Actigraphic Motor Asymmetries During Sleep

Transcription

1 INSTRUMENTATION AND METHODOLOGY Actigraphic Motor Asymmetries During Sleep Cristiano Violani, Paola Testa, and Maria Casagrande Dipartimento di Psicologia - Università di Roma La Sapienza Summary: Much evidence indicates that during sleep there is a repatterning of motor asymmetries with a relative advantage of the left hand (ie, the left hand moves more than the right). This could be due to the ability of the right hemisphere in operating at levels of reduced arousal (arousal hypothesis) or to its superior spatial abilities (motor specificity hypothesis), or it could indicate a greater need for sleep in the left hemisphere (homeostatic hypothesis). Since only the latter hypothesis predicts that the repatterning should be present in the first part of sleep (ie, when the homeostatic processes are more pronounced), the present study evaluated whether actigraphic data are consistent with this prediction. Sixteen right-handed college students wore actigraphs (AMI 16K) on both upper and lower limbs for about 56 hours. Factorial ANOVAs were carried out on side (left vs right) and part (first vs second) of the recording period during sleep and waking. During waking, the right hand showed more intense motor activity as compared to the left. During sleep, in the first part of the night, the right hand lost this advantage, while in the second part of the night it regained its superiority. Since this repatterning was specific for hand movements and no difference was found in overall motor activity and in arousal between the two parts of the sleep period, the results are interpreted as consistent with the homeostatic hypothesis. Key words: Hemispheric asymmetry; motor activity; actigraphy MULLER-LIMROTH was the first to describe a motor asymmetry during sleep; in his words, tickling the nose causes right handed people to make a defensive movement with the left hand and vice versa (cited in 1). Later, Jovanovic 1 analyzed the distribution of the laterality of the electromyographic movements of the hands in 15 righthanded subjects and in 5 left-handers; he found that during sleep the nondominant hand was two times more active than the dominant one, independent of the phase of the REM-NREM cycle. More recently, Muller-Limroth s early findings were replicated by a study 2 in which tactile stimuli were administered by touching the subject s nose with a paintbrush both during stage 1 and stage 2 sleep and during waking with subjects supine with eyes closed and both Accepted for publication April, 1998 Address correspondence and requests for reprints to Prof. Cristiano Violani, Metodologia delle Scienze del Comportamento, Dipartimento di Psicologia - Università di Roma La Sapienza, via dei Marsi Roma, ITALY hands equally free to move. Results revealed that righthand predominance in responding to annoying stimuli observed during waking declined across the sleep-onset period. These results are in keeping with data showing a predominance of the left hand during the sleep-onset period both in video-recorded hand movements and in manual responses to auditory stimuli. 3,4 In a first study, Lauerma and colleagues 3 evaluated the laterality of motor activity by infrared video recordings, distinguishing between organized (ie, a hand touching the subject s face, head and neck) and nonorganized movements (ie, incomplete hand and finger movements), and found a 2.5:1 ratio between the nondominant and the dominant hand, which was present in each sleep stage; furthermore, the predominance of the nondominant hand was independent of body position and involved only organized movements, while nonorganized movements were symmetrical. The latter difference has been explained by the authors as due to a different neuronal organization of the two types of movements, nonorganized movements depending on the disinhibition of brainstem 472

2 Table 1. Mean (± sd) sleep staging per hour during the night in which wrist actigraphy was recorded 1 st h 2 nd h 3 rd h 4 th h 5 th h 6 th h Minutes in stage ± ±.65.62± ± ± ±2.17 Minutes in stage ± ± ± ± ± ±9.25 Minutes in SWS 25.62± ± ± ± ± ±6.61 Minutes in stage REM 0±0 4± ± ± ± ±8.35 Minutes of W 10.86± ± ± ± ± ±8.07 Number of MT.38±.36.57±.46.67±.51.71± ±.57.91±.98 Number of MA 9.29± ± ± ± ± ±6.78 h: hour; SWS: slow-wave sleep; W: wake; MT: movement time; MA: movement arousal. and spinal systems, and organized movements on cortical structures involving the somatic-sensitive and prefrontal cortex. More recently, Lauerma and colleagues 4 studied both the lateralization of motor responses made in order to interrupt annoying acoustic stimuli presented during the sleeponset period, and the lateralization of motor activity recorded by actigraphs worn on both wrists for 13.5 hours. Responses to the acoustic stimuli showed significant righthand advantage in all sleep stages except in stage 1, while actigraphic data showed a slight