34

TITLE:

Overnight retention of emotional memories is influenced by BDNF Val66Met but not 5-HTTLPR

AUTHORS:

Marcus O. Harrington[footnoteRef:1]a*, Kristel Klausa, Mariliis Vahtb, Jaanus Harrob, Kyla Penningtona, Simon J. Durranta [1: Present address: Department of Psychology, University of York, Heslington, York, YO10 5DD, United Kingdom]

AUTHOR AFFILIATIONS:

a School of Psychology, College of Social Science, University of Lincoln.

Sarah Swift Building, Brayford Wharf East, Lincoln, Lincolnshire, LN5 7AY, United Kingdom

b Division of Neuropsychopharmacology, Department of Psychology, University of Tartu. Ravila 14A, 50411, Tartu, Estonia

* CORRESPONDING AUTHOR:

Department of Psychology, University of York. Heslington, York, YO10 5DD, United Kingdom

Email: marcus.harrington@york.ac.uk

RUNNING TITLE:

Genes, sleep, and emotional memory

Abstract

Emotional memory may be modulated by BDNF Val66Met and 5-HTTLPR polymorphisms. However, the influence of these genetic variants on the overnight retention of emotional memories has not been investigated in humans. Thirty-six healthy female students were selected to participate in this study based on 5-HTTLPR genotype status (L’/L’, L’/S’, S’/S’). Participants were also genotyped for BDNF Val66Met (Val/Val, Met carriers). We measured recognition performance for positive, neutral and negative images before and after overnight sleep. We found a significant interaction between BDNF Val66Met genotype group and image valence on post-sleep recognition performance. This interaction was driven by greater memory for negative and positive images, relative to neutral images, in Met carriers. We also found that longer Rapid Eye Movement (REM) sleep duration predicted greater post-sleep recognition performance for negative images in Met carriers, but not in Val homozygotes. We observed no influence of 5-HTTLPR polymorphisms on post- sleep recognition performance for positive, neutral or negative images. Our findings support a modulatory role for BDNF Val66Met in overnight emotional memory retention in females. We discuss the implications of this finding for understanding the influence of BDNF Val66Met on depression vulnerability.

Keywords: affective memory; brain-derived neurotrophic factor; depression; rapid eye movement (REM) sleep; serotonin transporter

1. Introduction

A wealth of psychological research demonstrates that emotionally salient memories are better remembered than their neutral or mundane counterparts [1,2]. This phenomenon is more robust when remembering occurs after several hours or days, as opposed to minutes [3–7]. These findings suggest that enhanced remembering of emotionally salient memories is mediated, at least in part, by the privileged consolidation of emotional memories [8].

Sleep is centrally involved in memory consolidation processes [9–12]. Importantly, sleep prioritises the consolidation of emotionally salient memories, relative to neutral memories [8,13–16]. Rapid Eye Movement (REM) sleep, in particular, plays an active and selective role in negative emotional memory consolidation [17–19].

REM sleep is associated with increased activity in several limbic, paralimbic and cortical regions associated with waking emotional memory processing, including the amygdala, medial prefrontal cortex, and entorhinal cortex [20–22]. Coherence between these structures during REM sleep may reflect the processing and consolidation of emotional memories [23,24]. REM sleep is thought to selectively consolidate memories which elicit amygdala activity during encoding [25,26]. Indeed, several neuroimaging studies have demonstrated that emotional stimuli evokes a greater amygdala response than neutral stimuli [27–30].

Amygdala reactivity to negative emotional stimuli may be influenced by a common Variable Number Tandem Repeat (VNTR) polymorphism at site 5-HTTLPR in the SLC6A4 gene which codes for the 5-HT transporter. Specifically, carriers of the Short (S) allelic variant within the 5-HTTLPR, which is associated with decreased transcriptional efficiency of the promoter [31], have been shown to exhibit greater amygdala reactivity to negative emotional faces [32–34], and other threatening stimuli [35], relative to Long (L) allele homozygotes. The 5-HTTLPR polymorphism has also been linked to fear conditioning in humans [36,37], suggesting an influence of the 5-HTTLPR on emotional memory [38]. However, to our knowledge, the effect of the 5-HTTLPR on the overnight retention of emotional memories has not yet been investigated. Given that the 5-HTTLPR influences amygdala reactivity to emotional stimuli [32–35], which is believed to modulate REM sleep- related emotional memory consolidation [25,26], it is plausible that the 5-HTTLPR may impact post-sleep recognition performance for emotional memories encoded the previous evening.

Although most research linking emotional cognition to the 5-HTTLPR have compared groups of S allele carriers with non-carriers, S alleles may have additive effects on emotional reactivity in homozygotes. Indeed, individuals homozygous for the S allele have been shown to exhibit greater emotional reactivity to a laboratory stressor [39], and are more likely to become depressed in response to stressful life events [40], than heterozygotes and L allele homozygotes.

Besides SLC6A4, another gene that may influence the overnight retention of emotional memories is the brain-derived neurotrophic factor (BDNF) gene, which encodes the BDNF protein. A recent study in rats demonstrated that BDNF, along with other genes which play key roles in long-term synaptic plasticity, is upregulated in the hippocampus during REM sleep [41]. Specifically, it was shown that BDNF expression in regions of the ventral and dorsal hippocampus more than doubled during rebound REM sleep in REM sleep deprived rats, relative to typical REM sleep in control animals. BDNF expression in the amygdala and hippocampus has been found to be critical for memory consolidation [42,43]. Indeed, inhibiting BDNF signalling in the amygdala has been shown to impair the consolidation of fear conditioning and fear extinction [44–46]. Collectively, these findings suggest that the BDNF gene may play a crucial role in REM sleep- related memory retention.

A common Single Nucleotide Polymorphism (SNP) has been identified in the coding exon of the human BDNF gene at position 66 (Val66Met). BDNF Val66Met modulates human memory [47,48] and hippocampal volume [49,50; however see 51,52], and may exert these effects by influencing intracellular trafficking and activity-dependent secretion of BDNF [51]. It has been shown that BDNF Val66Met interacts with sleep to influence post-sleep learning ability in healthy participants [52]. Furthermore, BDNF Val66Met genotype status has been shown to interact with cortisol response to exercise in modulating emotional memory performance [53]. Similarly to the 5-HTTLPR, BDNF Val66Met has also been linked to emotional reactivity in the amygdala and hippocampus. Specifically, relative to Val allele homozygotes, carriers of the Met allele exhibit stronger activations in these structures in response to emotional stimuli [54–56]. However, to our knowledge, the effects of BDNF Val66Met on the overnight retention of emotional memories has not yet been studied in humans. Given that the BDNF gene is upregulated in the hippocampus during REM sleep [41], a sleep stage associated with the selective consolidation of emotionally salient memories [17–19], it is plausible that BDNF Val66Met may affect overnight emotional memory retention.

The current study aimed to investigate the respective influences of the 5-HTTLPR and BDNF Val66Met on recognition performance for positive, neutral and negative images before and after overnight sleep. Existing research implies that both the 5-HTTLPR and BDNF Val66Met may modulate the overnight retention of emotional memories; however, there is a lack of direct evidence to support this to date. We hypothesised that post-sleep recognition performance for emotional images would be greater than post-sleep recognition performance for neutral images. We also predicted that post-sleep recognition performance for negative images would be influenced by both 5-HTTLPR and BDNF Val66Met genotype status. Specifically, we thought that the 5-HTTLPR S allele, and the BDNF Val66Met Met allele, would be associated with greater post-sleep recognition performance for negative images relative to neutral images. Finally, we predicted that greater REM sleep duration would be associated with greater post-sleep recognition performance for negative images, and that this effect would also be modulated by 5-HTTLPR and BDNF Val66Met genotype status. Most existing research exploring the relationship between emotional memory and the 5-HTTLPR and BDNF Val66Met has focused exclusively on comparisons between memory for negative and neutral items. Accordingly, we refrained from making predictions about the influence of genotype on the overnight retention of positive memories.

2. Materials and Methods

2.1. Participants

One hundred female student volunteers were genotyped for BDNF Val66Met, and two functional polymorphisms of the SLC6A4 gene – the 5-HTTLPR VNTR polymorphism and the rs25531 SNP. Participants were then grouped according to 5-HTTLPR genotype status (L’/L’, L’/S’, S’/S’). The rare, low expressing rs25531 SNP Lg alleles were classified as S’, in congruence with previous studies [57–59]. In the 100 volunteers who were genotyped, S’/S’ was the least frequent 5-HTTLPR genotype. As such, all S’/S’ carriers were invited to take part in this study. Eleven 5-HTTLPR S’/S’ carriers accepted our invitation and successfully completed the study. We then recruited at least 11 participants from the other two 5-HTTLPR genotype groups. The researcher who collected behavioural data was blind to the recruitment strategy and genotypic data. Using this sampling method, we recruited a total of 36 participants for this study (see Table 1 for participants’ demographic data). The same 36 participants were used to examine both genotypes (i.e. 5-HTTLPR and BDNF Val66Met). We chose to exclusively test females to eliminate gender as a potential confounding factor [60–62]. We selected participants based on 5-HTTLPR genotype status, rather than BDNF Val66Met genotype status, because 5-HTTLPR S’/S’ carriers were rarer than BDNF Met carriers in the sample that was genotyped (S’/S’ = 20, Met carriers = 28). Accordingly, selecting participants per 5-HTTLPR genotype status, rather than BDNF Val66Met genotype status, increased the likelihood of recruiting a reasonable number of participants in each genotype group (i.e. 5-HTTLPR: L’/L’, L’/S’, S’/S’; BDNF: Val/Val, Met carriers).

