Boosting Slow Oscillatory Activity Using tDCS ... - Brain Stimulation

Brain Stimulation 9 (2016) 730–739

Contents lists available at ScienceDirect

Brain Stimulation j o u r n a l h o m e p a g e : w w w. b r a i n s t i m j r n l . c o m

Boosting Slow Oscillatory Activity Using tDCS during Early Nocturnal Slow Wave Sleep Does Not Improve Memory Consolidation in Healthy Older Adults Sven Paßmann a,b,*, Nadine Külzow a,b, Julia Ladenbauer a,b, Daria Antonenko a,b, Ulrike Grittner c,d, Sascha Tamm e, Agnes Flöel a,b,d,** a

Department of Neurology, Charité University Hospital Berlin, Charitéplatz 1, 10117 Berlin, Germany NeuroCure Cluster of Excellence, Charité University Hospital Berlin, Charitéplatz 1, 10117 Berlin, Germany c Department for Biostatistics and Clinical Epidemiology, Charité University Hospital Berlin, Charitéplatz 1, 10117 Berlin, Germany d Center for Stroke Research, Charité University Hospital Berlin, Charitéplatz 1, 10117 Berlin, Germany e Department of Psychology, Free University Berlin, Habelschwerdter Alle 45, 14195 Germany b

A R T I C L E

I N F O

Article history: Received 16 November 2015 Received in revised form 21 April 2016 Accepted 26 April 2016 Available online 28 April 2016 Keywords: Transcranial slow oscillating stimulation Older subjects Consolidation Declarative memory Visuo-spatial task

A B S T R A C T

Background: Previous studies have demonstrated an enhancement of hippocampal-dependent declarative memory consolidation, associated slow wave sleep (SWS) and slow wave activity (SWA) after weak slow oscillatory stimulation (so-tDCS) during early non-rapid eye movement sleep (NREM) in young adults. Recent studies in older individuals could not confirm these findings. However, it remained unclear if this difference was due to variations in study protocol or to the age group under study. Objective/Hypothesis: Here, we asked if so-tDCS promotes neurophysiological events and associated sleepdependent memory in the visuo-spatial domain in older adults, using a stimulation protocol that closely resembled the one employed in young adults. Methods: In a randomized, placebo-controlled single-blind (participant) crossover study so-tDCS (0.75 Hz; max. current density 0.522 mA/cm2) vs. sham stimulation was applied over the frontal cortex of 21 healthy older subjects. Impact of stimulation on frequency band activity (linear mixed models), two declarative and one procedural memory tasks (repeated measures ANOVA) and percentage of sleep stages (comparison of means) was assessed. Results: so-tDCS, as compared to sham, increased SWA and spindle activity immediately following stimulation, accompanied by significantly impaired visuo-spatial memory consolidation. Furthermore, verbal and procedural memory remained unchanged, while percentage of NREM sleep stage 4 was decreased over the entire night (uncorrected). Conclusion: so-tDCS increased SWA and spindle activity in older adults, events previously associated with stimulation-induced improved consolidation of declarative memories in young subjects. However, consolidation of visuo-spatial (primary outcome) and verbal memories was not beneficially modulated, possibly due to decline in SWS over the entire night that may have prevented and even reversed immediate beneficial effects of so-tDCS on SWA. © 2016 Elsevier Inc. All rights reserved.

Introduction Sleep is a critical mediator of memory consolidation [1]. Robust empirical evidence, although mostly obtained from healthy young

* Corresponding author. Tel.: +49 30 450 560 395; fax: +49 30 450 7 539 939. E-mail address: [email protected] (S. Paßmann). ** Corresponding author. Tel.: +49 30 450 560 284; fax: +49 30 450 7 560 284. E-mail address: agnes.fl[email protected] (A. Flöel). 1935-861X/© 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brs.2016.04.016

subjects, demonstrates beneficial effects of nocturnal sleep on subsequent consolidation of hippocampal-dependent declarative memory. Especially slow oscillatory activity (50% of items in the visuo-spatial task. Thus, 21 subjects remained (11 male; mean age 65 ± 1, mean years of education 17 ± 4) for final analysis. They did not differ from excluded participants in sex, age and education. For sleep stage analyses of the entire night data from two further subjects had to be excluded due to incomplete EEG recordings. Baseline assessments In addition, participants underwent a standard neuropsychological test battery assessing verbal memory (German version of Auditory Verbal Learning Test (AVLT [22]), working memory [23], and executive functions (Stroop color-word-test [24]). The affective state at the time of testing was determined by the Positive and Negative Affective Scale (PANAS [25]). Subjective and objective sleep habits were recorded by the German version of Pittsburgh Sleep Quality Index (PSQI [26]), German version of Epworth Sleepiness Scale (ESS [27]), German version of Morningness–EveningnessQuestionnaire (d-MEQ [28]) and the Essener questionnaire on age and sleepiness (“Essener Fragenbogen Alter und Schläfrigkeit” [29]). Sleep diaries and actigraphy (Actigraph GT3X+, LLC Pensacola, USA) were used to monitor regular sleep–wake-cycles one week prior to each experimental night. To ensure similar conditions between experimental nights (so-tDCS vs. sham), participants were further asked to avoid caffeinated drinks or alcohol during the last 1.5 h before starting each experimental night and to maintain normal sleep duration the night before. Adherence to these instructions was evaluated before starting the experimental night, using a questionnaire which also gathered current wellbeing (yes or no) as well as a subjective rating of sleep quality of the last night (scale ranging from 0 – miserable to 6 – excellent).

