Effect of Voluntary Attention on Auditory Processing During REM Sleep Madoka Takahara, MSc1,2; Hiroshi Nittono, PhD1; Tadao Hori, PhD1 Department of Behavioral Sciences, Faculty of Integrated Arts and Sciences, Hiroshima University, of Higashi-Hiroshima, Japan; 2Research Fellow of the Japan Society for the Promotion of Science, Higashi-Hiroshima, Japan 1
Measurements and Results: In the tonic period of REM sleep (the period without REM), P200 and P400 were elicited by deviant stimuli, with scalp distributions maximal at central and occipital sites, respectively. The P400 in REM sleep showed larger amplitudes in the ATTENTIVE condition, whereas the P200 amplitude did not differ between the 2 conditions. No effects on ERPs due to attention were observed during stage 2 sleep. Conclusions: The instruction to pay attention to external stimuli during REM sleep influenced the late positive potentials. Thus electrophysiologic evidence of voluntary attention during REM sleep has been demonstrated. Keywords: REM sleep, event-related potentials, late positive potentials, behavioral response, voluntary control, attention Citation: Takahara M; Nittono H; Hori T. Effect of voluntary attention on auditory processing during REM sleep. SLEEP 2006;29(7):975-982.
Study Objectives: The study investigates whether there is an effect of voluntary attention to external auditory stimuli during rapid eye movement (REM) sleep in humans by measuring event-related potentials (ERPs). Design: Using a 2-tone auditory-discrimination task, a standard 1000-Hz tone and a deviant 2000-Hz tone were presented to participants when awake and during sleep. In the ATTENTIVE condition, participants were requested to detect the deviant stimuli during their sleep whenever possible. In the PASSIVE sleep condition, participants were only exposed to the tones. ERPs were measured during REM sleep and compared between the 2 conditions. Setting: All experiments were conducted at the sleep laboratory of Hiroshima University. Participants: Twenty healthy university student volunteers. Interventions: N/A.
event-related potentials (ERP) provide a useful tool for sleep research because they do not require overt behavioral responses. ERP is the averaged brain response temporally associated with a given stimulus. Numerous studies have investigated auditory processing in the awake and sleeping brain using ERP techniques. ERP studies have clearly established that there are systems for monitoring the environment that are active during sleep, and this monitoring fluctuates according to the sleep stages. This is particularly true for REM sleep, during which the level of information processing has been found to be close to that during wakefulness. A mismatch negativity (MMN)3,4 and a late positive potential resembling the P3005-9 of wakefulness have been reported during REM sleep. Cote et al recorded P300 during REM sleep using intensity-deviant stimuli (90 or 100 dB/SPL).7-9 They suggested that a cognitive processing system associated with the P300 during wakefulness is also active during REM sleep. REM sleep sometimes has 2 periods, tonic and phasic. The tonic period can be distinguished from the phasic one by the way transient events, such as REMs, tend to occur in clusters separated by episodes of relative quiescence.10 The most commonly used marker of phasic activity in human REM sleep is bursts of REM. Sallinen et al reported that sensitivity to changes in the external environment diminishes during the phasic period, as indicated by smaller ERPs during the phasic period than during the tonic period of REM sleep.11 In our previous study,12 we found that P200 and P400 ERPs, which were different from waking P300, could be elicited by presenting pitch-deviant 2-tone oddball stimuli (60 dB/SPL) during the tonic period of REM sleep that contains no REM. The P200 was dominant in the frontocentral area, whereas the P400 was found in the occipital area. During NREM sleep stage 2, P200 (P2) and the late positive potential (known as P420, P3, P430, or P450) were elicited by the rare deviant stimuli using the oddball paradigm.13-16 The N300 or N350, which were observed about 300 milliseconds after stimulus onset, were not related to the stimulus discrimination but to the presence of K-complexes following it.17 Thus, P200 is elicited throughout the sleep period, regardless of the stimulus modality,18 and reflects the beginning process of the subsequent stimulus processing.
