JOURNAL OFAvailable online at www.sciencedirect.com

PSYCHIATRIC

RESEARCHJournal of Psychiatric Research 42 (2008) 653–658

www.elsevier.com/locate/jpsychires

Declarative and procedural memory consolidation during sleepin patients with borderline personality disorder

Orla P. Hornung *, Francesca Regen, Claudia Warnstedt, Ion Anghelescu,Heidi Danker-Hopfe, Isabella Heuser, Claas-Hinrich Lammers

Department of Psychiatry and Psychotherapy, Charite-University Medicine Berlin, Campus Benjamin Franklin, Eschenallee 3, 14050 Berlin, Germany

Received 26 April 2007; received in revised form 5 July 2007; accepted 5 July 2007

Abstract

Borderline personality disorder (BPD) is characterized by changes in subjective and objective measures of sleep quality. As recentfindings point to the importance of sleep in memory consolidation, sleep-related memory consolidation was investigated in 15 femaleBPD patients (mean age 26.1 ± 6.1 years) and 15 female healthy controls (mean age 25.6 ± 6.8 years). Before and after the study night,declarative and procedural memory performance was tested by a paired associate list and a mirror tracing task. Subjective sleep qualitywas assessed by a sleep questionnaire, objective sleep quality was measured by a portable sleep recording device. During the study nightthe restorative value of sleep was significantly reduced in BPD patients (p < 0.001), while objective sleep quality showed a trend for longerREM sleep duration (p = 0.054). No significant differences were found regarding overnight performance improvement in the declarativeand procedural memory tasks. Present findings suggest that declarative and procedural memory consolidation during sleep is intact inBPD patients.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Borderline personality disorder; Sleep; Memory consolidation; Declarative and procedural memory

1. Introduction

Sleep in patients with borderline personality disorder(BPD) is characterized by changes in sleep architectureand sleep continuity parameters. With regard to sleeparchitecture, changes in rapid eye movement (REM) sleepsuch as shortening of REM latency, increases in REM den-sity and longer durations of REM sleep have been observedrepeatedly in BPD patients (Battaglia et al., 1993, 1999;Asaad et al., 2002). Moreover, prolonged sleep onsetlatency, reduced total sleep time and sleep efficiency,greater percentages of wake after sleep onset as well asreduced amounts of stage 2 non-REM (NREM) sleepand slow wave sleep (SWS) have been reported for BPDpatients (De la Fuente et al., 2001, 2004; Philipsen et al.,

0022-3956/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jpsychires.2007.07.001

* Corresponding author. Tel.: +49 30 84458771, fax: +49 30 84458233.E-mail address: orla.hornung@charite.de (O.P. Hornung).

2005). Finally, subjective sleep quality has been found tobe markedly impaired in patients with BPD (Philipsenet al., 2005).

Declarative memory, i.e., memory for facts and events,critically depends on the integrity of the hippocampus,while procedural memory, i.e., memory for skills and hab-its, relies primarily on the striatum (Squire and Zola, 1996).Previous research in healthy young adults indicates thatdeclarative memory benefits from early nocturnal sleep,when slow wave sleep predominates, whereas proceduralmemory is enhanced through late nocturnal sleep, whenREM sleep prevails (Plihal and Born, 1997; Plihal et al.,1999). Similarly, selective REM sleep deprivation wasfound to impair procedural memory consolidation com-pared to selective deprivation of SWS in healthy youngadults (Karni et al., 1994). Other findings suggest that anoptimal level of memory consolidation is only reached ifSWS precedes REM sleep during the course of sleep (Gais

654 O.P. Hornung et al. / Journal of Psychiatric Research 42 (2008) 653–658

et al., 2000; Stickgold et al., 2000). In this context, it is ofinterest to note that patients with BPD are known to exhi-bit cognitive dysfunctions, which can be related to alteredbrain activities (Beblo et al., 2006; Fertuck et al., 2006;Lange et al., 2005; Ruocco, 2005). These neurocognitivedeficits could possibly interact with processes of sleep-related memory consolidation in BPD patients, which hasnot yet been investigated.