732

S. Paßmann et al. / Brain Stimulation 9 (2016) 730–739

The study was approved by the Ethics Committee of the Charité University Hospital Berlin, Germany and was conducted in accordance with the declaration of Helsinki. All subjects received a small reimbursement and gave written informed consent prior to the study. Experimental procedure After an adaptation night to habituate participants with the experimental set-up for sleep monitoring and to introduce the memory tests (comprising short versions of each memory tasks) the two experimental nights followed a balanced cross-over design for stimulation or sham condition separated by 2–3 weeks to prevent carry-over effects. All sleep sessions took place at the sleep laboratory of the Free University Berlin, Germany. For each sleep session, participants arrived at the sleep laboratory at 8.00 p.m. Adherence to previously issued instructions (no caffeine or alcohol intake within the last 1.5 h before testing, regular sleep time the night before) was confirmed by a questionnaire. All participants followed these instructions. Afterwards the subjects were prepared for EEG recordings and stimulation electrodes were attached. They performed the computerized visuo-spatial (primary outcome) and two additional memory tasks (word-pair, finger-sequence tapping) using one of two different (parallel) versions for each experimental condition and task. The subjects went to bed 30 min after the end of memory testing and were asked to sleep until 7.30 a.m. in order to keep conditions comparable for post sleep memory testing between individuals. It was not allowed to read or use other aid to fall asleep. In case of

being awake earlier, participants were instructed to stay in bed until 7.30 a.m. Approximately 20 min after getting up, and a short standardized meal without caffeinated drinks, delayed recall on memory tasks was tested (see Fig. 2). During nocturnal sleep so-tDCS or sham stimulation was applied (10 vs. 11 received so-tDCS vs. sham on the 1st night; for details see below). Subjects were blinded for stimulation condition during the complete study and were asked on the last morning after completing all study-related procedures whether they had felt any sensations or pain and whether they were able to guess when the stimulation had been applied. Before learning in the evening and delayed recall the next morning, attention (Testbatterie zur Aufmerksamkeitsprüfung V2.3, TAP [30]), activation (Visual Analogue Scale, VAS for tension [31]), sleepiness (Tiredness Symptoms Scale, TSS [32]; VAS for tiredness) and affective state (PANAS) were tested using standard psychometric assessments to control for possible confounding effects. Slow oscillatory stimulation (so-tDCS) The stimulation protocol employed in this study is closely resembled those by Marshall et al. [10]. Anodal current was delivered by a battery driven device (DC-Stimulator, neuroConn GmbH Ilmenau, Germany) via sintered Ag/AgCl electrodes (8 mm in diameter), bilaterally at frontal electrode sites F3 and F4 according to the international EEG 10–20 system. Induced current oscillating sinusoidally between 0 and 260 μA at a frequency of 0.75 Hz resulted in a maximum current density of 0.522 mA/cm2. Resistance was kept always below 5 kΩ. Anodes were mounted into an EASY cap (Falk

Figure 2. Study design. Upper panel: Overview. Subjects performed a verbal, a visuo-spatial (blue squares; declarative) and a procedural memory task (black squares) in the indicated order following psychometric control tests (black square). During the first sleep cycle of a regular nocturnal sleep of at least 8.5 hours (11.00 p.m. to 7.30 a.m.) either so-tDCS or sham stimulation was applied (within-subject design, randomized order). Next morning, retrieval and psychometric control tests were tested in the same order. Bottom panel: Comparison of stimulation protocols. (a) Paßmann et al. (present study): Red bars indicate 5 min blocks of so-tDCS/sham, with each bar placed below the respective sleep stage in the example hypnogram. Note that in the protocol of Paßmann et al. the stimulation blocks started 4 min after sleep stage 2 onset, discontinued in this example as the subject moves into sleep stage 1 or wakes up, and resumed after subjects again enters sleep stage 2. (b) Marshall et al. (2006): stimulation blocks started 4 min after sleep stage 2 onset. Once started, stimulation was conducted via a fixed protocol of 5 blocks of stimulation separated by 1-min stimulation-free interval, regardless of subject’s current sleep stage during intervals. REM: rapid eye movement sleep (vertical black bar in the hypnogram); S1–S4: NREM sleep stages 1–4.

S. Paßmann et al. / Brain Stimulation 9 (2016) 730–739

Minow Services, Munich, Germany) and referenced to the mastoid electrodes (ipsilateral). In the sham condition stimulation electrodes were placed as in the stimulation condition, but the tDCS device remained off. so-tDCS was applied starting four minutes after the subject had entered stable NREM sleep stage 2 (no transitions back to sleep stage 1 or wakefulness) and was applied five times, each as a 5-min epoch of stimulation, separated by 1-min intervals, free of stimulation and artifacts. Due to strong and long lasting artifacts in our unfiltered online EEG signal, the marker for the beginning of the 1-min stimulation-free interval was always set manually 40 s after the end of the stimulation. Contrary to Marshall et al. [10] and in accordance with the protocol applied by Antonenko et al. [33], sleep was monitored online before and after stimulation. This procedure takes into account higher sleep fragmentation in older adults and ensured that the subjects were at least in NREM sleep stage 2 at time of stimulation. The stimulation was delayed when subjects were awake or in sleep stage 1 until subjects had again entered NREM sleep stage 2. The allocation of stimulation was prepared by a member of the research group who was neither involved in testing nor was in contact with the participants. The investigator administering the behavioral tasks as well as the so-tDCS application was blinded for condition only until the participant fell asleep in the first experimental night. EEG recordings, sleep monitoring and analysis EEG data (ground: FCz, impedance 50% of items in the visuospatial task. Verbal and procedural memory task Similar to previous studies [14,36] we used a paired associate word list learning paradigm (PAL) which was applied prior to the visuo-spatial task. However, we now created two new lists with 40 semantically related, non-emotional German nouns and 4 additional buffer items to prevent primacy/recency effects (for more details see supplementary material). Both lists follow an A-B/A-C scheme using a category word (A) and a semantically related instance word (B or C) in a randomized order for experimental nights. All verbal responses were recorded (voice recorder) and stored for later offline analyses. For testing procedural memory, a finger-sequence tapping task (adapted from Ref. 37) was used. The subjects were instructed to repeat a presented five element-sequence (e.g., 4-2-3-1-4) with the non-dominant hand as fast and accurate as possible within a 30 s trial using four numeric keys on the keyboard (for more details see supplementary material). Offline sleep-stage scoring Sleep stage analyses comprised both the entire night (scoring based on 30s epochs) and the five 1-min stimulation-free intervals after each stimulation period as well as the 1-min poststimulation interval (scoring based on 10s epochs). Sleep records were visually scored according to standard criteria [38] in NREM sleep stages 1, 2, 3, and 4 and REM sleep using program SchlafAus v1.5 (Steffen Gais, Lübeck; Germany). Epochs during stimulation were not scored due to heavy stimulation signal artifacts. Likewise, corresponding epochs in the sham condition were not scored, in order to obtain comparable time and proportions of sleep stages. Statistical analyses Spectral power Linear mixed models [39] were calculated separately for all frequency bands of interest and electrode sites (see above) based on this model:

yij = β0 + TIMEx1ij + ORDERx2 j + STIMx3ij + BASELINEx4 j + TIME ∗ STIMx5ij + u0 j + ε ij where yij is the specific frequency band (log-transformed) value at time point i of subject j, β0 is the fixed effect intercept, x1ij is the value for time point centered (values: −2, −1, 0, 1, 2) at time point i in subject j, x2 j is the value for ORDER of stimulation (1: first night so-tDCS, second night sham; 0: first night sham, second night sotDCS) in subject j, x3ij is the value for STIMULATION (values 1: stimulation 0: sham) for time point i and subject j, x4 j is the value for BASELINE (log-transformed) of the analyzed frequency band for subject j, x5ij is the value for the interaction for STIMULATION and TIME for time point i and subject j, u0 j is the residual or random 2 ), ε ij is the effect for the intercept for subject j (mean 0, variance σ u0 error term for time point i and subject j (mean 0, variance σ ε2 ).

For immediate stimulation-induced effects (primary outcome) we analyzed the first five 1-min stimulation-free intervals following each stimulation period (factor TIME) which were level-one units nested in subjects (level-two units). Random intercept models tested differences between the two stimulation conditions (so-tDCS, sham). To adjust for baseline differences in each frequency band, the baseline interval was included as a covariate (BASELINE), respectively. The interaction Stimulation × Time (STIM × TIME) assessed whether the slopes of the curves differed between the stimulation conditions. Additionally, the models were adjusted for sequence of stimulations (first so-tDCS or first sham; ORDER). Reported effects of stimulation are based on model-based estimations and post hoc tests. Effects of time-, baseline-, STIM × TIME-interaction- and sequence-values are based on regression coefficients (subsequently shortened to β). To evaluate prolonged offline impact of sotDCS in an exploratory approach, differences in power values between so-tDCS and sham were assessed by paired-samples t-tests at the 1-min interval 60 min post stimulation.

Memory tasks Overnight memory effects (accuracy) in the primary outcome (visuo-spatial memory task) were analyzed by repeated measures analysis of variance (rmANOVA) including the within-subject factors STIM (so-tDCS vs. sham) and TIME (before vs. after sleep). A Greenhouse–Geisser correction for degrees of freedom was used, if indicated. Additional exploratory outcomes (verbal and procedural memory) were analyzed using separate STIM × TIME rmANOVAs.

Sleep stages The impact of so-tDCS on mean time spent in sleep stages (in %; exploratory approach) was analyzed for the entire night as well as for the selected 1-min EEG intervals after each 5-min stimulation period (averaged over the five 1-min stimulation-free intervals), and for the 1-min post-stimulation interval either by pairedsamples t-test or by Wilcoxon signed-rank tests if indicated. Moreover, bivariate correlations to evaluate associations between stimulation induced effects (so-tDCS minus sham) in sleep stages (entire nights) and declarative memory performance were assessed by Pearson’s correlations. Tension and tiredness (assessed via VAS), sleepiness (assessed via TSS), and affective state (assessed via PANAS) were analyzed in separate STIM × TIME rmANOVAs. To statistically control for possible differences between experimental nights in psychometric baseline tests or reports in sleep quality the night before, these scores were incorporated as covariates in ANCOVAs if necessary. All statistical analyses were conducted using SPSS Statistics 22.0 (Statistical Package for the Social Sciences for Windows, IBM Corp., USA). A two-sided significance level α was set to 0.05 in all analyses. Reported effect size is partial η2 (subsequently shortened to η2). Given multiple testing for primary parameters of interest (visuospatial task, six memory-related EEG frequency bands, i.e., SO prefrontal and frontal, slow spindle activity prefrontal and frontal, fast spindle activity central and centro-parietal; so-tDCS compared to sham), all p-values were corrected using the Holm– Bonferroni step-down approach [40]. Only corrected p-values (pc) of less than 0.05 were considered significant. Impact of so-tDCS on memory-related EEG activity such as SO as well as sleep spindle frequency bands and recognition performance in the visuo-spatial memory task was defined as primary outcome. All other tests and comparisons were related to secondary hypotheses, p-values related to secondary hypotheses should be interpreted in a framework of exploratory analysis.

S. Paßmann et al. / Brain Stimulation 9 (2016) 730–739

735

Table 1 Results of separate mixed models analysis for each pre-specified frequency band additionally adjusted for ORDER and STIM × TIME. So-tDCS induced effects for 1-min intervals averaged across interval 1–5 immediately following stimulation. Main effect STIM is expressed as model based post-hoc estimates (mean, standard error) and main effects TIME and BASELINE are expressed as mixed model regression coefficients and standard errors. N = 21. Immediate stimulation induced effects interval 1–5

Measures, N

STIM Mean diff (SE)

p

pc

Time ß (SE)

p

ß (SE)

Baseline p

SO prefrontal SO frontal SlSp prefrontal Slsp frontal FsSp central FsSp centroparietal

204 183 204 183 183 172

0.131 (0.039) 0.112 (0.045) 0.094 (0.029) 0.065 (0.029) 0.086 (0.028) 0.073 (0.030)

0.001 0.013 0.001 0.002 0.003 0.018

0.007 0.013 0.007 0.010 0.012 0.039

0.057 (0.019) 0.067 (0.019) −0.026 (0.014) −0.017 (0.013) −0.046 (0.013) −0.039 (0.014)