Participants were required to have no history of sleep, neurological, endocrine or psychiatric disorders, as assessed through self-report, and were required to be free of long-term medication (except for the contraceptive pill). Participants were asked to abstain from alcohol, caffeine, and other drugs for 24 h prior to testing. Participants were paid £30 for their participation. Each participant gave informed consent for this study, which was approved by the School of Psychology Research Ethics Committee at the University of Lincoln.

2.2. Genotyping

Table 1. Summary of participants’ demographic data

Mean values are presented with SEM in parenthesis. Higher scores equate to higher levels of depressive symptoms or sleep quality impairment. Abbreviations: BDI-II, Beck Depression Inventory; PSQI, Pittsburgh Sleep Quality Index. *Of the 11 Met carriers, 10 were heterozygotes (i.e. Val/Met) and one was homozygous for the Met allele (i.e. Met/Met).

Genetic material was collected from saliva samples using the Oragene® collection kit (OG-500; DNA Genotek Inc., Kanata, Canada). Participants were paid £10 for providing a saliva sample. DNA extraction from saliva samples was performed using Oragene® prepIT L2P (DNA Genotek Inc.; http://www.dnagenotek.com), per the manufacturer’s protocol. Genotyping was performed blind to all phenotypic data. All DNA samples were successfully genotyped.

Genotyping for triallelic classification was performed in accordance with previous research [63], as described elsewhere [64]. Regarding BDNF Val66Met, participants were grouped as either Met carriers (i.e. Val/Met, Met/Met) or Val homozygotes (i.e. Val/Val) for data analysis, consistent with existing research [52,65]. For details about 5-HTTLPR and BDNF Val66Met genotyping see Supplementary Material 1.

2.3. Stimuli

Four hundred and eighty images were selected from the International Affective Picture System (IAPS) [66]. IAPS images range from photographs depicting happy relationships and appetizing food, to everyday scenes, to images of injury, violence and famine, 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 either a “positive”, “neutral” or “negative” emotion category, each containing 160 images. Pairwise comparisons demonstrated that there was a significant difference in the mean IAPS valence rating (i.e. the valence ratings from the IAPS database) between all emotion categories (positive: [6.98 ± 0.05], neutral: [5.05 ± 0.02], negative: [2.94 ± 0.05]; all pairwise p < .001). There was also a significant difference in the mean IAPS arousal rating (i.e. the arousal ratings from the IAPS database) for the neutral category relative to the positive and negative categories (positive: [5.27 ± 0.06], neutral: [3.34 ± 0.05], negative: [5.27 ± 0.05]; both p < .001), however, we ensured that there was no significant difference in the mean arousal rating between the positive and negative categories [p = .988].

Selected images were divided into four equal sets of 120 (40 positive, 40 neutral, 40 negative), which were matched as closely as possible for IAPS valence and arousal ratings. During learning phases, participants viewed two of these four sets of images which served as the targets during subsequent recognition tests (see Section 2.5 for details about learning phases and recognition tests). During the immediate recognition test (see Section 2.4 for information about the study protocol) participants viewed one of the two sets of images that they viewed during the learning phase (targets) intermixed with one of the two remaining sets of 120 images (foils). During the delayed recognition test participants viewed the other one of the two sets of images that they viewed during the learning phase, intermixed with the other remaining set of foil images. The image sets used for encoding and recognition tests were counterbalanced across 5-HTTLPR genotype groups (i.e. each combination of image sets was viewed an equal number of times by each 5-HTTLPR genotype group), and each image set appeared an equal number of times in learning phases and recognition tests.

2.4. Experimental protocol

The experimental protocol is summarised in Fig 1. Participants arrived at the Lincoln Sleep and Cognition Laboratory at 8 PM. They were first asked to read an information sheet outlining the study procedure, and invited to ask any questions. This was followed by completion of a written consent form. Participants were then asked to complete the Beck Depression Inventory (BDI-II), a self-report questionnaire commonly used in cognitive research [67–69] to measure depressive symptom severity with high reliability and validity in both healthy and depressed samples [70]. Participants then completed the Pittsburgh Sleep Quality Index (PSQI), a self-report questionnaire commonly used in sleep research [71–73] to measure subjective sleep quality over the preceding month [74].

The learning phase began at approximately 8.30 PM (see Section 2.5 for details about learning phases and recognition tests), and was immediately followed by completion of the Stanford Sleepiness Scale (SSS) [75] and a 5 min version of the Psychomotor Vigilance Task (PVT) [76], as described earlier [77]. The SSS is a self-report questionnaire which measures subjective sleepiness on a scale from one (alert, wide awake) to seven (fighting sleep). The PVT measures reaction time to the presentation of a known stimulus, and is commonly used in sleep research to measure objective alertness [19,78,79]. The immediate recognition test started at approximately 9 PM, immediately after completion of the PVT, and lasted approximately 30 min. Participants were then given opportunity to prepare for bed, before electrodes and a patient unit were attached for full polysomnography (see Section 2.6.2 for details about polysomnography recording). Lights were turned out at 10.30 PM, and participants were awoken the following morning at 6.30 AM, providing a sleep opportunity of 8 h.

Fig 1. Pictorial representation of study protocol. All participants began a learning phase at 8.30 PM which was followed by an immediate recognition test. Electrodes were then attached for polysomnography (PSG) monitoring. Participants then slept for 8 h between 10.30 PM and 6.30 AM. They were then given 30 min to recover from sleep inertia, before completing a delayed recognition test. Abbreviations: REM, Rapid Eye Movement sleep; N1 – N3, stages of non-REM sleep.

Once awoken, participants were given 30 min to recover from sleep inertia [13,72,80], during which the electrodes and patient unit were detached. Participants were then asked to complete the SSS and PVT, immediately followed by the delayed recognition test. Finally, participants were given a debrief which included an explanation of the purpose of the experiment, and were invited to ask questions and give feedback. The study protocol was carried out blind to genotypic data.

2.5. Learning phases and recognition tests

The learning phases and recognition tests are summarised in Fig 2. The learning phase began with participants viewing a black screen with a white central fixation cross for 500 ms. The first image was then presented in the middle of the screen for 1000 ms (image size: 15 cm x 11 cm). Participants were then instructed to rate the image for emotional valence on a scale from one (very negative) to nine (very positive) using corresponding keys. After providing their valence rating, participants were instructed to rate the image for emotional arousal on a scale from one (low arousal) to nine (high arousal). Following submission of the participants’ arousal rating, the black screen with the white central fixation cross reappeared, followed by the second image. This pattern continued until all 240 target images had been viewed and rated. Participants were instructed to provide their emotion ratings quickly and spontaneously, and were made aware that they would be tested on their ability to recognise the images before and after the overnight sleep interval.

The recognition tests followed the same structure as the learning phases. However, during the recognition tests, participants were required to make a remember/know/new judgement for each image using corresponding keys (i.e. R, K or N), before providing their emotion ratings. Participants were instructed to provide a remember (R) judgement if they could consciously recollect seeing that specific image during the learning phase. They were asked to provide a know (K) judgement if they felt that the image was familiar, but could not consciously recollect details about its previous occurrence. They were asked to provide a new (N) judgement if they felt that they had not seen the image during the learning phase. Remembering reflects recollection of the episodic details of an item. Conversely, knowing reflects familiarity in the absence of recollection [81]. These different types of memory are thought to be underpinned by separate neural processes [82,83]. In this study, we were particularly interested in studying recollection (R), rather than familiarity (K; see Section 2.7.1 for more information). We instructed participants to provide their memory judgements as quickly and accurately as possible. Images were presented in a random order during the learning phase and the recognition tests. Emotion ratings and memory judgements were not subjected to time constraints.

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 a memory rating to indicate whether they could recollect seeing the image before (R response), they thought the image was familiar (K response), or they thought the image was new (N response).

2.6. Equipment

2.6.1. Experimental task

Stimulus presentation and data collection used custom-written scripts running in SuperLab 5TM (Cedrus Corp, San Pedro, CA) on a Toshiba Satellite laptop with a 15.6 ” screen. Participant responses were recorded using the laptop keyboard.

2.6.2. Polysomnography

Overnight sleep monitoring was carried out at the University of Lincoln Sleep and Cognition Laboratory using an Embla© N7000 polysomnography system. Silver-silver chloride (Ag–AgCl) electrodes were attached using EC2© electrode cream after the scalp was cleaned with NuPrep© exfoliating agent. Scalp electrodes were attached at six standard locations according to the international 10 - 20 system [84]: C3, C4, F3, F4, O1, and O2 - each referenced to the contralateral mastoid (A1 and A2). Left and right electrooculogram, left, right and upper chin electromyogram, and a ground electrode were also attached. In addition, the Patient Unit was attached to record physiological signals including movement and respiration. All electrodes were verified to have a connection impedance of < 5 000 Ω. All signals were digitally sampled at a rate of 200 Hz.

2.7. Data analysis

2.7.1. 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 [85,86]. Therefore, we made this the primary focus of our investigation. 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 [87]), a signal detection process widely used in memory studies to account for response bias [88]. This was done separately for positive, neutral and negative image trials for each participant.