In the present study, declarative and procedural memoryconsolidation during sleep was investigated in female BPDpatients and healthy controls. The study protocol wasrestricted to female participants, because about 76% ofBPD patients are female (Widiger and Weissman, 1991).It was hypothesized that sleep-related memory consolida-tion in BPD patients differs from that of healthy controlsbased on the changes in sleep characteristics observed inthis patient group. More specifically, it was expected thatthe increase in REM sleep observed in BPD patients wouldlead to a significant improvement in overnight proceduralmemory consolidation compared to healthy controls.Based on the reductions of SWS reported for BPD patients,it was hypothesized that overnight declarative memoryconsolidation would be significantly impaired. To ourknowledge, this is the first study to investigate sleep-relatedmemory consolidation in patients with BPD.

2. Methods

2.1. Participants

Fifteen female BPD patients and fifteen healthy con-trols, matched for age, sex and years of education, wereincluded in the study. All BPD patients fulfilled theDSM-IV diagnostic criteria for BPD and were recruitedas inpatients from the Department of Psychiatry and Psy-chotherapy, Charite-University Medicine Berlin, CampusBenjamin Franklin. Intake of any medication known toaffect sleep characteristics was stopped at least three daysprior to study participation. Healthy controls were mainlyrecruited by newspaper advertisements and were paid fortheir participation. A telephone screening was conducted

Table 1Demographic data and descriptive results from self-report measures of depres

BPD patients

Age (years) n = 1526.1 ± 6.1

Education (years) n = 1512.6 ± 1.3

MWT-A (IQ) n = 15110.1 ± 12.4

BDI (total score) n = 1424.7 ± 9.0

SCL-90 R (global severity index) n = 1376.3 ± 5.5

PSQI (sleep quality index) n = 1410.8 ± 3.9

MWT-A: ‘‘Mehrfachwahl-Wortschatz-Intelligenztest’’ Version A (premorbidChecklist-90-Revised, global severity index (t-values); PSQI: Pittsburgh Sleep

to exclude any volunteers with psychiatric disorders, psy-chiatric or psychotherapeutic treatment, sleep disordersor substance abuse. Written informed consent wasobtained prior to the study. The present study was con-ducted in accordance with the Declaration of Helsinkiand the study protocol was approved by the ethics commit-tee of the Charite-University Medicine Berlin, CampusBenjamin Franklin. For demographic data and descriptiveresults from self-report measures of depression, psycho-pathological symptoms and sleep quality see Table 1.

Please note that three of the fifteen patients had acomorbid diagnosis of major depression. Furthermore,nine patients had received antidepressant medication priorto study participation, including citalopram, paroxetine,sertraline, venlafaxine and trimipramine.

2.2. Study design

All participants took part in declarative and proceduralmemory tasks before and after the study night. The mem-ory tasks were carried out between 8.30 PM and 9:00 PMthe evening before the study night and between 7:30 AMand 8:00 AM the morning thereafter. All participants wereasked to go to bed at their regular bedtimes after the even-ing testing. While BPD patients slept in the hospital,healthy controls spent the night at home and returnedthe following morning. All participants were requestednot to smoke at least one hour before the test sessionsand not to consume any caffeine at least four hours aheadof testing. Prior to the study, all participants had beeninstructed on how to apply the portable sleep recordingdevice. After the evening testing, a sleep questionnairewas handed out to the participants.

2.3. Measures of sleep

Subjective sleep quality was assessed by a standardizedsleep questionnaire the morning after the study night (Hoff-mann et al., 1997). Participants were asked to estimate theduration of sleep onset latency (time between bedtime andoccurrence of first sleep stage different from stage 1 NREM

sion, psychopathological symptoms and sleep quality (M ± SD)

Healthy controls Statistics

n = 15 t(28) = 0.227ns25.6 ± 6.8

n = 15 t(28) = 0.081ns12.5 ± 0.9

n = 15 t(28) = �0.050ns110.3 ± 9.3

n = 15 t(27) = 6.477p < 0.0017.1 ± 5.2

n = 15 t(26) = 5.811p < 0.00150.9 ± 14.9

n = 14 t(27) = 5.749p < 0.0014.4 ± 1.7

intelligence); BDI: Beck Depression Inventory; SCL-90-R: SymptomQuality Index (0 = maximum sleep quality).

O.P. Hornung et al. / Journal of Psychiatric Research 42 (2008) 653–658 655

sleep), total sleep time (sleep period time minus time spentawake) and wake after sleep onset (time spent awake fromsleep onset to final awakening). Furthermore, the question-naire included a five-point rating scale regarding the restor-ative value of sleep.

Objective sleep quality was assessed by a one-channel,portable sleep electroencephalogram (EEG) recordingdevice using neuronal networks techniques for automaticsleep stage classification (QUISI�; Baumgart-Schmittet al., 2002). The sleep parameters investigated were the fol-lowing: sleep onset latency, total sleep time, wake aftersleep onset, NREM sleep stages 1 and 2, SWS, REM sleepand REM latency (time between sleep onset and firstappearance of REM sleep).

2.4. Memory tasks

Evening and morning declarative memory performancewas tested by a paired associate word list. The eveningbefore the study night, a list of 34 paired associates wasread out to the participants at a rate of 3 s per word pair.Thereafter, participants were asked to recall the targetwords of the paired associate list, whenever the corre-sponding cue words were read out, and direct feedbackwas given with regard to the correctness of the answers.Afterwards the list of paired associates was read out againand a second recall trial followed. The morning after thestudy night, a third recall trial was conducted in the samefashion and after the list of paired associates had been readout again a fourth recall trial followed. To rule out primacyand recency effects, the first and last two paired associatesof the list remained the same across the trials serving asbuffer items, which were not included in the statistical anal-yses. The remaining 30 paired associates were presented inrandomized order in each trial to further exclude serialposition effects. For statistical analyses, the number oferrors, i.e. words not recalled, was averaged within eveningas well as morning trials resulting in one evening and onemorning mean error score.

Evening and morning procedural memory performancewas tested by a mirror tracing task. In this task, a light-sen-sitive stylus was used to draw along black lines of geomet-ric figures, which could only be viewed through a mirror.The evening before the study night, a simple practice figurein the shape of a star was drawn seven times in order toacquire a basic level of the mirror tracing skill. Thereafter,six more complex test figures in the shape of a human beingwere presented consecutively. The outer lines of the test fig-ures were 0.8 cm wide. The morning after the study night,participants were asked to draw the practice figure andeach of the six test figures one more time. Procedural mem-ory performance was measured by recording number oferrors, i.e., number of offline instances, performance timeand error time, i.e., time offline, for each test figure. Forstatistical analyses, average scores of performance mea-sures within the evening and morning test figures were

calculated resulting in one evening and one morning meanscore for each of the three performance measures.

2.5. Statistical analysis

Statistical analyses were conducted with SPSS 13.0 forWindows. t-Tests for independent samples were appliedto investigate group differences between BPD patientsand healthy controls. Paired sample t-tests were used toinvestigate differences between evening and morning mem-ory performance within the participants. Missing data weretreated by listwise deletion in each specific analysis and thesignificance level was set to p < 0.05 (two-tailed). Due tothe small sample sizes, data were also analysed by nonpara-metric tests, which did not change the pattern of resultspresented in the following section of this paper.

3. Results

Table 2 displays subjective and objective measures ofsleep during the study night for patients with BPD andhealthy controls. With regard to subjective sleep quality,no significant differences were found for estimations ofsleep onset latency, total sleep time and wake after sleeponset between BPD patients and healthy controls. In con-trast, the restorative value of sleep was significantlyreduced in BPD patients compared with healthy controls.Objective sleep quality did not differ significantly betweenthe two groups regarding sleep onset latency, total sleeptime, wake after sleep onset, stages 1 and 2 NREM sleep,SWS and REM latency. It is of interest to note that BPDpatients showed a trend for longer REM sleep durations.

Before focussing on the results with regard to group dif-ferences, please note that a significant overnight memoryperformance improvement was observed for all measuresof declarative and procedural memory performance withinthe complete study sample (see Fig. 1). Patients with BPDand healthy controls did not differ significantly with regardto evening or morning memory performance in any of themeasures of declarative and procedural memory perfor-mance investigated, as illustrated in Table 3. In addition,no significant group differences were found regarding thechanges from evening to morning memory performance.

4. Discussion

Contrary to our hypotheses, we did not find any signif-icant group differences with regard to objective measures ofsleep quality in the present study, except for a trend forlonger REM sleep duration in BPD patients. The two studygroups differed significantly with regard to the restorativevalue of sleep, which was reduced in BPD patients. Neitherin a declarative, nor in a procedural memory task did sleep-related, overnight performance improvement of BPDpatients differ significantly from that of healthy controls.The results of the present study suggest that declarativeand procedural memory consolidation during sleep is intact

Table 2Subjective and objective measures of sleep during the study night (M ± SD)

BPD patients Healthy controls Statistics

Subjective sleep quality

Sleep onset latency (min) n = 13 n = 15 t(26) = 1.245ns44.3 ± 32.9 30.3 ± 27.4

Total sleep time (min) n = 15 n = 15 t(28) = �0.283ns375.0 ± 80.0 383.0 ± 74.6

Wake after sleep onset (min) n = 12 n = 14 t(24) = 1.029ns17.9 ± 23.2 9.9 ± 16.6

Restorative value of sleepa n = 15 n = 14 t(27) = 4.379p < 0.0014.0 ± 1.0 2.7± 0.5

Objective sleep quality

Sleep onset latency (min) n = 9 n = 12 t(19) = �0.256ns31.6 ± 15.2 33.3 ± 15.5

Total sleep time (min) n = 9 n = 12 t(19) = 1.353ns385.3 ±57.9 345.3 ± 73.1

Wake after sleep onset (min) n = 9 n = 12 t(19) = 1.031ns22.2 ± 22.3 14.3 ± 12.3

Stage 1 NREM (min) n = 9 n = 12 t(19) = �0.279ns24.2 ± 9.8 26.6 ± 25.0

Stage 2 NREM (min) n = 9 n = 12 t(19) = �0.459ns138.8 ± 45.2 148.6 ± 50.7

Stage SWS (min) n = 9 n = 12 t(19) = 0.873ns95.3 ± 30.3 81.9 ± 37.9

Stage REM (min) n = 9 n = 12 t(19) = 2.057p = 0.054120.8 ± 38.2 85.3 ± 39.9

REM latency (min) n = 9 n = 12 t(19) = �0.208ns57.6 ± 44.7 61.4 ± 39.0

a Restorative value of sleep was determined by a five-point rating scale (1 = very high to 5 = very low).

Dec

lara

tive

Mem

ory

Task

Num

ber o

f Erro

rs

0

5

10

15

20

25

30

*****

Proc

edur

al M

emor

y Ta

skN

umbe

r of E

rrors

0

10

20

30

40

50

***

Proc

edur

al M

emor

y Ta

skPe

rform

ance

Tim

e (s

econ

ds)

0

20

40

60

80

100

120

140

160

***

Proc

edur

al M

emor

y Ta

skEr

ror T

ime

(sec

onds

)

0

2

4

6

8

10

12

14

a b

c d

Fig. 1. Evening and morning memory performance measures within the complete study sample (M ± SD). Evening (black bars), morning (grey bars).Declarative memory task: number of errors (a), procedural memory task: number of errors (b), performance time (c), error time (d). **p < 0.01,***p < 0.001.

