortical arousals (PLMA >10 [45, 46]).

Procedure

All participants underwent a neuropsychological assessment to confirm eligibility and measure baseline cognition. Participants were then issued activity monitors (Actiwatch-2, Philips Respironics, USA) and sleep diaries, and completed 1 week of at-home actigraphy monitoring of habitual sleep/wake habits. Following at home monitoring, participants were admitted to the laboratory for a 2-night sleep and circadian assessment. On both nights, participants were given an 8-hour sleep opportunity beginning at their habitual bed time. Participants were monitored with electroencephalography (EEG) to record nocturnal sleep. Night one served as a sleep disorder screening and adaptation night. On evening two, participants were kept in dim light (<10 lux), provided saliva samples for circadian analysis, and had their sleep recorded for analysis.

Neuropsychological assessment

The Memory Functioning Questionnaire (MFQ) [47] was used to assess subjective memory complaints. The MFQ contains 64 items rated on a 7-point Likert scale, with higher scores representing better memory, or less perceived difficulties. The MFQ was split into five sub-scales: Retrospective Functioning (5 items) quantifies current memory performance compared to the past (e.g., how is your memory compared to 5 years ago); Frequency of Forgetting (28 items) presents 28 everyday scenarios and asks how often they are a problem (i.e., remembering names, faces, book chapters etc.); Seriousness of Forgetting (18 items) asks how serious of a problem a memory failure in 18 scenarios is, when it occurs; Remembering Past Events (4 items) asks how well life events are remembered (e.g., how well do you remember things that happened ten years ago); and Mnemonics Usage (8 items) asks how often participants rely on various techniques to remember things, such as keeping an appointment book.

Participants completed a neuropsychological test battery to establish objective cognitive performance and to ensure eligibility for SCD. The following tests were administered: Mini-Mental State Examination –Second Edition (MMSE) [48] was used to assess global cognition and for descriptive purposes; The National Adult Reading Test (NART) [49] assessed premorbid intellectual functioning; the Comprehensive Trail Making Test (CTMT) [50] assessed processing speed and set-shifting; and the California Verbal Learning Test 2 (CVLT) assessed verbal learning and memory [51]. Objective cognitive deficits were defined as an MMSE <24, a CTMT Composite Score less than 35 (representing 1.5 SD from age-matched norms), or a CVLT Composite Score (Z score) of less than –1.5. CVLT Composite was calculated as the average scaled scores on Trial 5 Total Recall, and Short and Long Delay Free and Cued Recall, representing an overall memory deficit of more than 1.5 SD below age- and sex-matched norms.

Participants also completed a range of questionnaires to ascertain sleep complaints and psychiatric symptoms. The Geriatric Depression Scale (GDS) [52] was used to assess depressive symptoms. The Depression Anxiety & Stress Scale (DASS-42) [53] was also used to assess depressive symptoms, anxiety symptoms, and stress levels. Daytime sleepiness was assessed using the Epworth Sleepiness Scale (ESS) [54], and subjective sleep quality was assessed with the Pittsburgh Sleep Quality Index (PSQI) [55]. Diurnal preference was assessed using the Horne-Ostberg Morningness-Eveningness Questionnaire (MEQ) [56]. Medical history and a list of current medications was also provided by each participant.

At-home sleep monitoring

For one week prior to laboratory admission, participants’ habitual sleep/wake habits were monitored with wrist-worn activity monitors (Actiwatch-2, Philips Respironics, USA) and sleep diaries. Habitual bed and wake time were manually calculated using a combination of actigraphy and sleep diaries.

Overnight sleep monitoring with polysomnography (PSG)

Nocturnal PSG recordings were recorded on two consecutive nights in the laboratory (Compumedics Grael, Melbourne, Vic, Australia) using a bilateral 18–channel electroencephalography (EEG) montage (Fp1, Fp2, F3, F4, C3, C4, P3, P4, Po3, Po4, P7, P8, O1, O2, Fz, Cz, Pz, Oz) with all electrodes referenced to the contralateral mastoid (i.e., C3-M2; C4-M1). Two electrooculographic channels (left and right outer canthi); and three electromyographic (sub-mentalis) channels were used to score eye movements and muscle activity. EEG data were sampled at 510 Hz.

In addition to EEG, EOG, and EMG, thoracic and abdominal respiratory effort, airflow (via nasal-oral thermocouple and nasal airflow recording), finger pulse oximetry, body position, and bilateral leg movements were monitored on night one for diagnostic purposes. All sleep data were scored by certified scorers. Sleep stages, arousals, apneas and hypopneas, and periodic limb movements (PLMs) were scored according to current AASM criteria [44]. Sleep EEG on night two was scored for sleep staging and arousals according to AASM criteria [44]. Sleep onset latency (min), total sleep time (min), minutes spent in each stage, and wake after sleep onset (min) were all calculated.

Salivary melatonin samples

Salivary melatonin was assessed to determine DLMO. Participants were kept in dim light (<10 lux) and hourly saliva samples (Salivette, Sarstedt, Germany) were collected from five hours prior to, and up until habitual bed time. Participants were required to remain seated and refrain from consuming any food or liquid for 15 min prior to each sample. Samples were immediately centrifuged and frozen at –20°C.

Melatonin was assayed in 200μL duplicate saliva samples by double antibody RIA (Buhlmann Laboratories, Switzerland) using previously published methods [57]. The lowest detectable level of melatonin was 4.3 pmol/ml. DLMO was calculated using linear interpolation between samples, and was defined as the time when the salivary melatonin level crossed and remained above 10 pmol/ml [58]. For participants whose melatonin samples did not cross the 10 pmol/ml threshold, a relative threshold of +2 S.D. of the mean of the first 3 samples for that participant was calculated [57]. The relationship between the time of melatonin onset and habitual bed time (HBT), referred to as the phase angle of entrainment, was calculated by subtracting DLMO time from HBT.