0.004 0.001 0.072 0.178 0.72). Delayed stimulation-induced effects (1-min interval 60 min post stimulation; exploratory approach). Stimulation-induced changes were only significant at prefrontal electrode site (so-tDCS: mean 0.78 ± 0.38; sham: mean 0.47 ± 0.52, t (19) = 2.44 p = 0.025)

indicating higher SO activity after so-tDCS compared to sham. Differences between the conditions at frontal electrode site were not significant (so-tDCS: mean 0.87 ± 0.40; sham: mean 0.55 ± 0.59, t(15) = 1.19, p = 0.301). In sum, the power in SO was significantly enhanced immediately after stimulation (corrected). Moreover, power in SO was enhanced in the 1-min interval 60 min after the end of the fifth stimulation (exploratory approach; uncorrected). Spindle activity Immediate stimulation-induced effects (interval 1–5 immediately following each stimulation period, primary outcome). Mixed models analyses resulted in stimulation-induced changes for SlSp (prefrontal: p = 0.001, pc = 0.007; frontal: p = 0.002, pc = 0.010) as well as for FsSp (central: p = 0.003, pc = 0.012; centro-parietal: p = 0.018, pc = 0.039) indicating an increase in power in the spindle frequency bands after so-tDCS compared to sham (see Table 1 and Fig. 4). A negative effect for TIME emerged for FsSp (central: p < 0.001; centro-parietal: p = 0.006) suggesting a decrease in activity over time. No such effects were observed for SlSp at any electrode site (all p-values > 0.072). No other significant effects (main effect ORDER, interaction effect STIM × TIME) were found by so-tDCS in none of the reported electrode sites and frequency bands (all p-values > 0.23). Delayed stimulation-induced effects (1-min interval 60 min post stimulation; exploratory approach). Paired-samples t-tests revealed no significant differences at any electrode sites and frequency band (all p-values > 0.372). In sum, externally imposed so-tDCS significantly increased power in slow as well as fast spindle frequency band in 1-min EEG intervals immediately following a stimulation period (corrected), but not the delayed 1-min EEG interval 60 min post stimulation (exploratory approach, uncorrected).

Figure 4. So-tDCS enhances EEG power within the slow oscillation band as well as within slow and fast spindle frequency band. EEG topographic plots of both prefrontal and frontal slow oscillatory (0.5–1 Hz) and slow spindle activity (8–12 Hz) as well as both central and centro-parietal fast spindle activity (12–15 Hz) averaged across the 1-min stimulation free intervals 1–5 for so-tDCS and sham condition. ROIs are indicated by the dashed squares (prefrontal: FP1, AFz, FP2; frontal: FC1, Fz, FC2; central: C3, Cz, C4; centro-parietal: CP1, Cz, CP2). N = 21.

Behavioral tasks Visuo-spatial memory task (primary outcome). The STIM × TIME rmANOVA on PC yielded a significant interaction effect (F(1,20) = 5.08, p = 0.036, pc = 0.039, η2 = 0.203) indicating worse visual memory overnight performance after a night with so-tDCS compared to sham. Main effects of TIME or STIM were not significant (all p-values > 0.152). An additional conducted analysis revealed no significant effect between both error types (false alarms and misses) between so-tDCS and sham condition (all p-values > 0.101). But in general more misses occurred after sleep compared to the evening before as indicated by the main effect TIME (F(1,20) = 5.85, p = 0.025, η2 = 0.226). Memory of picture location did not significantly differ between the conditions (F(1,20) = 1.93 p = 0.180, η2 = 0.088), but in general less correct location decisions were made the next morning (main effect TIME: F(1,20) = 7.96, p = 0.011, η2 = 0.285; see Fig. 5), independently of the stimulation.

736

S. Paßmann et al. / Brain Stimulation 9 (2016) 730–739

Figure 5. Memory performance in declarative and procedural memory tasks before (immediate) and after (delayed) sleep in the so-tDCS vs. sham condition. (a) Recognition performance (PC in %; proportion of hits and correctly rejected pictures) in the visuo-spatial task. A significant stimulation effect emerged with decreased picture recognition performance following so-tDCS compared to sham condition (p = 0.036; pc = 0.039). (b) Retention performance (in %; correct indicated locations in regard to correct indicated pictures) in the visuo-spatial memory task, showing no significant stimulation effect. (c) Recall performance (in %) in the verbal memory tasks. No significant stimulation effect was evident for verbal memory performance. (d) Correct tapped cycles in the procedural memory task for so-tDCS (black bars) and sham stimulation (white bars). No significant stimulation effect was evident for procedural performance. Data are expressed as means ± SE. *p < 0.05 (N = 21).

F(1,20) = 0.123, p = 0.730, η2 = 0.006). However, as expected an increment of overnight performance was observed independently of the stimulation (F(1,20) = 10.33, p = 0.004, η2 = 0.341). In sum, so-tDCS applied during early nocturnal sleep impaired recognition memory for neutral pictures (primary outcome; corrected), whereas memory performance for locations, for word-pairs and for finger-sequence tapping was not affected (exploratory approach; uncorrected).

Verbal memory task. Number of correctly remembered wordpairs as well as number of inferences did not significantly differ between so-tDCS and sham condition (STIM × TIME: all p-values> 0.791). However, on average verbal memory declined overnight as indicated by a decrement of correct remembered word-pairs (F(1,20) = 23.63, p < 0.001, η2 = 0.542), independently of the stimulation. Likewise, an increment of number of inferences was observed (F(1,20) = 10.54, p = 0.004, η2 = 0.345) on average independently of the stimulation.

Sleep stages Table 2 summarizes the time asleep (exploratory approach; uncorrected) in the particular sleep stages for so-tDCS and sham

Procedural memory task. No differences in number of correct tapped cycles between so-tDCS compared to sham were found (STIM × TIME:

Table 2 Sleep architecture dependent on stimulation condition. Top panel indicates mean (±standard deviation) percentages (%) of sleep period time (SPT). Note that 2 participants were excluded from sleep stage analysis for entire night due to incomplete EEG recordings (N = 19). Bottom panel indicates mean percentage of time in the different sleep stages averaged over the 1-min stimulation free intervals 1–5 as well as the stimulation free interval 60 min after the end of the last stimulation block. N = 21. so-tDCS Mean of SPT (SD) Entire night, N = 19 WASO (in %) 11.8 (8.8) NREM stage 1 (in %) 12.9 (5.7) NREM stage 2 (in %) 48.6 (6.7) NREM stage 3 (in %) 6.2 (3.7) NREM stage 4 (in %) 1.0 (1.7) SWS (in %) 7.2 (4.8) REM (in %) 12.2 (5.1) Stimulation-free intervals 1–5 (immediate effects), N = 21 WASO (in sec) 17.1 (33.6) NREM stage 1 (in seconds) 42.4 (38.3) NREM stage 2 (in seconds) 177.1 (53.5) NREM stage 3 (in seconds) 41.9 (31.6) NREM stage 4 (in seconds) 9.5 (11.6) SWS (in seconds) 51.5 (41.3) REM (in seconds) 0.5 (2.2) Stimulation-free interval 60 (delayed effect), N = 21 WASO (in seconds) 0 (0) NREM stage 1 (in seconds) 14.8 (18.9) NREM stage 2 (in seconds) 36.2 (2.2) NREM stage 3 (in seconds) 5.7 (10.8) NREM stage 4 (in seconds) 1.4 (3.6) SWS (in seconds) 7.14 (13.1) REM (in seconds) 1.9 (6.0)

SHAM MD 9.7 12.4 48.7 5.6 0.3 30.0 12.4

Mean of SPT (SD) 8.2 (6.2) 12.5 (5.1) 51.0 (7.8) 6.3 (3.5) 1.6 (2.3) 7.9 (4.9) 13.5 (4.7)

p MD 6.2 10.9 49.4 6.1 0.5 33.0 13.8

0.138 0.649 0.193 0.861 0.046# 0.345 0.342

0.0 30.0 170.0 50.0 0.0 60.0 0.0

2.8 (13.1) 49.5 (44.4) 179 (46.7) 48.1 (39.6) 14.8 (22.1) 62.9 (58.5) 0.5 (2.2)

0.0 30.0 180.0 40.0 0.0 40.0 0.0

0.096# 0.488# 0.895 0.418 0.234# 0.283 1.00#

0.0 1.0 5.0 0.0 0.0 0.0 0.0

2.9 (13.1) 28.1 (28.1) 22.9 (25.2) 3.3 (7.3) 1.9 (6.8) 5.2 (12.9) 0.9 (4.4)

0.0 1.0 1.0 0.0 0.0 0.0 0.0

0.317# 0.064# 0.053 0.258# 0.705# 0.550# 0.564#

WASO = wake after sleep onset; SWS = summarized NREM sleep stage 3 + 4; SPT = sleep period time; SD = standard deviation; MD = median. Comparisons are based on pairedsamples t-test otherwise mentioned. # Wilcoxon signed-rank test due to skewed distribution.

S. Paßmann et al. / Brain Stimulation 9 (2016) 730–739

condition for the entire night as well as in 1-min stimulus-free intervals immediately following the 5 stimulation periods and in poststimulation interval. Entire night. Pairwise comparisons of the sleep parameters yielded a significant stimulation effect in NREM sleep stage 4 (z = −1.99; p = 0.046) indicating a decrement after so-tDCS compared to sham. No significant correlation was found between the amount of decrement in SP4 and the difference in declarative memory performance (primary: visuo-spatial PR: r = 0.290, p = 0.228; secondary: word pair: r = 0.039, p = 0.874). No other significant differences in any sleep parameter were evident (all p-values > 0.13).

737

Discussion This interventional study demonstrated that in healthy older adults, application of so-tDCS during early nocturnal NREM sleep boosted EEG power in slow oscillatory as well as spindle frequency bands compared to sham. Behaviorally, these electrophysiological events were accompanied by decreased visuo-spatial memory performance. Exploratory analysis revealed no differences between stimulation conditions for overnight retention of verbal and procedural memories. However, NREM sleep stage 4 over the entire night was reduced after a night with so-tDCS (uncorrected).

Impact of stimulation on brain activity and memory consolidation Immediate stimulation-induced effects (averaged across interval 1–5 following stimulation). On average, time spent in particular sleep stages did not significantly differ between the conditions (all p-values > 0.23), but time awake increased slightly after so-tDCS compared to sham (z = −1.67; p = 0.096). Delayed stimulation-induced effects (60 min post stimulation). Pairwise comparisons of sleep parameters revealed no statistical difference between sleep stages (all p-values > 32), but a trend for NREM sleep stage 1 (z = −1.36; p = 0.064) and NREM sleep stage 2 (t(21) = 2.06; p = 0.053) suggesting that the subjects spent less time in NREM sleep stage 1 and more in NREM sleep stage 2 60 min after so-tDCS compared to sham. In sum, trends for increased arousal immediately after so-tDCS stimulation were observed, indicated by less NREM1 and more NREM 2 sleep 60 minutes after the end of the last stimulation period. Further, after so-tDCS the percentage of NREM sleep stage 4 across the entire night was reduced. Perception of stimulation The participants tolerated the stimulation well. Post experimental debriefing revealed that four participants were able to guess correctly the so-tDCS night. The remaining subjects stated not to know in which experimental night the stimulation had been applied. No subject felt pain due to stimulation, but two participants reported to have felt a tingling sensation, which in fact occurred in the sham night. Note that more subjects showed delayed beginnings (n = 9) of the following stimulation block under so-tDCS condition compared to sham (n = 2) due to lighter sleep after a stimulation block: A Pearson’s chi-squared test revealed significantly more delayed stimulation blocks under so-tDCS than under sham condition (χ2 (1) = 6.035, p = 0.014). The exploratory analysis revealed for self-reported affective state (PANAS) and sleepiness pre and post sleep no significant difference between the conditions (STIM × TIME: all p-values > 0.086). Overall, subjects reported a less positive affective state the next morning compared to the evening before, independently of the stimulation condition as indicated by a main effect of TIME (PANAS positive subscale: F(1,20) = 5.17, p = 0.034, η2 = 0.205). However, on average, the subjects felt more relaxed after a night under sham compared to so-tDCS as revealed by a significant STIM × TIME interaction for the VAS-tension scale (F(1,20) = 5.07, p = 0.036, η2 = 0.202), but without affecting retention performance as revealed by the corresponding ANCOVA (STIM × TIME) taking into account the difference (so-tDCS minus sham) of VAS tension scale as covariate (F(1,20) = 0.621, p = 0.440, η2 = 0.032). With regard to reported sleep quality of the night prior to experimental nights we found no significant difference between both so-tDCS and sham condition (STIM mean ± SD: 4.2 ± 1.0; SHAM mean ± SD: 3.9 ± 1.1; t(20) = 1.227, p = 0.234).

Electrophysiological modifications induced by so-tDCS such as increased frontal SO power (0.5–1 Hz) and enhanced activity in frontal slow spindle bands (8–12 Hz) are in line with previous studies on night-time stimulation conducted with young subjects [10,14] and on daytime-nap stimulation in older subjects [16,17], but contrary to Sahlem et al. [11] in young and Eggert et al. [15] in old healthy adults. This difference could be due to morphology (square wave form [11]) or mode (ramping [15]) of the so-tDCS pulse. For instance, the ramping modes at the beginning and the end of each stimulation interval might have precluded short-lasting stimulationdependent entrainments of the slow oscillatory activity. Similar to Marshall et al. [10], Eggert et al. did not control each stimulus-free interval for ongoing sleep stage [15], an issue of particular relevance in older adults with higher sleep fragmentation [12]. Given that oscillatory stimulation effects are strongly dependent on ongoing network activity and brain state [41–43], the stimulation might thus have been applied during periods of waking or NREM sleep stage 1, accounting for the lack of intended effect on sleep physiology. However, although improved neurophysiological sleep parameters are assumed to correlate with better memory consolidation, declarative memory performance was not improved by so-tDCS, in contrast to previous studies [10,16,43]. In fact, impaired visuospatial memory consolidation after a night of so-tDCS compared to sham was found. Given that neither Eggert and colleagues [15] nor Marshall et al. [10] assessed visuo-spatial memory, this result cannot be directly compared to previous studies. Consolidation of verbal memories was not altered by so-tDCS in our older subjects, as had been shown in young adults [10,14] though. Overall, our behavioral findings are thus in line with a previous report of Eggert et al. in older adults [15]. Note that we used exclusively non-emotional German nouns for the verbal memory task. Thus, our results might reflect the different mechanisms which are involved during encoding and during sleep-dependent consolidation of emotional and non-emotional stimuli [44–46]. Also, word-pairs in the present study were more strongly semantically related. Since the role of the hippocampus formation is to index and bind together sparse cortical representations [47], weaker associations may carry higher sleep-dependent benefits as indicated in previous studies [48]. Thus, pre-semantic knowledge of word-pairs might have masked stimulation-induced effects on consolidation [49,50], as seen previously in young adults [10,14]. Since Eggert et al., who used a mixture of neutral and emotional words, were not able to demonstrate improved memory consolidation after so-tDCS in older adults, their and our results argue against nocturnal stimulation-induced improvements of declarative memories in older adults, at least with currently employed stimulation protocols. Results from the exploratory analysis indicate that several changes in sleep parameters observed after so-tDCS vs. sham in the present study may serve to explain non-significant or even detrimental effects on behavior.

738

S. Paßmann et al. / Brain Stimulation 9 (2016) 730–739

Firstly, albeit not statistically significant, more time spent awake after sleep onset was noted during stimulation-free intervals immediately following so-tDCS and contributes to the delayed onset of the next stimulation block in some subjects. These findings are in line with Eggert et al. [15] and may indicate a stimulationinduced increase in sleep fragmentation for early nocturnal NREM sleep in older adults, possibly accounting for the impairment of consolidation. In addition, some authors have reported delayed SWS onset in the elderly [51]. Thus, given that so-tDCS effects seem to depend on the ongoing brain state [43], external stimulation starting four minutes after NREM sleep stage 2-onset might be too early in this age group, leading to increased arousal and awakenings rather than increased SWS. Secondly, we found a reduction of NREM sleep stage 4 for the entire night which might also be a consequence of higher sleep fragmentation induced by so-tDCS and might have contributed to the absent or even negative impact on physiological processes that support memory consolidation. No correlation between primary outcome and SP4 were found suggesting that memory deficits were not associated in a simple linear fashion with reduction in SP4 of entire night. Note that due to technical limitations, i.e., strong stimulation artifacts, sleep stages could not be assessed during stimulation intervals. However, characterizing sleep stages during stimulation would help to provide a more comprehensive view of so-tDCS effects on sleep physiology, an issue to be addressed in future studies. Thirdly, we observed slightly increased frontal SO-power 60 min after the end of the last stimulation as well as less NREM sleep stage 1 and increased NREM sleep stage 2 after so-tDCS compared to sham. Note that in a previous study demonstrating beneficial memory effects after boosting SO, no such delayed effects were reported [43].

differential findings need to be further explored, one intriguing possibility is that so-tDCS leads to disturbances in other sleep parameters, most importantly increments in time spent awake after sleep onset, and decrements in NREM stage 4 sleep during the entire night, accounting for the decline in function. Thus, for a nocturnal stimulation in older individuals, ageadjusted stimulation protocols may need to be developed, i.e., delivering stimulation only during SWS stable for longer periods of time (e.g., 10 min instead of 4 min), a hypothesis to be tested in future studies. Also, so-tDCS delivered during napping may circumvent detrimental effects of stimulation on later stages of night-time sleep, as indicated in previous studies [16,17]. Future studies should directly compare night-time with nap stimulation in older adults to address this issue. Acknowledgements This work was supported by grants from the Deutsche Forschungsgemeinschaft (Fl 379-8/1, Fl 379-10/1; Fl 379-11/1, and DFG-Exc 257) and the Bundesministerium für Bildung und Forschung (FKZ 01EO0801, 01GQ1424A, 01GQ1420B). We thank Lena Reich and Aileen Hakus for their help with data acquisition as well as Tom Lübstorf for sleep scoring. Furthermore, we are grateful that Maren Cordie and Sandra Ackermann supported us in questions concerning sleep scoring and that Steffen Gais provided us the program SchlafAus. Furthermore, we would like to thank all subjects for their participation. Appendix. Supplementary material Supplementary data to this article can be found online at doi:10.1016/j.brs.2016.04.016.

Limitations Several study limitations should be noted. Firstly, the sample size was small, thus results should be replicated in larger cohorts. However, subjects were well-characterized at baseline, and were tested in a counterbalanced within-subject design, rendering confounding factors between subjects (i.e., inter-individual variability in sleep architecture [52]) unlikely to account for the current findings. The number of subjects was comparable to previous reports though [10,15]. Secondly, even though our stimulation protocol was almost identical to Marshall and colleagues, slight differences like control for ongoing sleep stage were implemented, and may possibly account for differential behavioral outcomes. However, we believe that these changes should, if anything, have biased the results toward improved memory consolidation, because they resulted in stimulation parameters that were more closely adapted to older adults’ sleep physiology. Thirdly, while recognition performance in the visuo-spatial memory task was significantly impaired by so-tDCS, retention of locations was unaffected. This might be due to the poor performance level in retrieving locations (on average only 35% of locations were correctly assigned after sleep) in our sample of older adults rendering it difficult to detect so-tDCS induced differences between conditions. Conclusion We conclude that our study with older adults could not confirm enhanced memory consolidation by nocturnal so-tDCS. This is in line with previous studies [11,15], but contrary to the findings of Marshall et al. [10] in young subjects as well as during a nap in older adults [16,17]. While the exact mechanisms underlying such

References [1] Rasch B, Born J. About sleep’s role in memory. Physiol Rev 2013;93:681–766. doi:10.1152/physrev.00032.2012. [2] Schabus M, Gruber G, Parapatics S, Sauter C, Klösch G, Anderer P, et al. Sleep spindles and their significance for declarative memory consolidation. Sleep 2004;27:1479–85. [3] Diekelmann S, Born J. The memory function of sleep. Nat Rev Neurosci 2010;11:114–26. doi:10.1038/nrn2762. [4] Rattenborg NC, Martinez-Gonzalez D, Roth TC, Pravosudov VV. Hippocampal memory consolidation during sleep: a comparison of mammals and birds. Biol Rev Camb Philos Soc 2011;86(3):658–91. doi:10.1111/j.1469-185X.2010.00165.x. [5] Massimini M, Huber R, Ferrarelli F, Hill S, Tononi G. The sleep slow oscillation as a traveling wave. J Neurosci 2004;24:6862–70. doi:10.1523/JNEUROSCI.131804.2004. [6] Steriade M, McCormick DA, Sejnowski TJ. Thalamocortical oscillations in the sleeping and aroused brain. Science 1993;262:679–85. [7] Steriade M, Contreras D, Amzica F. Synchronized sleep oscillations and their paroxysmal developments. Trends Neurosci 1994;17:199–208. doi:10.1016/ 0166-2236(94)90105-8. [8] Kuo M-F, Nitsche MA. Effects of transcranial electrical stimulation on cognition. Clin EEG Neurosci 2012;43:192–9. doi:10.1177/1550059412444975. [9] Kuo M-F, Paulus W, Nitsche MA. Therapeutic effects of non-invasive brain stimulation with direct currents (tDCS) in neuropsychiatric diseases. Neuroimage 2014;85:948–60. doi:10.1016/j.neuroimage.2013.05.117. [10] Marshall L, Helgadóttir H, Mölle M, Born J. Boosting slow oscillations during sleep potentiates memory. Nature 2006;444:610–13. doi:10.1038/nature05278. [11] Sahlem GL, Badran BW, Halford JJ, Williams NR, Korte JE, Leslie K, et al. Oscillating square wave transcranial direct current stimulation (tDCS) delivered during slow wave sleep does not improve declarative memory more than sham: a randomized sham controlled crossover study. Brain Stimul 2015;06:doi:10.1016/j.brs.2015.01.414. [12] Hornung OP, Danker-Hopfe H, Heuser I. Age-related changes in sleep and memory: commonalities and interrelationships. Exp Gerontol 2005;40:279–85. doi:10.1016/j.exger.2005.02.001. [13] Fogel S, Martin N, Lafortune M, Barakat M, Debas K, Laventure S, et al. NREM sleep oscillations and brain plasticity in aging. Front Neurol 2012;3:1–7. doi:10.3389/fneur.2012.00176. [14] Marshall L, Mölle M, Hallschmid M, Born J. Transcranial direct current stimulation during sleep improves declarative memory. J Neurosci 2004;24:9985–92. doi:10.1523/JNEUROSCI.2725-04.2004.

S. Paßmann et al. / Brain Stimulation 9 (2016) 730–739

[15] Eggert T, Dorn H, Sauter C, Nitsche MA, Bajbouj M, Danker-Hopfe H. No effects of slow oscillatory transcranial direct current stimulation (tDCS) on sleepdependent memory consolidation in healthy elderly subjects. Brain Stimul 2013;6:938–45. doi:10.1016/j.brs.2013.05.006. [16] Westerberg CE, Florczak SM, Weintraub S, Mesulam M-M, Marshall L, Zee PC, et al. Memory improvement via slow-oscillatory stimulation during sleep in older adults. Neurobiol Aging 2015;36:2577–86. doi:10.1016/j.neurobiolaging .2015.05.014. [17] Ladenbauer J, Külzow N, Passmann S, Antonenko D, Grittner U, Tamm S, et al. Brain stimulation during an afternoon nap boosts slow oscillatory activity and memory consolidation in older adults. Washington, DC: Alzheimer’s Association International Conference; 2015. [18] Morris JC, Heyman A, Mohs RC, Hughes J, van Belle G, Fillenbaum G, et al. The consortium to establish a registry for Alzheimer’s disease neuropathology confirmation of the. Neurology 1989;39:1159–65. [19] Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189–98. 0022–3956(75)90026-6 [pii]. [20] Hautzinger M, Bailer M, Worall H, Keller F. Beck depression inventory (BDI). Huber Bern, Switz: 1994. [21] Spielberger CD, Gonzales HP, Taylor CJ, Anton WD, Agaze B, Test GRR. Anxiety inventory: preliminary professional manual. Palo Alto, CA Consult Psychol Press 1980. [22] Helmstaedter C, Lendt M, Lux S. VLMT: verbal learning and memory test. Göttingen, Ger: Beltz Test; 1980. [23] Wechsler D. Wechsler adult intelligence scale – fourth edition (WAIS-IV). San Antonio, TX: NCS Pearson; 2008. [24] Van der Elst W, Van Boxtel MPJ, Van Breukelen GJP, Jolles J. The Stroop color-word test: influence of age, sex, and education; and normative data for a large sample across the adult age range. Assessment 2006;13:62–79. doi:10.1177/1073191105283427. [25] Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol 1988;54:1063–70. doi:10.1037/0022-3514.54.6.1063. [26] Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res 1989;28:193–213. doi:10.1016/0165-1781(89)90047-4. [27] Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 1991;14:540–5. doi:10.1016/j.sleep.2007.08.004. [28] Griefahn B, Künemund C, Bröde P, Mehnert P. Zur validität der deutschen übersetzung des morningness-eveningness-questionnaires von horne und östberg. Somnologie 2001;5:71–80. doi:10.1046/j.1439-054X.2001.01149.x. [29] Frohnhofen H, Bibl M, Nickel B, Popp R. Der essener fragebogen alter und schläfrigkeit (EFAS) – ein neues assessmentinstrument zur messung von tagesschläfrigkeit bei alten menschen. Eur J Ger 2010;12:84–9. [30] Zimmermann P, Fimm B. Test for attentional performance (TAP). Herzogenrath: PsyTest; 1995. p. 76–7. [31] Luria RE. The validity and reliability of the Visual Analogue Mood scale. J Psychiatr Res 1975;12:51–7. doi:10.1016/0022-3956(75)90020-5. [32] Bes F, Wilde-Frenz J, Schulz H. Application of the Tiredness Symptoms Scale during sleep deprivation. Sleep-Wake Res Netherlands 1992;3:19–20. [33] Antonenko D, Diekelmann S, Olsen C, Born J, Mölle M. Napping to renew learning capacity: enhanced encoding after stimulation of sleep slow oscillations. Eur J Neurosci 2013;37:1142–51. doi:10.1111/ejn.12118. [34] Lang PJ, Bradley MM, Cuthbert BN. International affective picture system (IAPS): affective ratings of pictures and instruction manual. Tech Rep A-8, 2008.

739

[35] Schneider TR, Engel AK, Debener S. Multisensory identification of natural objects in a two-way crossmodal priming paradigm. Exp Psychol 2008;55:121–32. doi:10.1027/1618-3169.55.2.121. [36] Plihal W, Born J. Effects of early and late nocturnal sleep on declarative and procedural memory. J Cogn Neurosci 1997;doi:10.1162/jocn.1997.9.4.534. [37] Walker MP, Brakefield T, Morgan A, Hobson JA, Stickgold R. Practice with sleep makes perfect: sleep-dependent motor skill learning. Neuron 2002;35:205–11. doi:10.1016/S0896-6273(02)00746-8. [38] Rechtschaffen A, Kales A. A manual of standardised terminology, techniques, and scoring system for sleep stages of human subjects. Los Angeles: UCLA Brain Inf Serv; 1968. [39] Verbeke G, Molenberghs G. Linear mixed models for longitudinal data, 2000. [40] Holm S. A simple sequential rejective multiple test procedure. Scand J Stat 1979;6:65–70. [41] Kanai R, Chaieb L, Antal A, Walsh V, Paulus W. Frequency-dependent electrical stimulation of the visual cortex. Curr Biol 2008;18:1839–43. doi:10.1016/ j.cub.2008.10.027. [42] Mölle M, Bergmann TO, Marshall L, Born J. Fast and slow spindles during the sleep slow oscillation: disparate coalescence and engagement in memory processing. Sleep 2011;34:11. doi:10.5665/sleep.1290. [43] Kirov R, Weiss C, Siebner HR, Born J, Marshall L. Slow oscillation electrical brain stimulation during waking promotes EEG theta activity. Proc Natl Acad Sci USA 2009;106:15460–5. doi:10.1073/pnas.0904438106. [44] Kensinger EA, Corkin S. Memory enhancement for emotional words: are emotional words more vividly remembered than neutral words? Mem Cognit 2003;31:1169–80. doi:10.3758/BF03195800. [45] Wagner U, Degirmenci M, Drosopoulos S, Perras B, Born J. Effects of cortisol suppression on sleep-associated consolidation of neutral and emotional memory. Biol Psychiatry 2005;doi:10.1016/j.biopsych.2005.05.008. [46] Wagner U, Born J. Memory consolidation during sleep: interactive effects of sleep stages and HPA regulation. Stress 2008;11:28–41. doi:10.1080/ 10253890701408822. [47] Winocur G, Moscovitch M. Memory transformation and systems consolidation. J Int Neuropsychol Soc 2011;17:766–80. doi:10.1017/S1355617711000683. [48] Drosopoulos S, Schulze C, Fischer S, Born J. Sleep’s function in the spontaneous recovery and consolidation of memories. J Exp Psychol Gen 2007;136(2):169–83. doi:10.1037/0096-3445.136.2.169. [49] Arbuckle TY, Cooney R, Milne J, Melchior A. Memory for spatial layouts in relation to age and schema typicality. Psychol Aging 1994;9:467. doi:10.1037/ 0882-7974.9.3.467. [50] Castel AD. Memory for grocery prices in younger and older adults: the role of schematic support. Psychol Aging 2005;20:718–21. doi:10.1037/08827974.20.4.718. [51] Gigli GL, Placidi F, Diomedi M, Maschio M, Silvestri G, Scalise A, et al. Sleep in healthy elderly subjects: a 24-hour ambulatory polysomnographic study. Int J Neurosci 1996;85:263–71. doi:10.3109/00207459608986687. [52] Van Dongen HPA, Vitellaro KM, Dinges DF. Individual differences in adult human sleep and wakefulness: leitmotif for a research agenda. Sleep 2005;28:479– 96.

Web Reference Deutscher Wortschatz. Universität Leipzig. [accessed 13.10.15].