Our main analyses were two-way mixed measures ANOVAs, with image valence (positive, neutral, negative) and either 5-HTTLPR genotype group (S’/S’, S’/L’, L’/L’) or BDNF Val66Met genotype group (Val/Val, Met carriers) as the independent variables, and immediate or delayed recognition performance as the dependent variable. Significant effects were explored using one-way ANOVAs and t-tests as appropriate. Interactions between 5-HTTLPR and BDNF Val66Met genotype groups were not examined due to the relatively small sample size in this study. We predicted that there would be a significant main effect of emotion on delayed recognition performance, but not immediate recognition performance, where emotional images would be better remembered than neutral images. We also expected there to be a significant interaction between image valence and both 5-HTTLPR and BDNF Val66Met genotype group on delayed recognition performance. We hypothesised that these interactions would be driven by greater recognition performance for negative images in 5-HTTLPR S’ allele homozygotes, and BDNF Val66Met Met allele carriers.

2.7.2. Sleep data analysis

Each participant’s sleep data was divided into 30 s epochs and independently scored by two trained sleep researchers (level of agreement: mean = 85.25 %; SD = 7.81 %) using REM Logic © 1.1 according to standardised criteria [89]. 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. Additionally, REMs occurring during REM sleep phases were scored in order to calculate REM density and the total number of REMs. Each eye movement which was detectable above the background noise, exhibited a rapid time course, and appeared simultaneously on both right and left electrooculogram channels was counted, regardless of the amplitude of the eye movement [90–92]. REM density was calculated as the total number of REMs / REM sleep duration (mins).

Given that REM sleep is implicated in emotional memory consolidation [17–19], we investigated the relationship between REM sleep duration and post-sleep recognition performance, separately for each emotion category, using planned parametric correlation. To examine the influence of the 5-HTTLPR and BDNF Val66Met on this relationship, the analysis was performed separately for each genotype group (i.e. 5-HTTLPR: L’/L’, L’/S’, S’/S’; BDNF Val66Met: Val/Val, Met carriers). We predicted that greater REM sleep duration would be associated with greater post-sleep recognition performance for negative images, particularly in 5-HTTLPR S’ allele homozygotes and BDNF Val66Met Met allele carriers.

3. Results

3.1. Participant demographics

The participants’ demographic characteristics are shown in Table 1. Observed 5-HTTLPR and BDNF Val66Met allele frequencies did not differ significantly from Hardy-Weinberg equilibrium in the wider pool of 100 participants who were genotyped for this study [5-HTTLPR: X2 = 0.01, p > .05; BDNF Val66Met: X2 = 0.79, p > .05]. Of note, the ratio of BDNF Val66Met Met carriers to Val homozygotes was not even across the 5-HTTLPR genotype groups. Indeed, Met carriers accounted for 3 out of 11 L’/L’ participants, 8 out of 14 S’/L’ participants, and 0 out of 11 S’/S’ participants.

To compare demographic characteristics between the three 5-HTTLPR genotype groups we conducted one-way between-subjects ANOVAs with BDI-II score, PSQI score, and age as the dependent variables. There was a marginally significant age difference between the 5-HTTLPR genotype groups [F(2, 33) = 2.61, p = .089, ηp² = 0.137], however, the mean age difference between the oldest and the youngest group was very modest (1.09 years; see Table 1). Importantly, there were no significant differences between the 5-HTTLPR genotype groups regarding BDI-II score [F(2, 33) = 1.24, p = .301, ηp² = 0.070] or PSQI score [F(2, 33) = 0.93, p = .406, ηp² = 0.053].

Independent-samples t-tests comparing demographic characteristics between BDNF Val66Met genotype groups revealed no significant differences between the BDNF Val66Met genotype groups regarding age [t(34) = 0.19, p = .849], BDI-II score [t(34) = 0.82, p = .417] or PSQI score [t(34) = 1.53, p = .135].

3.2. Alertness

Psychomotor vigilance and subjective sleepiness was not influenced by 5-HTTLPR or BDNF Val66Met genotype status. For data and analyses, see Supplementary Material 2.

3.3. Sleep parameters

Sleep parameter data is available in Table 2. We utilized one-way ANOVAs to compare all sleep parameters presented in Table 2 between the 5-HTTLPR genotype groups. The analyses revealed that none of these sleep parameters differed between the 5-HTTLPR genotype groups [all p ≥ .155]. Similarly, independent-samples t-tests showed that the sleep parameters were also comparable between the BDNF Val66Met genotype groups [all p ≥ .431].

3.4. Subjective ratings of valence and arousal

Subjective valence and arousal ratings collected during the learning phase were in accordance with the normative ratings from the IAPS from which they were sourced, and were not significantly influenced by 5-HTTLPR or BDNF Val66Met genotype status. For data and analyses, see Supplementary Material 3.

3.5. Recognition performance

3.5.1. Immediate recognition performance

Immediate recognition performance data is available in Fig 3. To examine the influence of 5-HTTLPR genotype group and image valence on immediate recognition performance we conducted a three (5-HTTLPR genotype group: L’/L’, L’/S’, S’/S’) x three (image valence: positive, neutral negative) mixed-measures ANOVA with immediate recognition performance (d’) as the dependent variable. There was no significant main effect of image valence [F(2, 66) = 0.12, p = .892, ηp² = 0.003] or 5-HTTLPR genotype group [F(2, 33) = 0.83, p = .447, ηp² = 0.048] on immediate recognition performance, and no significant interaction between image valence and 5-HTTLPR genotype group [F(4, 66) = 0.19, p = .944, ηp² = 0.011].

Table 2. Summary of sleep parameters

Mean values are presented with SEM in parenthesis. Abbreviations: N1, N2, stages of non-REM sleep; SWS, Slow-Wave Sleep, REM, Rapid Eye Movement sleep; TST, Total Sleep Time; REMs, Rapid Eye Movements.

Fig 3. Mean immediate recognition performance (d’) for recollection (R) responses separately for each genotype group and image valence. Error bars represent SEM.

With regards to BDNF Val66Met, there was similarly no significant main effect of image valence [F(2, 68) = 0.15, p = .864, ηp² = 0.004] or BDNF Val66Met genotype group [F(1, 34) = 0.12, p = .736, ηp² = 0.003] on immediate recognition performance, and no significant interaction between image valence and BDNF Val66Met genotype group [F (2, 68) = 0.21, p = .812, ηp² = 0.006].

3.5.2. Delayed recognition performance

Delayed recognition performance data is available in Fig 4. To examine the influence of 5-HTTLPR genotype group and image valence on sleep-related memory consolidation we conducted a three (5-HTTLPR genotype group: L’/L’, L’/S’, S’/S’) x three (image valence: positive, neutral negative) mixed-measures ANOVA with delayed recognition performance (d’) as the dependent variable. A significant main effect of image valence was revealed [F(2, 66) = 3.45, p = .038, ηp² = 0.095]. Post-hoc paired-samples t-tests demonstrated that negative images were recognised more accurately than both positive images [t(35) = 2.06, p = .047] and neutral images [t(35) = 2.62, p = .013]. However, the ANOVA revealed no significant main effect of 5-HTTLPR genotype group [F(2, 33) = 1.01, p = .377, ηp² = 0.057], and no significant interaction between 5-HTTLPR genotype group and image valence [F(4, 66) = 1.73, p = .155, ηp² = 0.095]. These results suggest that negative images were better retained following a night of sleep than both positive and neutral images, but 5-HTTLPR genotype group did not influence overnight memory retention.

Examining the influence of BDNF Val66Met genotype group and image valence on delayed recognition performance (d’) also revealed a significant main effect of image valence [F(2, 68) = 5.05, p = .009, ηp² = 0.129]. Furthermore, there was no significant main effect of BDNF Val66Met genotype group [F(1, 34) = 0.16, p = .693, ηp² = 0.005]. However, a significant interaction between BDNF Val66Met genotype group and image valence was revealed [F(2, 68) = 3.53, p = .035, ηp² = 0.094]. To explore this interaction, we conduced one-way within-subjects ANOVAs with image valence as the independent variable, separately for each BDNF Val66Met genotype group. A significant main effect of image valence was revealed in Met carriers [F(2, 20) = 6.66, p = .006, ηp² = 0.400], but not in Val homozygotes [F(2, 48) = 1.90, p = .160, ηp² = 0.073]. The main effect of image valence on delayed recognition performance in Met carriers was driven by greater recognition performance for both negative and positive images, relative to neutral images [t(10) = 3.97, p = .003, t(10) = 2.92, p = .015, respectively]. These emotion-specific effects were not found in Val homozygotes [both p ≥ .326]. These findings suggest that overnight retention of emotional images, relative to neutral images, was greater in Met carriers than Val homozygotes.

Fig 4. Mean delayed recognition performance (d’) for recollection (R) responses separately for each genotype group and image valence. Error bars represent SEM. *, p ≤ .015.

These findings may show that sleep supported the retention of emotional images, relative to neutral images, more readily in BDNF Val66Met Met carriers than Val homozygotes.

3.6. Sleep stage correlations

To examine the relationship between REM sleep duration and post-sleep recognition performance (d’), and the influence of 5-HTTLPR and BDNF Val66Met on this relationship, we performed planned parametric correlations, separately for each genotype group and image valence. In 5-HTTLPR heterozygotes (L’/S’), longer REM sleep duration was associated with greater delayed recognition performance for negative images [r = .616, p = .019] (Fig 5d) and positive images [r = .594, p = .025], but not neutral images [p = .085]. REM sleep duration was not associated with delayed recognition performance in any emotion categories for participants homozygous for the 5-HTTLPR L’ allele [all p ≥ .265] or S’ allele [all p ≥ .271] (see Fig 5c and Fig 5e for correlations with negative images).

Regarding BDNF Val66Met, longer REM sleep duration was associated with greater delayed recognition performance for negative images in Met carriers [r = .613, p = .045] (Fig 5b), but this relationship was not observed for positive or neutral images [both p ≥ .082]. Conversely, REM sleep duration was associated with greater delayed recognition performance for both positive images [r = .415, p = .039] and neutral images [r = .404, p = .045], but not negative images [p = .579] (Fig 5a), in Val homozygotes.

In sum, these findings suggest that REM sleep was associated with the retention of emotional images (positive and negative) in 5-HTTLPR L’/S’ carriers. For BDNF Val66Met Met carriers, REM sleep may have selectively supported the retention of negative images, but not positive or neutral images. Conversely, the inverse was true of Val homozygotes.

Fig 5. Scatterplots showing correlations between REM sleep duration (mins) and delayed recognition performance (d’) for recollection (R) responses of negative images, separately for each genotype group. Error bars represent SEM. Solid regression lines: p ≤ .045, dashed regression lines: p ≥ .331.

4. Discussion

We investigated the influence of the 5-HTTLPR and BDNF Val66Met on recognition performance for positive, neutral and negative images before and after a night of sleep in female participants. We hypothesised that these polymorphisms would have emotion-specific effects on post-sleep recognition performance. In support of our hypothesis, our data revealed that recognition performance for emotional images was modulated by BDNF Val66Met following a night of sleep. Specifically, Met carriers were shown to exhibit greater recognition performance for both positive and negative images, relative to neutral images. These effects were not observed in Val homozygotes. Contrary to our hypothesis, we found no effects of the 5-HTTLPR on post-sleep recognition performance for emotional images.

A wealth of evidence demonstrates that emotional salience enhances memory [1,2]. The benefit of emotion on remembering becomes more robust as the interval between learning and remembering increases [3–7], suggesting that the emotional memory advantage is achieved by modulating consolidation processes. Sleep supports memory consolidation [9–12], and particularly favours the consolidation of emotional memories [8,13–16].

Here, we similarly show that emotion enhances remembering. Specifically, negative images were remembered better than neutral images during the post-sleep recognition test across all participants (i.e. pooled across genotype groups). Recognition performance was not, however, influenced by emotion during the pre-sleep recognition test. These findings support the notion that emotion enhances memory by modulating slow consolidation processes, and may offer some support for a selective role of sleep in the retention of emotional memories. However, it is unknown whether this effect would be seen following a similar interval of wakefulness.

Previous studies have demonstrated an influence of BDNF Val66Met on human memory, where Met carriers exhibit memory impairments [47,48]. Less well studied, however, is the effect of BDNF Val66Met on emotional memory. Nonetheless, an fMRI investigation examined the influence of BDNF Val66Met on hippocampal encoding activity and subsequent recognition performance for emotional and neutral words in healthy participants [54]. It was revealed that Met carriers exhibited greater hippocampal activity during the encoding of negative emotional words, relative to both Val homozygotes and neutral words. However, there was no influence of BDNF Val66Met on recognition performance for emotional or neutral words following a 10 min retention interval. The increase in hippocampal reactivity to negative emotional stimuli observed by Molendijk et al. [54] supports prior research in anxious and depressed adolescents, which showed that emotional facial expressions elicit greater activity in the anterior hippocampus and amygdala in Met carriers, relative to Val homozygotes [55]. Healthy Met carriers have also been shown to exhibit greater amygdala reactivity to unpleasant images than Val homozygotes [56]. Although the mechanism underlying increased amygdala reactivity to emotional stimuli in Met carriers is unclear, the Met allele has been shown to impair synaptic plasticity in the medial prefrontal cortex [93], which may suppress top-down regulation of the amygdala [94–96].

Studies demonstrating a modulatory role of BDNF Val66Met on activity in limbic regions associated with emotion and learning may help to explain the greater recognition performance for positive and negative images, relative to neutral images, exhibited by Met carriers in this study. Sleep-related memory consolidation is thought to be enhanced by amygdala and hippocampal activations during learning [25,26]. For example, research in rats demonstrates that brief activations of the basolateral amygdala enhances memory for novel objects following a 24 h retention interval [97], an effect which is dependent on interactions between the amygdala and the hippocampus [98]. Similarly, recent research in humans has shown that direct electrical amygdala stimulation during encoding enhances next-day recognition of emotionally neutral object images [99]. In these studies, amygdala activations did not influence memory during the immediate test phase, suggesting that amygdala and hippocampal activations benefit memory by enhancing consolidation processes. These findings may explain why recognition performance was greater for emotional images, relative to neutral images, in our Met carriers, whilst accounting for the lack of an effect of BDNF Val66Met on emotional memory during the immediate recognition test in this study, and during a 10 min recognition test in a previous study [54]. In sum, it is possible that amygdala reactivity to the emotional images could have been enhanced for the Met carriers in this study. This amplified activation of the amygdala during encoding could have primed subsequent consolidation processes during the overnight sleep interval. Hence, we see greater post-sleep recognition performance for emotionally salient images, relative to neutral images, in the Met carriers but not the Val homozygotes.

Our results suggest that BDNF Val66Met may modulate the relationship between REM sleep and emotional memory consolidation. More specifically, we found that REM sleep duration correlated positively with post-sleep recognition performance for negative images in Met carriers, but not in Val homozygotes. Conversely, REM sleep duration was associated with greater post-sleep recognition performance for positive and neutral images, but not negative images, in Val homozygotes. This may suggest that the benefit of REM sleep on the retention of negative emotional memories is amplified by the Met allele, which could explain why post-sleep recognition performance for negative images, relative to neutral images, was greater in Met carriers but not Val homozygotes. However, it is important to note that Met carriers also exhibited greater post-sleep recognition performance for positive images, relative to neutral images, but REM sleep duration was not associated with post-sleep recognition performance for positive images in this genotype group.

Nonetheless, a study in rats showed that hippocampal BDNF expression levels increase during REM rebound sleep following sleep deprivation [41]. Furthermore, a study of two-way active-avoidance learning in rats showed that learning increased pontine-wave density during subsequent REM sleep, which correlated positively with BDNF protein levels in the dorsal hippocampus [100]. Given that REM sleep is reliably implicated in the consolidation of emotionally salient memories [17–19], these studies may support a role for BDNF in REM sleep- related emotional memory consolidation. However, further research is required to elucidate the specific influence of BDNF Val66Met in this context.

Contrary to our hypothesis, we did not detect any influence of the 5-HTTLPR on emotional memory performance. The 5-HTTLPR S allele has been associated with increased amygdala reactivity to emotionally salient stimuli [32–35], which is thought to enhance subsequent sleep-related consolidation processes [25,26]. Interestingly, however, meta-analyses highlight the controversy regarding the relationship between the 5-HTTLPR and amygdala function. For example, whilst some meta-analyses have been able to show a small 5-HTTLPR–amygdala association, they have also demonstrated statistical power issues in the majority of previous studies [35,101]. Furthermore, an analysis of unpublished studies in addition to the 34 studies analysed by Murphy et al. [101] showed no significant 5-HTTLPR–amygdala relationship [102], suggesting that the potential association may have been overstated due to publication bias. These findings suggest that the 5-HTTLPR may not modulate neural mechanisms underlying emotional memory processes, supporting the absence of an effect of 5-HTTLPR on recognition performance in this study. However, we cannot completely rule out the possibility that our negative findings regarding the 5-HTTLPR may also be due to the relatively small sample size used in this study.

Our correlational analyses did reveal that the relationship between REM sleep and post-sleep emotional memory performance was modulated by the 5-HTTLPR. Unexpectedly, greater REM sleep predicted greater retention of negative and positive images in L’/S’ heterozygotes, but not L’/L’ or S’/S’ homozygotes. This finding may suggest an ‘inverted-U’ relationship between 5-HTT expression and REM sleep- related emotional memory retention. Given that the present study is, to our knowledge, the first to examine the influence of the 5-HTTLPR on REM sleep- related emotional memory retention, it is impossible to assess the validity of this interpretation. However, this result highlights the importance of analysing homozygote and heterozygote S allele carriers separately; in many studies, these groups are collapsed into a single ‘S allele carrier’ group. It is relevant to mention that a large proportion of the L’/S’ genotype group was made up of BDNF Val66Met Met carriers (8 out of 14; see Section 3.1). Accordingly, it is possible that the relationship between REM sleep and post-sleep recognition performance of emotional images was driven by Met carriers. However, longer REM sleep duration predicted greater post-sleep recognition performance for negative images, but not positive images, in Met carriers. Therefore, it is unlikely that this explanation provides a full account of this observed effect.

4.1. Implications

BDNF Val66Met has been implicated in multiple neurological disorders, including schizophrenia, panic disorder, PTSD, and MDD [103–106]. We will focus here on MDD, as our results may have important implications for understanding the mechanisms underlying the relationship between BDNF Val66Met and MDD. Several studies investigating the BDNF Val66Met– MDD relationship have demonstrated that the Met allele is associated with increased MDD risk [104,107,108]. However, other research has failed to show a relationship between BDNF Val66Met and vulnerability to MDD [109], and one study has suggested that Val homozygotes are more susceptible to MDD than Met carriers [110]. Nonetheless, the Met allele has also been associated with increased depressive symptoms [111], greater suicidal behaviour [112,113], and heightened chronicity of MDD [114].