INTRODUCTION PEOPLE’S AWARENESS OF THEIR SURROUNDINGS IS LOWER DURING SLEEP THAN WHEN AWAKE. HOWEVER, EVEN DURING SLEEP, THERE IS A CERTAIN level of perception occurring, as shown by the fact that light or sound may wake people up. In fact, perceptual thresholds change throughout the night, as has been shown in studies of the “depth of sleep.” For example, Okuma et al1 calculated the rate of recognition and behavioral response to flicker lights and click stimuli. The most notable results were found during rapid eye movement (REM) sleep, during which the rate of behavioral response was much lower than during stage 1 sleep, even though the electroencephalographic (EEG) patterns were similar. The accuracy of recognition did not seem to be consistent with the behavioral responses during REM sleep. Much of this inconsistency has to do with “the confounding effect” and “the incorporation process.” Williams2 found a remarkable improvement in accuracy of behavioral response during REM sleep when the participant’s motivation was manipulated by using the threat of electric shocks, whereas no changes were observed during non-rapid eye movement (NREM) sleep. The importance of the stimuli and the intensity of motivation probably influence the rate of behavioral response during REM sleep. A difficulty with studies of perception in REM sleep is that the muscles are in a state of atonia, which limits behavioral responses and subjective reports. However, techniques using evoked or Disclosure Statement This was not an industry supported study. Drs. Takahara, Nittono, and Hori have indicated no financial conflicts of interest. Submitted for publication August 31, 2005 Accepted for publication April 1, 2006 Address correspondence to: Tadao Hori, PhD, Department of Behavioral Sciences, Faculty of Integrated Arts and Sciences, Hiroshima University, 17-1 Kagamiyama, Higashi-Hiroshima, 739-8521, Japan; Tel: 81 82 424 6580; Fax: 81 424 0759; E-mail:
[email protected] SLEEP, Vol. 29, No. 7, 2006
975
Voluntary Attention During REM Sleep—Takahara et al
People sometimes want to go to sleep while still paying attention to what is happening around them, for example, to detect a night-time arrival or the distress of a child. Voluntarily paying attention to external stimuli during sleep could affect information processing and may be reflected by evoked potentials. During wakefulness, P300 is usually elicited by rare target stimuli in a sequence of standard stimuli (oddball paradigm) when participants are asked to discriminate a target from standard stimuli; P300 is not elicited when a participant fails to detect the target stimulus or ignores the stimuli.19 Thus, P300 reflects a detection process for a specific stimulus. The present research focused on the brain response during REM sleep to deviant tones significantly different from standard tones. We predict that, during REM sleep, the P400 amplitudes might increase in response to deviance in task-related tones, as has previously been found during wakefulness P300, when participants are instructed to detect the deviant stimuli.20 The purpose of the present study was to investigate the cognitive processing of sound during REM sleep. Specifically, we were interested in any effect that voluntary attention might have on the late positive potential during REM sleep. To that end, we investigated the late positive potentials during wakefulness, REM sleep, and NREM sleep stage 2, using an oddball paradigm for stimulus detection. We also investigated how voluntary attention to deviant stimuli would affect REM sleep, whether or not the effect of taskrelevance on ERPs could be observed in a manner similar to that observed with waking P300. Pitch-deviant oddball stimuli were presented with identical probability in 2 conditions on separate nights while participants slept: on 1 night, participants were given a stimulus discrimination task (ATTENTIVE), whereas, on the other night, the stimuli were presented as auditory stimuli only (PASSIVE). ERP responses were compared between the 2 conditions. We hypothesized that an effect of voluntary attention would be observed on the late positive potential during REM sleep and that the amplitude would increase according to attention to the stimuli.
air-conditioned, sound-proofed room. After 1 night of adaptation, each participant took part in the ATTENTIVE and PASSIVE conditions over 2 nonconsecutive nights to minimize an after effect of the former experimental night. The order of conditions was counterbalanced across participants. Two types of tones, 1000 Hz (p = .9, standard) and 2000 Hz (p = .1, deviant), generated by Cool Edit 96 (Syntrillium Software Corp., Scottsdale, AZ), with a duration of 50 milliseconds, including 10 milliseconds rise-fall time, 60 dB / SPL were presented binaurally through fixed inserted earphones (Sony, MDR-EX70SL, Shinagawa-ku, Japan) using a digital compact disc player (Sony, MXD-D1). An interstimulus interval of 1450 milliseconds was used because it was commonly employed in previous studies with the oddball paradigm and was chosen so as not to elicit K-complexes.21 The 2 types of tones (standard and deviant) were presented in a pseudorandom order. Each block contained 20 deviant and 180 standard tones (for a total of 5 minutes). Deviant tones were separated from each other by at least 5 seconds because only a behavioral response made within 5 seconds after a deviant-stimulus onset was judged as valid. In the ATTENTIVE condition, participants were instructed to respond to deviant tones by lifting their right or left index finger, whereas, in the PASSIVE condition, they were told to not do anything. All behavioral responses were recorded by a personal computer (Epson, PC-486HX, Suwa, Japan) with the elapsed times from the deviant tones. A simple switch on a board covered with cloth was attached to the participant’s right or left forearm during the sleep session, and signals were sent to the personal computer whenever the index finger was off the switch. On each experimental night, participants had a 15-minute session involving 3 blocks of stimuli while sitting awake in a reclining chair. Then they got into bed and fell asleep. In both the ATTENTIVE and PASSIVE conditions, tones were constantly presented throughout the night. In only the ATTENTIVE condition, were participants requested to constantly keep their attention on the tones and to try to respond whenever they heard the deviant tones. They were told they would win some prizes of their choice if they made the largest number of responses among the participants during sleep. Although behavioral responses were also observed during their wake times after sleep onset, all participants slept well and had no difficulties to go back to sleep.