656 O.P. Hornung et al. / Journal of Psychiatric Research 42 (2008) 653–658

Table 3Declarative and procedural memory performance before and after thestudy night (M ± SD)

BPD patients Healthy controls Statistics

Declarative memory task

Number of errors

Evening n = 15 n = 15 t(28) = 0.000ns12.6 ± 3.0 12.6 ± 4.6

Morning n = 15 n = 15 t(28) = 0.125ns5.0 ± 3.7 4.9 ± 3.6

Change n = 15 n = 15 t(28) = �0.122ns7.5 ± 4.5 7.7 ± 2.8

Procedural memory task

Number of errors

Evening n = 15 n = 15 t(28) = 0.749ns18.3 ± 17.4 14.1 ± 12.9

Morning n = 15 n = 15 t(28) = �0.045ns9.8 ± 9.2 10.0 ± 16.7

Change n = 15 n = 15 t(28) = 1.310ns8.5 ± 10.4 4.1 ± 7.8

Performance time

Evening n = 15 n = 15 t(28) = 0.806ns86.4 ± 24.9 79.6 ± 21.2

Morning n = 15 n = 15 t(28) = 0.666ns71.0 ± 32.3 64.6 ± 19.3

Change n = 15 n = 15 t(28) = 0.065ns15.4 ± 17.0 15.1 ± 10.4

Error time

Evening n = 15 n = 15 t(28) = 0.574ns4.6 ± 5.7 3.5 ± 4.0

Morning n = 15 n = 15 t(28) = 0.439ns2.3 ± 2.5 1.9 ± 3.1

Change n = 15 n = 15 t(28) = 0.588ns2.2 ± 3.5 1.6 ± 1.6

O.P. Hornung et al. / Journal of Psychiatric Research 42 (2008) 653–658 657

in BPD patients, even though subjective sleep quality ismarkedly impaired in this patient group.

Increased amounts of REM sleep have been reportedrepeatedly for BPD patients when compared to healthycontrols (De la Fuente et al., 2001, 2004; Asaad et al.,2002). Nevertheless, in the present study the trend forlonger durations of REM sleep in BPD patients did notaffect overnight procedural memory consolidation, eventhough previous research suggests that REM sleep playsa beneficial role in this context (Karni et al., 1994; Plihaland Born, 1997; Plihal et al., 1999; Gais et al., 2000; Stick-gold et al., 2000). A possible explanation for this findingwould be that the differences in REM sleep durationobserved between BPD patients and healthy controls werenot sufficient to produce such an effect. Considering thesmall sample sizes, it could also be cautiously argued thatREM sleep showed a trend to be increased in patients withBPD and the improvement in procedural memory perfor-mance was twice as good, however not significant, whichfits the original hypothesis. An alternative explanation con-cerns the choice of memory tasks. There is increasing evi-dence that sleep, and in particular REM sleep, facilitatesthe consolidation of emotional memories (Wagner et al.,2001, 2002, 2005). Since BPD is strongly associated withaffective symptoms, the consolidation of emotional memo-ries may be of much higher relevance for the disorder than

the consolidation of declarative and procedural memories.Therefore, further research is needed to investigate sleep-related consolidation of emotionally toned material inBPD.

SWS has been found to be reduced in BPD patients (Dela Fuente et al., 2001, 2004). However, this finding couldnot be replicated by others (Philipsen et al., 2005). In thepresent study, we did not observe any significant differencein SWS between patients with BPD and healthy controlsindicating that changes in SWS are not very robust in thispatient group. Due to this lack of difference, it is not sur-prising that overnight declarative memory consolidationdid not differ between BPD patients and healthy controlsin the present study as was expected.

In the present study, we could confirm the markedimpairment in subjective sleep quality in patients withBPD reported previously (Philipsen et al., 2005). However,inspite of this impairment in subjective sleep quality,patients with BPD showed intact sleep-related declarativeand procedural memory consolidation. This finding indi-cates that subjective sleep quality is not a crucial factorfor sleep-related memory consolidation. The potentialimpact of subjective sleep quality on sleep-related memoryconsolidation has not yet been investigated systematically.Therefore, although this is a very interesting finding, itneeds to be replicated.