Although BDNF Val66Met is likely to modulate risk for affective disorders, the mechanisms underlying this association remain unclear. Our results suggest that BDNF Val66Met may modulate MDD risk by increasing the strength of emotional memory traces. Indeed, greater recognition performance for negative emotional stimuli has been observed in MDD patients [115]. Emotional memory biases such as these may play a causal role in the onset and maintenance of MDD [69,116,117]. In our study, Met carriers exhibited greater recognition performance for negative and positive images, rather than a specific bias for negative emotional memories. Nonetheless, it is conceivable that enhanced emotional memory, positive and negative, could be a risk factor for MDD, particularly for individuals who have experienced severe trauma. Intriguingly, general population studies have shown that childhood adversity such as childhood sexual and emotional abuse has a greater impact on depressive symptoms and lifetime depression risk in Met carriers than Val homozygotes [118,119]. Our results suggest that the Met allele may increase the risk of developing MDD following childhood adversity by increasing the strength of memories associated with such adversities. Of course, further research is required to support this notion.

4.2. Limitations

One potential limitation of this study is the lack of a waking control group. Indeed, without a comparison to a wake condition it remains unclear whether the observed effects relate to sleep-related consolidation processes, or are simply time-dependent, and would have emerged over a similar interval of wakefulness. However, the significant positive correlations between REM sleep and post-sleep recognition performance for emotion-specific image categories (see Section 3.6) offers some reassurance that the effects of BDNF Val66Met on delayed recognition may be related to sleep-associated consolidation processes. Moreover, no valid control group is available to directly compare the effects of time and sleep. Indeed, results from an overnight sleep deprivation condition would be limited by severe sleepiness during the delayed recognition test. A daytime wakefulness condition would be limited by circadian effects on cognition and hippocampal BDNF expression [120]. Nonetheless, further research is required to differentiate the influences of sleep and time on BDNF Val66Met modulated emotional memory consolidation.

It is also important to reiterate here that the sample for this study was composed entirely of female participants. As noted in Section 2.1, we chose to recruit an exclusively female sample to eliminate gender as a potential confounding factor [60–62]. However, it is relevant to mention that the extent to which our findings can be generalized to males is unclear. Future research should examine whether the influence of BDNF Val66Met on the overnight retention of emotionally salient memories is gender-specific.

A further limitation of this study relates to differences in alertness between the immediate and delayed recognition tests (see Supplementary Material 2). Participants exhibited faster RTs during the PVT before the immediate recognition test, relative to the delayed recognition test. It is plausible that these differences in alertness could have influenced recognition performance. Importantly however, reaction time was not influenced by either 5-HTTLPR or BDNF Val66Met genotype group, meaning that the emotion-specific effects of BDNF Val66Met on recognition performance in the delayed recognition test was not influenced by between-group differences in alertness. Furthermore, it is unlikely that decreased alertness would have emotion-specific effects on recognition performance. Finally, scores on the SSS were comparable between the recognition tests, suggesting that participants did not feel less alert during the delayed recognition test, relative to the immediate recognition test. Nonetheless, future research may benefit from increasing the duration of the recovery interval between waking and the delayed recognition test.

4.3. Conclusion

Our study has shown that the Met allele of BDNF Val66Met increases recognition performance for emotionally positive and negative images, relative to neutral images, following an overnight sleep interval. Furthermore, we found that greater REM sleep duration predicted greater post-sleep recognition performance for negative images in Met carriers, but not in Val homozygotes. We did not observe any effect of the 5-HTTLPR on pre- or post- sleep recognition performance for positive, neutral or negative images. However, there was a positive correlation between REM sleep duration and post-sleep recognition performance for emotional images (positive and negative) in L’/S’ carriers, but not in S’/S’ or L’/L’ homozygotes.

Our observation that the Met allele of the BDNF Val66Met was associated with increased overnight retention of emotional memories, relative to neutral memories, may have implications for our understanding of the BDNF Val66Met- MDD relationship. Specifically, the Met allele may increase vulnerability to MDD by increasing the strength of emotionally salient memory traces. Although further studies are required to test this prediction, it may elucidate the mechanisms underlying the influence of BDNF Val66Met on MDD vulnerability following childhood adversity.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Competing interests

Declarations of interest: none.

Acknowledgments

The authors would like to thank Cameron J. Burn for his assistance with participant recruitment and collection of saliva samples.

References

[1]E.A. Kensinger, Remembering emotional experiences: the contribution of valence and arousal., Rev. Neurosci. 15 (2004) 241–251. doi:10.1515/REVNEURO.2004.15.4.241.

[2]J.L. McGaugh, The amygdala modulates the consolidation of memories of emotionally arousing experiences., Annu. Rev. Neurosci. 27 (2004) 1–28. doi:10.1146/annurev.neuro.27.070203.144157.

[3]A.K. Anderson, Y. Yamaguchi, W. Grabski, D. Lacka, Emotional memories are not all created equal: Evidence for selective memory enhancement, Learn. Mem. 13 (2006) 711–718. doi:10.1101/lm.388906.

[4]F. Dolcos, K.S. LaBar, R. Cabeza, Remembering one year later: Role of the amygdala and the medial temporal lobe memory system in retrieving emotional memories, Proc. Natl. Acad. Sci. 102 (2005) 2626–2631. doi:10.1073/pnas.0409848102.

[5]E. Levonian, Retention over time in relation to arousal during learning: An explanation of discrepant results, Acta Psychol. (Amst). 36 (1972) 290–321. doi:10.1016/0001-6918(72)90013-3.

[6]T. Sharot, E.A. Phelps, How arousal modulates memory: Disentangling the effects of attention and retention, Cogn. Affect. Behav. Neurosci. 4 (2004) 294–306. doi:10.3758/CABN.4.3.294.

[7]E.L. Walker, R.D. Tarte, Memory storage as a function of arousal and time with homogeneous and heterogeneous lists, J. Verbal Learning Verbal Behav. 2 (1963) 113–119. doi:10.1016/S0022-5371(63)80075-4.

[8]P. Hu, M. Stylos-Allan, M.P. Walker, Sleep facilitates consolidation of emotional declarative memory., Psychol. Sci. 17 (2006) 891–8. doi:10.1111/j.1467-9280.2006.01799.x.

[9]J.M. Ellenbogen, J.D. Payne, R. Stickgold, The role of sleep in declarative memory consolidation: passive, permissive, active or none?, Curr. Opin. Neurobiol. 16 (2006) 716–22. doi:10.1016/j.conb.2006.10.006.

[10]R. Stickgold, Sleep-dependent memory consolidation., Nature. 437 (2005) 1272–8. doi:10.1038/nature04286.

[11]T. Abel, R. Havekes, J.M. Saletin, M.P. Walker, Sleep, Plasticity and Memory from Molecules to Whole-Brain Networks, Curr. Biol. 23 (2013) 1–32. doi:10.1097/MPG.0b013e3181a15ae8.Screening.

[12]B. Rasch, J. Born, About sleep’s role in memory, Physiol. Rev. 93 (2013) 681–766. doi:10.1152/Physrev.00032.2012.

[13]J.D. Payne, A.M. Chambers, E.A. Kensinger, Sleep promotes lasting changes in selective memory for emotional scenes., Front. Integr. Neurosci. 6 (2012) 108. doi:10.3389/fnint.2012.00108.

[14]U. Wagner, M. Hallschmid, B. Rasch, J. Born, Brief Sleep After Learning Keeps Emotional Memories Alive for Years, Biol. Psychiatry. 60 (2006) 788–790. doi:10.1016/j.biopsych.2006.03.061.

[15]U. Wagner, S. Gais, J. Born, Emotional memory formation is enhanced across sleep intervals with high amounts of rapid eye movement sleep., Learn. Mem. 8 (2001) 112–119. doi:10.1101/lm.36801.

[16]J.D. Payne, R. Stickgold, K. Swanberg, E.A. Kensinger, Sleep preferentially enhances memory for emotional components of scenes, Psychol. Sci. 19 (2008) 781–788. doi:10.1111/j.1467-9280.2008.02157.x.

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

[18]C.D. Wiesner, J. Pulst, F. Krause, M. Elsner, L. Baving, A. Pedersen, A. Prehn-Kristensen, R. Göder, The effect of selective REM-sleep deprivation on the consolidation and affective evaluation of emotional memories., Neurobiol. Learn. Mem. 122 (2015) 131–141. doi:10.1016/j.nlm.2015.02.008.

[19]S. Groch, K. Zinke, I. Wilhelm, J. Born, Dissociating the contributions of slow-wave sleep and rapid eye movement sleep to emotional item and source memory, Neurobiol. Learn. Mem. 122 (2015) 122–130. doi:10.1016/j.nlm.2014.08.013.

[20]P. Maquet, J. Péters, J. Aerts, G. Delfiore, C. Degueldre, A. Luxen, G. Franck, Functional neuroanatomy of human rapid-eye-movement sleep and dreaming., Nature. 383 (1996) 163–166. doi:10.1038/383163a0.

[21]Y. Nir, G. Tononi, Dreaming and the brain: from phenomenology to neurophysiology., Trends Cogn. Sci. 14 (2010) 88–100. doi:10.1016/j.tics.2009.12.001.

[22]E.A. Nofzinger, M.A. Mintun, M. Wiseman, D.J. Kupfer, R.Y. Moore, Forebrain activation in REM sleep: An FDG PET study, Brain Res. 770 (1997) 192–201. doi:10.1016/S0006-8993(97)00807-X.

[23]I. Hutchinson, S. Rathore, The role of REM sleep theta activity in emotional memory, Front. Psychol. 6 (2015) 1–15. doi:10.3389/fpsyg.2015.01439.

[24]M.P. Walker, E. van der Helm, Overnight therapy? The role of sleep in emotional brain processing., Psychol. Bull. 135 (2009) 731–748. doi:10.1037/a0016570.