METHOD Participants Twenty university students in a behavioral psychology class, 10 women and 10 men, volunteered to participate (20 to 24 years old, mean age = 21.5 ± 0.25 years). All participants reported that they were good sleepers, right-handed, nonsmokers, and free from any medication that may have an effect on arousal levels, including antidepressant or antipsychotic drugs. None had a reported history of psychiatric or neurologic disorders. Their auditory abilities were confirmed as normal using an audiometer (RION, AA-56, Kokubunji, Japan). On the experimental days, participants were instructed to wake up at their usual time and were told to abstain from excessive exercise, taking naps, and consuming alcohol, caffeine, or other arousal influencing substances. Each participant was informed about the purpose and procedure of the study and provided a signed consent form. They were aware that they were free to discontinue the study whenever they wished. The protocol was approved by the Human Ethics Committee of the Faculty of Integrated Arts and Sciences of Hiroshima University.
EEG Recordings EEG recordings were made during the awake and sleeping sessions. The sleeping session corresponded to each participant’s reported usual bedtime and awaking time (usually between the times of 11:00 PM and 1:00 AM and 7:00 and 8:00 AM, respectively). The beginning time of the waking session was then back calculated. Thus, the recorded lengths of the sleeping sessions were between 6.5 and 9 hours. EEGs were recorded from 21 electrode sites (Fp1, Fp2, F7, F8, Fz, F3, F4, T3, T4, Cz, C3, C4, T5, T6, Pz, P3, P4, O1, O2, M2, M2) as referenced from the tip of the nose (amplifier: NEC, BIOTOP6R12-4, Minato-ku, Japan; EEG recording software: KISSEI COMTEC Corp., Matsumoto, Japan, BIMUTAS II; personal computer: DELL, OptiPlex GX1, Kawasaki, Japan). Vertical and horizontal electrooculograms and submental electromyograms were also recorded. Time constants were 3.2 seconds for EEG and electrooculogram and 0.003 seconds for electromyogram. The data were digitized at 200 Hz. A 12-bit A/D converter
Procedure Participants came to the university sleep laboratory approximately 2 hours before their usual bedtime. All the experiment sessions were conducted in an electrically shielded, light-controlled, SLEEP, Vol. 29, No. 7, 2006
976
Voluntary Attention During REM Sleep—Takahara et al
was used to attain the required resolution. A high-pass filter was set at 60 Hz. Electrode impedance was below 5 kΩ.
Table 1—Sleep Parameters of Each Experiment Night
Analysis Polysomnography data were classified by 2 experienced scorers (interscorer agreement above 90%) using standard sleep-staging criteria based on 30-second epochs.22,23 As in our prior study,12 the REM stage was further classified into phasic period (with 1 or more REMs) and tonic period (without any REMs) every 3 seconds. Only the epochs confirmed by the agreement of the 2 scorers were analyzed. ERP waveforms were calculated by averaging the epoch between 100 milliseconds before and 900 milliseconds after the onset of the tone. The baseline was determined as the mean voltage between -100 to 0 milliseconds prior to a tone. For the sleep session, tones presented during REM sleep and sleep stage 2 were analyzed. Epochs with K-complexes were excluded from further analysis. Any epochs in the EEGs, electrooculograms, or electromyograms containing amplitudes in excess of ± 100 μV were also excluded from further analysis. At least 20 artifact-free EEGs were averaged for each participant. The number of behavioral responses made during the sleeping trials was very small; therefore, these trials were excluded from further analysis. The most positive peaks of the P200 during sleep were identified as occurring between 150 and 250 milliseconds after the onset of the tone. The late positive potentials during wakefulness and sleep had their most positive peaks occurring between 250 and 600 milliseconds after the tone onset. For each participant, sleep stage 2 data were divided into 2 halves, with each half corresponding to the first and second half of their total sleep time. Each sleep parameter was compared between the ATTENTIVEand PASSIVE-condition nights using Student paired t-tests. For each ERP potential, the mean was calculated using the amplitudes between each peak latency ± 20 milliseconds, and the data were analyzed using a repeated-measures analysis of variance (ANOVA). The P300 data during wakefulness were analyzed using a 3way ANOVA with the factors of Condition (ATTENTIVE, PASSIVE) × Stimulus (standard, deviant) × Electrode Site (Fz, Cz, Pz, O1, O2). Since no difference was observed between the first and second half of the night, the stage-2 sleep data were pooled, and a 3-way ANOVA with the factors of Condition × Stimulus × Electrode Site was carried out. The ERPs measured during REM sleep were tested with a 4-way ANOVA for Condition × Period (REM tonic, REM phasic) × Stimulus × Electrode Site. Any significant results were corrected with a Huynh-Feldt ε correction, and, as a posthoc test, comparisons were made using Tukey honest significant difference test. A .05 level of significance was used for all comparisons.
Time in bed, min Total sleep time, min Sleep efficiency, % Sleep stage, min 1 61.2 ± 4.76 2 192.1 ± 7.53 3 32.8 ± 2.40 4 42.8 ± 4.51 REM Wake time after sleep onset, min Arousals, no. Sleep latency, min REM latency, min Sleep cycle, no.