left-side superiority (mean motor activity percentage was 1.16 for the left hand and 1.13 for the right hand) during sleep as opposed to waking, in which mean motor activity was for the left hand and for the right hand. Most of the data so far considered converge in showing a repatterning of motor asymmetry during wake-sleep transition and/or during sleep. This is in keeping with recent results 5 showing that, during the sleep-onset period, there is a state-specific relative advantage of the right hemisphere which is revealed by a stronger impairment of the right hand, both in reaction time to auditory stimuli and in sustaining endogenous motor programs (such as performing a finger-tapping task). The relative superiority of the left hand during sleep could be due either to the ability of the right hemisphere in operating at levels of reduced arousal, as suggested by its superiority in sustaining attention (eg, 6,7) and in performing tasks requiring nonfocal attention, 8,9 or it could depend on a more pronounced homeostatic deactivation of the left hemisphere. In this view, as the left hemisphere in righthanded humans is more engaged in mediating the relationships with others (language) and with the environment (control of rapid and complex sequences of movements) during waking, it could accumulate a greater sleep debt compared with the right hemisphere and, therefore, would cease controlling motor activities in a quicker and more pronounced way than the right hemisphere. A third explanation for the relative superiority of the left hand movements during sleep was proposed by Lauerma and his collaborators 4 ; in their view, it is due specifically to the spontaneous organized movements related to arranging bedclothing and making postural shifts, which imply the superior spatial abilities of the right hemisphere. The aim of the present study was to evaluate whether actigraphic data can reveal variations of lateralization of motor activity during sleep that are consistent with one of these hypotheses. Both the reduced-arousal hypothesis and the motorspecificity hypothesis would be supported if the righthand impairment is present throughout the entire sleep period. Conversely, the homeostatic hypothesis would be supported by finding that a repatterning of motor hemispheric asymmetry is present and more pronounced in the first part of the night, ie, when the homeostatic process is more pronounced. In other words, the homeostatic hypothesis predicts that a change in lateralization favoring the movements of the left hand should be present only during the first hours of sleep. In the present study we assessed whether the hypothesized change of motor laterality during sleep occurs only in the first part of the night or persists during the whole sleep period, and we controlled whether the change of motor laterality from waking to sleep affects both upper and lower limbs or is specific of hand movements. METHOD Subjects Sixteen healthy right-handed male undergraduates (with hand preference ³.85, as assessed by means of a Lateral Preference Questionnaire, 10 ) aged (mean =24.75; sd=2.42), served as paid volunteers in the study. For 1 week, subjects were required to complete a sleep questionnaire at home every day upon morning awakening. Only subjects who reported normal sleep duration and schedule with a monophasic sleep period from 23:00 to 07:00 ± 1 hour without any sleep, medical, or psychiatric disorders, were recruited for the study. 473

3 Apparatus Actigraphic recording. Actigraphic recordings were obtained using four Motion Logger 16K actigraphs (Ambulatory Monitoring Inc., Ardsley, NY). Actigraphs were initialized for zero crossing (mode 18, internal device code), with a 1-minute epoch. Data were edited to mark the time spent in bed and were analyzed through the Action 2.1 software. Polygraphic recording. EEG was recorded from two monopolar locations (C3-A2, C4-A1); two horizontal EOG were recorded from the left and right outer canthus, respectively, referred to A2 and A1; and one bipolar vertical EOG, recorded from electrodes located 3 cm above and below the right eye pupil. EEG and EOG parameters were recorded with a time constant of 0.03 seconds. All recordings were in AC. Central EEG (C3-A2), EMG, and horizontal and vertical EOG were used to visually score sleep stage, using the standard criteria. 11 PROCEDURE Subjects spent 3 nights in a sleep laboratory, in a sound-attenued air-conditioned sleep room, with standard polygraphic recordings. At the beginning of the first night (about 20:00), each subject wore an actigraph on each wrist or ankle for about 12 hours to get used to it. Data collected during this first night were not analyzed. From the end of the first night and for 48 consecutive hours (from 09:00), half of the subjects wore an actigraph on each wrist during the first 24 hours and an actigraph on each ankle during the following 24 hours, and the other half wore the actigraphs in an inverse order. Subjects were free to spend their daytime hours outside of the sleep lab, but were asked to avoid intense motor activity (such as jogging, playing soccer or tennis); they were also instructed to use the actigraph event marker to signal times when they were in a vehicle or they had removed the actigraph. Prior to each recording period, to ensure that the actigraphs sensitivity was identical and constant throughout the study, the instruments were tested with the pendulum technique as described by Tryon. 