This study has several limitations. First of all, sleep wasrecorded by a portable sleep recording device, which givesan impression of sleep architecture and allows for objectiveverification of sleep disturbances, but can certainly notreplace standard polysomnography in the sleep laboratory.Furthermore, due to technical problems a high amount ofmissing data was produced for objective measures of sleepresulting in small sample sizes, especially for BPD patients,requiring a more cautious interpretation of the statisticalfindings.

It is important to note that three of the fifteen patientshad a comorbid diagnosis of major depression. This is ofrelevance as depression is also associated with sleepchanges, which could affect overnight memory consolida-tion (Riemann et al., 2001). However, when repeating theanalyses on sleep differences excluding these three patients,the result pattern remained the same and can therefore notbe attributed to comorbidity of major depression. Anotherlimitation of the study concerns the antidepressant premed-ication, which nine patients had received prior to studyparticipation. It is well known that the discontinuation oftreatment with REM sleep suppressing antidepressantscan cause a REM rebound. Therefore, the relative increaseof REM sleep in BPD patients could simply represent aREM rebound phenomenon. We tested this hypothesisand compared REM sleep duration between those patientswho had discontinued antidepressant treatment and thosewho had not received any antidepressant premedication.We did not find any significant differences between thesetwo groups with regard to REM sleep duration. Hence,the relative increase of REM sleep in BPD patients cannot

658 O.P. Hornung et al. / Journal of Psychiatric Research 42 (2008) 653–658

be explained by REM rebound due to antidepressantpremedication.

Finally, if we had found a significant difference betweenBPD patients and healthy controls with regard to overnightmemory consolidation, we would not have been able toexclude that such differences also occur during wakefulnesswithin the present study design. Hence, future studydesigns in this research context should include periods ofwakefulness.

To our knowledge, this is the first study to investigatesleep-related memory consolidation in patients with BPD.Future research in this context should focus on other mem-ory systems such as emotional memory and further examinequalitative components of REM sleep. Previous research inolder adults suggests that qualitative aspects of REM sleepare crucial for plasticity-related processes during sleep andrecent findings in young adults indicate that cortisol playsa role in the consolidation of emotional memories (Wagneret al., 2005; Hornung et al., 2007). Hence, these aspects needto be considered in future research.

Conflict of interest

All authors declare that they have no conflict of interest.

Contributors

All authors contributed to and have approved the finalmanuscript.

Role of funding source

Funding for this study was provided by the Charite-University Medicine Berlin; the Charite-University Medi-cine Berlin had no further role in study design; in thecollection, analysis and interpretation of data; in the writ-ing of the report; and in the decision to submit the paperfor publication.

Acknowledgements

We would like to thank Jan Born and his research groupfor providing us with the standardized paired associateword list material and the test figures of the mirror tracingtask.

References

Asaad T, Okasha T, Okasha A. Sleep EEG findings in ICD-10 borderlinepersonality disorder in Egypt. Journal of Affective Disorders2002;71:11–8.

Battaglia M, Ferini-Strambi L, Smirne S, Bernardeschi L, Bellodi L.Ambulatory polysomnography of never-depressed borderline subjects:A high-risk approach to rapid eye movement latency. BiologicalPsychiatry 1993;33:326–34.

Battaglia M, Ferini Strambi L, Bertella S, Bajo S, Bellodi L. First-cycleREM density in never-depressed subjects with borderline personalitydisorder. Biological Psychiatry 1999;45:1056–8.

Baumgart-Schmitt R, Wenzel A, Danker-Hopfe H, Herrmann WM.Genetic programming approach for the optimal selection of combina-tions of neuronal networks to classify sleep stages by QUISI. Methodsand Findings in Experimental and Clinical Pharmacology2002;24:27–32.