Data analysis

Data was obtained from 25 participants. Of this, data loss was minimal: actigraphy data was lost for N = 3 (12.4%) participants due to device failure. As sleep diary- and actigraphy-determined bed and wake times were highly correlated (r = 0.955, p < 0.001), habitual bed and wake time was calculated from the sleep diary for these participants. Polysomnography was lost for N = 2 participants due to technical difficulties. DLMO was calculated for N = 18, with no melatonin detected within the sensitivity of the assay for N = 7. DLMO may have occurred after bedtime for these participants. There was no significant difference in age, years of education, MMSE score, proportion of males to females or percentage of participants on medication between the 18 participants with a calculated DLMO, and the N = 7 without. Participants with no DLMO data had near-significantly less depressive symptoms (t(23) = 2.05, p = 0.052) and stress symptoms (t(23) = 2.05, p = 0.052) on the DASS compared to those with DLMO data. There was no difference in anxiety symptoms or depressive symptoms measured by the GDS. One participant did not complete the mnemonics scale for the MFQ.

Data was analyzed in SPSS V23. Non-normally distributed variables were log-transformed prior to analysis. Pearson’s correlations were conducted to examine relationships between subjective memory complaints, demographics, psychiatric outcomes, and sleep and circadian timing outcomes. Partial correlations between memory complaints and sleep and circadian timing outcomes were conducted where a demographic or neuropsychiatric outcome had a significant relationship with either variable. Where there was missing data, analysis was conducted on the available data only.

RESULTS

Participant demographics

Participants were aged between 60 and 79, with a mean age of 69.97 years (S.D. = 5.33). Seventy-two percent of participants were female, and participants were highly educated, with on average 13.68 years of education, and an estimated IQ of 118.49 (NART). No participants had any objective memory deficits: mean MMSE score was 28.1, with no participants scoring below 24. Verbal memory was intact, with the group scoring 1 SD above norms, and no individuals scoring <1 SD below norms. Processing speed was also intact, with the sample average equal to age-matched norms (CTMT Composite M = 49.61). One participant scored slightly below norms on the CTMT; however, as they were within expected normative ranges on all other tests, they were left in the sample. Depressive symptomology was low, with 19 participants (79%) scoring below 10 on the GDS (normal) and 6 participants scoring between 10 and 20, suggesting mild depressive symptoms. Similar depressive scores were seen on the DASS. Anxiety and stress symptoms were also low, with 20 participants (83%) scoring within the “normal” range for each. Three participants (12%) had elevated scores on all three outcomes. Neuropsychiatric and neuropsychological scores are displayed in Table 1. Medication use was high, with N = 17 (68%) of participants regularly using at least 1 prescription medication. The most common medication type was statins, with N = 7 (28%) of participants reporting current use. Those taking statin medications did not differ from those who were not on any outcome (p > 0.366). Two participants (8%) reported current anti-depressant use. Two participants (8%) had a history of major depression, with no reports of a depressive episode in the past 10 years.

Table 1

Demographic, neuropsychiatric, and neuropsychological descriptive data

	Range	Mean	S.D.
Age (y)	62–79	69.97	5.33
Sex (m:f)	−	7:18	−
Education (y)	8–21	13.68	3.56
Body Mass Index (kg/m²)	21–32	26.01	3.57
Neuropsychiatric Data
Geriatric Depression Scale (/30)	0–15	5.84	4.10
DASS –Depression Scale	0–17	4.88	5.13
DASS –Anxiety Scale	0–17	3.64	5.04
DASS –Stress Scale	0–21	7.48	6.78
MFQ –Total (/448)	184–395	286.08	48.78
MFQ - Retrospective Functioning (/35)	8–34	15.28	6.34
MFQ - Remembering Past Events (/28)	13–28	19.56	4.86
MFQ - Frequency of Forgetting (/196)	78–177	139.68	25.06
MFQ - Seriousness of Forgetting (/126)	40–117	82.52	22.40
MFQ - Mnemonic Use (/56)	12–56	25.75	10.30
Neuropsychological Data
NART (Estimated IQ)	109–124	118.49	3.45
Mini-Mental State Examination (/30)	25–30	28.12	1.45
CTMT Composite Score (T score)	32–60	49.61	7.81
CVLT Composite Score (Z score)	−0.3–2.4	1.02	0.67
CTMT Trail 1 (s)	25–105	50.83	20.04
CTMT Trail 5 (s)	43–189	81.00	34.32
CVLT Total Learning (total words trial 1–5; /80)	44–73	58.68	7.27
CVLT Short Delay Free Recall (total words /16)	7–16	12.60	2.47
CVLT Long Delay Free Recall (total words /16)	7–16	13.40	2.20

MFQ, Memory Functioning Questionnaire; NART, National Adult Reading Test; CTMT, Comprehensive Trail Making Test; CVLT, California Verbal Learning Test.

Sleep and circadian timing outcomes

Participants scored themselves as either moderately morning type (50%) or “neither type” (50%). There were no evening or extreme morning types. Participants rated their sleep quality as poor overall, with 67% scoring >5 on the PSQI. Daytime sleepiness however was low, with only 17% scoring >10 on the ESS. Based on actigraphy, participants had an average habitual bed time of 22:54, and a wake time of 7:07. Participants spent 8.32 h in bed, with a total sleep time of 7.08 (sleep efficiency M = 85.14; SD = 6.22). On average, participants had a DLMO of 20:40, which was on average 2 h prior to HBT (see Fig. 1).

Fig.1

Habitual sleep timing and dim light melatonin onset (DLMO) at the group and individual level. Panel A shows group mean (▴) and SD, as well as individual DLMO times (∘) relative to clock time; Panel B shows mean and SD of DLMO (▴) and average sleep timing from habitual bed time (HBT) to habitual wake time (grey block); and Panel C shows group mean (▴) and SD, as well as individual phase angles (∘) between DLMO and HBT.

Based on polysomnographically determined sleep, participants slept, on average, 413 min (6 h 53 min), with an average sleep efficiency of 86.1% (S.D. = 6.61). Amount of time spent in SWS was highly variable, ranging from 0.5–187 min. All sleep and circadian parameters, including sleep architecture are summarized in Table 2.

Table 2

Sleep and circadian timing descriptive data

	Range	Mean	S.D.
Subjective Sleep Outcomes
Horne-Ostberg MEQ	43 –74	61.00	9.45
Pittsburgh Sleep Quality Index	0 –13	6.88	3.63
Epworth Sleepiness Scale	0 –17	6.60	3.70
Sleep and Circadian Timing
Habitual Bed Time (hh:mm)	21:29 –00:24	22:54	0:44
Habitual Wake Time (hh:mm)	5:48 –8:11	7:07	0:41
DLMO (hh:mm)	18:29 –22:59	20:55	1:20
Habitual Bed Time Phase Angle (h)	0.11 –3.77	2.14	1.09
Polysomnography Outcomes
Total Sleep Time (min)	341 –457	413.54	32.02
Sleep Onset Latency (min)	0.5 –27	11.35	6.29
WASO (min)	13 –115	55.22	28.79
NREM 1 (min)	17 –83	45.46	16.03
NREM 2 (min)	133 –257	187.96	38.15
NREM 3/SWS (min)	0.5 –187	83.52	48.26
REM (min)	58 –158	96.61	23.54
AHI	0.4 –19	7.70	5.37
PLM Index	0 –77	17.98	21.54

MEQ, Morningness-Eveningness Questionnaire; WASO, wake after sleep onset; AHI, Apnea-Hypopnea Index; PLM, Periodic Limb Movement; NREM, non-rapid eye movement; DLMO, dim light melatonin onset.

Sleep and circadian timing outcomes were not significantly correlated with age, estimated IQ, or depressive, anxiety, or stress symptoms. As expected, SWS (min) was correlated with sex (r = 0.52, p = 0.012) and years of education (r = 0.62, p = 0.002). No other correlations with covariates were found.

Association between subjective memory complaints and demographics, psychiatric symptoms, and neuropsychological performance

Subjective memory complaints as measured by the MFQ sub-scales were not correlated with age, sex, BMI, years of education, or anxiety symptoms measured by the DASS. Stress symptoms were significantly correlated with MFQ Total (r = –0.50, p = 0.011) and Frequency of Forgetting (r = –0.53, p = 0.006), such that higher stress was linked to more frequent forgetting in daily life. Retrospective Functioning was significantly correlated with Estimated IQ (r = –0.48, p = 0.016) and depressive symptoms on the GDS (r = –0.45, p = 0.026), such that higher IQ and depressive symptoms were related to worse perceived memory decline. Subjective complaints (total or subscales) were not correlated with objective performance on the MMSE or CTMT. CVLT Composite Score, however, was correlated with MFQ Total (r = 0.47, p = 0.026); and Remembering Past Events was correlated with CVLT Long Delay Recall (r = 0.46, p = 0.021; See Fig. 2).

Fig.2

Relationship between subjective memory complaints and objective memory performance on the CVLT. Panel A shows with increasing MFQ total score (less complaints), CVLT composite score increases (better memory performance). Panel B shows participants with less trouble remembering past events recalled more words during the long delay trial of the CVLT.

Association between subjective memory complaints, and sleep and circadian timing outcomes

Table 3 shows all relationships between subjective memory complaints and sleep and circadian timing outcomes. For MFQ Total, partial correlations controlling for stress symptoms revealed that greater overall memory complaints were associated with increased total sleep time (r = –0.51, p = 0.016), and reduced wake after sleep onset (r = 0.49, p = 0.02). Increased memory complaints overall were also associated with an earlier melatonin onset (DLMO), and a wider phase angle, although these did not reach statistical significance (p < 0.10).

Table 3

Correlation coefficients (r) for sleep and circadian timing outcomes and subjective memory as measured by the MFQ Subscales

	MFQ Total^†	MFQ Retrospective Functioning ^†	MFQ Past Events	MFQ Frequency^†	MFQ Seriousness	MFQ Mnemonics
Sleep and Circadian Timing
Habitual Bed Time (hh:mm)	0.24	0.26	0.23	0.24	0.08	0.01
Habitual Wake Time (hh:mm)	0.36	−0.01	0.05	0.16	0.43^*	0.16
DLMO (hh:mm)	0.47^∧	0.58^*	0.65^**	0.31	0.42^∧	0.51^*
Habitual Bed Time Phase Angle (h)	−0.42^∧	−0.57^*	−0.69^**	0.20	−0.48^*	−0.66^**
Polysomnography Outcomes
Total Sleep Time (min)	−0.51^*	−0.53^*	−0.19	−0.17	−0.48^*	−0.11
Sleep Onset Latency (min)	0.23	0.12	0.19	0.12	0.32	−0.22
WASO (min)	0.49^*	0.53^*	0.17	0.15	0.44^*	0.16
SWS (min)^†	−0.18	−0.36	−0.03	0.10	−0.23	0.10
REM (min)	−0.07	0.00	0.01	−0.07	−0.12	0.07