12 Actigraphic and polygraphic lights-off were synchronized by asking the subjects to press the actigraph event marker the moment the lights were turned off and the polysomnographic recording was started. Sleep onset was at about 23:30, ranging from 23:00 to 24:00. Data Analysis Polysomnographic recordings were scored by two expert scorers according to standard criteria. 11 The sleep onset was defined by the appearance of the first epoch of stage-1 sleep. Actigraphic sleep onset was defined a posteriori on the Figure 1a Figure 1b Figure 1. (a) Sum of motor activity for right and left wrists in the first and in the last part of sleep. (b) Number of 1-minute epochs with motor activity greater than 10 units for right and left wrists in the first and last part of sleep. basis of the polygraphic criteria, as coincident to the beginning of the first epoch of stage 1 sleep. For each actigraphic recording, we calculated the sum of motor activity for each hour; waking periods marked because the subject was in a vehicle or because the actigraph was removed were not considered. During sleep, in order to distinguish the contribution of small and large movements and to estimate the number of the latter, wealso counted the number of 1-minute epochs with motor activity greater than 10 units. Since the quantity of motor 474

4 activity recorded during sleep and waking was obviously incomparable, the data were analyzed by ANOVAs considering side (right vs left) and hour of sleep (the first 3 hours vs the last 3 hours) separately as a function of the limbs (wrists-ankles) and of condition (wake-sleep). RESULTS Sleep Staging For the night during which the actigraphs were on the wrists, total bed time mean was ±28.6 minutes, total sleep time was ±36.6, sleep efficiency was 91.19±3 and stage-1 latency ±8.8. Table 1 reports sleep staging per hour. One-way ANOVAs did not reveal any significant time-of-night effects in stage W, MT, and number of movement arousals (all p>.10). Wrist Activity During Sleep A2 2 3 factorial ANOVA on the sum of motor activity per hour with wrist (right vs left), part (first vs second),and hours (first, second, third) as factors revealed no main significant effects (F<1), while the interaction wrist by part was significant (F1,15=7.15; p<.02). Planned comparisons showed that, while during the first part of sleep there was no significant difference between right wrist (M= ±282.66) and left wrist (M=372.58±337.24), during the second part of the night the difference was significant (F1,15=7.94; p<.01), with the right wrist showing greater motor activity (M=387.90±314.80) with respect to the left (M=275.79±284.76; p<.05; see Figure 1a). Wrist-by-hour (F<1), hour-by-part (F<1), and wrist-by-hour-by-part (F=1.44) interactions were not significant. In order to assess the contribution of large movements, the number of 1-minute epochs with motor activity greater than 10 units was submitted to a factorial ANOVA with wrist (right vs left), part (first vs second) and hours (first, second, third) as factors, which revealed no significant main effects (F<1), although the interaction wrist by part was not significant (F1,15=3.77; p<.07). Planned comparisons showed that although during the first part of the sleep the difference between right wrist (M=20.77±12.35) and left wrist (M=20.15±10.58) was not significant, during the second part of the night the difference was significant (F1,15=7.86; p<.01), with the right wrist showing more prolonged periods of activity (M=22.79±10.84) with respect to the left (M=19.19±9.37) (see Figure 1b). Wrist - by-hour (F<1), hour-by-part (F<1), and wrist-by-hour-bypart (F=1.13) interactions were not significant. Wrist Activity During Waking A2 2 3 factorial ANOVA on the sum of motor activity per hour with wrist (right vs left), part (first vs second), and hours (first, second, third) as factors showed a main effect for the wrist (F1,15=5.62; p<.03), indicating that there was greater motor activity for the right wrist (M= ± ) as compared to the left wrist (M= ± ). Other effects or interactions were not significant (F<1). ANOVAs carried out on the number of epochs with wrist activity greater than 10 units during waking did not show any main effect or interaction. Ankle Activity During Waking and During Sleep ANOVAs carried out on ankle movements both during sleep and during waking did not show any main effect or interaction (F<1). DISCUSSION During waking, actigraphic data reveal a significant asymmetry; this does not concern the number of epochs in which there are movements, but instead the quantity of movements, which, as expected, is greater in the right wrist. This difference disappears during the first part of the night, with a slight (nonsignificant) greater motor activity for the left wrist. During the second part of the night, the right wrist regains its superiority. These data are consistent with those reported by Lauerma, 3,4 suggesting a repatterning of the lateralization of motor activity during sleep. They are also consistent with our previous results showing an inversion of motor asymmetries during the sleep-onset period. 