Changes in the Waking EEG as a Consequence of Sleep and Sleep Deprivation

Transcription

1 Sleep. 15(6): American Sleep Disorders Association and Sleep Research Society Changes in the Waking EEG as a Consequence of Sleep and Sleep Deprivation M. Corsi-Cabrera, J. Ramos, C. Arce, M. A. Guevara, M. Ponce-de Leon and I. Lorenzo Facultad de Psicologfa, Escuela Nacional Preparatoria, Universidad Nacional Aut6noma de Mexico, Mexico, D.F., Mexico Summary: Electroencephalographic (EEG) activity was monopolariy recorded during resting wakefulness in 10 volunteers under the following conditions: at night before going to sleep, at night before total sleep deprivation, in the morning after waking, in the morning after sleep deprivation and at night after having slept during the day. Absolute and relative power and inter- and intrahemispheric correlation were established. After diurnal and nocturnal sleep as compared to sleep deprivation, we obtained the following significant results: interhemispheric correlations were higher; intrahemispheric correlations were lower; absolute power of alpha2, beta I and beta2 was lower; and relative power of alpha2 and beta2 was lower. EEG changes as a consequence of sleep or lack of sleep are dependent on prior sleep and/or wakefulness and not on circadian phase. EEG activity during wakefulness is a sensitive parameter and a useful tool to assess the consequences of sleep on brain functional organization. Key words: EEG-Interhemispheric correlation-sleep deprivation-circadian effects-sleep. Studies on sleep deprivation have shown a progressive deterioration of physiological (1) and psychological functions over accumulating hours of sleep loss, as well as over periods of fragmented sleep. Current data report significant loss of daytime functions, such as the ability to sustain performance and discriminate signals, memory, mood, motivation, cognitive functions and electroencephalographic (EEG) activity. Sleep, on the contrary, has a general restitutive and reorganizing effect (2-11). However, studies on EEG activity during wakefulness as a consequence of previous sleep are very few. Spectral analysis of the EEG gives information on frequency distribution. Coherence or correlation analysis between two EEG signals gives information about functional relationships between two cortical regions (12). Changes have been noted in coherence and/or correlation values during coma (13) and sleep (14-16), during cognitive functions in normal subjects and in other conditions such as schizophrenia, arteriosclerosis and senile dementia (17-19). Therefore, the EEG (and particularly coherence and correlation analysis) can be considered a sensitive parameter to assess the consequences of sleep and lack of sleep on the waking EEG, and it can be a useful tool to study the effects of sleep on the functional organization of the brain. Accepted for publication July Address correspondence and reprint requests to Dr. Maria Corsi Cabrera, Margaritas 64, San Angel Inn, Mexico D.F. 0 I 060, Mexico. 550 In a previous work, it was found that normal sleep has an active effect on the relative power and the interhemispheric correlation of subsequent wakefulness. After normal sleep, interhemispheric correlation between homologous derivations was significantly higher than pre sleep values, whereas after sleep deprivation it was lower (20). The importance of counting on an objective and easily obtainable parameter of sleep action on brain functional organization, such as the EEG, led us to try to replicate the previous findings and to explore whether there could be a circadian effect superposed to these results. In this experiment, the EEG was recorded with eyes open and closed during resting wakefulness before and after nocturnal and diurnal sleep. In addition, intrahemispheric correlations and absolute power were assessed. METHODS Ten adult volunteers, all graduates or professionals between 23 and 27 years old, participated in the experiment. Subjects were right-handed with no central nervous system disorder or use of medications known to affect sleep or EEG. They were free of sleep complaints and had normal sleep habits, as assessed by a questionnaire. Experimental sessions were recorded only after habitual sleep schedules.

2 WAKING EEG AFTER SLEEP AND SLEEP LOSS 551 ew 'ig. [ N Z 52 f- OJ ::! OJ :I: Co </'> ai :I: OJ f ::; 110 ;;, _4. A.. o...,.. O'. <l"\.. ' " de Ita. the ta A \ ' ", \,\ '4. \,\ 6 \ alphal alpha2 betal be ta2 0. o.. o, A.. -o.. J:) B a) absolute power (AP), by means of a Fast Fourier Transform for the following bands: delta , theta , alpha , alpha , beta and beta ; b) relative power (RP) was calculated for the same bands considering as 100% the total power between 1.6 and Hz; c) interhemispheric correlations (INTERr) (Pearson product-moment coefficients between successive amplitude values) between homologous derivations; and d) intrahemispheric correlations (INTRAr) within left and right derivations were obtained for the same frequency bands. Samples were digitally filtered before processing. For statistical purposes, AP and RP were log transformed (21,22) and correlation values were transformed to Fisher's Z scores. Differences among conditions were tested for each band separately by twoway analyses of variance for repeated measurements (conditions x derivations). The level of significance was set at p < Pairwise comparisons were performed with Tukey's Student's t test. de Ita theta alpha! ajpha2 betal beta2 FIG. 1. Main effect of interhemispheric correlation ofeeg activity with eyes closed transformed to Fisher's Z scores. The lines connect mean values for each frequency band. A: In the morning after normal sleep (e); in the morning after sleep deprivation (0); in the evening after having slept during the day (l'.). B: Values for the morning after sleep (e) and the morning after sleep deprivation (0) expressed relative to the averaged value of baseline nights (=100% horizontal line). * below the abscissae indicates frequency bands showing significant changes among conditions. EEG was recorded during resting wakefulness with eyes open and closed under five conditions: 1) at night before going to sleep, 2) in the morning after sleep, 3) at night before sleep deprivation, 4) in the morning after one night of sleep deprivation and 5) at night after having slept during the day. Morning samples were recorded between 7:00 and 9:00 a.m. and night samples between 9:00 and i 1 :00 p.m. Sleep and sleep deprivation were separated by several weeks, and the order was counterbalanced among subjects. EEG was monopolarly recorded according to the International System at C3, C4, T3, T4, P3, P4, 01 and 02 referred to ipsilateral earlobes on a Grass 8-16E polygraph, set to pass frequencies of 1-35 Hz. Ten artifact-free samples of seconds from each condition and derivation were fed to an analog/digital converter (12 bits precision) of a PC computer at a sampling rate of 125 Hz. Samples were visually inspected, excluding those sections showing artifacts and drowsiness or stage 1 sleep. The following parameters were obtained and averaged across samples of the same subject and condition: RESULTS There were no significant differences in INTERr, AP and RP between the two baseline nights (sleep and sleep deprivation nights). Interhemispheric correlation Analyses of variance of the waking EEG with eyes closed (conditions x derivations) showed significant main effects for both factors for all frequency bands. The main effects for condition were significant at p < for theta, alpha2, beta 1, beta2 and for the full band; at p < 0.0 I for alpha I, and at p < 0.02 for delta. The main effects for derivation were significant at p < for all frequency bands. Interaction was significant only for beta2. Sleep vs. deprivation As can be observed in Fig. la, interhemispheric correlations were significantly higher in the morning after nocturnal sleep and at night after diurnal sleep than in the morning after sleep deprivation. Lower correlations were observed for the betal and beta2 bands than for slower frequencies (no statistical comparison). Recovery There were no significant differences after having slept during the night or during the day. A few hours Sleep, Vol. 15, No.6, 1992

3 552 M. CORSI-CABRERA ET AL. of sleep were enough to increase r values to the same level as after a night of normal sleep. Sleep vs. presleep values As shown in Fig. 1 B, INTERr values were higher after sleep and lower after sleep loss than presleep values. Derivations INTERr values were significantly lower at temporal derivations for all frequency bands and conditions. Analyses of variance with eyes open showed signif icant main effects for both factors at all frequency bands except for the condition effect for the delta band (p < 0.01 for condition effect for alpha2 and p < for all the rest). INTERr values with eyes open were also higher after sleep than after sleep deprivation. Intrahemispheric correlation