Condition
Passive 430.0 ± 8.83 403.6 ± 9.24 93.9 ± 1.08a
54.6 ± 5.06 193.8 ± 5.65 35.5 ± 3.18 49.4 ± 5.08 65.8 ± 6.09 27.6 ± 4.92
70.4 ± 3.95 21.6 ± 4.47b
21.3 ± 1.98 7.9 ± 2.69 114.2 ± 10.84 4.5 ± 0.17
15.8 ± 2.34a 4.0 ± 0.77b 120.8 ± 12.23 4.7 ± 0.15
Data are presented as mean ± SEM. REM refers to rapid eye movement sleep. a Statistically significant difference between the 2 conditions (ATTENTIVE / PASSIVE) at p < .05. b Statistically significant difference between the 2 conditions (ATTENTIVE / PASSIVE) at p < .1.
on the ATTENTIVE night were observed during the sleep-onset period and the wake time following sleep onset. None of the counted behavioral responses during sleep were associated with microarousals according to the American Sleep Disorders Association report.24 There were 1 to 20 (mean = 5.5 ± 1.13) responses to deviant tones during stage 2 sleep and 0 to 19 (mean = 1.8 ± 0.80) during REM sleep across the participants. Thirteen participants responded during the REM tonic period (n = 1-16), whereas 4 responded during the REM phasic period (n = 1-3). ERP data P300 in wakefulness The grand average waveforms recorded during wakefulness for each condition are shown in Figure 1. Parietal P300s were elicited by deviant tones. There was a Condition × Stimulus × Electrode Site interaction (F4,76 = 11.93, ε = .627, p < .001) for the P300 amplitudes. The P300 was larger in the ATTENTIVE than in the PASSIVE condition (F1,19 = 43.22, p < .001). There was an effect of Stimulus in the 2-way ANOVA for Stimulus × Electrode Site for each condition. The P300 amplitudes were larger in response to deviant tones than to standard tones in both conditions (ATTENTIVE: F1,19 = 73.29, p < .001; PASSIVE: F1,19 = 13.79, p < .01). There was also a significant effect of Electrode Site in the ATTENTIVE condition (F4,76 = 18.68, ε = .443, p < .001); the posthoc test revealed a larger amplitude for site Pz than for Fz, O1, and O2. A significant interaction between Stimulus × Electrode Site was observed for PASSIVE P300 (F4,76 = 3.00, ε = .508, p < .05); the posthoc test revealed a larger amplitude for Pz than for other electrode sites.
RESULTS Sleep Parameters and Behavioral Data Table 1 shows the sleep parameters for each condition. Sleep was less efficient (t19 = 2.14, p < .05) and there were more frequent arousals (t19 = 2.84, p < .05) on the ATTENTIVE-condition nights than on the PASSIVE-condition nights. There was a tendency for longer wake times after sleep onset (t19 = 1.81, p < .10) and sleep latencies (t19 = 1.42, p < .10) on the ATTENTIVE nights. All participants responded to the tones with almost perfect accuracy when tested while awake. Most behavioral responses SLEEP, Vol. 29, No. 7, 2006
Attentive 432.5 ± 8.52 396.9 ± 9.84 91.8 ± 1.40
ERPs During REM Sleep Figure 2 shows the grand ERP average waveforms recorded 977
Voluntary Attention During REM Sleep—Takahara et al
during the tonic and phasic periods of REM sleep for each condition. Fronto-central P200 and occipital late positive potentials were observed in the tonic period of REM sleep. In the phasic period of REM sleep, only fronto-central P200 was observed.
����������� ���������
�������
ANOVA of P200 Amplitudes A significant main effect of Period for P200 amplitudes recorded during REM sleep reveals that the P200 recorded during the tonic period was larger than during the phasic period of REM sleep (F1,19 = 46.49, p < .001). The P200 measured during REM sleep also showed a significant State × Stimulus × Electrode Site interaction (F4,76 = 4.15, ε = .319, p < .01). Because there was no effect of condition (ATTENTIVE vs PASSIVE), the data were pooled and a 2-way Stimulus × Electrode Site ANOVA was done for each period of REM sleep (tonic and phasic). A significant Stimulus × Electrode Site interaction was observed during the tonic period of REM sleep (F4,76 = 37.84, ε = .319, p < .001), with significantly larger amplitudes for deviant tones (thick lines) than for standard tones (dotted lines) at Fz, Cz, and Pz (Fz: F1,19 = 45.90, p < .001; Cz: F1,19 = 19.04, p < .001; Pz: F1,19 = 5.96, p < .05). The posthoc test revealed more positive amplitudes for the Fz, Cz, and Pz than for the O1 and O2 sites. A significant Stimulus × Electrode Site interaction was also observed during the phasic period of REM sleep (F4,76 = 37.84, ε = .319, p < .001), and the P200 amplitudes were significantly larger for deviant tones than for standard tones at Fz and Cz (Fz: F1,19 = 16.06, p < .001; Cz: F1,19 = 5.27, p < .05). Table 2 shows mean amplitudes of P200 in response to the deviant tones in each state for each condition. The P200 amplitudes did not differ between the 2 conditions. During REM sleep, the P200 amplitude was larger during the tonic period than during the phasic period.
� �� �
��
��
��
ANOVA of Late Positive Potential (P400) Amplitudes
� ���
�� +5 μV
�������
The P400 measured during the tonic periods of REM sleep showed a significant effect due to Condition (F1,19 = 4.41, p < .05). Comparison between both conditions of the P400 response to deviant tones at the O1 site is displayed in Figure 3. The P400 amplitude was larger in the ATTENTIVE than in the PASSIVE condition. Since there was also a significant interaction of Condition × Stimulus (F1,19 = 4.71, p < .05) and Stimulus × Electrode Site (F4,76 = 21.23, p < .001), a Stimulus × Electrode Site 2-way ANOVA was carried out for each condition. In the ATTENTIVE condition, a significant Stimulus × Electrode Site interaction was observed (F4,76 = 10.28, ε = .790, p < .001). The P400 amplitude was significantly more positive in response to deviant tones than
900 ms
��������
Figure 1—Event-related potential waveforms during wakefulness for each condition.
Table 2—Amplitudes of P200 to Deviant Tones During NREM Stage 2 Sleep and REM Sleep for Each Condition
Stage 2 REM tonic REM phasic
A P A P A P
Fz 8.31±0.760 8.10±0.869 7.46±0.865 6.23±0.825 1.13±0.670 1.67±0.666
Cz 10.48±0.954 10.33±1.035 7.75±1.029 7.04±1.016 1.52±0.739 1.41±0.989
Electrode Site Pz 7.30±0.730 7.10±0.786 4.80±0.832 4.92±0.805 0.71±0.774 1.24±0.926
O1 1.84±0.568 1.58±0.603 1.21±0.657 1.76±0.568 -0.36±0.816 0.79±0.891
O2 2.13±0.527 2.21±0.708 1.15±0.663 1.27±0.763 0.14±0.677 1.63±1.249
Data are presented as mean μV ± SEM. A: ATTENTIVE condition; P: PASSIVE condition. SLEEP, Vol. 29, No. 7, 2006
978
Voluntary Attention During REM Sleep—Takahara et al
���������
���������
����������
�������
���������
�������
� �� �
��
�� � ��� � ���
��
�� +5 μV
� ���
�������
900 ms
��������
Figure 2—Event-related potential waveforms of tonic (without rapid eye movement (REM)) and phasic (with REM) periods of REM sleep for each condition.
to standard tones at the Pz, O1, and O2 sites (Pz: F1,19 = 9.71, p < .01; O1: F1,19 = 17.49, p < .001; O2: F1,19 = 19.72, p < .001), and the P400 amplitudes were significantly larger at the Pz, O1, and O2 sites, compared with the Fz and Cz sites. In the PASSIVE condition, a significant Stimulus × Electrode interaction was observed (F4,76 = 16.24, ε = .508, p < .001). The P400 amplitudes were significantly more positive in response to deviant tones than to standard tones at the O1 site (F1,19 = 5.15, p < .05). The posthoc test showed that the P400 was larger at the Pz, O1, and O2 than at the Fz and Cz sites. No significant late potential was observed during the phasic period of REM sleep. Table 3 shows mean amplitudes of late positive potentials (P300, P400, and P450) in response to the deviant tones in each state for each condition. During sleep, an effect of voluntary attention was observed on the P400 amplitude during the tonic period of REM sleep.
sleep stage 2. The P200 and the late positive potential of P450 were elicited in response to deviant tones. The occipital late positive potential measured during stage 2 sleep resembled the late positive potentials in REM sleep (P400). Regarding the direct comparison between P400 and P450 amplitudes, a Component (P400, P450) × Condition × Electrode Site repeated-measures 3way ANOVA revealed a significant interaction of Component × Electrode Site (F4,76 = 5.00, ε = .301, p < .05). So the late positive potential measured in stage 2 sleep should be labeled as P450 to distinguish it from REM sleep P400. There was no effect of condition (ATTENTIVE or PASSIVE) on potentials measured during stage 2 sleep. ANOVA of P200 Amplitudes The P200 amplitude showed a significant effect of Stimulus (F1,19 = 68.36, p < .001) and was significantly larger in response to deviant tones than to standard tones. There was also a significant effect due to Electrode Site (F4,76 = 28.33 , ε = .362, p < .001), with significantly more positive P200 amplitudes at Fz, Cz, and Pz than at O1 and O2.
ERPs During Stage 2 Sleep Figure 4 shows the grand mean of waveforms measured during SLEEP, Vol. 29, No. 7, 2006
979
Voluntary Attention During REM Sleep—Takahara et al
Table 3—Amplitudes of Late Positive Potentials to Deviant Tones During Wakefulness, NREM Stage 2 Sleep, and the Tonic Period of REM Sleep for Each Condition Electrode site Fz Cz Pz O1 O2 Wakefulness (P300) Stage 2 (P450) REM tonic (P400)
A P A P A P
11.62 ± 1.345a 2.70 ± 0.935 -0.58 ± 0.997 -0.58 ± 0.924 -1.71 ± 0.939a -3.39 ± 0.668
14.43 ± 1.815a 2.91 ± 1.101 0.82 ± 1.193 0.81 ± 1.059 -0.91 ± 0.996a -3.33 ± 0.711
16.17 ± 1.663a 4.50 ± 0.997 -0.33 ± 0.883 -0.15 ± 0.724 2.60 ± 1.022a 0.36 ± 0.691
9.29 ± 1.211a 2.36 ± 0.813 1.45 ± 0.911 2.01 ± 0.648 4.13 ± 0.902a 2.47 ± 0.800
9.65 ± 1.356a 3.06 ± 0.834 2.00 ± 0.971 2.71 ± 0.690 4.47 ± 0.869a 2.29 ± 0.853
Data are presented as mean μV ± SEM. A: ATTENTIVE condition; P: PASSIVE condition.
�������
���������
�������
� �� �
��
N 300
��
� ��� Figure 3—The effect of attention on event-related potentials to deviant tones during rapid eye movement (REM) tonic period (O1 site).
��
ANOVA of Late Positive Potential (P450) Amplitudes The P450 amplitude showed a significant effect of Stimulus (F1,19 = 19.56, p < .001) and was larger in response to deviant, compared with standard, tones. There was also a significant effect of Electrode Site (F4,76 = 5.73, ε = .319, p < .001), and the P450 was significantly more positive at O1 and O2 than at the Fz site and at O2 as compared with the Pz site.
�� � ���
Topographies of Late Positive Potentials
�������
Amplitude distribution maps at each potential peak latency are shown in Figure 5. Waking P300 was dominant in the parietal area. The P400 in the tonic REM period and the P450 in stage 2 sleep dominated the temporal-occipital area, but the P450 also had a central distribution. So these late positive potentials for each sleep stage had different scalp distributions.
900 ms
��������
Figure 4—Event-related potential waveforms of non-rapid eye movement (NREM) stage 2 sleep for each condition.
were slightly impaired, compared with the PASSIVE condition, sleep efficacy was still above 90%. Thus, the ATTENTIVE night sleep was not seriously impaired, as compared with the PASSIVE night sleep, and, so, further analyses were carried out on the present study data. There was a smaller number of responses during the phasic period of REM sleep than during the tonic period, and this is consistent with previous studies.11,25 People often experience vivid and emotional dreams during bursts of REM (that is, during the
DISCUSSION Sleep Parameters and Behavioral Data Although the sleep parameters in the ATTENTIVE condition SLEEP, Vol. 29, No. 7, 2006
+5 μV
980
Voluntary Attention During REM Sleep—Takahara et al
evident during REM sleep, even though REM sleep is characterized by a deactivation in the frontoparietal association cortexes, as measured by regional cerebral blood flow in areas that normally participate in the higher-order analysis and integration of neural information.30 This is consistent with the findings of Williams2 using an avoidance paradigm that showed that only during REM sleep was there a notably higher rate of correct behavioral responses. Our study may provide the first electrophysiologic evidence of voluntary control measured during REM sleep during which behavioral responses are absent. The sleep parameters indicated that sleep on the ATTENTIVE nights was lighter than on the PASSIVE nights. This might raise the question of whether the present finding of an attentional effect on P400 during tonic REM sleep was caused by the lighter sleep on the ATTENTIVE night. However, the present study divided REM sleep by 3 seconds, and the comparison about voluntary attention was performed for the same REM sleep period, not in relationship to arousal indexes. Moreover, the amplitude difference between the ATTENTIVE and PASSIVE nights was observed only for the P400 component. If the lighter sleep on the ATTENTIVE nights affected the ERP amplitudes, the P200 amplitude would also have been influenced because P200 increases with deepening sleep.18 Also, participants’ own names or novel sounds can naturally raise their attention to stimuli.31,32 Using pure tones in our study was intended to allow us to investigate the fundamental effect of voluntary attention according to the waking P300 paradigm. During the phasic period of REM sleep, the mean amplitude of P400 was not different between deviant tones and standard tones. However, there appears to be a strong negative shift in the later latency, indicating that it might not be possible to have a sufficiently positive deflection. However, P400 during the phasic period of REM sleep can be observed, as we saw in our previous study using a single-stimulus paradigm to increase stimulus salience.33 In comparison, the results from this present study reflect a lower responsiveness during the phasic period of REM sleep. On the other hand, REMs and dreaming are often associated, as has been claimed by the scanning hypothesis,34 and the high probability of dreaming that has often been reported during bursts of REM (phasic period of REM sleep).35 Nevertheless, further investigation is necessary to determine whether the observed lower responsiveness is due to a relatively higher perceptual threshold specific to this period of REM sleep or to limited resources of attention. ERP waveforms during stage 2 sleep were almost the same as have been found in previous research.13-16 Late positive potential P450 was not a reflection of voluntary attention. However, this does not mean that stimuli were totally ignored during stage 2 sleep because the behavioral response level was close to or greater than that in REM sleep. An effect due to attention was reflected by the N550, which is 1 component of the K-complex; the amplitude was altered depending on the probability of presentation or task relevance during the detection of target stimuli. 36 Since the P450 did not include any effect due to attention, the potential must instead include an exogenous component. Further investigation is necessary to elucidate the functional differences between P450 during stage sleep 2 and P400 during tonic-REM sleep. P200 was observed both during stage 2 sleep and REM sleep. P200 shows a clear change relatively earlier than other components during the sleep-onset period and is continuously elicited
Wakefulness P300
(325 ms)
REM-tonic P400
(370 ms)
Stage 2 P450
(445 ms)
Figure 5—Peak voltage distribution maps of late positive potentials during each state (ATTENTIVE condition). The iso-contour lines are separated by 5 μV for wakefulness, 2 μV for rapid eye movement (REM) tonic period and stage 2 sleep, respectively. The blacker the area, the more positive voltages the potential shows. Peak latencies are in parentheses.