[25]M.O. Harrington, K. Pennington, S.J. Durrant, The “affect tagging and consolidation” (ATaC) model of depression vulnerability, Neurobiol. Learn. Mem. 140 (2017) 43–51. doi:10.1016/j.nlm.2017.02.003.

[26]K.A. Bennion, J.D. Payne, E.A. Kensinger, Selective Effects of Sleep on Emotional Memory: What Mechanisms Are Responsible?, Transl. Issues Psychol. Sci. 1 (2015) 79–88.

[27]F. Kumfor, M. Irish, J.R. Hodges, O. Piguet, Discrete Neural Correlates for the Recognition of Negative Emotions: Insights from Frontotemporal Dementia, PLoS One. 8 (2013) e67457. doi:10.1371/journal.pone.0067457.

[28]R. Stark, M. Zimmermann, S. Kagerer, A. Schienle, B. Walter, M. Weygandt, D. Vaitl, Hemodynamic brain correlates of disgust and fear ratings, Neuroimage. 37 (2007) 663–673. doi:10.1016/j.neuroimage.2007.05.005.

[29]E.A. Kensinger, D.L. Schacter, Processing emotional pictures and words: effects of valence and arousal., Cogn. Affect. Behav. Neurosci. 6 (2006) 110–126. doi:10.3758/CABN.6.2.110.

[30]P. Vrticka, L. Lordier, B. Bediou, D. Sander, Human amygdala response to dynamic facial expressions of positive and negative surprise., Emotion. 14 (2014) 161–169. doi:10.1037/a0034619.

[31]K.P. Lesch, D. Bengel, A. Heils, S.Z. Sabol, B.D. Greenberg, S. Petri, J. Benjamin, C.R. Müller, D.H. Hamer, D.L. Murphy, Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region., Science (80-. ). 274 (1996) 1527–1531. doi:10.1126/science.274.5292.1527.

[32]U. Dannlowski, C. Konrad, H. Kugel, P. Zwitserlood, K. Domschke, S. Schöning, P. Ohrmann, J. Bauer, M. Pyka, C. Hohoff, W. Zhang, B.T. Baune, W. Heindel, V. Arolt, T. Suslow, Emotion specific modulation of automatic amygdala responses by 5-HTTLPR genotype., Neuroimage. 53 (2010) 893–898. doi:10.1016/j.neuroimage.2009.11.073.

[33]T.B. Lonsdorf, A. Golkar, K.M. Lindstöm, P. Fransson, M. Schalling, A. Ohman, M. Ingvar, 5-HTTLPR and COMTval158met genotype gate amygdala reactivity and habituation., Biol. Psychol. 87 (2011) 106–112. doi:10.1016/j.biopsycho.2011.02.014.

[34]L. Pezawas, A. Meyer-Lindenberg, E.M. Drabant, B.A. Verchinski, K.E. Munoz, B.S. Kolachana, M.F. Egan, V.S. Mattay, A.R. Hariri, D.R. Weinberger, 5-HTTLPR polymorphism impacts human cingulate-amygdala interactions: a genetic susceptibility mechanism for depression., Nat. Neurosci. 8 (2005) 828–34. doi:10.1038/nn1463.

[35]M.R. Munafò, S.M. Brown, A.R. Hariri, Serotonin Transporter (5-HTTLPR) Genotype and Amygdala Activation: A Meta-Analysis, Biol. Psychiatry. 63 (2008) 852–857. doi:10.1016/j.biopsych.2007.08.016.

[36]H. Garpenstrand, P. Annas, J. Ekblom, L. Oreland, M. Fredrikson, Human fear conditioning is related to dopaminergic and serotonergic biological markers., Behav. Neurosci. 115 (2001) 358–364. doi:10.1037/0735-7044.115.2.358.

[37]L.G. Crişan, S. Panǎ, R. Vulturar, R.M. Heilman, R. Szekely, B. Drugǎ, N. Dragoş, A.C. Miu, Genetic contributions of the serotonin transporter to social learning of fear and economic decision making, Soc. Cogn. Affect. Neurosci. 4 (2009) 399–408. doi:10.1093/scan/nsp019.

[38]R.M. Todd, D.J. Palombo, B. Levine, A.K. Anderson, Genetic differences in emotionally enhanced memory., Neuropsychologia. 49 (2011) 734–44. doi:10.1016/j.neuropsychologia.2010.11.010.

[39]I.H. Gotlib, J. Joormann, K.L. Minor, J. Hallmayer, HPA Axis Reactivity: A Mechanism Underlying the Associations Among 5-HTTLPR, Stress, and Depression, Biol. Psychiatry. 63 (2008) 847–851. doi:10.1016/j.biopsych.2007.10.008.

[40]K.S. Kendler, J.W. Kuhn, J. Vittum, C.A. Prescott, B. Riley, The Interaction of Stressful Life Events and a Serotonin Transporter Polymorphism in the Prediction of Episodes of Major Depression, Arch. Gen. Psychiatry. 62 (2005) 529. doi:10.1001/archpsyc.62.5.529.

[41]P. Ravassard, A.M. Hamieh, G. Malleret, P. Salin, Paradoxical sleep: A vigilance state to gate long-term brain plasticity?, Neurobiol. Learn. Mem. 122 (2015) 4–10. doi:10.1016/j.nlm.2014.11.013.

[42]M. Mizuno, K. Yamada, J. He, A. Nakajima, T. Nabeshima, Involvement of BDNF Receptor TrkB in Spatial Memory Formation, Learn. Mem. 10 (2003) 108–115. doi:10.1101/lm.56003.

[43]P. Bekinschtein, M. Cammarota, J.H. Medina, BDNF and memory processing, Neuropharmacology. 76 (2014) 677–683. doi:10.1016/j.neuropharm.2013.04.024.

[44]A.L. Mahan, K.J. Ressler, Fear Conditioning, Synpatic Plasticity, and the Amygdala: Implications for Posttraumatic Stress Disorder, Trends Neurosci. 35 (2013) 24–35. doi:10.1016/j.tins.2011.06.007.Fear.

[45]J.P. Chhatwal, L. Stanek-Rattiner, M. Davis, K.J. Ressler, Amygdala BDNF signaling is required for consolidation but not encoding of extinction., Nat. Neurosci. 9 (2006) 870–872. doi:10.1038/nn1718.

[46]L.M. Rattiner, M. Davis, C.T. French, K.J. Ressler, Brain-Derived Neurotrophic Factor and Tyrosine Kinase Receptor B Involvement in Amygdala-Dependent Fear Conditioning, J. Neurosci. 24 (2004) 4796–4806. doi:10.1523/JNEUROSCI.5654-03.2004.

[47]K.M. Kennedy, E.D. Reese, M.M. Horn, A.N. Sizemore, A.K. Unni, M.E. Meerbrey, A.G. Kalich, K.M. Rodrigue, BDNF val66met polymorphism affects aging of multiple types of memory, Brain Res. 1612 (2015) 104–117. doi:10.1016/j.brainres.2014.09.044.

[48]J.P. Kambeitz, S. Bhattacharyya, L.M. Kambeitz-Ilankovic, I. Valli, D.A. Collier, P. McGuire, Effect of BDNF val66met polymorphism on declarative memory and its neural substrate: A meta-analysis, Neurosci. Biobehav. Rev. 36 (2012) 2165–2177. doi:10.1016/j.neubiorev.2012.07.002.

[49]F. Harrisberger, K. Spalek, R. Smieskova, A. Schmidt, D. Coynel, A. Milnik, M. Fastenrath, V. Freytag, L. Gschwind, A. Walter, T. Vogel, K. Bendfeldt, D.J.F. De Quervain, A. Papassotiropoulos, S. Borgwardt, The association of the BDNF Val66Met polymorphism and the hippocampal volumes in healthy humans: A joint meta-analysis of published and new data, Neurosci. Biobehav. Rev. 42 (2014) 267–278. doi:10.1016/j.neubiorev.2014.03.011.

[50]B. Cao, I.E. Bauer, A.N. Sharma, B. Mwangi, T. Frazier, L. Lavagnino, G.B. Zunta-Soares, C. Walss-Bass, D.C. Glahn, F. Kapczinski, D.A. Nielsen, J.C. Soares, Reduced hippocampus volume and memory performance in bipolar disorder patients carrying the BDNF val66met met allele, J. Affect. Disord. 198 (2016) 198–205. doi:10.1016/j.jad.2016.03.044.

[51]M.F. Egan, M. Kojima, J.H. Callicott, T.E. Goldberg, B.S. Kolachana, A. Bertolino, E. Zaitsev, B. Gold, D. Goldman, M. Dean, B. Lu, D.R. Weinberger, The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function, Cell. 112 (2003) 257–269. doi:10.1016/S0092-8674(03)00035-7.

[52]N. Gosselin, L. De Beaumont, K. Gagnon, A.-A. Baril, V. Mongrain, H. Blais, J. Montplaisir, J.-F. Gagnon, S. Pelleieux, J. Poirier, J. Carrier, BDNF Val66Met Polymorphism Interacts with Sleep Consolidation to Predict Ability to Create New Declarative Memories., J. Neurosci. 36 (2016) 8390–8. doi:10.1523/JNEUROSCI.4432-15.2016.

[53]D. Keyan, R. a. Bryant, Role of BDNF val66met polymorphism in modulating exercised-induced emotional memories, Psychoneuroendocrinology. 77 (2017) 150–157. doi:10.1016/j.psyneuen.2016.12.013.

[54]M.L. Molendijk, M.J. van Tol, B.W.J.H. Penninx, N.J.A. van der Wee, A. Aleman, D.J. Veltman, P. Spinhoven, B.M. Elzinga, BDNF val66met affects hippocampal volume and emotion-related hippocampal memory activity, Transl. Psychiatry. 2 (2012) e74. doi:10.1038/tp.2011.72.

[55]J.Y.F. Lau, D. Goldman, B. Buzas, C. Hodgkinson, E. Leibenluft, E. Nelson, L. Sankin, D.S. Pine, M. Ernst, BDNF gene polymorphism (Val66Met) predicts amygdala and anterior hippocampus responses to emotional faces in anxious and depressed adolescents, Neuroimage. 53 (2010) 952–961. doi:10.1016/j.neuroimage.2009.11.026.

[56]C. Montag, M. Reuter, B. Newport, C. Elger, B. Weber, The BDNF Val66Met polymorphism affects amygdala activity in response to emotional stimuli: Evidence from a genetic imaging study, Neuroimage. 42 (2008) 1554–1559. doi:10.1016/j.neuroimage.2008.06.008.

[57]A.N. Banducci, M. Gomes, L. Macpherson, C.W. Lejuez, M.N. Potenza, J. Gelernter, A Preliminary Examination of the Relationship Between the 5-HTTLPR and Childhood Emotional Abuse on Depressive Symptoms in 10-12-Year-Old Youth, Psychol. Trauma Theory, Res. Pract. Policy. 6 (2014) 1–7.

[58]T. Outhred, P. Das, C. Dobson-Stone, K.L. Felmingham, R.A. Bryant, P.J. Nathan, G.S. Malhi, A.H. Kemp, Impact of 5-HTTLPR on SSRI Serotonin Transporter Blockade during Emotion Regulation: A Preliminary fMRI Study, J. Affect. Disord. 196 (2016) 11–19. doi:10.1016/j.jad.2016.02.019.

[59]N. Sadeh, S. Javdani, J.J. Jackson, E.K. Reynolds, M.N. Potenza, J. Gelernter, C.W. Lejuez, E. Verona, Serotonin transporter gene associations with psychopathic traits in youth vary as a function of socioeconomic resources., J. Abnorm. Psychol. 119 (2010) 604–609. doi:10.1037/a0019709.

[60]L.M. Monteggia, B. Luikart, M. Barrot, D. Theobold, I. Malkovska, S. Nef, L.F. Parada, E.J. Nestler, Brain-Derived Neurotrophic Factor Conditional Knockouts Show Gender Differences in Depression-Related Behaviors, Biol. Psychiatry. 61 (2007) 187–197. doi:10.1016/j.biopsych.2006.03.021.

[61]A.E. Autry, M. Adachi, P. Cheng, L.M. Monteggia, Gender-Specific Impact of Brain-Derived Neurotrophic Factor Signaling on Stress-Induced Depression-Like Behavior, Biol. Psychiatry. 66 (2009) 84–90. doi:10.1016/j.biopsych.2009.02.007.

[62]R. Jiang, M.A. Babyak, B.H. Brummett, I.C. Siegler, C.M. Kuhn, R.B. Williams, Brain-derived neurotrophic factor (BDNF) Val66Met polymorphism interacts with gender to influence cortisol responses to mental stress, Psychoneuroendocrinology. 79 (2017) 13–19. doi:10.1016/j.psyneuen.2017.02.005.

[63]H.C. Anchordoquy, C. McGeary, L. Liu, K.S. Krauter, A. Smolen, Genotyping of three candidate genes after whole-genome preamplification of DNA collected from buccal cells, Behav. Genet. 33 (2003) 73–78. doi:10.1023/A:1021007701808.

[64]K. Tomson, L. Merenäkk, H.-M. Loit, J. Mäestu, J. Harro, The relationship between serotonin transporter gene promoter polymorphism and serum lipid levels at young age in a longitudinal population-representative study, Prog. Neuro-Psychopharmacology Biol. Psychiatry. 35 (2011) 1857–1862. doi:10.1016/j.pnpbp.2011.08.004.

[65]C. Klein, S. Bodenmann, W. Berger, P. Brugger, H. Landolt, The BDNF Val66Met polymorphism modulates sleep intensity: EEG frequency-and state-specificity, Sleep. 35 (2012) 335–344.

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

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

[68]A.H.C. Hindash, N. Amir, Negative Interpretation Bias in Individuals with Depressive Symptoms, Cognit. Ther. Res. 36 (2011) 502–511. doi:10.1007/s10608-011-9397-4.

[69]J.M. Newby, T. Lang, A. Werner-Seidler, E. Holmes, M.L. Moulds, Alleviating distressing intrusive memories in depression: A comparison between computerised cognitive bias modification and cognitive behavioural education, Behav. Res. Ther. 56 (2014) 60–67. doi:10.1016/j.brat.2014.03.001.

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

[71]S.J. Durrant, S.A. Cairney, C. McDermott, P.A. Lewis, Schema-conformant memories are preferentially consolidated during REM sleep., Neurobiol. Learn. Mem. 122 (2015) 41–50. doi:10.1016/j.nlm.2015.02.011.

[72]J.S. Rihm, B. Rasch, Replay of conditioned stimuli during late REM and stage N2 sleep influences affective tone rather than emotional memory strength., Neurobiol. Learn. Mem. (2015). doi:10.1016/j.nlm.2015.04.008.

[73]L. Mascetti, A. Foret, J. Schrouff, V. Muto, V. Dideberg, E. Balteau, C. Degueldre, C. Phillips, A. Luxen, F. Collette, V. Bours, P. Maquet, Concurrent synaptic and systems memory consolidation during sleep., J. Neurosci. 33 (2013) 10182–90. doi:10.1523/JNEUROSCI.0284-13.2013.

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

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

[76]G.D. Roach, D. Dawson, N. Lamond, Can a shorter psychomotor vigilance task be used as a reasonable substitute for the ten-minute psychomotor vigilance task?, Chronobiol. Int. 23 (2006) 1379–87. doi:10.1080/07420520601067931.

[77]R.T. Wilkinson, D. Houghton, Field test of arousal: a portable reaction timer with data storage., Hum. Factors. 24 (1982) 487–493. doi:10.1177/001872088202400409.

[78]M. Gilson, G. Deliens, R. Leproult, A. Bodart, A. Nonclercq, R. Ercek, P. Peigneux, REM-Enriched Naps Are Associated with Memory Consolidation for Sad Stories and Enhance Mood-Related Reactivity., Brain Sci. 6 (2015) 1–18. doi:10.3390/brainsci6010001.

[79]L.M. Talamini, L.F. Bringmann, M. de Boer, W.F. Hofman, Sleeping worries away or worrying away sleep? Physiological evidence on sleep-emotion interactions., PLoS One. 8 (2013) e62480. doi:10.1371/journal.pone.0062480.

[80]S.A. Cairney, S.J. Durrant, J. Hulleman, P.A. Lewis, Targeted Memory Reactivation During Slow Wave Sleep Facilitates Emotional Memory Consolidation, Sleep. 37 (2014) 701–7. doi:10.5665/sleep.3572.

[81]K.R. Mickley, E.A. Kensinger, Emotional valence influences the neural correlates associated with remembering and knowing., Cogn. Affect. Behav. Neurosci. 8 (2008) 143–152. doi:10.3758/CABN.8.2.143.

[82]I.G. Dobbins, N.E.A. Kroll, A.P. Yonelinas, Dissociating familiarity from recollection using rote rehearsal., Mem. Cogn. 32 (2004) 932–944. doi:10.3758/BF03196871.

[83]A.P. Yonelinas, The nature of recollection and familiarity: A review of 30 years of research, J. Mem. Lang. 46 (2002) 441–517. doi:10.1006/jmla.2002.2864.

[84]R.W. Homan, J. Herman, P. Purdy, Cerebral location of international 10-20 system electrode placement., Electroencephalogr. Clin. Neurophysiol. 66 (1987) 376–382. doi:10.1016/0013-4694(87)90206-9.

[85]V. Sterpenich, G. Albouy, M. Boly, G. Vandewalle, A. Darsaud, E. Balteau, T.T. Dang-Vu, M. Desseilles, A. D’Argembeau, S. Gais, G. Rauchs, M. Schabus, C. Degueldre, A. Luxen, F. Collette, P. Maquet, Sleep-related hippocampo-cortical interplay during emotional memory recollection, PLoS Biol. 5 (2007) 2709–2722. doi:10.1371/journal.pbio.0050282.

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

[87]N.A. Macmillan, H.L. Kaplan, Detection theory analysis of group data: estimating sensitivity from average hit and false-alarm rates., Psychol. Bull. 98 (1985) 185–199. doi:10.1037/0033-2909.98.1.185.

[88]N.A. Macmillan, C.D. Creelman, Detection Theory: A User’s Guide, Psychology press, 2004.

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

[90]S.B.S. Khalsa, D.A. Conroy, J.F. Duffy, Sleep- and circadian-dependent modulation of REM density, Sleep Res. 11 (2002) 53–59.

[91]E. Aserinsky, Relationship of rapid eye movement density to the prior accumulation of sleep and wakefulness., Psychophysiology. 10 (1973) 545–558. doi:10.1111/j.1469-8986.1973.tb00804.x.

[92]C.C. Hong, J.C. Gillin, B.M. Dow, J. Wu, M.S. Buchsbaum, Localized and lateralized cerebral glucose metabolism associated with eye movements during REM sleep and wakefulness: a positron emission tomography (PET) study., Sleep. 18 (1995) 570–80.