Analyses of variance with eyes closed showed sig nificant main effects for condition and derivation for all frequency bands (p < 0.02 for the condition effect for alpha2 and p < 0.00 I for all the rest). INTRAr values were significantly higher after sleep deprivation than after nocturnal and diurnal sleep for all frequency bands. INTRAr values of all frequency bands were similar after diurnal and after nocturnal sleep, except for the delta band (Fig. 2). Derivations The highest intrahemispheric correlations, for both hemispheres, were observed between central-parietal derivations for delta, theta, beta1, beta2 and the full band, and between central-temporal derivations for alfa1 and alfa2. Temporal-occipital and central-occipital derivations showed the lowest correlation values at both hemispheres. In most cases, pairs involving the parietal derivation showed higher correlations, whereas pairs involving the occipital derivation showed lower values. The same results were obtained with eyes open. Analyses of variance with eyes open showed significant main effects for condition and derivation for all frequency bands at p < INTRAr values were significantly higher after sleep deprivation than afte:r sleep. Absolute power Absolute power with eyes closed showed significant main effects for both factors for all frequency bands N 0( -' "" Cl. " 0.4 g; z de Ita. theta alphal alpha2 betal beta2 FIG. 2 A. Main effect of intrahemispheric correlation of EEG activity with eyes closed transformed to Fisher's Z scores. The lines connect mean values for each frequency band: in the morning after normal sleep (e); in the morning after sleep deprivation (0); in the evening after having slept during the day (6). except for the condition effect for alpha1 (p < for all the rest). Interactions were not significant. Sleep vs. deprivation Alpha2, beta1 and beta2 AP was significantly higher after sleep deprivation than after nocturnal and diurnal sleep. The theta band showed significantly higher AP after sleep deprivation in comparison to diurnal sleep only (Fig. 3A). Recovery There were no significant differences in AP between nocturnal and diurnal sleep. As in the case oflnterr, a few hours of sleep were enough for AP to return to the same levels observed after nocturnal sleep. Sleep vs. presleep The AP of all frequency bands was lower after sleep than pre sleep values. After sleep deprivation, alpha2, beta1 and beta2 AP was higher than before sleep (see Fig.3B). Derivations AP was significantly lower at the temporal cortex for all frequency bands. AP with eyes open showed significant condition and derivation effects for all frequency bands except for the derivation effect for beta2 (p < 0.03 for condition effect for alpha 1 and p < for all the rest). AP was significantly higher after sleep deprivation than after nocturnal and diurnal sleep for all frequency bands. Sleep, Vol. 15, No.6, 1992

4 WAKING EEG AFTER SLEEP AND SLEEP LOSS A 40 A "" o 13.0 c< _ 12.5 s; ::> 12.0 D ;; "-.. "-..:.- "\,!, > ;: 0 «10 9. i/"\ '\ :::; ::> de I t a de I t a theta./' " " " theta " alphal ".. / 'd alphal,1/ alpha2,/, - betal beta2 B //P.,.., o alpha2 beta 1 beta2 FIG. 3. Main effect of absolute power with eyes closed log transformed. The lines connect mean values for each frequency band. A: In the morning after normal sleep (e); in the morning after sleep deprivation (0); in the evening after having slept during the day (b.). B: Values for the morning after sleep (e) and the morning after sleep deprivation (0) expressed relative to the average value of baseline nights (= 100% horizontal line). * below the abscissae indicates frequency bands showing significant changes among conditions. Relative power Analyses of variance with eyes closed showed significant main effects for both factors at p < 0.00 I for all frequency bands except for the condition effect for theta and betal and the derivation effect for alphal. Sleep vs. deprivation Alpha2 and beta2 RP was significantly higher while delta and alpha! RP was significantly lower after sleep deprivation than after nocturnal and diurnal sleep (Fig. 4A). Recovery There were no significant differences between nocturnal and diurnal sleep. Sleep vs. presleep Sleep values in comparison to pre sleep values did not induce significant changes in the proportional con- :::; > de Ita ;:..-" 80 «.' o 70 delta " the ta,0. theta a J phaj ', --.0 alphal alpha2 alpha2 beta 1 be ta I beta2 B beta2 FIG. 4. Main effect of relative power with eyes closed. The lines connect mean values for each frequency band. A: In the morning after normal sleep (e); in the morning after sleep deprivation (0); in the evening after having slept during the day (.6.). B: Values for the morning after sleep (e) and the morning after sleep deprivation (0) expressed relative to the averaged value of baseline nights (= 100% horizontal line). * below the abscissae indicates frequency bands showing significant changes among conditions. tribution of each band to total power. However, sleep deprivation produced significantly higher proportion of alpha2 and beta2 and lower proportion of delta (see Fig. 4B). The differential contribution of each band becomes more apparent after sleep deprivation. Relative power with eyes open showed significant condition effects for alphal and alpha2 only. Alpha RP was lower after deprivation than after nocturnal or diurnal sleep. DISCUSSION The present results confirm previous findings of an active effect of sleep on the EEG activity of subsequent wakefulness (20). Sleep resulted in higher interhemispheric correlation, lower intrahemispheric correlation and lower absolute power for all frequency bands as compared to pre sleep values. With sleep deprivation, opposite results were obtained: lower interhemispheric correlation and higher intrahemispheric correlation for all frequency bands, and higher absolute power on the faster bands of the spectrum (alpha2, beta I and beta2). The changes induced by sleep on the waking EEG seem Sleep. Vol. 15. No

5 554 M. CORSI-CABRERA ET AL. to be influenced by prior sleep and/or wakefulness and not by circadian phase, much like sleep delta power is influenced by previous wakefulness (23). Although the increase in interhemispheric correlation after sleep was observed in all frequency bands and derivations, suggesting a mechanism affecting EEG activity as a whole, sleep showed a stronger relative effect on betal and beta2, whereas sleep deprivation affected all frequency bands in a similar proportion. The results obtained are reinforced by the facts that interhemispheric correlation showed good stability, there were no significant differences between both bas: line nights and results had the same trend with open and closed eyes. These are in agreement with the stability previously reported both for coherence in rats and rabbits (24) and correlation in women for 12 sessions during a month (Corsi-Cabrera et ai., in preparation). Interhemispheric correlation was significantly higher and intrahemispheric correlation was significantly lower after sleep as compared to presleep values. These results suggest that sleep increases the temporal coupling between both hemispheres and enhances local functional differentiation within each hemisphere, whereas sleep deprivation tends to produce a loss of interhemispheric coupling and a more homogeneous organization within each hemisphere. In the present experiment, it is not possible to link with certainty the changes in the waking EEG after sleep or sleep deprivation with the motivational or cognitive deterioration found after sleep deprivation. However, several interesting correlations can be considered: 1) absolute power is significantly higher in populations with poorer performance in cognitive tests (25); 2) inter- and intrahemispheric relationships are altered in physiologic states characterized by some sort of dysfunction, such as coma (13), schizophrenia (26) or depression (27); 3) inter- and intrahemispheric correlation is lower during the perimenstrual phase (28); and 4) failure in cognitive task solution is correlated with lower interhemispheric correlation (29). The present findings show that the waking EEG varies as a (unction of prior sleep or wakefulness, suggesting an organizing effect of sleep on the EEG, whereas the accumulating hours of wakefulness have a deteriorating effect. This hypothesis is presently being tested by measuring EEG and task performance at different durations of wakefulness. It has been reported that delta absolute power is increased during the first hours of sleep as a function of accumulating hours of prior wakefulness (30). Accordingly, it could be hypothesized that the pressure for delta would also be observed in the waking EEG. However, we found no significant differences between delta power after sleep and after sleep deprivation. Results observed in relative power suggest that a particular physiological state is characterized not only by changes in absolute power but also by different proportions of each band. In addition to the decrease of absolute power observed after sleep, the proportional contribution of each band was different. The proportion of delta and alpha 1 was higher after sleep, whereas the proportion of alpha2 and beta2 was higher after sleep deprivation. It seems as ifthe faster components of the spectrum are enhanced by the lack of sleep, whereas the slower components are more prominent