5 The fact that the disappearance of the right-wrist advantage is present only in the first half of the sleep period (when SWS and the restorative functions of sleep are more pronounced) is consistent with the homeostatic hypothesis, assuming that during waking a greater need for sleep is built into the left hemisphere. This homeostatic interpretation of behavioral data is in keeping with recent EEG data showing that unilateral activation of the left somatosensory cortex during wakefulness results in an increase in power density in the delta frequency in the left hemisphere during the first hour of subsequent sleep. 13 The different pattern of wrist lateralization observed is not due either to an increase of waking in the second part of sleep (minutes in wake stage were not different across the sleep period) nor to any aspecific increase of motor activity. Right-wrist advantage in the second part of sleep might be specific to REM sleep; unfortunately, due to the loss of some paper recordings, we cannot test this hypothesis. Actigraphic recordings with large (1-minute) time windows do not permit us to establish the nature of the movements contributing to the asymmetry. It is, however, noteworthy that the repatterning was shown also when small movements were excluded by considering only time win- 475

5 dows with more intense motor activity. Changes in asymmetry during sleep concerned movements of the wrists and not those of the ankles, which does not contradict the homeostatic hypothesis, since no significant difference was found for the ankles either in waking or in sleep. The lack of asymmetry in the lower-limb movements does not support the possibility that the repatterning of laterality of wrist movements depends on the side on which a subject sleeps. The present data show that during sleep there is a repatterning of motor asymmetries which is confined to the first hours. The nonsignificant difference in our data allows us to conclude that, during the first hours of sleep, the superiority of the right hand is lost. However, converging evidence from other studies 3-5 leads us to hypothesize that actually there is a nondominant hand superiority which might be revealed by more sensitive methods. Further studies allowing a more precise categorization of movements (through actigraphy with small time windows and/or videopoligraphy) are needed to establish the size of the phenomenon and whether it is specific to some category of movements. ACKNOWLEDGMENTS Study supported in part by fundings MURST60%92 Ateneo. Suggestions and comments by two anonymous reviewers are gratefully acknowledged. REFERENCES 1. Jovanovic UJ, ed. Normal sleep in man. Stuttgart: Hippokrates Verlag, Violani C, Casagrande M, Cinelli A, Testa P. Motor asymmetries in removing annoying stimuli during the transition from wakefulness to sleep. J Sleep Res 1992;1(Suppl. 1): Lauerma H, Lehtinen I, Lehtinen P, Korkeila JA, Toivonen S, Vaahtoranta K, Holmstrom R. Laterality of motor activity during normal and disturbed sleep. Biol Psychiatry 1992;32: Lauerma H, Kaartinen J, Polo O, Sallinen M, Lyytinen H.Asymmetry of instructed motor response to auditory stimuli during sleep. Sleep 1994;17(5): Casagrande M, Violani C, De Gennaro L, Braibanti P, Bertini M. Which hemisphere falls asleep first? Neuropsychologia 1995;33(7): Heilman KM & Van Den Abel T. Right hemispheric dominance for mediating cerebral activation. Neuropsychologia 1979;17: Posner MI & Petersen SE. The attention system of the human brain. Annu Rev Neurosi 1990;13: Peters M. Attentional asymmetries during concurrent bimanual performance. Q J Exp Psychol 1981;33A: Peters M. Constraints in the performance of bimanual tasks and their expression in unskilled and skilled subjects. Q J Exp Psychol 1985; 37A: Salmaso D, Longoni AM. Problems in the assessment of hand preference. Cortex 1985;21: Rechtshaffen A, Kales A, eds. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Brain Information Service/Brain Res Institute, University of California at Los Angeles, Tryon WW Activity measurements in psychology and medicine. New York: Plenum Press, Kattler H, Dijk D, Borbély AA. Effect of unilateral somatosensory stimulation prior to sleep on the sleep EEG in humans. J Sleep Res 1994;3: