1

Abstract

Sleep has been shown to play a crucial role in the consolidation of emotionally

salient memories. However, the influence of sleep, and Sleep Deprivation (SD), on

emotional memory consolidation in depressive individuals remains elusive. For this

experiment we recruited two groups of healthy students, one reporting mild-to-severe

depressive symptoms, and another reporting minimal/no depressive symptoms

(assessed using the Beck Depression Inventory; BDI-II). We measured recognition

performance for positive, neutral and negative images before and after a 12 h

overnight retention interval, during which participants either remained awake in the

laboratory or returned home to sleep normally. We found a significant depressive

symptomatology group x sleep condition x image valence interaction on memory

consolidation across the 12 h retention interval [F(2, 98) = 3.12, p = .049, ηp² =

0.060]. We also found that depressive participants who slept normally consolidated

significantly more negative and neutral images across the 12 h retention interval than

depressive participants who were sleep deprived [t(24) = 2.35, p = .028, t(24) = 2.79,

p = .010, respectively]. Our preliminary results indicate that SD may impair the

consolidation of negative and neutral memories in depressive participants, but not in

participants reporting minimal/no depressive symptoms.

Keywords: sleep deprivation; emotional memory; depression; REM sleep

2

1. Introduction

Individuals reporting symptoms of depression such as despair and

hopelessness exhibit greater recall and recognition performance for negative

emotional stimuli, relative to individuals reporting minimal/no depressive symptoms

(Everaert, Duyck, & Koster, 2014; Fattahi Asl, Ghanizadeh, Mollazade, & Aflakseir,

2015; Howe & Malone, 2011). This emotional memory bias may increase

vulnerability to the development and recurrence of Major Depressive Disorder (MDD;

Beck, 1967; Everaert, Koster, & Derakshan, 2012; Gotlib & Joormann, 2010). In

healthy participants, emotionally salient memories are better remembered than their

neutral or mundane counterparts (Kensinger, 2004; McGaugh, 2004). Emotionally

enhanced memory may be facilitated by sleep, which selectively consolidates

memories which elicit emotional arousal (Payne, Chambers, & Kensinger, 2012;

Payne, Stickgold, Swanberg, & Kensinger, 2008).

Empirical evidence suggests that Rapid Eye Movement (REM) sleep, in

particular, plays a crucial role in the consolidation of emotional memories (Groch,

Zinke, Wilhelm, & Born, 2015; Nishida, Pearsall, Buckner, & Walker, 2009; Payne et

al., 2012; Wagner, Gais, & Born, 2001; Wiesner et al., 2015). For example, in

healthy students selective REM Sleep Deprivation (SD) has been shown to impair

the consolidation of negative emotional images, but not emotionally neutral images,

to a greater extent than selective Slow-Wave Sleep (SWS) deprivation (Wiesner et

al., 2015). Furthermore, correlational analyses reveal that, for healthy individuals,

greater consolidation of emotionally negative images is predicted by longer duration

of time spent in REM sleep during retention intervals (Nishida et al., 2009; Payne et

al., 2012). Similarly, split-night studies in healthy participants have shown that 3h of

3

late night REM-rich sleep fosters the selective consolidation of emotionally negative

stimuli more abundantly than 3 h of early night sleep where REM sleep is less

plentiful (Groch et al., 2015; Wagner et al., 2001). Importantly, although it is clear

that REM sleep is centrally involved in emotional memory consolidation, SWS has

also been shown to contribute to this process in some studies (e.g. Cairney, Durrant,

Power, & Lewis, 2015; Sabine Groch et al., 2011). Moreover, contrary to the bulk of

the existing literature, some investigations have failed to detect a specific role for

REM sleep in emotional memory consolidation (Ackermann, Hartmann,

Papassotiropoulos, de Quervain, & Rasch, 2015; Cellini, Torre, Stegagno, & Sarlo,

2016; Morgenthaler et al., 2014).

Experiments investigating the relationship between sleep and memory in

depressive participants have mostly shown that sleep-associated consolidation is

impaired in MDD patients. For example, healthy controls have been shown to exhibit

greater overnight improvements on a motor memory task than MDD patients (Dresler

et al., 2011; Dresler, Kluge, Genzel, Schüssler, & Steiger, 2010; Nishida,

Nakashima, & Nishikawa, 2016). Moreover, MDD patients who performed poorly on

a declarative visual memory task exhibited shorter REM sleep duration and total

sleep time than patients who performed well on the task (Göder et al., 2007). The

association between sleep and emotional memory in depressive participants has

received comparatively less empirical attention. Nonetheless, we recently revealed

that late night REM-rich sleep promotes the consolidation of negative emotional

images more readily in participants reporting mild-to-moderate depressive symptoms

than participants reporting minimal/no symptoms of depression (Harrington,

Johnson, Croom, Pennington, & Durrant, 2018).

4

The current study aimed to investigate the short- and long- term effect of post-

learning sleep, or SD, on recognition performance for positive, neutral and negative

images in samples of healthy participants reporting either minimal/no depressive

symptoms or mild-to-severe depressive symptoms. These two groups of participants

are referred to below as the low Beck Depression Inventory (BDI-II) score group and

the higher BDI-II score group, respectively. Although it has been shown that SD, and

selective REM SD, impairs the consolidation of emotionally negative information in

healthy participants (Tempesta, De Gennaro, Natale, & Ferrara, 2015; Wiesner et

al., 2015); the effect of SD on emotional memory consolidation in participants

reporting depressive symptoms remains unexplored. We predicted that SD would

impair the consolidation of negative emotional memories, and that this effect would

be more robust in the higher BDI-II score group than the low BDI-II score group.

2. Materials and Methods

2.1. Participants

Recruitment was based on BDI-II (Beck, Steer, & Brown, 1996) scores

obtained from a prescreen of 128 student volunteers who signed up using an online

research participation system. Fifty-six participants were selected based on their

prescreen BDI-II scores and assigned to one of two depressive symptomatology

groups: a low BDI-II score group and a higher BDI-II score group (see Section 2.1.1

for group criteria). Both depressive symptomatology groups were then randomly sub-

divided into a further two separate sleep condition groups, a Sleep Deprivation (SD)

condition and a Normal Sleep (NS) condition (see Section 2.3 for information about

sleep conditions). Participants were required to have no history of sleep,

5

neurological, endocrine or psychiatric disorders as assessed through self-report.

Participants were also free of long-term medication (except for the female

contraceptive pill), and were native English speakers.

Of the selected participants, two were excluded from final analyses for failing

to meet the eligibility criteria at the first subsequent test session (i.e. they were

originally assigned to the higher BDI-II score group, but reported a BDI-II score

below the cut-off boundary for this group at the first subsequent test session). Data is

reported from the remaining 54 participants (see Table 1): 14 in the higher BDI-II

score SD group (four male, ten female) aged 19 – 23, 12 in the higher BDI-II score

NS group (three male, nine female) aged 18 – 23, 14 in the low BDI-II score SD

group (six male, eight female) aged 18 – 22, and 14 in the low BDI-II score NS group

(four male, ten female) aged 18 – 25.

Table 1. Summary of participants’ demographic data, separately for each depressive symptomatology group and sleep condition group

Low BDI-II score group Higher BDI-II score group

NS group SD group NS group SD group

n 14 14 12 14

Gender (m/f) 4/10 6/8 3/9 4/10

Age (years) 19.86 (0.62) 19.43 (0.33) 19.92 (0.45) 20.14 (0.36)

BDI-II 1.86 (0.50) 2.43 (0.62) 21.50 (2.01) 20.50 (1.27)

PSQI 3.64 (0.53) 4.29 (0.47) 8.17 (0.92) 8.00 (0.79)

Mean values are presented with SEM in parenthesis. Higher scores indicate greater depressive symptoms or sleep quality impairment. Abbreviations: NS, Normal Sleep; SD, Sleep Deprivation; BDI-II, Beck Depression Inventory; PSQI, Pittsburgh Sleep Quality Index.

6

Participants were asked to abstain from alcohol, caffeine, and other drugs for

24 h prior to test sessions. Participants were rewarded for their participation with

course credits. Each participant gave informed consent for this study, which was

approved by the School of Psychology Research Ethics Committee at the University

of Lincoln.

2.1.1. Depressive Symptomatology Groups

The BDI-II measures depressive symptom severity with 21 items rated on a

scale from zero to three. This self-report measure has good reliability and validity in

both healthy and depressed samples (Beck et al., 1996) and is widely used in

cognitive research (Everaert et al., 2014; Hindash & Amir, 2011; Newby, Lang,

Werner-Seidler, Holmes, & Moulds, 2014). Prescreen data showed that BDI-II scores

ranged from zero to 52, with 92 individuals reporting minimal/no (BDI-II cut off range:

0 - 13), 16 mild (BDI-II cut off range: 14 - 19), 13 moderate (BDI-II cut off range: 20 -

28), and seven severe symptom levels (BDI-II cut off range: 29 - 63). We invited

individuals reporting BDI-II scores ≥ 14, and a matching number of individuals who

reported the lowest BDI-II scores at the prescreen, to participate in this study.

Individuals who accepted our invitation, reported BDI-II scores ≤ 13 at both the

prescreen and the first test session, and successfully completed the study, make up

the low BDI-II score group (BDI-II score: mean = 2.14, range = 0 - 6). Individuals who

accepted our invitation, reported BDI-II scores ≥ 14 at both the prescreen and the

first test session, and successfully completed the study, make up the higher BDI-II

score group (BDI-II score: mean = 20.96, range = 14 - 35). BDI-II scores did not

7

fluctuate between the prescreen and the first test session to the extent that they no

longer fit the ranges outlined above.

2.2. Stimuli

Three hundred and sixty images were selected from the International Affective

Picture System (IAPS; Lang, Greenwald, Bradley, & Hamm, 1993). IAPS images

range from photographs depicting happy relationships, to everyday scenes, to

images of injury and violence, and each are rated on nine-point scales for emotional

valence (1 = negative, 5 = neutral, 9 = positive) and arousal (1 = calm, 9 = exciting).

Images were selected based on their valence and arousal ratings, and placed into

one of three emotion categories: “positive”, “neutral” or “negative”, each of which

contained 120 images. Pairwise comparisons demonstrated that there was a

significant difference in the mean IAPS valence rating (i.e. normative valence ratings

from the IAPS database) between all emotion categories (positive: [7.31 ± 0.04;

mean ± SEM], neutral: [5.04 ± 0.02], negative: [2.42 ± 0.05]; all pairwise p < .001).

There was also a significant difference in the mean IAPS arousal rating (i.e.

normative arousal ratings from the IAPS database) for the neutral category relative

to the positive and negative categories (positive: [5.5 ± 0.06], neutral: [3.33 ± 0.06],

negative: [5.5 ± 0.06]; both p < .001), however we ensured that there was no

significant difference in the mean normative arousal rating between the positive and

negative categories [p = .886].

Selected images were divided into two equal sets of 180 (60 positive, 60

neutral, 60 negative), which were matched as closely as possible for IAPS valence

and arousal ratings. During learning phases, each participant viewed one of these

8

two sets of images which served as the targets during subsequent recognition tests

(see Section 2.4 for details regarding learning phases and recognition tests). Each of

the two sets of images was further subdivided into three equal sets of 60 (20

positive, 20 neutral, 20 negative), which were also matched as closely as possible

for IAPS valence and arousal ratings. During each of the three recognition tests

(immediate recognition, 12 h recognition, 7 d recognition; see Section 2.3 for

information about the study protocol), participants viewed one of the three sets of 60

images which they saw during the learning phase (targets) intermixed with one of the

three sets of 60 images from the other set of 180 (foils). The image sets used for

encoding and recognition tests were counterbalanced across depressive

symptomatology groups and sleep condition groups as closely as possible (i.e. each

set of learning phase images and recognition test images was viewed approximately

an equal number of times by each depressive symptomatology group and each

sleep condition group). The image sets used for recognition tests were also viewed

approximately an equal number of times at each recognition test (i.e. immediate

recognition, 12 h recognition, 7 d recognition).

2.3. Experimental Protocol

The experimental protocol is summarised in Fig 1. Participants took part in

one of two experimental conditions, a SD condition or a NS condition. All participants

arrived at the Lincoln Sleep and Cognition Laboratory at 8 PM. The study protocol

began with participants completing the BDI-II and the Pittsburgh Sleep Quality Index

(PSQI; a self-report measure of subjective sleep quality over the last month; Buysse,

Reynolds, Monk, Berman, & Kupfer, 1989). This was immediately followed by a

9

learning phase and immediate recognition test (see Section 2.4 for details) which

lasted approximately 1 h in total. Participants assigned to the NS group then

returned home where they were asked to be in bed from 10.30 PM to 7.30 AM and to

sleep normally, before returning to the laboratory at 8 AM the following morning. This

group of participants recorded their sleep using portable, in-home PSG devices

(Sleep ProfilerTM; Levendowski et al., 2017; see Section 2.5.2 for details).

Participants assigned to the SD group remained awake in the laboratory under the

supervision of the researcher until the following morning. During this time

participants were free to eat, drink, read, listen to music, communicate with the

researcher, or use the computer. Participants were not permitted to consume

alcohol, caffeine, or other drugs during this time.

Fig 1. Pictorial representation of the study protocol. Participants in the SD condition began a learning phase at 8 PM which was followed by an immediate recognition test. They then remained awake in the laboratory until 8 AM the following morning, before completing a 12 h recognition test. Participants in the NS condition began a learning phase at 8 PM which was followed by an immediate recognition test. They then returned home to sleep normally between 10.30 PM and 7.30 AM, before returning to the laboratory for a 12 h recognition test at 8 PM. These participants recorded their sleep using a Sleep ProfilerTM In-home EEG Sleep Monitor. Participants in both conditions returned to the laboratory at 8 PM, one week after the learning phase, to complete a 7 d recognition test. Abbreviations: PSG, Polysomnography; SD, Sleep Deprivation; NS, Normal Sleep.

10

At 8 AM (12 h after the learning phase) all participants completed a 12 h

recognition test which lasted approximately 30 min and provided a measurement of

consolidation processes which occurred during the preceding 12 h of sleep or

wakefulness. One week after the learning phase, participants returned to the

laboratory at 8 PM where they completed a 7 d recognition test which also lasted

approximately 30 min. The purpose of this one week follow-up was to examine the

impact of SD on long-term consolidation processes. The 7 d recognition test was

carried out at the same time of day as the immediate recognition test to control for

the impact of circadian rhythms on cognitive performance (Blatter & Cajochen, 2007;

Kyriacou & Hastings, 2010).

2.4. Learning Phases and Recognition Tests

The immediate recognition test was administered immediately after the

learning phase. Prior to each recognition test participants indicated their level of

sleepiness on a scale from one (alert, wide awake) to seven (fighting sleep) using

the Stanford Sleepiness Scale (SSS; Hoddes, Dement, & Zarcone, 1972). The two

delayed recognition tests (i.e. 12 h and 7 d) were identical to the immediate

recognition test, except that they contained different target and foil images (see

Section 2.2 for more information about image sets).

The learning phase and recognition tests are summarised in Fig 2. At the

beginning of the learning phase participants saw a black screen with a white central

fixation cross for 500 ms. They were then presented with the first image (size: 15 cm

x 11 cm) for 1000 ms which appeared in the centre of the screen. Following the

image, participants were prompted to provide a valence rating for the image on a

11

scale from one (very negative) to nine (very positive) using corresponding keys.

Once a valence rating had been submitted, participants were prompted to rate the

image for emotional arousal on a scale from one (low arousal) to nine (high arousal).

After the arousal rating had been submitted, participants were again presented with

the fixation cross followed by the second image. This pattern continued until every

image had been viewed and rated. Participants were instructed to provide their

emotion ratings quickly and spontaneously, and were informed prior to the learning

phase that their memory for the images would be tested immediately after learning

and again after both 12 h and 7 d.

Fig 2. Pictorial representation of the learning phase and recognition tests. During learning phases and recognition tests participants were required to rate each image on a scale from one to nine in terms of emotional valence (1 = very negative, 9 = very positive) and emotional arousal (1 = low arousal, 9 = high arousal). During recognition tests, participants were additionally required to provide remember (R), know (K), or new (N) response.

12

The recognition tests were identical to the learning phase except that before

providing their emotion ratings participants were required to make a remember/

know/ new judgement for the image using corresponding keys (i.e. R, K or N). For

the R/K/N procedure, participants were instructed to provide a remember (R)

response if they could consciously recollect seeing that specific image during the

prior learning phase. They were asked to provide a know (K) response if they felt the

image was familiar, but they could not consciously recollect details about its previous

occurrence. They were asked to press the new (N) key if they felt that they had not

seen the image during the prior learning phase. These instructions were presented

both orally and as text on the computer monitor before the task commenced. A

“remember” response indicates recollection of the episodic details of an item,

whereas a “know” response reflects item familiarity in the absence of recollection

(Mickley & Kensinger, 2008); these different forms of remembering are thought to be

supported by different neural processes (Dobbins, Kroll, & Yonelinas, 2004;

Yonelinas, 2002). For this study, we were particularly interested in examining

remember responses (see Section 2.6.1 for more information). Participants were

asked to provide their memory judgements as quickly and accurately as possible.

There was no time constraint for memory judgments or emotion ratings in the

learning phase or recognition tests.

2.5. Equipment

2.5.1. Experimental Task

Stimulus presentation and data collection used custom-written scripts running

in SuperLab 5TM (Cedrus Corp, San Pedro, CA) on a Windows desktop computer

13

with a 15.6” screen. Participant responses were recorded using the computer

keyboard.

2.5.2. Polysomnography

Overnight sleep monitoring of participants in the NS group was carried out in

the homes of participants using a Sleep ProfilerTM in-home EEG Sleep Monitor,

which provides data comparable to lab-based PSG (Lucey et al., 2016). Disposable

electrodes were attached at two standard locations according to the international 10

- 20 system (Homan, Herman, & Purdy, 1987): Fp1 and Fp2, as well as upper and

lower chin electromyogram and a reference electrode placed at Fpz. All electrodes

were verified to have a connection impedance of < 5 000 Ω. All signals were digitally

sampled at a rate of 256 Hz.

2.6. Data Analysis

2.6.1. Behavioural Data Analysis

Although our experimental task provided an index of both recollection (R

responses) and familiarity (K responses) recognition performance (see Fig 2), we

were particularly interested in recollection as this has previously shown emotion-

specific sensitivity to sleep (Cairney et al., 2015; Harrington et al., 2018). Therefore,

we made this the primary focus of this manuscript. Nonetheless, analyses pertaining

to familiarity recognition performance, and combined recollection and familiarity

recognition performance, are available in Supplement 1 and Supplement 2,

respectively.

14

Data from recollection trials were first converted to d’ scores (d’ = Z (hits / hits

+ misses) – Z (false alarms / false alarms + correct rejections); extreme proportions

of zero or one were replaced with values of “1 / (2N)” and “1 – 1 / (2N)”, respectively,

where N equals the number of trials upon which the proportion is based; Macmillan &

Kaplan, 1985), a signal detection process widely used in memory studies to account

for response bias (Macmillan & Creelman, 2005). This was done separately for

positive, neutral and negative image trials for each participant. In accordance with

previous research (Cairney et al., 2015; Groch et al., 2015; Nishida et al., 2009), to

measure the time course of memory consolidation we created a difference measure,

which we refer to as ‘behavioural consolidation’, by subtracting d’ scores at

immediate recognition testing from d’ scores at 12 h and 7 d recognition testing. The

resulting behavioural consolidation scores provide indices of the emotional memory

processes taking place across the 12 h and 7 d retention intervals, respectively.

Our main analysis used a two (depressive symptomatology group: higher BDI-

II score group, low BDI-II score group) x two (sleep condition group: SD, NS) x three

(image valence: positive; neutral; negative) mixed-measures ANOVA with

behavioural consolidation as the dependent variable. Our primary hypothesis was

that there would be a significant three-way ‘depressive symptomatology group’ x

‘sleep condition group’ x ‘image valence’ interaction, driven by impaired

consolidation of negative images in the higher BDI-II score SD group.

2.6.2. Sleep Data Analysis

Each participant’s sleep data was divided into 30 s epochs and scored

manually by trained sleep researchers using REM Logic© 1.1 according to the AASM

15

standardised criteria (Iber, Ancoli-Israel, Chesson, & Quan, 2007). Sleep data was

scored from a channel created by referencing Fp1 to Fp2 (for more information see:

Lucey et al., 2016). This gave the duration and proportion of each sleep stage (N1,

N2, SWS, and REM), as well as measures of sleep efficiency and total sleep time.

3. Results

3.1. Participant Demographics

The participants’ demographic characteristics are shown in Table 1. To

compare demographic characteristics between depressive symptomatology groups

(higher BDI-II score group, low BDI-II score group) we conducted independent-

samples t-tests, which confirmed that the higher BDI-II score group had significantly

higher BDI-II scores than the low BDI-II score group [t(52) = 16.18, p < .001]. The

higher BDI-II score group also reported significantly higher PSQI scores than the low

BDI-II score group [t(52) = 6.08, p < .001], indicating poorer sleep quality in the

higher BDI-II score group, as expected (Baglioni et al., 2011; Palagini et al., 2013).

There was no significant age difference between the two depressive

symptomatology groups [t(52) = 0.88, p = .382].

To examine whether demographic characteristics differed between sleep

condition groups (i.e. SD group, NS group), the independent-samples t-tests

reported above were repeated, this time comparing sleep condition groups,

separately for each depressive symptomatology group (i.e. higher BDI-II score

group, low BDI-II score group). There was no significant difference between the SD

group and the NS group in BDI-II score, PSQI score, or age in either the higher BDI-

II score group [all p ≥ .668] or the low BDI-II score group [all p ≥ .374]. These results

16

suggest that between- sleep condition group differences in depressive symptom

severity, sleep quality or age can reasonably be discounted as a source of variability

in recognition performance.

3.2. Alertness

Subjective sleepiness scores are shown in Table 2. To examine the effect of

recognition test (immediate, 12 h, 7 d), depressive symptomatology group, and sleep

condition group on subjective sleepiness, we conducted a three (time: immediate, 12

h, 7 d) x two (depressive symptomatology group: higher BDI-II score group, low BDI-

II score group) x two (sleep condition group: SD group, NS group) mixed-measures

ANOVA on SSS scores. A significant main effect of time was revealed [F(2, 98) =

45.82, p < .001, ηp² = 0.483], which was largely driven by a significant interaction

between time and sleep condition group [F(2, 98) = 24.91, p < .001, ηp² = .337]. The

ANOVA revealed no other significant main effects or interactions [p ≥ .124],

suggesting that between- depressive symptomatology group differences in alertness

at any of the three test sessions can reasonably be discounted as a source of

variability in recognition performance.

17

Table 2. Subjective sleepiness scores before each recognition test, separately for each depressive symptomatology group and sleep condition group

Low BDI-II score group Higher BDI-II score group

NS group SD group NS group SD group

SSS immediate 2.29 (0.13) 2.29 (0.19) 3.00 (0.28) 3.29 (0.22)

SSS 12 h 2.21 (0.21) 4.86 (0.35) 3.58 (0.42) 4.93 (0.40)

SSS 7 d 2.07 (0.20) 1.77 (0.17) 2.83 (0.39) 2.50 (0.25)

Mean values are presented with SEM in parentheses. Higher values indicate greater subjective sleepiness. Abbreviations: NS, Normal Sleep; SD, Sleep Deprivation; SSS, Stanford Sleepiness Scale.

To explore the interaction between time and sleep condition, we compared

SSS scores between the two sleep conditions, separately for each recognition test

(immediate, 12 h, 7 d), using independent-samples t-tests. It was revealed that

relative to the NS condition, participants in the SD condition reported significantly

greater subjective sleepiness before the 12 h recognition test [t(52) = 5.59, p < .001].

There was no difference in SSS scores between the two sleep condition groups

before the immediate recognition test [t(52) = 0.72, p = .473] or the 7 d recognition

test [t(52) = 1.01, p = .316]. These results confirm that, unsurprisingly, participants in

the SD group felt more tired before the 12 h recognition test, following a night of SD.

Importantly, however, there was no correlation between 12 h SSS scores and

behavioural consolidation of positive, neutral or negative images for participants who

completed the SD condition [all p ≥ .490], suggesting that decreased subjective

alertness is unlikely to underlie any between-condition differences in recognition

performance.

3.3. Sleep Parameters

18

Sleep parameter data obtained from the higher BDI-II score NS group and the

low BDI-II score NS group are shown in Table 3. To compare sleep parameters

between the two depressive symptomatology groups in the NS condition we

conducted independent-samples t-tests on total sleep time, sleep efficiency, and

duration and proportion of sleep stages N1, N2, SWS and REM. It was revealed that

there was no significant difference between the two depressive symptomatology

groups in any of these sleep parameters [all p ≥ .124]. These results suggest that

sleep structure on the night of testing was comparable between participants with

higher BDI-II scores and participants with low BDI-II scores. Correlational analyses

examining the relationship between sleep parameters and post-sleep recognition

performance are available in Supplement 3.

Table 3. Sleep parameter data, separately for each depressive symptomatology group

Low BDI-II score group Higher BDI-II score group P-value

N1 (min)

N1 (%)

5.75 (1.51)

1.41 (0.36)

5.92 (0.79)

1.51 (0.18)

.927

.826

N2 (min)

N2 (%)

208.61 (9.49)

51.13 (1.81)

184.25 (15.94)

47.16 (2.55)

.187

.207

SWS (min)

SWS (%)

91.79 (5.47)

22.54 (1.30)

87.42 (5.67)

22.70 (1.16)

.585

.927

REM (min)

REM (%)

103.68 (8.73)

24.93 (1.51)

110.75 (8.90)

28.64 (1.79)

.577

.124

TST (min) 413.64 (14.16) 388.33 (22.53) .337

Efficiency (%) 94.53 (2.35) 91.78 (2.67) .445

Mean values are presented with SEM in parenthesis. Abbreviations: REM, Rapid Eye Movement sleep; N1 and N2, stages of non-REM sleep; SWS, Slow-Wave Sleep; TST, Total Sleep Time; BDI-II, Beck Depression Inventory.

19

3.4. Subjective Ratings of Valence and Arousal

Subjective valence and arousal ratings obtained from participants during the

learning phase of this study are shown in Table 4. The subjective valence and

arousal ratings provided by six participants - four from the low BDI-II score SD group,

one from the higher BDI-II score NS group, and one from the low BDI-II score NS

group - were not recorded due to a technical fault. The valence and arousal ratings

obtained from the remaining participants during the learning phase were analysed

using two (depressive symptomatology group: higher BDI-II score group, low BDI-II

score group) x two (sleep condition group: SD group, NS group) x three (image

valence: positive, neutral, negative) mixed-measures ANOVAs. Given that sleep

quality differed significantly between the two depressive symptomatology groups,

mean-centered PSQI score was included as a covariate in these analyses, and all

subsequent analyses where appropriate. A significant main effect of image valence

was revealed for both valence ratings [F(1.12, 48.24) = 634.33, p < .001, ηp² = 0.937,

Greenhouse-Geisser corrected] and arousal ratings [F(1.36, 58.42) = 88.28, p

< .001, ηp² = 0.672, Greenhouse-Geisser corrected]. The ANOVA revealed no other

significant main effects or interactions [all p ≥ .234], suggesting that between-

depressive symptomatology group and sleep condition group differences in

subjective emotional responses to images can reasonably be discounted as a source

of variability in recognition performance.

20

Table 4. Subjective ratings of valence and arousal obtained from learning phases, separately for each depressive symptomatology group, sleep condition group, and image valence

Low BDI-II score group Higher BDI-II score group

NS group SD group NS group SD group

Positive images

Valence 6.82 (0.16) 6.48 (0.26) 6.79 (0.20) 6.81 (0.17)

Arousal 4.33 (0.39) 4.42 (0.46) 4.36 (0.51) 4.51 (0.34)

Neutral images

Valence 5.03 (0.06) 5.04 (0.08) 5.03 (0.06) 4.97 (0.09)

Arousal 2.41 (0.36) 2.66 (0.44) 2.48 (0.45) 2.86 (0.35)

Negative images

Valence 2.27 (0.12) 2.61 (0.15) 2.44 (0.18) 2.57 (0.20)

Arousal 5.64 (0.35) 5.34 (0.65) 5.53 (0.55) 5.63 (0.30)

Mean values are presented with SEM in parenthesis. Abbreviations: BDI-II, Beck Depression Inventory; NS, Normal Sleep; SD, Sleep Deprivation.

Paired-samples t-tests comparing valence and arousal ratings between image

valence categories revealed that negative images were rated as significantly more

negative and greater in emotional arousal than both neutral images (valence [t(47) =

27.47, p <.001]; arousal [t(47) = 11.64, p < .001]) and positive images (valence [t

(47) = 26.31, p <.001]; arousal [t(47) = 4.86, p < .001]). Positive images were also

rated as significantly more positive and greater in emotional arousal than neutral

images (valence [t(47) = 20.36, p < .001]; arousal [t(47) = 14.72, p < .001]). These

results suggest that the stimuli evoked the expected emotional response from

participants.

Conflicting findings exist regarding the effects of sleep, and sleep deprivation,

on changes in subjective emotional reactivity (for review see: Tempesta, Socci, De

21

Gennaro, & Ferrara, 2017). Accordingly, we sought to investigate whether the sleep

condition groups and depressive symptomatology groups in this study differed with

regard to their change in valence and arousal ratings between the encoding session

and each of the recognition tests. These analyses are reported in Supplement 4.

Briefly, changes in emotionality ratings between the encoding and recognition tests

were not influenced by depressive symptomatology group, sleep condition or image

valence, and there were no interactions between these factors.

3.5. Immediate Recognition Performance

As described in Section 2.6.1, the results presented in this section focus

exclusively on recollection responses. Immediate recognition performance data is

available in Table 5. For unstandardized hit-rates and false alarm- rates at

immediate recognition see Supplement 5.

Table 5. Immediate recognition performance (d’), separately for each depressive symptomatology group, sleep condition group, and image valence

Low BDI-II score group Higher BDI-II score group

NS group SD group NS group SD group

Positive images 3.63 (0.23) 3.49 (0.27) 3.89 (0.23) 3.42 (0.22)

Neutral images 3.84 (0.19) 3.67 (0.29) 3.97 (0.16) 4.22 (0.22)

Negative images 4.07 (0.14) 3.69 (0.27) 3.88 (0.26) 3.68 (0.20)

Mean values are presented with SEM in parenthesis. Abbreviations: BDI-II, Beck Depression Inventory; NS, Normal Sleep; SD, Sleep Deprivation.

To test for any effects of depressive symptomatology group, sleep condition

group, or image valence on immediate recognition performance we conducted a two

22

(depressive symptomatology group: higher BDI-II score group, low BDI-II score

group) x two (sleep condition group: SD group, NS group) x three (image valence:

positive, neutral, negative) mixed-measures ANOVA on immediate recognition

scores (d’). The analysis showed a significant main effect of image valence [F(2, 98)

= 3.46, p = .035, ηp² = 0.066], which was driven by significantly greater recognition

performance for neutral images (3.92 ± 0.11) relative to positive images (3.60 ± 0.12)

[t(53) = 2.94, p = .005]. No other significant main effects or interactions were

revealed [p ≥ .290].

3.6. Behavioural Consolidation

As described in Section 2.6.1, to measure the consolidation processes which

occurred across the 12 h sleep or wake interval we created a measure which we

refer to as behavioural consolidation. Behavioural consolidation is calculated as the

difference between d’ scores at immediate and 12 h recognition tests [d’ 12 h

recognition – d’ immediate recognition], or immediate and 7 d recognition tests [d’ 7

d recognition – d’ immediate recognition]. Memory performance data for the 12 h and

7 d recognition tests is available in Supplement 6.

To investigate the effect of depressive symptomatology group, sleep condition

group, and image valence on memory consolidation we performed a two (depressive

symptomatology group: higher BDI-II score, low BDI-II score) x two (sleep condition

group: SD group, NS group) x three (image valence: positive, neutral, negative)

mixed-measures ANOVA on behavioural consolidation scores. This was done

separately for behavioural consolidation at 12 h and 7 d. Mean-centered PSQI score

was included as a covariate in these analyses. In accordance with the hypothesis,

23

we found a significant three-way interaction between depressive symptomatology

group, sleep condition, and image valence on 12 h behavioural consolidation scores

[F(2, 98) = 3.12, p = .049, ηp² = 0.060]. This interaction was also marginally

significant for 7 d behavioural consolidation [F(2, 98) = 2.79, p = .066, ηp² = 0.054]. A

significant main effect of sleep condition group on 12 h behavioural consolidation

[F(1, 49) = 6.10, p = .017, ηp² = 0.111] was also revealed, where participants in the

NS group exhibited greater overall behavioural consolidation than participants in the

SD group (NS group [-0.63 ± 0.10], SD group [-0.93 ± 0.14]). No other main effects

or interactions were revealed [all p ≥ .135]. Behavioural consolidation data are

shown in Fig 3.