DLMO, dim light melatonin onset; WASO, wake after sleep onset; SWS, slow wave sleep; REM, rapid eye movement; ^†Covariates controlled; ^∧p < 0.10 ^*p < 0.05 ^**p < 0.01

For Retrospective Functioning, partial correlations controlling for Estimated IQ and depressive symptoms, revealed that greater memory decline compared to previous years was significantly associated with an earlier DLMO (r = 0.57, p = 0.019), a wider phase angle (r = –0.57, p = 0.022), greater total sleep time (r = –0.53, p = 0.014), and reduced WASO (r = 0.53, p = 0.014).

For Remembering Past Events, correlations showed that poorer ability to remember episodic memories was associated with an earlier DLMO (r = 0.65, p = 0.003), and a wider phase angle (r = –0.69, p = 0.002). No sleep outcomes were associated with remembering past events.

An increased Seriousness of Forgetting was associated with a wider phase angle (r = –0.48, p = 0.044), an earlier habitual wake time (r = 0.43, p = 0.031), longer total sleep time (r = 0.48, p = 0.021), and reduced WASO (r = 0.44, p = 0.038).

Finally, increased Mnemonics Use was associated with an earlier DLMO (r = 0.51, p = 0.037), and a wider phase angle (r = –0.66, p = 0.004). There were no significant correlations with Frequency of Forgetting. Figure 3 shows individual profiles relating to the timing of DLMO and the habitual sleep period ranked by memory complaint for each of the sub-scales. This highlights that as severity of memory complaint increased, DLMO advanced, resulting in a wider phase angle with bed time.

Fig.3

Relationships between Memory Functioning Questionnaire Total, and sub-scales, and melatonin onset (▾), habitual sleep episode (grey bars), and the phase relationship between DLMO and habitual bed time (the distance between the triangle and the grey bar) for each participant. For each sub-scale, participants are ranked from lowest complaint (top of the figure) to highest complaint (bottom of the figure).

DISCUSSION

Our data show that advanced circadian timing is associated with subjective memory complaints, particularly decreased memory of autobiographical events and increased use of mnemonic strategies in everyday life. We also show decreased sleep fragmentation and longer total sleep time was associated with reports of more decline over time, and greater seriousness attributed to everyday memory failures. This is the first time characterization of circadian timing and sleep outcomes, and the association with subjective memory complaints, has been examined in older adults with SCD.

The most consistent finding from the current study was that a wider phase angle (time difference between DLMO and habitual bed time) was associated with greater memory complaints; this was driven by an advance of DLMO, rather than a change to HBT. Advanced circadian timing is well documented in healthy aging and is exacerbated in MCI [23]. Furthermore, this advance has been associated to poorer verbal memory outcomes in MCI patients [23]. Of particular note, advanced circadian timing was strongly associated with mnemonic use in everyday life, suggesting that individuals with advanced circadian phase may rely more on compensatory strategies to combat memory failures. As mnemonic use has been shown to be increased in SCD [59], and is predictive of future cognitive decline [60], our data add further evidence to the potential of advanced circadian phase representing an early biomarker for dementia, and importantly provides a potential target for intervention, i.e., chronotherapy. While patients with MCI exhibit earlier bed times and DLMO relative to controls, we did not find any associations between sleep-wake behaviors and subjective memory complaints. Moreover, we observed relatively normal bedtimes, with participants retiring to bed, on average, at 11:00 pm. While we are unable to attribute causality between circadian timing and subjective memory complaints, previous work characterizing the impact of circadian misalignment on oxidative stress, neurogenetics, and gene expression (see [61] for review), along with a plethora of studies linking fragmentation in actigraphic rest-activity rhythms with increased risk of cognitive decline [12 , 63], suggest that circadian disruption may be a driving factor in neurodegenerative progression.

Advanced circadian timing may be due to a multitude of factors, including neurodegeneration in the SCN [64 –66]; abnormalities in entrainment to the light dark cycle, caused by reduced light exposure due to eye degeneration [67] or lifestyle choices; or age-related changes to the pineal gland, which synthesizes melatonin [24]. Light is the major external time cue for entraining the clock to the external environment. While morning light leads to advances in circadian timing [68], it is currently unknown whether habitual light exposure patterns are altered in individuals with SCD and MCI. Examining the association between light exposure and circadian timing in these groups, and whether the impact of a light intervention known to delay the clock (e.g., evening bright light exposure [69]) would elucidate the role of light on circadian timing in SCD, and inform intervention strategies to delay cognitive decline and progression to dementia.

Individuals who reported greater memory decline compared to previous points in their lives, and who attributed more seriousness to their memory complaints had greater total sleep time, and more consolidated sleep (less WASO) according to PSG-defined sleep. While this appears counterintuitive, it does replicate previous work utilizing actigraphy [41]. Two possible explanations lend themselves to these findings. Firstly, improved sleep quality may be a compensatory response to greater daytime cognitive effort to conserve memory function [41], particularly since participants with SCD exhibit increased prefrontal activation during encoding tasks with no change in memory performance, relative to controls [70]. While we found no relationship between sleep outcomes and reliance on mnemonic tools, we did observe long bouts of SWS in some participants (>20% of TST), comparable to that seen in a young, healthy adult [71], which may reflect a compensatory response to increased cortical effort throughout the waking day; although this would require verification from future work. The second explanation relates to the differential relationship between sleep consolidation and complaint subtypes. Individuals with poorer sleep did not report any difference in the frequency of problems or the ability to remember past events, but did attribute less seriousness to those problems and felt the decline was not as severe compared to previous years. This may represent a continuum of progression whereby individuals with poorer sleep have become accustomed to any associated cognitive issues, and thus express less concern when reflecting on everyday memory failure. While many prospective studies have investigated the impact of sleep fragmentation on objective cognitive performance [11, 12], a longitudinal study design assessing subjective and objective decline, and parallel changes in sleep parameters, would yield much insight in this respect.

As both sleep quality and quantity are influenced by circadian timing, we do not believe that our counterintuitive sleep results are a consequence of our circadian timing results. Advanced circadian timing, which was associated with greater memory complaints, results in the melatonin peak occurring earlier within the sleep period, which can increase sleep fragmentation and awakenings in the second half of the sleep period (due to the lack of a sleep-promoting circadian signal, and reduced homeostatic pressure for sleep [22]). If advanced circadian timing was leading to increased sleep fragmentation, we would expect increased fragmentation and reduced total sleep time to be associated with increased memory complaints, not decreased. As circadian timing and sleep do not operate independently of each other, it is still likely that these outcomes are interacting and thus together might have a greater association to subjective complaints. Our study, however, lacks the statistical power to investigate these potential interactions.

SCD may encompass a wide range of changes or impairments in semantic and episodic memory, working memory, and even attention and executive function. Previous work utilized questionnaires which ask individuals to rate everyday memory failures in comparison to a previous time in their lives [41, 42]. The differential relationships between circadian timing and sleep fragmentation (as well as covariates such as intelligence, depressive symptoms and stress) on the complaint subtypes measured by the MFQ highlights the importance of examining the multi-faceted nature of SCD in future studies. Perhaps surprisingly, the frequency subscale was the only subscale to show no relationship with sleep or circadian timing outcomes. Unlike any other sub-scale, Frequency of Forgetting was highly correlated with levels of stress, suggesting it may be more related to psychosocial factors and daily stressors than the other subscales. While the MFQ examines various facets of memory complaints, including some previously identified as sub-themes [43], it does not assess other complaints across other cognitive domains, which may explain why subjective complaints were not strongly associated with neuropsychological performance.

While our study fills an important gap about the role of sleep and circadian timing in SCD, it is not without limitations. Due to the extensive time commitment required from participants, preliminary nature of the work, and the strict exclusion criteria, this study comprised a small dataset of relatively healthy participants with SCD. As such, we did not correct for multiple comparisons, and it is likely that the study was underpowered to detect some relationships between sleep and cognitive outcomes. As can be seen in Table 3, only relationships of r > 0.42 reached statistical significance, representing medium to large effects, and explaining at least 18% of the variance within each analysis; our findings are therefore unlikely due to chance. Loss of power was further exacerbated by the inability to detect melatonin onset in seven participants. As melatonin samples were only collected up to HBT, it is possible that DLMO occurred after HBT in these individuals. Due to this, and the commonly reported decrease in amplitude in AD [24], future work should aim to capture the entire melatonin curve, to examine the impact of circadian phase and amplitude. While the strict exclusion criteria employed in this study means that the results seen here are not likely due to other comorbidities or confounds, it also means that these results may not generalize to the wider SCD population—particularly in relation to mood and sleep disorders. As sleep-disordered breathing is associated with an increased risk of cognitive impairment [72], and people with MCI who showed disordered sleep demonstrated reduced functional connectivity compared to those without disordered sleep [73], further investigation into the impact of sleep disorders on individuals with SCD is warranted. Finally, our sample was also comprised of predominantly female participants. While we found no effect of sex (except for SWS), and the male:female ratio is in line with previous studies [41, 42], it is possible that these relationships may be altered in a larger sample of males, who show higher rates of sleep disordered breathing, and reduced SWS compared to females [5, 74].

In summary, this study is the first to show associations between SCD and advanced circadian timing, suggesting that circadian timing may be an early target for intervention for age-related memory complaints. We also show differential associations between sleep fragmentation and SCD subtype, with poorer sleep typically being associated with less concern over memory complaints. SCD represents an ideal opportunity for interventions to slow or delay neurodegenerative changes leading to cognitive decline and dementia. Given that advanced circadian timing was associated with subjective memory problems previously associated with a doubling of dementia risk, our findings support the potential for targeting the circadian timing system (chronotherapy) in this critical period for intervention.

Footnotes

ACKNOWLEDGMENTS

The authors wish to thank the study participants for their time and effort, L. Thomson for assisting with sleep scoring, the technical staff and students at the Monash Sleep and Circadian Medicine Laboratory for their assistance with data collection, and M. Salkeld at the Adelaide Research Assay Facility for salivary melatonin analysis.

This study was funded by the Mason Foundation, ANZ Trustees (grant number CT23151 awarded to CA). JEM and AJS are supported by an Australian Government Research Training Program Scholarship. JEM is also supported by a scholarship from the NHMRC Centre of Research Excellence Neurosleep. SN is supported by an NHMRC Dementia Leadership Fellowship.

Authors’ disclosures available online ().

References

Alzheimer’s Disease International (2015), World Alzheimer report 2015: The global impact of dementia, an analysis of prevalence, incidence, cost, and trends. London.

Norton

, Matthews

, Barnes

, Yaffe

, Brayne

(2014) Potential for primary prevention of Alzheimer’s disease: An analysis of population-based data. Lancet Neurol 13, 788–794.

Spira

, Gottesman

(2017) Sleep disturbance: An emerging opportunity for Alzheimer’s disease prevention? Int Psychogeriatr 29, 529–531.

Mander

, Winer

, Jagust

, Walker

(2016) Sleep: A novel mechanistic pathway, biomarker, and treatment target in the pathology of Alzheimer’s disease? Trends Neurosci 39, 552–566.

Mander

, Winer

, Walker

(2017) Sleep and human aging. Neuron 94, 19–36.

Prinz

, Vitaliano

, Vitiello

, Bokan

, Raskind

, Peskind

, Gerber

(1982) Sleep, EEG and mental function changes in senile dementia of the Alzheimer’s type. Neurobiol Aging 3, 361–370.

Vitiello

, Prinz

(1989) Alzheimer’s disease. Sleep and sleep/wake patterns. Clin Geriatr Med 5, 289–299.

Vitiello

, Prinz

, Williams

, Frommlet

, Ries

(1990) Sleep disturbances in patients with mild-stage Alzheimer’s disease. J Gerontol 45, M131–M138.

Spira

, Chen-Edinboro

, Wu

, Yaffe

(2014) Impact of sleep on the risk of cognitive decline and dementia. Curr Opin Psychiatry 27, 478–483.

10.

Hahn

, Wang

, Andel

, Fratiglioni

(2014) A change in sleep pattern may predict Alzheimer disease. Am J Geriatr Psychiatry 22, 1262–1271.

11.

Blackwell

, Yaffe

, Laffan

, Ancoli-Israel

, Redline

, Ensrud

, Song

, Stone

(2014) Associations of objectively and subjectively measured sleep quality with subsequent cognitive decline in older community-dwelling men: The MrOS sleep study. Sleep 37, 655–663.

12.

Lim

ASP

, Kowgier

, Yu

, Buchman

, Bennett

(2013) Sleep fragmentation and the risk of incident Alzheimer’s disease and cognitive decline in older persons. Sleep 36, 1027–1032.

13.

Yaffe

, Laffan

, Harrison

, Redline

, Spira

, Ensrud

, Ancoli-Israel

, Stone

(2011) Sleep-disordered breathing, hypoxia, and risk of mild cognitive impairment and dementia in older women. JAMA 306, 613–619.

14.

Sprecher

, Bendlin

, Racine

, Okonkwo

, Christian

, Koscik

, Sager

, Asthana

, Johnson

, Benca

(2015) Amyloid burden is associated with self-reported sleep in nondemented late middle-aged adults. Neurobiol Aging 36, 2568–2576.

15.

Sprecher

, Koscik

, Carlsson

, Zetterberg

, Blennow

, Okonkwo

, Sager

, Asthana

, Johnson

, Benca

, Bendlin

(2017) Poor sleep is associated with CSF biomarkers of amyloid pathology in cognitively normal adults. Neurology 89, 445–453.

16.

Mander

, Rao

, Lu

, Saletin

, Lindquist

, Ancoli-Israel

, Jagust

, Walker

(2013) Prefrontal atrophy, disrupted NREM slow waves and impaired hippocampal-dependent memory in aging. Nat Neurosci 16, 357–364.

17.

Varga

, Ducca

, Kishi

, Fischer

, Parekh

, Koushyk

, Yau

, Gumb

, Leibert

, Wohlleber

, Burschtin

, Convit

, Rapoport

, Osorio

, Ayappa

(2016) Effects of aging on slow-wave sleep dynamics and human spatial navigational memory consolidation. Neurobiol Aging 42, 142–149.

18.

Saper

, Lu

, Chou

, Gooley

(2005) The hypothalamic integrator for circadian rhythms. Trends Neurosci 28, 152–157.

19.

Duffy

, Dijk

D-J

, Klerman

, Czeisler

(1998) Later endogenous circadian temperature nadir relative to an earlier wake time in older people. Am J Physiology 275, R1478–R1487.

20.

Mattis

, Sehgal

(2016) Circadian rhythms, sleep, and disorders of aging. Trends Endocrinol Metabol 27, 192–203.

21.

Monk

, Buysse

, Reynolds

3rd , Kupfer

, Houck

(1995) Circadian temperature rhythms of older people. Exp Gerontol 30, 455–474.

22.

Dijk

, Duffy

, Kiel

, Shanahan

, Czeisler

(1999) Ageing and the circadian and homeostatic regulation of human sleep during forced desynchrony of rest, melatonin and temperature rhythms. J Physiol 516, 611–627.

23.

Naismith

, Hickie

, Terpening

, Rajaratnam

, Hodges

, Bolitho

, Rogers

, Lewis

(2014) Circadian misalignment and sleep disruption in mild cognitive impairment. J Alzheimers Dis 38, 857–866.

24.

Y-H

, Swaab

(2005) The human pineal gland and melatonin in aging and Alzheimer’s disease. J Pineal Res 38, 145–152.

25.

Wang

, Lim

, Chiang

, Hsieh

, Lo

, Schneider

, Buchman

, Bennett

, Hu

, Saper

(2015) Suprachiasmatic neuron numbers and rest-activity circadian rhythms in older humans. Ann Neurol 78, 317–322.

26.

Tranah

, Blackwell

, Stone

, Ancoli-Israel

, Paudel

, Ensrud

, Cauley

, Redline

, Hillier

, Cummings

, Yaffe

(2011) Circadian activity rhythms and risk of incident dementia and mild cognitive impairment in older women. Ann Neurol 70, 722–732.

27.

Van Someren

, Hagebeuk

, Lijzenga

, Scheltens

, de Rooij

, Jonker

, Pot

, Mirmiran

, Swaab

(1996) Circadian rest-activity rhythm disturbances in Alzheimer’s disease. Biol Psychiatry 40, 259–270.

28.

Cummings

, Doody

, Clark

(2007) Disease-modifying therapies for Alzheimer disease: Challenges to early intervention. Neurology 69, 1622–1634.

29.

Landry

, Liu-Ambrose

(2014) Buying time: A rationale for examining the use of circadian rhythm and sleep interventions to delay progression of mild cognitive impairment to Alzheimer’s disease. Front Aging Neurosci 6, 325.

30.

Sperling

, Mormino

, Johnson

(2014) The evolution of preclinical Alzheimer’s disease: Implications for prevention trials. Neuron 84, 608–622.

31.

Jessen

, Amariglio

, Van Boxtel

, Breteler

, Ceccaldi

, Chételat

, Dubois

, Dufouil

, Ellis

, Van Der Flier

, Glodzik

, Van Harten

, De Leon

, McHugh

, Mielke

, Molinuevo

, Mosconi

, Osorio

, Perrotin

, Petersen

, Rabin

, Rami

, Reisberg

, Rentz

, Sachdev

, De La Sayette

, Saykin

, Scheltens

, Shulman

, Slavin

, Sperling

, Stewart

, Uspenskaya

, Vellas

, Visser

, Wagner

(2014) A conceptual framework for research on subjective cognitive decline in preclinical Alzheimer’s disease. Alzheimers Dement 10, 844–852.

32.

Mitchell

, Beaumont

, Ferguson

, Yadegarfar

, Stubbs

(2014) Risk of dementia and mild cognitive impairment in older people with subjective memory complaints: Meta-analysis. Acta Psychiatr Scand 130, 439–451.

33.

Striepens

, Scheef

, Wind

, Meiberth

, Popp

, Spottke

, Kölsch

, Wagner

, Jessen

(2011) Interaction effects of subjective memory impairment and ApoE4 genotype on episodic memory and hippocampal volume. Psychol Med 41, 1997–2006.

34.

Meiberth

, Scheef

, Wolfsgruber

, Boecker

, Block

, Träber

, Erk

, Heneka

, Jacobi

, Spottke

, Walter

, Wagner

, Hu

, Jessen

(2015) Cortical thinning in individuals with subjective memory impairment. J Alzheimers Dis 45, 139–146.

35.

Peter

, Scheef

, Abdulkadir

, Boecker

, Heneka

, Wagner

, Koppara

, Klöppel

, Jessen

(2014) Gray matter atrophy pattern in elderly with subjective memory impairment. Alzheimers Dement 10, 99–108.

36.

Wolfsgruber

, Polcher

, Koppara

, Kleineidam

, Frölich

, Peters

, Hüll

, Rüther

, Wiltfang

, Maier

, Kornhuber

, Lewczuk

, Jessen

, Wagner

(2017) Cerebrospinal fluid biomarkers and clinical progression in patients with subjective cognitive decline and mild cognitive impairment. J Alzheimers Dis 58, 939–950.

37.

Okamura

, Ura

, Miyamae

, Sugiyama

, Niikawa

, Ito

, Awata

(2016) Excessive daytime sleepiness is related to subjective memory impairment in late life: A cross-sectional community-based study. Psychogeriatrics 16, 196–201.

38.

Clarnette

, Almeida

, Forstl

, Paton

, Martins

(2001) Clinical characteristics of individuals with subjective memory loss in Western Australia: Results from a cross-sectional survey. Int J Geriatr Psychiatry 16, 168–174.

39.

Kang

, Yoon

, Lee

, Kim

, Lee

, Han

, Kim

(2017) Subjective memory complaints in an elderly population with poor sleep quality. Aging Ment Health 21, 532–536.

40.

Gamaldo

, Wright

, Aiken-Morgan

, Allaire

, Thorpe

Jr , Whitfield

(2017) The association between subjective memory complaints and sleep within older African American adults. J Gerontol B Psychol Sci Soc Sci. doi: 10.1093/geronb/gbx069.

41.

Cavuoto

, Ong

, Pike

, Nicholas

, Bei

, Kinsella

(2016) Better objective sleep quality in older adults with high subjective memory decline. J Alzheimers Dis 53, 943–953.

42.

Lauriola

, Esposito

, Delli Pizzi

, de Zambotti

, Londrillo

, Kramer

, Rabinovici

, Tartaro

(2017) Sleep changes without medial temporal lobe or brain cortical changes in community-dwelling individuals with subjective cognitive decline. Alzheimers Dement 13, 783–791.

43.

Buckley

, Saling

, Frommann

, Wolfsgruber

, Wagner

(2015) Subjective cognitive decline from a phenomenological perspective: A review of the qualitative literature. J Alzheimers Dis 48, S125–S140.

44.

AASM (2014) The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications, American Academy of Sleep Medicine, Westchester.

45.

Claman

, Ewing

, Redline

, Ancoli-Israel

, Cauley

, Stone

(2013) Periodic leg movements are associated with reduced sleep quality in older men: The MrOS sleep study. J Clin Sleep Med 9, 1109–1117.

46.

Claman

, Redline

, Blackwell

, Ancoli-Israel

, Surovec

, Scott

, Cauley

, Ensrud

, Stone

(2006) Prevalence and correlates of periodic limb movements in older women. J Clin Sleep Med 2, 438–445.

47.

Gilewski

, Zelinski

, Schaie

(1990) The Memory Functioning Questionnaire for assessment of memory complaints in adulthood and old age. Psychol Aging 5, 482–490.

48.