after sleep. EEG activity during wakefulness is a sensitive parameter and a useful tool to assess the consequences of sleep and of lack of sleep on brain functional organization. Acknowledgement: This study was financed in part by grant 90 III 3666 from CONACYT, Mexico. REFERENCES 1. Rechtschaffen A, Bergmann BM, Everson CA, Kishida CA, Gilliland MA. Sleep deprivation in the rat: integration and discussion of the findings. Sleep 1989;12: Naitoh P, Kales E, Kollar J, Smith C, Jacobson A. Electroencephalographic activity after prolonged sleep loss. Electroenceph Clin NeurophysioI1969;27: Naitoh P, Pasnau RO, Kollar EJ. Psychophysiological changes after prolonged deprivation of sleep. Bioi Psychiat 1971 ;3: Campbell K, Charbonneau S, Beaudoin R. Evoked potentials correlates of total sleep deprivation. Sleep Res 1980;9: Kjellberg A. Sleep deprivation and some aspects of performance: motivation, comment and conclusions. Waking and Sleeping 1977;1: Home JA, Anderson NR, Wilkinson RT. Effects of sleep deprivation on signal detection measures of vigilance; implications for sleep function. Sleep 1983;6: Mullaney DJ, Kripke DF, Fleck PA, Johnson LC. Sleep loss and nap effects on sustained continuous performance. Psychophysiol 1983;20: Bonnet MH. Sleep restoration as a function of periodic awakening, movement, or electroencephalographic change. Sleep 1987; 10: Bonnet MH. Infrequent periodic sleep disruption: effects on sleep, performance and mood. Physiol Behav 1989;45: Babkoff H, Mikulincer M, Caspy T, Kempinski D, Sing HB. The topology of performance curves during 72 hours of sleep loss: a memory and search task. Q J Exp Psychol 1988;40A: II. Mikulincer M, Babkoff H, Caspy T. The effects of 72 hours of sleep loss on psychological variables. Br J PsychoiI989;80: Shaw Jc. Correlation and coherence analysis of the EEG: a tutorial review. Int J Psychophysiol1984; 1 : Grindel OM. Optimal level of EEG coherence and its role in evaluation of the state of human brain functions. Neurosci Behav Physio/ 1982; 12: Dumermuth G, Lange B, Lehmann D, Meier CA, Kinhelmann R, Molinari L. Spectral analysis of all-night sleep EEG in healthy adults. Eur NeuroI1983;22: IS. Banquet JP. Inter and intrahemispheric relationships of the EEG during sleep in man. Electroenceph Clin NeurophysioI1983;55: Corsi-Cabrera M, Meneses S, Molina E. Correlaci6n interhemis- Sleep. Vol. 15. No

6 WAKING EEG AFTER SLEEP AND SLEEP LOSS 555 ferica y acoplamiento temporal de la actividad electrica cortical durante la vigilia, la etapa II y el suefio parad6jico en el hombre. Revista Mexicana de Psicologia 1987;4: French CCh, Beaumont JG. A critical review ofeeg coherence studies of hemisphere function. I nt J Psychophysiol1984; I : Ford MR, Goethe JW, Dekker DK. EEG coherence and power changes during a continuous movement task. Int J PsychophysioI1986;99-1l Corsi-Cabrera M, Herrera P, Malvido M. Correlation between EEG and cognitive abilities: sex differences. Int J Neurosci 1989; 45; Corsi-Cabrera M, Ramos J, Meneses S. Effect of normal sleep and sleep deprivation on interhemispheric correlation during subsequent wakefulness in man. Electroenceph Clin NeurophysioI1989;72: John ER, Ahn H, Prichep L, Trepetin M, Brown D, Kaye H. Developmental equations for the EEG. Science 1980;210: Gasser T, Bacher P, Mocks J. Transformation toward the normal distribution of broad band spectral parameters of the EEG. Electroenceph and Clin Neuroph 1982;53: Dijk DJ, Beersma DGM, Daan S. EEG power density during nap sleep: reflection of an hourglass measuring the duration of prior wakefulness. J BioI Rhythms 1987;2: Bullock TH, McClune Me. Lateral coherence of the electrocorticogram: a new measure of brain synchrony. Electroencephalograph Clin Physiol 1989;73: Giannitrapani D. The coherence of the EEG in normal and schi?:ophrenic subjects. Electroenceph Clin Neurophysiol 1980; 49:108P. 26. Flor-Henry P, Koles ZJ. Studies in depression, mania and normals: evidence for partial shifts of laterality in the affective psychosis. Adv BioI Psychiat 1980;4: Solis-Ortiz S, Corsi-Cabrera M, Ramos J, Arce C, Guevara MA. Correlatos electroencefalognificos del ciclo menstrual. Rev Cubana Invest Biomed 1991;10: Arce C, Corsi-Cabrera M, Ramos J, Guevara MA. Correlatos electroencefalognificos de la habilidad espacial en sujetos con puntajes extremos. Rev Cubana Inves Biomed 1991;10: Corsi-Cabrera M, Gutierrez S, Ramos J, Arce e. Interhemispheric correlation of EEG activity during successful and unsuccessful cognitive performance. Int J neuroscie 1988;39: Borbely AA, Baumann F, Brandeis D, Strauch I, Lehmann D. Sleep deprivation: effect on sleep stages and EEG power density in man. Electroenceph Clin Neurophysiol1981 ;51 : Sleep. Vol. 15. No