To explore the significant depressive symptomatology group x sleep condition

group x image valence interaction, we performed two (sleep condition group: SD

group, NS group) x three (image valence: positive, neutral, negative) mixed-

measures ANOVAs on 12 h behavioural consolidation, separately for each

depressive symptomatology group. The ANOVAs revealed a significant main effect

of image valence in the higher BDI-II score group [F(2, 48) = 3.51, p = .038, ηp² =

0.128], which was driven by significantly greater consolidation of negative images (-

0.64 ± 0.18) relative to neutral images (-1.23 ± 0.15) [t(25) = 2.86, p = .008]. A

significant main effect of sleep condition group on 12 h behavioural consolidation

was also revealed in the higher BDI-II score group [F(1, 24) = 4.49, p = .045, ηp² =

0.158], where overall behavioural consolidation (i.e. regardless of image valence)

was greater in the NS group (-0.80 ± 0.15) than the SD group (-1.20 ± 0.22). No

other significant main effects or interactions were revealed [all p ≥ .114].

To test our hypothesis that SD would impair the consolidation of negative

memories in participants reporting mild-to-severe depressive symptoms, we

24

performed a priori independent-samples t-tests comparing behavioural consolidation

scores at 12 h between sleep condition groups for participants in the higher BDI-II

score group, separately for each image valence. In further support of our hypothesis,

it was revealed that the higher BDI-II score NS group exhibited significantly greater

behavioural consolidation of negative images at 12 h, relative to the higher BDI-II

score SD group [t(24) = 2.35, p = .028]. The t-tests also showed significantly greater

behavioural consolidation of neutral images at 12 h in the higher BDI-II score NS

group, relative to the higher BDI-II score SD group [t(24) = 2.79, p = .010]. Notably,

neither of these effects were observed in the low BDI-II score group [both p ≥ .447].

However, an additional a priori t-test showed that the low BDI-II score NS group

consolidated significantly more positive memories than the low BDI-II score SD

group [t(26) = 2.14, p = .042]. The consolidation of positive images was not

influenced by sleep condition for participants in the higher BDI-II score group [t(24) =

0.13, p = .895]. These findings suggest that SD may impair the consolidation of

negative and neutral memories for participants reporting mild-to-severe depressive

symptoms.

25

Fig 3. Mean behavioural consolidation [d’ 12 h or 7 d recognition – d’ immediate recognition] for recollection (R) responses separately for each depressive symptomatology group, sleep condition group, image valence, and recognition test. Error bars represent SEM. *, p < .05. Abbreviations: BDI-II, Beck Depression Inventory; NS, Normal Sleep; SD, Sleep Deprivation.

To examine differences in sleep- and sleep deprivation- related consolidation

between the two depressive symptomatology groups we conducted a two

(depressive symptomatology group: higher BDI-II score group, low BDI-II score

group) x three (image valence: positive, neutral, negative) mixed-measures ANOVA

on 12 h behavioural consolidation, separately for each sleep condition group. Mean-

centered PSQI score was included as a covariate in these analyses. The analysis

revealed a marginally significant depressive symptomatology group x image valence

interaction in the NS condition [F(2, 46) = 2.55, p = .089, ηp² = 0.100], which was

driven by greater consolidation of positive images in the low BDI-II score NS group,

relative to the higher BDI-II score NS group [t(24) = 2.13, p = .043]. The ANOVA

revealed no other significant main effects or interactions [all p ≥ .153].

26

4. Discussion

We investigated the effect of sleep, and SD, on the consolidation of

emotionally positive, neutral and negative images in participants reporting either

mild-to-severe depressive symptoms or minimal/no depressive symptoms. We

hypothesized that SD would impair the consolidation of negative emotional

memories, particularly in the higher BDI-II score group. Offering some support to our

hypothesis, we found that a night of SD, relative to a night of sleep, significantly

impaired the consolidation of negative and neutral images after 12 h in participants

reporting mild-to-severe depressive symptoms. Importantly, these effects were not

observed in participants reporting minimal/no depressive symptoms.

Research has shown that in healthy participants overnight sleep supports the

consolidation of negative emotional memories, relative to an equal duration of

daytime wakefulness (Payne et al., 2008) or overnight SD (Tempesta et al., 2015).

However, to our knowledge, our results are the first to demonstrate that the

detrimental effect of SD on the consolidation of negative and neutral memories is

more robust in participants reporting mild-to-severe depressive symptoms than

participants reporting minimal/no depressive symptoms.

A multitude of recent studies in healthy participants have demonstrated the

active and selective role of REM sleep in the processing and consolidation of

negative emotional memories (Groch et al., 2015; Nishida et al., 2009; Payne et al.,

2012; Wagner et al., 2001; Wiesner et al., 2015). REM sleep is thought to support

the consolidation of emotionally salient memories through the reactivation of, and

coherence between, neural regions implicated in emotional memory processing

27

during wake, in particular the amygdala, entorhinal cortex and medial prefrontal

cortex (Hutchinson & Rathore, 2015; Maquet et al., 1996; Nir & Tononi, 2010). By

inhibiting this pattern of neural activation associated with REM sleep, SD may have

impaired the processing and consolidation of negative emotional images in this

study. Indeed, a recent study found that healthy students selectively deprived of

REM sleep exhibited significantly less overnight consolidation of negative emotional

images, relative to participants selectively deprived of SWS (Wiesner et al., 2015),

supporting the interpretation that REM sleep suppression, rather than SD per se,

impaired the consolidation of negative emotional images in this study. Importantly,

however, this interpretation alone does not explain why the detrimental effect of SD

on negative emotional memory consolidation in the higher BDI-II score group was

absent in the low BDI-II score group.

One possible explanation for this effect is that REM sleep facilitated emotional

memory consolidation more readily in the higher BDI-II score group than the low

BDI-II score group. Indeed, in a recent study we found that the consolidation of

emotional memories during REM sleep was greater in participants reporting mild-to-

moderate depressive symptoms than participants reporting minimal/no depressive

symptoms (Harrington et al., 2018). However, in this study, negative memory

consolidation did not differ significantly between the higher BDI-II score NS group

and the low BDI-II score NS group. In addition, no relationship was observed

between REM sleep duration and recognition performance for negative images in

either depressive symptomatology group (see Supplement 3). Further research

including larger sample sizes are required to understand why the detrimental effect

of SD on negative emotional memory consolidation, relative to a night of sleep, is

more robust in participants reporting greater depressive symptoms. Moreover,

28

additional experiments are required to investigate whether this effect is specifically

linked to REM sleep. Such studies will benefit from the inclusion of a selective REM

SD condition, and a sleep group that sleeps in the laboratory under constant

observation by a researcher. Both conditions should include a comprehensive

polysomnographic assessment. Moreover, the current study lacks a control group

wherein which the participants are tested individually to prevent interactions amongst

participants which could have influenced the results.

Besides our hypothesis-related finding that SD impaired the consolidation of

negative emotional images in our higher BDI-II score group, our data also showed

that SD, relative to a night of sleep, impaired the consolidation of neutral images in

our higher BDI-II score group after 12 h. Considering the plentiful evidence that SWS

is critically involved in the consolidation of hippocampus-based, emotionally neutral

declarative memories (Stickgold, 2005), this finding is not surprising. However, it is

unclear why this effect was found in our higher BDI-II score group, and not in our low

BDI-II score group, especially considering that sleep-related consolidation of

emotionally neutral memories has been shown to be impaired in MDD patients

(Dresler et al., 2011, 2010; Göder et al., 2007; Nishida et al., 2016). Intriguingly,

however, although overnight motor memory consolidation was impaired in MDD

patients relative to controls in the Nishida et al. (2016) study, it was reported that

greater SWS duration predicted greater consolidation in participants with MDD, but

not control participants (Nishida et al., 2016). These findings corroborate a recent

study of ours, which showed that SWS supported the consolidation of neutral images

more readily in individuals reporting mild-to-moderate depressive symptoms than

individuals reporting minimal/no depressive symptoms (Harrington et al., 2018).

These findings suggest that the consolidation of neutral memories may be more

29

dependent on SWS in depressive individuals, and explains why SD, relative to a

night of sleep, had a detrimental effect on the consolidation of neutral images in our

higher BDI-II score group but not our low BDI-II score group. However, it is important

to consider that the differences in neutral memory consolidation between the NS and

SD groups may not be related specifically to SWS. Similarly, REM sleep in particular

may not subserve differences in negative memory consolidation. Indeed, both effects

could be related to the same process.

Our results also showed that overnight sleep supported the consolidation of

positive memories more readily in the low BDI-II score NS group than the higher

BDI-II score NS group. The effect of sleep on the consolidation of positive images

has been scarcely investigated, (however, see: Cairney, Durrant, Power, & Lewis,

2015; Harrington et al., 2018; Sterpenich et al., 2007) so it is difficult to explain why

sleep consolidated positive memories more effectively for individuals reporting

minimal/no depressive symptoms. One plausible explanation for this effect relates to

the mechanisms underlying emotional memory encoding. It has been proposed that

the amygdala activity elicited by encoding an emotional memory may act as an

‘emotional tag’, leading to the prioritised consolidation of these memories during

subsequent sleep intervals (Bennion, Payne, & Kensinger, 2015). Neuroimaging

research reveals that healthy participants exhibit greater amygdala responses to

positive stimuli (masked happy human facial expressions) than MDD patients

(Stuhrmann et al., 2013; Suslow et al., 2010). In this study, the positive images may

have elicited greater amygdala activity in the low BDI-II score group than the higher

BDI-II score group, which led to greater consolidation of positive images during sleep

in the low BDI-II score group. Further research employing fMRI is required to confirm

this interpretation.

30

In sum, the behavioural consolidation data obtained from this study suggests

that for participants reporting mild-to-severe depressive symptoms, sleep

preferentially preserved memory for negative and neutral items. Conversely, memory

for positive items was preserved across a night of sleep, relative to a night of SD, for

participants reporting minimal/no symptoms of depression. Although SD did not

exclusively suppress the consolidation of negative images in the higher BDI-II score

group (i.e. the consolidation of neutral images was also impaired), the ability of SD to

prevent negative memory consolidation in these individuals may have implications

for our understanding of SD therapy in MDD.

4.1. Implications

Growing clinical evidence supports the efficacy of SD therapy as an

antidepressant intervention for MDD (for reviews see: Benedetti & Colombo, 2011;

Wu & Bunney, 1990). Remarkably, SD has been reported to produce a clinical

improvement in depressive symptoms within a single 24 h period for approximately

60 % of MDD patients (Berger, van Calker, & Riemann, 2003; Leibenluft & Wehr,

1992; Wirz-Justice & Van Den Hoofdakker, 1999; Wu & Bunney, 1990). However,

this effect is typically transient, with approximately 80 % of SD responders

experiencing a return of depressive symptoms following subsequent recovery sleep

or daytime naps (Riemann, Wiegand, Lauer, & Berger, 1993).

Several researchers have attempted to explain the neurobiological basis of

this phenomenon (e.g. Brunner, Dijk, & Borbély, 1993; Wolf et al., 2015; Wu &

Bunney, 1990), however, no consensus has been reached on the mechanisms

underlying the therapeutic effect of SD in MDD. Nonetheless, in accordance with the

31

predictions of our Affect Tagging and Consolidation (ATaC) model (Harrington,

Pennington, & Durrant, 2017), we propose that SD may improve depressive

symptoms by suppressing the processing and consolidation of negative emotional

memories which typically occurs during REM sleep (Groch et al., 2015; Nishida et

al., 2009; Payne et al., 2012; Wagner et al., 2001; Wiesner et al., 2015). Our finding

that SD significantly impairs the consolidation of negative emotional images in

participants reporting mild-to-severe depressive symptoms may provide some

support for this notion. However, given that we utilized the BDI-II as our measure of

depressive symptom severity, which assesses depressive symptoms over the

preceding two weeks, it is impossible to determine whether the suppression of

negative emotional memory consolidation was associated with improvements in

depressive symptoms in this study. It is of course plausible that the impaired

emotional memory consolidation elicited by SD in the higher BDI-II score group

represents an effect which is independent of the therapeutic response to SD

observed in MDD patients. Future research is required to examine whether

impairments in negative emotional memory consolidation associated with SD,

relative to a night of sleep, correlate with improvements in depressive symptoms.

Such studies will require the use of clinically depressed MDD patients (confirmed

using structured clinical interview for DSM-5; SCID-5) and measures of current

depression severity that are sensitive to changes across a single SD interval such as

self-ratings and external ratings from psychiatrists who are blind to the experimental

conditions (e.g. Elsenga & van den Hoofdakker, 1982).

In addition to elucidating the mechanisms underlying SD therapy in MDD,

such research may also improve our understanding of the mechanisms of action of

several antidepressant medications. It is well reported that most antidepressant

32

drugs considerably inhibit REM sleep, thus increasing REM sleep latency or

decreasing REM sleep density and duration (Mayers & Baldwin, 2005; Murck et al.,

2003; Palagini et al., 2013; Steiger & Kimura, 2010; Thase, 2006; Vogel, Buffenstein,

Minter, & Hennessey, 1990). This finding lead researchers to propose that REM

sleep suppression may be an essential component of any effective form of MDD

therapy (Vogel, McAbee, Barker, & Thurmond, 1977). However, this notion is

contested by studies which report that a few pharmaceuticals effective in the

treatment of MDD such as the tricyclic antidepressant trimipramine, the

norepinephrine-dopamine reuptake inhibitor bupropion, and the serotonin reuptake

enhancer tianeptine, do not inhibit REM sleep (Murck et al., 2003; Nofzinger et al.,

1995; Sonntag et al., 1996). Although it is clear that REM sleep suppression is not

an exclusive mechanism by which antidepressants improve depressive symptoms,

the effect of pharmaceuticals which inhibit REM sleep on the overnight consolidation

of negative emotional memories is certainly one possible mechanism.

4.2. Limitations

Our analyses of SSS scores revealed that during the 12 h recognition test,

participants in the SD group reported greater subjective sleepiness than participants

in the NS group. Whilst this finding was expected, it could be argued that the

emotion-specific effects of SD on memory consolidation were due to reduced

alertness, rather than reduced sleep-dependent memory consolidation. However,

there is no inherent reason for tiredness in SD to impair negative and neutral but not

positive memory, so we do not believe this is the most parsimonious explanation of

our results. Indeed, 12 h SSS scores did not correlate with 12 h behavioural

33

consolidation of positive, neutral or negative images for participants who completed

the SD condition.

Similarly, it is plausible that the data from our 12 h recognition test may have

been influenced by the effect of circadian rhythms on cognitive performance and

memory formation (Gerstner & Yin, 2010; Schmidt, Collette, Cajochen, & Peigneux,

2007). Although we carried out the 7 d recognition test at the same time of day as

the immediate recognition test, it would have been impossible to match for time of

day between the immediate recognition test and the 12 h recognition test without

depriving our participants of a further 12 h of sleep. Nonetheless, as both groups

were tested at the same time of day throughout the study, circadian effects cannot

easily explain the differences we see between groups.

Our t-tests revealed significant differences in behavioural consolidation for

negative and neutral images between the SD and NS higher BDI-II score groups.

However, it is important to note that no significant interaction was found in the

preceding sleep condition x image valence ANOVA. Accordingly, the finding that SD

impairs the consolidation of negative and neutral memories in participants reporting

depressive symptoms should be regarded as preliminary, as opposed to solid

empirical evidence. Further studies utilizing larger sample sizes are required to

ascertain the impact of SD on emotional memory consolidation in depressive

participants.

Finally, between-group comparisons of our two depressive symptomatology

groups revealed that sleep quality in the month preceding the first session was

significantly lower in the higher BDI-II score group, relative to the low BDI-II score

group. This result is unsurprising, considering that sleep disturbances are reported

34

by up to 90 % of MDD patients (Baglioni et al., 2011; Palagini et al., 2013).

Importantly, sleep quality has been shown to influence emotional memory

processing (Tempesta, De Gennaro, Natale, & Ferrara, 2015). Nevertheless, we see

differences between our two higher BDI-II score groups which cannot be due to

these differences in sleep quality. Moreover, we included PSQI score as a covariate

in our analyses where appropriate, ensuring that our findings were not confounded

by between-group differences in sleep quality.