[93]S.S. Pattwell, K.G. Bath, R. Perez-Castro, F.S. Lee, M. V Chao, I. Ninan, The BDNF Val66Met polymorphism impairs synaptic transmission and plasticity in the infralimbic medial prefrontal cortex, 32 (2012) 2410–2421. doi:10.1523/JNEUROSCI.5205-11.2012.The.

[94]J. Peters, L.M. Dieppa-Perea, L.M. Melendez, G.J. Quirk, Induction of Fear Extinction with Hippocampal-Infralimbic BDNF, Science (80-. ). 328 (2010) 1288–1290. doi:10.1126/science.1186909.

[95]F. Soliman, C.E. Glatt, K.G. Bath, L. Levita, M. Rebecca, S.S. Pattwell, D. Jing, N. Tottenham, D. Amso, H. Leah, H.U. Voss, G. Glover, D.J. Ballon, C. Liston, T. Teslovich, T. Van Kempen, F.S. Lee, B.J. Casey, A Genetic Variant BDNF Polymorphism Alters Extinction Learning in Both Mouse and Human, 327 (2010) 863–866. doi:10.1126/science.1181886.A.

[96]I. Ninan, Synaptic regulation of affective behaviors; Role of BDNF, Neuropharmacology. 76 (2014) 684–695. doi:10.1016/j.neuropharm.2013.04.011.

[97]D.I. Bass, K.N. Partain, J.R. Manns, Event-specific enhancement of memory via brief electrical stimulation to the basolateral complex of the amygdala in rats., Behav. Neurosci. 126 (2012) 204–208. doi:10.1037/a0026462.

[98]D.I. Bass, Z.G. Nizam, K.N. Partain, A. Wang, J.R. Manns, Amygdala-mediated enhancement of memory for specific events depends on the hippocampus., Neurobiol. Learn. Mem. 107 (2014) 37–41. doi:10.1016/j.nlm.2013.10.020.

[99]C.S. Inman, J.R. Manns, K.R. Bijanki, D.I. Bass, S. Hamann, D.L. Drane, R.E. Fasano, C.K. Kovach, R.E. Gross, J.T. Willie, Direct electrical stimulation of the amygdala enhances declarative memory in humans, Proc. Natl. Acad. Sci. 115 (2018) 98–103. doi:10.1073/pnas.1714058114.

[100]J. Ulloor, S. Datta, Spatio-temporal activation of cyclic AMP response element-binding protein, activity-regulated cytoskeletal-associated protein and brain-derived nerve growth factor: A mechanism for pontine-wave generator activation-dependent two-way active-avoidance memor, J. Neurochem. 95 (2005) 418–428. doi:10.1111/j.1471-4159.2005.03378.x.

[101]S.E. Murphy, R. Norbury, B.R. Godlewska, P.J. Cowen, Z.M. Mannie, C.J. Harmer, M.R. Munafò, The effect of the serotonin transporter polymorphism (5-HTTLPR) on amygdala function: a meta-analysis., Mol. Psychiatry. 18 (2013) 512–520. doi:10.1038/mp.2012.19.

[102]J.A. Bastiaansen, M.N. Servaas, J.B.C. Marsman, J. Ormel, I.M. Nolte, H. Riese, A. Aleman, Filling the gap: relationship between the serotonin-transporter-linked polymorphic region and amygdala activation., Psychol. Sci. 25 (2014) 2058–66. doi:10.1177/0956797614548877.

[103]M. Kheirollahi, E. Kazemi, S. Ashouri, Brain-Derived Neurotrophic Factor Gene Val66Met Polymorphism and Risk of Schizophrenia: A Meta-analysis of Case–Control Studies, Cell. Mol. Neurobiol. 36 (2016) 1–10. doi:10.1007/s10571-015-0229-z.

[104]Y. Pei, A.K. Smith, Y. Wang, Y. Pan, J. Yang, Q. Chen, W. Pan, F. Bao, L. Zhao, C. Tie, Y. Wang, J. Wang, W. Zhen, J. Zhou, X. Ma, The brain-derived neurotrophic-factor (BDNF) val66met polymorphism is associated with geriatric depression: a meta-analysis., Am. J. Med. Genet. Part B Neuropsychiatr. Genet. 159 (2012) 560–6. doi:10.1002/ajmg.b.32062.

[105]K. Chen, N. Wang, J. Zhang, X. Hong, H. Xu, X. Zhao, Q. Huang, Is the Val66Met polymorphism of the brain-derived neurotrophic factor gene associated with panic disorder? A meta-analysis., Asia-Pacific Psychiatry. 9 (2015) 1–7. doi:10.1111/appy.12228.

[106]L. Zhang, D.M. Benedek, C.S. Fullerton, R.D. Forsten, J.A. Naifeh, X.X. Li, X.Z. Hu, H. Li, M. Jia, G.Q. Xing, K.N. Benevides, R.J. Ursano, PTSD risk is associated with BDNF Val66Met and BDNF overexpression, Mol. Psychiatry. (2014). doi:10.1038/mp.2012.180.

[107]K.E. Hutchison, G. Carey, Variation in Brain-Derived Neurotrophic Factor (BDNF) Gene is Associated with Symptoms of Depression, J. Affect. Disord. 115 (2009) 215–219. doi:10.1016/j.jad.2008.08.016.Variation.

[108]J.P. Hwang, S.J. Tsai, C.J. Hong, C.H. Yang, J.F. Lirng, Y.M. Yang, The Val66Met polymorphism of the brain-derived neurotrophic-factor gene is associated with geriatric depression, Neurobiol. Aging. 27 (2006) 1834–1837. doi:10.1016/j.neurobiolaging.2005.10.013.

[109]X. Liu, Y. Xu, S. Jiang, D. Cui, Y. Qian, K. Jiang, Family-based association study between brain-derived neurotrophic factor gene and major depressive disorder of Chinese descent, Psychiatry Res. 169 (2009) 169–172. doi:10.1016/j.psychres.2008.06.014.

[110]L. Ribeiro, J. V Busnello, R.M. Cantor, F. Whelan, P. Whittaker, The brain-derived neurotrophic factor rs6265 ( Val66Met ) polymorphism and depression in Mexican-Americans, Neuroreport. 18 (2007) 1291–1293. doi:10.1097/WNR.0b013e328273bcb0.The.

[111]M.E. Czira, H. Wersching, B.T. Baune, K. Berger, Brain-derived neurotrophic factor gene polymorphisms, neurotransmitter levels, and depressive symptoms in an elderly population, Age (Omaha). 34 (2012) 1529–1541. doi:10.1007/s11357-011-9313-6.

[112]M. Sarchiapone, V. Carli, A. Roy, L. Iacoviello, C. Cuomo, M.C. Latella, M. Di Giannantonio, L. Janiri, M. De Gaetano, M.N. Janal, Association of polymorphism (Val66Met) of brain-derived neurotrophic factor with suicide attempts in depressed patients, Neuropsychobiology. 57 (2008) 139–145. doi:10.1159/000142361.

[113]C.C. Zai, M. Manchia, V. De Luca, A.K. Tiwari, N.I. Chowdhury, G.C. Zai, R.P. Tong, Z. Yilmaz, S. a Shaikh, J. Strauss, J.L. Kennedy, The brain-derived neurotrophic factor gene in suicidal behaviour: a meta-analysis., Int. J. Neuropsychopharmacol. 15 (2012) 1037–42. doi:10.1017/S1461145711001313.

[114]Y. Lee, S.W. Lim, S.Y. Kim, J.W. Chung, J. Kim, W. Myung, J. Song, S. Kim, B.J. Carroll, D.K. Kim, Association between the BDNF Val66Met Polymorphism and Chronicity of Depression., Psychiatry Investig. 10 (2013) 56–61. doi:10.4306/pi.2013.10.1.56.

[115]A. Fattahi Asl, A. Ghanizadeh, J. Mollazade, A. Aflakseir, 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 Res. 228 (2015) 223–227. doi:10.1016/j.psychres.2015.04.001.

[116]I.H. Gotlib, J. Joormann, Cognition and depression: Current status and future directions, Annu. Rev. Clin. Psychol. 6 (2010) 285–312. doi:10.1146/annurev.clinpsy.121208.131305.Cognition.

[117]A.T. Beck, Depression: Clinical, experimental, and theoretical aspects, Hoeber, New York, 1967.

[118]M. Aguilera, B. Arias, M. Wichers, N. Barrantes-Vidal, J. Moya, H. Villa, J. van Os, M.I. Ibáñez, M. a. Ruipérez, G. Ortet, L. Fañanás, Early adversity and 5-HTT/BDNF genes: new evidence of gene–environment interactions on depressive symptoms in a general population, Psychol. Med. 39 (2009) 1425. doi:10.1017/S0033291709005248.

[119]G. Juhasz, J.S. Dunham, S. McKie, E. Thomas, D. Downey, D. Chase, K. Lloyd-Williams, Z.G. Toth, H. Platt, K. Mekli, A. Payton, R. Elliott, S.R. Williams, I.M. Anderson, J.F.W. Deakin, The CREB1-BDNF-NTRK2 pathway in depression: Multiple gene-cognition- environment interactions, Biol. Psychiatry. 69 (2011) 762–771. doi:10.1016/j.biopsych.2010.11.019.

[120]C.A. Martin-Fairey, A.A. Nunez, Circadian modulation of memory and plasticity gene products in a diurnal species, Brain Res. 1581 (2014) 30–39. doi:10.1016/j.brainres.2014.07.020.