phasic period of REM sleep)26 and feel very sleepy when awakened from such states.27 Therefore, it would be much harder to physically respond during the phasic period of REM sleep. ERP Data The results from the ERP topographies showed that, during the tonic period of REM sleep, the P400 had an occipital distribution, as compared with the P300 during wakefulness. This distribution is similar to the P400 found during the sleep-onset period.28 Similar activation over the lateral-occipital area was evident in previous studies by Cote et al.7-9 This may reflect activation in the visual association area involved in the dreaming process during REM sleep. The P400 found during the tonic period of REM sleep reflects an effect of voluntary attention. The participants gave some attention to a specific stimulus while remaining asleep. As has been previously shown, during REM sleep, the activity occurring within the sleeping brain reflects a sensitivity to external stimuli, which suggests the involvement of higher cognitive processes, such as voluntary attention. In this study, attention did increase the P400 amplitude during the tonic period of REM sleep. This suggests that the instructions given to the participants to listen for deviant tones engaged their attention. This voluntary attention modulated the ERP amplitudes through a top-down approach, probably involving the attentional role of the thalamus in the process of extracting perceptual-specific information from a quantity of extraneous information.29 Note that mental control could be SLEEP, Vol. 29, No. 7, 2006
981
Voluntary Attention During REM Sleep—Takahara et al
throughout sleep.18 Consistent with this, the present study clearly observed P200 in response to both standard and deviant tones and in both the ATTENTIVE and PASSIVE conditions. P200 is probably some type of orienting response elicited by external stimuli. Moreover, the amplitudes were significantly different in the tonic period, compared with in the phasic period, of REM sleep. This also reflects the lower responsiveness during the phasic period of REM sleep. In summary, in the tonic period of REM sleep (the period without REM), deviant stimuli elicited P200 and P400 waves, with scalp distributions maximal at central and occipital sites, respectively. The P400 in REM sleep showed larger amplitudes in the ATTENTIVE condition, whereas the P200 amplitude did not differ between the 2 conditions. No effects on ERPs due to attention were observed for P200 and P450 during sleep stage 2. These data reveal that the P400 component during the tonic REM period reflects auditory attention processing. These findings are consistent with the fact that people are able to process auditory stimuli while remaining asleep, and voluntary attention may improve their information processing during REM sleep. It is possible that people use an environment-monitoring system specific to sleep, continuously paying attention to external stimuli so long as the stimuli do not impair people’s sleep quality too much.
sleep. Psychiatry Clin Neurosci 2002;56:257-8. 13. Nielsen-Bohlman L, Knight R, Woods D, Woodward K. Differential auditory processing continues during sleep. Electroencephalogr Clin Neurophysiol 1991;79:281-90. 14. Salisbury D, Squires NK, Ibel S, Maloney T. Auditory event-related potentials during stage 2 NREM sleep in humans. J Sleep Res 1992;1:251-7. 15. Winter O, Kok A, Kenemans L, Elton M. Auditory event-related potentials to deviant stimuli during drowsiness and stage 2 sleep. Electroencephalogr Clin Neurophysiol 1995;96:398-412. 16. Colrain IM, Parsia PD, Gora J. The impact of prestimulus EEG frequency on auditory evoked potentials during sleep onset. Can J Exper Psychol 2000;54:243-54. 17. Bastien C, Campbell KB. The evoked K-complex: All-or-none phenomenon? Sleep 1992;15:236-45. 18. Crowley KE, Colrain IM. A review of the evidence for P2 being an independent component process: Age, sleep and modality. Clin Neurophysiol 2004;115:732-44. 19. Hillyard SA, Squires KC, Bauer JW, Lindsay PH. Evoked potential correlates of auditory signal detection. Science 1971;172:1357-60. 20. Duncan-Johnson CC, Donchin E. On quantifying surprise: the variation of event-related potentials with subjective probability. Psychophysiology 1977;14:456-67. 21. Bastien C, Campbell K. Effects of rate of tone-pip stimulation on the evoked K-complex. J Sleep Res 1994;3:65-72. 22. Recheschaffen A, Kales A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Washington: Public Health Service, U.S. Government Printing Office; 1968. 23. Sleep Computing Committee of the Japanese Society of Sleep Research Society (JSSR): Hori T, Sugita Y, Koga E et al. Proposed supplements and amendments to ‘A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects’, the Rechtschaffen & Kales (1968) standard. Psychiatry Clin Neurosci 2001;55:305-10. 24. ASDA Report. EEG arousals: Scoring rules and examples. A preliminary report from the sleep disorders atlas task force of the American Sleep Disorders Association. Sleep 1992;15:173-84. 25. Price L, Kremen I. Variations in behavioral response threshold within the REM period of human sleep. Psychophysiology 1980;17:13340. 26. Berger RF, Oswald I. Eye movements during active and passive dreams. Science 1962;137:601. 27. Weinstein L, Schwartch D, Ellman SJ. The development of scales to measure the experience of self-participation in sleep. Sleep 1988;11:437-47. 28. Michida N, Ebata A, Tanaka H, Hayashi M, Hori T. Changes of amplitude and topographical characteristics of event-related potentials during the hypnagogic period. Psychiatry Clin Neurosci 1999;53:163-5. 29. Frith CD, Friston KJ. The role of the thalamus in “top down” modulation of attention to sound. Neuroimage 1996;4:210-15. 30. Braun AR, Balkin TJ, Wesensten NJ, et al. Regional cerebral blood flow throughout the sleep-wake cycle. An H215O PET study. Brain 1997;120:1173-97. 31. Pratt H, Berlad I, Lavie P. ‘Oddball’ event-related potentials and information processing during REM and non-REM sleep. Clin Neurophysiol 1999;110:53-61. 32. Perrin F, García-Larrea L, Mauguière F, Bastuji H. A differential brain response to the subject’s own name persists during sleep. Clin Neurophysiol 1999;110:2153-64. 33. Takahara M, Miyamoto O, Nittono H, Hori T. Auditory event-related potentials from a single-stimulus paradigm during rapid eye movement sleep. Sleep Biol Rhythms 2004;2:150-2. 34. Dement W, Kleitman N. Cyclic variations in EEG during sleep and their relation to eye movements, body motility and dreaming. Electroencephalogr Clin Neurophysiol 1957;9:673-90. 35. Antrobus J, Antorobus J, Fisher C. Discrimination of dreaming and nondreaming sleep. Arch Gen Psychiatry 1965;12:395-401. 36. Niiyama Y, Fushimi M, Sekine A, Hishikawa Y. K-complex evoked in NREM sleep is accompanied by a slow negative potential related to cognitive process. Electroencephalogr Clin Neurophysiol 1995;95:27-33.
ACKNOWLEDGMENTS This study was supported by a Grant-in-Aid of Japan Society for the Promotion of Science (JSPS) (No. 17005268), and Grantin-Aid for Scientific Research (C) (No.17605008). REFERENCES 1.
Okuma T, Nakamura K, Hayashi A, Fujimori M. Psychophysiological study on the depth of sleep in normal human subjects. Electroencephalogr Clin Neurophysiol 1966;21:140-7. 2. Williams HL. The problem of defining depth of sleep. In: Kety SS, Evarts EV, Williams HL, eds. Sleep and Altered States of Consciousness. Baltimore: Williams & Wilkins, 1967:277-87. 3. Loewy DH, Campbell KB, Bastien C. The mismatch negativity to frequency deviant stimuli during natural sleep. Electroencephalogr Clin Neurophysiol 1996;98:493-501. 4. Atienza M, Cantero JL. Complex sound processing during human REM sleep by recovering information from long-term memory as revealed by the mismatch negativity (MMN). Brain Res 2001;901:151-60. 5. Niiyama Y, Fujiwara R, Satoh N, Hishikawa Y. Endogenous components of event-related potential appearing during NREM stage 1 and REM sleep in man. Int J Psychophysiol 1994;17:165-74. 6. Bastuji H, Garcia-Larrea L, Franc C, Mauguiere, F. Brain processing of stimulus deviance during slow-wave and paradoxical sleep: A study of human auditory evoked responses using the oddball paradigm. J Clin Neurophysiol 1995;12:155-67. 7. Cote K, Campbell K. P300 to high intensity stimuli during REM sleep. Clin Neurophysiol 1999a;110:1345-50. 8. Cote K, Campbell K. The effects of varying stimulus intensity on P300 during REM sleep. NeuroReport 1999b;10:2313-8. 9. Cote K, Etienne L, Campbell K. Neurophysiological evidence for the detection of external stimuli during sleep. Sleep 2001;24:791803. 10. Carskadon MA, Dement WC. Normal human sleep: an overview. In: Kryger MH, Roth T, Dement WC, eds. Principles and Practice of Sleep Medicine, 4th ed. Philadelphia: Elsevier Saunders; 2005:1323. 11. Sallinen M, Kaartinen J, Lyytinen H. Processing of auditory stimuli during tonic and phasic periods of REM sleep as revealed by eventrelated brain potentials. J Sleep Res 1996;5:220-8. 12. Takahara M, Nittono H, Hori T. Comparison of the event-related potentials between tonic and phasic periods of rapid eye movement SLEEP, Vol. 29, No. 7, 2006
982
Voluntary Attention During REM Sleep—Takahara et al