Beblo T, Saavedra AS, Mensebach C, Lange W, Markowitsch HJ, Rau H,et al. Deficits in visual functions and neuropsychological inconsistencyin Borderline Personality Disorder. Psychiatry Research2006;145:127–35.

De la Fuente JM, Bobes J, Vizuete C, Mendlewicz J. Sleep-EEG inborderline patients without concomitant major depression: A com-parison with major depressives and normal control subjects. PsychiatryResearch 2001;105:87–95.

De la Fuente JM, Bobes J, Morlan I, Bascaran MT, Vizuete C, LinkowskiP, et al. Is the biological nature of depressive symptoms in borderlinepatients without concomitant Axis I pathology idiosyncratic? SleepEEG comparison with recurrent brief, major depression and controlsubjects. Psychiatry Research 2004;129:65–73.

Fertuck EA, Lenzenweger MF, Clarkin JF, Hoermann S, Stanley B.Executive neurocognition, memory systems, and borderline personal-ity disorder. Clinical Psychology Review 2006;26:346–75.

Gais S, Plihal W, Wagner U, Born J. Early sleep triggers memory for earlyvisual discrimination skills. Nature Neuroscience 2000;3:1335–9.

Hoffmann RM, Muller T, Hajak G, Cassel W. Abend-Morgenprotokollein Schlafforschung und Schlafmedizin – Ein Standardinstrument furden deutschsprachigen Raum. Somnologie 1997;1:103–9.

Hornung OP, Regen F, Danker-Hopfe H, Schredl M, Heuser I. Therelationship between REM sleep and memory consolidation in old ageand effects of cholinergic medication. Biological Psychiatry2007;61:750–7.

Karni A, Tanne D, Rubenstein BS, Askenasy JJ, Sagi D. Dependence onREM sleep of overnight improvement of a perceptual skill. Science1994;265:679–82.

Lange C, Kracht L, Herholz K, Sachsse U, Irle E. Reduced glucosemetabolism in temporo-parietal cortices of women with borderlinepersonality disorder. Psychiatry Research 2005;139:115–26.

Philipsen A, Feige B, Al-Shajlawi A, Schmahl C, Bohus M, Richter H,et al. Increased delta power and discrepancies in objective andsubjective sleep measurements in borderline personality disorder.Journal of Psychiatric Research 2005;39:489–98.

Plihal W, Born J. Effects of early and late nocturnal sleep on declarativeand procedural memory. Journal of Cognitive Neuroscience1997;9:534–47.

Plihal W, Pietrowsky R, Born J. Dexamethasone blocks sleep inducedimprovement of declarative memory. Psychoneuroendocrinology1999;24:313–31.

Riemann D, Berger M, Voderholzer U. Sleep and depression – resultsfrom psychobiological studies: An overview. Biological Psychology2001;57:67–103.

Ruocco AC. The neuropsychology of borderline personality disorder: Ameta-analysis and review. Psychiatry Research 2005;137:191–202.

Squire LR, Zola SM. Structure and function of declarative and nonde-clarative memory systems. Proceedings of the National Academy ofSciences USA 1996;93:13515–22.

Stickgold R, Whidbee D, Schirmer B, Patel V, Hobson JA. Visualdiscrimination task improvement: A multi-step process occurringduring sleep. Journal of Cognitive Neuroscience 2000;12:246–54.

Wagner U, Gais S, Born J. Emotional memory formation is enhancedacross sleep intervals with high amounts of rapid eye movement sleep.Learning and Memory 2001;8:112–9.

Wagner U, Fischer S, Born J. Changes in emotional responses to aversivepictures across periods rich in slow-wave sleep versus rapid eyemovement sleep. Psychosomatic Medicine 2002;64:627–34.

Wagner U, Degirmenci M, Drosopoulos S, Perras B, Born J. Effects ofcortisol suppression on sleep-associated consolidation of neutral andemotional memory. Biological Psychiatry 2005;58:885–93.

Widiger TA, Weissman MM. Epidemiology of borderline personalitydisorder. Hospital and Community Psychiatry 1991;42:1015–21.