4.3. Conclusion

Our study has shown that a night of SD, relative to a night of sleep,

significantly impairs the consolidation of negative and neutral images after 12 h in

participants reporting mild-to-severe depressive symptoms. Notably, these effects

were not found in participants reporting minimal/no depressive symptoms. Our

findings may support the notion that the antidepressant response to SD in MDD is

related to the suppression of REM sleep- related negative emotional memory

processing and consolidation (Harrington et al., 2017). Further research including

larger sample sizes, MDD patients, and a more clinically robust measure of

depressive symptom severity should examine whether the suppression of negative

emotional memory consolidation associated with a night of SD, relative to a night of

sleep, correlates with improvements in depressive symptoms.

Acknowledgements

35

This research did not receive any specific grant from funding agencies in the public,

commercial, or not-for-profit sectors.

36

References

Ackermann, S., Hartmann, F., Papassotiropoulos, A., de Quervain, D. J., & Rasch, B. (2015). No Associations between Interindividual Differences in Sleep Parameters and Episodic Memory Consolidation. Sleep, 38(6). https://doi.org/10.5665/sleep.4748

Baran, B., Pace-Schott, E. F., Ericson, C., & Spencer, R. M. C. (2012). Processing of emotional reactivity and emotional memory over sleep. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 32, 1035–1042. https://doi.org/10.1523/JNEUROSCI.2532-11.2012

Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical aspects. New York: Hoeber.

Beck, A. T., Steer, R., & Brown, G. (1996). Beck depression inventory-II: manual (78(2)). San Antonio: The Psychological Corporation.

Benedetti, F., & Colombo, C. (2011). Sleep deprivation in mood disorders. Neuropsychobiology, 64(3), 141–151. https://doi.org/10.1159/000328947

Bennion, K. A., Payne, J. D., & Kensinger, E. A. (2015). Selective Effects of Sleep on Emotional Memory: What Mechanisms Are Responsible? Translational Issues in Psychological Science, 1(1), 79–88.

Berger, M., van Calker, D., & Riemann, D. (2003). Sleep and manipulations of the sleep – wake rhythm in depression. Acta Psychiatrica Scandinavica, 108(s418), 83–91. https://doi.org/10.1034/j.1600-0447.108.s418.17.x

Blatter, K., & Cajochen, C. (2007). Circadian rhythms in cognitive performance: Methodological constraints, protocols, theoretical underpinnings. Physiology and Behavior, 90(2–3), 196–208. https://doi.org/10.1016/j.physbeh.2006.09.009

Brunner, D. P., Dijk, D. J., & Borbély, A. A. (1993). Repeated partial sleep deprivation progressively changes in EEG during sleep and wakefulness. Sleep, 16(2), 100–113.

Buysse, D. J., Reynolds, C. F., Monk, T. H., Berman, S. R., & Kupfer, D. J. (1989). The Pittsburgh Sleep Quality Index: A New Instrument for Psychiatric Practice and Research. Psychiatry Res, 28, 193–213. https://doi.org/10.1016/0165-1781(89)90047-4

Cairney, S. A., Durrant, S. J., Power, R., & Lewis, P. A. (2015). Complementary Roles of Slow-Wave Sleep and Rapid Eye Movement Sleep in Emotional Memory Consolidation. Cerebral Cortex, 25(6), 1565–1575. https://doi.org/10.1093/cercor/bht349

Cellini, N., Torre, J., Stegagno, L., & Sarlo, M. (2016). Sleep before and after learning promotes the consolidation of both neutral and emotional information regardless of REM presence. Neurobiology of Learning and Memory, 133, 136–144. https://doi.org/10.1016/j.nlm.2016.06.015

Dobbins, I. G., Kroll, N. E. A., & Yonelinas, A. P. (2004). Dissociating familiarity from recollection using rote rehearsal. Memory and Cognition, 32(6), 932–944. https://doi.org/10.3758/BF03196871

Dresler, M., Kluge, M., Genzel, L., Schüssler, P., & Steiger, A. (2010). Impaired off-

37

line memory consolidation in depression. European Neuropsychopharmacology, 20(8), 553–561. https://doi.org/10.1016/j.euroneuro.2010.02.002

Dresler, M., Kluge, M., Pawlowski, M., Schüssler, P., Steiger, A., & Genzel, L. (2011). A double dissociation of memory impairments in major depression. Journal of Psychiatric Research, 45(12), 1593–1599. https://doi.org/10.1016/j.jpsychires.2011.07.015

Elsenga, S., & van den Hoofdakker, R. H. (1982). Clinical effects of sleep deprivation and clomipramine in endogenous depression. Journal of Psychiatric Research, 17(4), 361–374. https://doi.org/10.1016/0022-3956(82)90041-3

Everaert, J., Duyck, W., & Koster, E. H. W. (2014). Attention, interpretation, and memory biases in subclinical depression: a proof-of-principle test of the combined cognitive biases hypothesis. Emotion, 14(2), 331–340. https://doi.org/10.1037/a0035250

Everaert, J., Koster, E. H. W., & Derakshan, N. (2012). The combined cognitive bias hypothesis in depression. Clinical Psychology Review, 32(5), 413–424. https://doi.org/10.1016/j.cpr.2012.04.003

Fattahi Asl, A., Ghanizadeh, A., Mollazade, J., & Aflakseir, A. (2015). Differences of biased recall memory for emotional information among children and adolescents of mothers with MDD, children and adolescents with MDD, and normal controls. Psychiatry Research, 228, 223–227. https://doi.org/10.1016/j.psychres.2015.04.001

Gerstner, J. R., & Yin, J. C. (2010). Circadian rhythms and memory formation. Nature Reviews Neuroscience, 11(18), 577–588. https://doi.org/10.1038/nrn2881

Göder, R., Fritzer, G., Hinze-Selch, D., Huchzermeier, C., Koch, J. M., Seeck-Hirschner, M., & Aldenhoff, J. B. (2007). Sleep in major depression: Relation to memory performance and outcome after interpersonal psychotherapy. Neuropsychobiology, 55(1), 36–42. https://doi.org/10.1159/000103574

Gotlib, I. H., & Joormann, J. (2010). Cognition and depression: Current status and future directions. Annual Review of Clinical Psychology, 6, 285–312. https://doi.org/10.1146/annurev.clinpsy.121208.131305.Cognition

Groch, S., Wilhelm, I., Diekelmann, S., Sayk, F., Gais, S., & Born, J. (2011). Contribution of norepinephrine to emotional memory consolidation during sleep. Psychoneuroendocrinology, 36, 1342–1350. https://doi.org/10.1016/j.psyneuen.2011.03.006

Groch, S., Zinke, K., Wilhelm, I., & Born, J. (2015). Dissociating the contributions of slow-wave sleep and rapid eye movement sleep to emotional item and source memory. Neurobiology of Learning and Memory, 122, 122–130. https://doi.org/10.1016/j.nlm.2014.08.013

Harrington, M. O., Johnson, J. M., Croom, H. E., Pennington, K., & Durrant, S. J. (2018). The influence of REM sleep and SWS on emotional memory consolidation in participants reporting depressive symptoms. Cortex, 99, 281–295. https://doi.org/10.1016/j.cortex.2017.12.004

Harrington, M. O., Pennington, K., & Durrant, S. J. (2017). The “affect tagging and consolidation” (ATaC) model of depression vulnerability. Neurobiology of Learning and Memory, 140, 43–51. https://doi.org/10.1016/j.nlm.2017.02.003

38

Hindash, A. H. C., & Amir, N. (2011). Negative Interpretation Bias in Individuals with Depressive Symptoms. Cognitive Therapy and Research, 36(5), 502–511. https://doi.org/10.1007/s10608-011-9397-4

Hoddes, E., Dement, W., & Zarcone, V. (1972). The development and use of the Stanford sleepiness scale (SSS). Psychophysiology, 9, 150. https://doi.org/10.1111/j.1469-8986.1972.tb00747.x

Homan, R. W., Herman, J., & Purdy, P. (1987). Cerebral location of international 10-20 system electrode placement. Electroencephalography and Clinical Neurophysiology, 66, 376–382. https://doi.org/10.1016/0013-4694(87)90206-9

Howe, M. L., & Malone, C. (2011). Mood-congruent true and false memory: effects of depression. Memory, 19(2), 192–201. https://doi.org/10.1080/09658211.2010.544073

Hutchinson, I., & Rathore, S. (2015). The role of REM sleep theta activity in emotional memory. Frontiers in Psychology, 6(1439), 1–15. https://doi.org/10.3389/fpsyg.2015.01439

Iber, C., Ancoli-Israel, S., Chesson, A., & Quan, S. F. (2007). The AASM manual for the scoring of sleep and associated events rules, terminology, and technical specifications. Westchester, IL: American Academy of Sleep Medicine.