Folstein

, Folstein

, White

, Messer

(2010) Psychological Assessment Resources. Mini-Mental State Examination - Second Edition.

49.

Nelson

(1982) National Adult Reading Test (NART): For the Assessment of Premorbid Intelligence in Patients with Dementia: Test Manual NFER-Nelson, Windsor, UK.

50.

Reynolds

(2002) Comprehensive trail making test (CTMT), Pro-Ed Austin, TX.

51.

Delis

, Kramer

, Kaplan

, Ober

(2000) California Verbal Learning Test Psychological Corporation, San Antonio.

52.

Yesavage

, Brink

, Rose

, Lum

, Huang

, Adey

, Leirer

(1982) Development and validation of a geriatric depression screening scale: A preliminary report. J Psychiatr Res 17, 37–49.

53.

Lovibond

, Lovibond

(1995) Manual for the Depression Anxiety Stress Scales, Psychology Foundation, Sydney.

54.

Johns

(1991) A new method for measuring daytime sleepiness: The Epworth Sleepiness Scale. Sleep 14, 540–545.

55.

Buysse

, Reynolds III

, Monk

, Berman

, Kupfer

(1989) The Pittsburgh sleep quality index: A new instrument for psychiatric practice and research. Psychiatry Res 28, 193–213.

56.

Horne

, Ostberg

(1976) A self-assessment questionnaire to determine morningness-eveningness in human circadian rhythms. Int J Chronobiol 4, 97–110.

57.

Voultsios

, Kennaway

, Dawson

(1997) Salivary melatonin as a circadian phase marker: Validation and comparison to plasma melatonin. J Biol Rhythms 12, 457–466.

58.

Murray

, Sletten

, Magee

, Gordon

, Lovato

, Bartlett

, Kennaway

, Lack

, Grunstein

, Lockley

, Rajaratnam

SMW

(2017) Prevalence of circadian misalignment and its association with depressive symptoms in delayed sleep phase disorder. Sleep 40, zsw002.

59.

Frankenmolen

, Overdorp

, Fasotti

, Claassen

JAHR

, Kessels

RPC

, Oosterman

(2017) Memory strategy use in older adults with subjective memory complaints. Aging Clin Exp Res 29, 1061–1065.

60.

Small

, La Rue

, Komo

, Kaplan

(1997) Mnemonics usage and cognitive decline in age-associated memory impairment. Int Psychogeriatr 9, 47–56.

61.

Musiek

, Holtzman

(2016) Mechanisms linking circadian clocks, sleep, and neurodegeneration. Science 354, 1004–1008.

62.

Musiek

, Bhimasani

, Zangrilli

, Morris

, Holtzman

, Ju

YES

(2018) Circadian rest-activity pattern changes in aging and preclinical Alzheimer disease. JAMA Neurol 75, 582–590.

63.

Lim

ASP

, Yu

, Costa

, Leurgans

, Buchman

, Bennett

, Saper

(2012) Increased fragmentation of rest-activity patterns is associated with a characteristic pattern of cognitive impairment in older individuals. Sleep 35, 633–640.

64.

Hofman

, Swaab

(1995) Influence of aging on the seasonal rhythm of the vasopressin-expressing neurons in the human suprachiasmatic nucleus. Neurobiol Aging 16, 965–971.

65.

Hofman

, Swaab

(1994) Alterations in circadian rhythmicity of the vasopressin-producing neurons of the human suprachiasmatic nucleus (SCN) with aging. Brain Res 651, 134–142.

66.

Swaab

, Fliers

, Partiman

(1985) The suprachiasmatic nucleus of the human brain in relation to sex, age, and senile dementia. Brain Res 342, 37–44.

67.

Wulff

, Foster

(2017) Insight into the role of photoreception and light intervention for sleep and neuropsychiatric behaviour in the elderly. Curr Alzheimer Res 14, 1022–1029.

68.

Khalsa

SBS

, Jewett

, Cajochen

, Czeisler

(2003) A phase response curve to single bright light pulses in human subjects. J Physiol 549, 945–952.

69.

Münch

, Scheuermaier

, Zhang

, Dunne

, Guzik

, Silva

, Ronda

, Duffy

(2011) Effects on subjective and objective alertness and sleep in response to evening light exposure in older subjects. Behav Brain Res 224, 272–278.

70.

Rodda

, Dannhauser

, Cutinha

, Shergill

, Walker

(2009) Subjective cognitive impairment: Increased prefrontal cortex activation compared to controls during an encoding task. Int J Geriatr Psychiatry 24, 865–874.

71.

Van Cauter

, Leproult

, Plat

(2000) Age-related changes in slow wave sleep and rem sleep and relationship with growth hormone and cortisol levels in healthy men. JAMA 284, 861–868.

72.

Leng

, McEvoy

, Allen

, Yaffe

(2017) Association of sleep-disordered breathing with cognitive function and risk of cognitive impairment: A systematic review and meta-analysis. JAMA Neurol 74, 1237–1245.

73.

McKinnon

, Duffy

, Cross

, Terpening

, Grunstein

, Lagopoulos

, Batchelor

, Hickie

, Lewis

SJG

, Shine

, Naismith

(2017) Functional connectivity in the default mode network is reduced in association with nocturnal awakening in mild cognitive impairment. J Alzheimers Dis 56, 1373–1384.

74.

Young

, Shahar

, Nieto

, Redline

, Newman

, Gottlieb

, Walsleben

, Finn

, Enright

, Samet

(2002) Predictors of sleep-disordered breathing in community-dwelling adults: The Sleep Heart Health study. Arch Int Med 8, 893–900.