Kensinger, E. A. (2004). Remembering emotional experiences: the contribution of valence and arousal. Reviews in the Neurosciences, 15, 241–251. https://doi.org/10.1515/REVNEURO.2004.15.4.241

Kyriacou, C. P., & Hastings, M. H. (2010). Circadian clocks: genes, sleep, and cognition. Trends in Cognitive Sciences, 14(6), 259–67. https://doi.org/10.1016/j.tics.2010.03.007

Lang, P. J., Greenwald, M. K., Bradley, M. M., & Hamm, A. O. (1993). Looking at pictures: affective, facial, visceral, and behavioral reactions. Psychophysiology, 30, 261–273. https://doi.org/10.1111/j.1469-8986.1993.tb03352.x

Leibenluft, E., & Wehr, T. A. (1992). Is sleep deprivation useful in the treatment of depression? The American Journal of Psychiatry, 149(2), 159–168. https://doi.org/10.1176/ajp.149.2.159

Levendowski, D. J., Ferini-Strambi, L., Gamaldo, C., Cetel, M., Rosenberg, R., & Westbrook, P. R. (2017). The accuracy, night-To-night variability, and stability of frontopolar sleep electroencephalography biomarkers. Journal of Clinical Sleep Medicine, 13(6), 791–803. https://doi.org/10.5664/jcsm.6618

Lucey, B. P., Mcleland, J. S., Toedebusch, C. D., Boyd, J., Morris, J. C., Landsness, E. C., … Holtzman, D. M. (2016). Comparison of a single-channel EEG sleep study to polysomnography. Journal of Sleep Research, 25(6), 625–635. https://doi.org/10.1111/jsr.12417

Macmillan, N. A., & Creelman, C. D. (2004). Detection Theory: A User’s Guide. Psychology press.

Macmillan, N. A., & Kaplan, H. L. (1985). Detection theory analysis of group data: estimating sensitivity from average hit and false-alarm rates. Psychological Bulletin, 98(1), 185–199. https://doi.org/10.1037/0033-2909.98.1.185

Maquet, P., Péters, J., Aerts, J., Delfiore, G., Degueldre, C., Luxen, A., & Franck, G. (1996). Functional neuroanatomy of human rapid-eye-movement sleep and

39

dreaming. Nature, 383, 163–166. https://doi.org/10.1038/383163a0McGaugh, J. L. (2004). The amygdala modulates the consolidation of memories of

emotionally arousing experiences. Annual Review of Neuroscience, 27, 1–28. https://doi.org/10.1146/annurev.neuro.27.070203.144157

Mickley, K. R., & Kensinger, E. A. (2008). Emotional valence influences the neural correlates associated with remembering and knowing. Cognitive, Affective & Behavioral Neuroscience, 8(2), 143–152. https://doi.org/10.3758/CABN.8.2.143

Morgenthaler, J., Wiesner, C. D., Hinze, K., Abels, L. C., Prehn-Kristensen, A., & Göder, R. (2014). Selective REM-sleep deprivation does not diminish emotional memory consolidation in young healthy subjects. PloS One, 9(2), e89849. https://doi.org/10.1371/journal.pone.0089849

Newby, J. M., Lang, T., Werner-Seidler, A., Holmes, E., & Moulds, M. L. (2014). Alleviating distressing intrusive memories in depression: A comparison between computerised cognitive bias modification and cognitive behavioural education. Behaviour Research and Therapy, 56(1), 60–67. https://doi.org/10.1016/j.brat.2014.03.001

Nir, Y., & Tononi, G. (2010). Dreaming and the brain: from phenomenology to neurophysiology. Trends in Cognitive Sciences, 14(2), 88–100. https://doi.org/10.1016/j.tics.2009.12.001

Nishida, M., Nakashima, Y., & Nishikawa, T. (2016). Slow sleep spindle and procedural memory consolidation in patients with major depressive disorder. Nature and Science of Sleep, 8, 63–72. https://doi.org/10.2147/NSS.S100337

Nishida, M., Pearsall, J., Buckner, R. L., & Walker, M. P. (2009). REM sleep, prefrontal theta, and the consolidation of human emotional memory. Cerebral Cortex, 19(5), 1158–1166. https://doi.org/10.1093/cercor/bhn155

Payne, J. D., Chambers, A. M., & Kensinger, E. A. (2012). Sleep promotes lasting changes in selective memory for emotional scenes. Frontiers in Integrative Neuroscience, 6, 108. https://doi.org/10.3389/fnint.2012.00108

Payne, J. D., Stickgold, R., Swanberg, K., & Kensinger, E. A. (2008). Sleep preferentially enhances memory for emotional components of scenes. Psychological Science, 19(8), 781–788. https://doi.org/10.1111/j.1467-9280.2008.02157.x

Riemann, D., Wiegand, M., Lauer, C. J., & Berger, M. (1993). Naps after total sleep deprivation in depressed patients: Are they depressiogenic? Psychiatry Research, 49(2), 109–120. https://doi.org/10.1016/0165-1781(93)90099-3

Schmidt, C., Collette, F., Cajochen, C., & Peigneux, P. (2007). A time to think: Circadian rhythms in human cognition. Cognitive Neuropsychology, 24(7), 755–789. https://doi.org/10.1080/02643290701754158

Sterpenich, V., Albouy, G., Boly, M., Vandewalle, G., Darsaud, A., Balteau, E., … Maquet, P. (2007). Sleep-related hippocampo-cortical interplay during emotional memory recollection. PLoS Biology, 5, 2709–2722. https://doi.org/10.1371/journal.pbio.0050282

Stickgold, R. (2005). Sleep-dependent memory consolidation. Nature, 437(7063), 1272–8. https://doi.org/10.1038/nature04286

Stuhrmann, A., Dohm, K., Kugel, H., Zwanzger, P., Redlich, R., Grotegerd, D., …

40

Dannlowski, U. (2013). Mood-congruent amygdala responses to subliminally presented facial expressions in major depression: associations with anhedonia. Journal of Psychiatry & Neuroscience, 38(4), 249–258. https://doi.org/10.1503/jpn.120060

Suslow, T., Konrad, C., Kugel, H., Rumstadt, D., Zwitserlood, P., Schöning, S., … Dannlowski, U. (2010). Automatic mood-congruent amygdala responses to masked facial expressions in major depression. Biological Psychiatry, 67(2), 155–160. https://doi.org/10.1016/j.biopsych.2009.07.023

Tempesta, D., De Gennaro, L., Natale, V., & Ferrara, M. (2015). Emotional memory processing is influenced by sleep quality. Sleep Medicine, 16(7), 862–870. https://doi.org/10.1016/j.sleep.2015.01.024

Tempesta, D., Socci, V., De Gennaro, L., & Ferrara, M. (2017). Sleep and Emotional Processing. Sleep Medicine Reviews, 1–13. https://doi.org/10.1016/j.smrv.2017.12.005

Wagner, U., Gais, S., & Born, J. (2001). Emotional memory formation is enhanced across sleep intervals with high amounts of rapid eye movement sleep. Learning & Memory, 8(2), 112–119. https://doi.org/10.1101/lm.36801

Wiesner, C. D., Pulst, J., Krause, F., Elsner, M., Baving, L., Pedersen, A., … Göder, R. (2015). The effect of selective REM-sleep deprivation on the consolidation and affective evaluation of emotional memories. Neurobiology of Learning and Memory, 122, 131–141. https://doi.org/10.1016/j.nlm.2015.02.008

Wirz-Justice, A., & Van Den Hoofdakker, R. H. (1999). Sleep deprivation in depression: What do we know, where do we go? Biological Psychiatry. https://doi.org/10.1016/S0006-3223(99)00125-0

Wolf, E., Kuhn, M., Norman, C., Mainberger, F., Maier, J. G., Maywald, S., … Nissen, C. (2016). Synaptic plasticity model of therapeutic sleep deprivation in major depression. Sleep Medicine Reviews, 30. https://doi.org/10.1016/j.smrv.2015.11.003

Wu, J. C., & Bunney, W. E. (1990). The biological basis of an antidepressant response to sleep deprivation and relapse: Review and hypothesis. American Journal of Psychiatry, 147, 14–21.

Yonelinas, A. P. (2002). The nature of recollection and familiarity: A review of 30 years of research. Journal of Memory and Language, 46(3), 441–517. https://doi.org/10.1006/jmla.2002.2864