Effects of sleep deprivation, lunch, and personality on performance, mood, and cardiovascular function

Physiology&Behavior,Vol. 54, pp. 967-972, 1993

0031-9384/93 $6.00 + .00 Copyright© 1993 PergamonPressLtd.

Printed in the USA.

Effects of Sleep Deprivation, Lunch, and Personality on Performance, Mood, and Cardiovascular Function ANDREW

SMITH 1 AND ANDREA

MABEN

Health Psychology Research Unit, School of Psychology, University of Wales College of Cardiff, P.O. Box 901, Cardiff CF1 3YG Wales R e c e i v e d 21 D e c e m b e r 1992 SMITH, A. AND A. MABEN. Effects of sleep deprivation, lunch, and personality on performance, mood, and cardiovascular function. PHYSIOL BEHAV 54(5) 967-972, 1993.--The present study examined the effects of sleep deprivation on performance, mood, and cardiovascular functioning in the late morning and early afternoon. The results showed that the sleep-deprived subjects felt less alert and detected fewer targets in a cognitive vigilance task. Selective impairments due to sleep deprivation were also observed in a logical reasoning task. There was little evidence to suggest that consumption of lunch altered these effects of sleep deprivation. However, individual differences in the effects of sleep deprivation were apparent, with trait anxiety being related to the changes in subjective alertness produced by sleep deprivation, and sleep-deprived extraverts showing greater performance impairments than the sleep-deprived introverts. Sleep deprivation

Lunch

Personality

Performance

THERE has been considerable interest in the effects of sleep deprivation on performance and mood (24). It is generally agreed that sleep deprivation will impair tasks that require sustained attention or where heavy demands are made on working memory and high response rates are required. It has also been shown that the effects of sleep deprivation may be modified by other factors (25). Generally, factors that increase alertness, such as noise, incentives, and stimulant drugs, reduce the effects of sleep deprivation. In contrast, factors that reduce alertness, such as fatigue and depressant drugs, increase the impairments associated with sleep deprivation. Circadian rhythms clearly modify the effects of sleep deprivation. For example, the effects of sleep deprivation are more pronounced at night, and performance during the first night without sleep can often be worse than performance the next day, even though the number of hours awake is greater then (l 1). It has been suggested that there is a bicircadian rhythm of sleepiness (9), which accounts for the drop in alertness found in the early afternoon. One would predict, therefore, that the effects of sleep deprivation should be more marked then than later in the afternoon. This issue was examined in the present study. Recent research has considered the postlunch dip in alertness and performance in detail, and the main results can be briefly summarised as follows. First, there is considerable evidence that subjective alertness is reduced after lunch, as is the ability to perform sustained

Requests for reprints should be addressed to Andrew Smith.

967

Mood

Cardiovascular function

attention tasks. These differences between late morning and early afternoon depend on consumption of a meal and do not just reflect endogenous circadian rhythms (i.e., they are not observed in subjects who have fasted (6,18,19). The extent of these postlunch changes in performance depend on the personality of the subject, the nature of the meal, the type of task carried out, and whether alertness is increased by stimulants after the meal. Subjects who have low levels of anxiety have been shown to have greater impairments in performance after lunch than anxious subjects (6,20). Postlunch impairments are not usually found with short tasks (4), and lunch may even improve aspects of tasks that do not require sustained attention (12). However, short tasks may be impaired following lunch if they are part of a long battery of tests (13,20). Meal size has been shown to influence the frequency of lapses of attention (21), whereas the nutrient composition of the lunch changes the type of attention that is impaired; high protein meals lead to increased distractability, and high carbohydrate meals slow reactions to stimuli in the periphery (16). Postlunch impairments can be reduced or removed by giving the subject catfeinated drinks (22) or by having the subject perform in noise (20). Other changes in the early afternoon do not depend on consumption of lunch ( 18,19). This suggests that postlunch changes have both an endogenous and meal-related component. Indeed, it may be the case that the potential for sleepiness is greater at this time, and this may express itself in certain forms. Consumption of the meal may change

968

SMITH AND MABEN

the state of the person from one of potential sleepiness to a real state of reduced arousal. This would explain why it is a combination of the meal and this particular time of day that reduces alertness and leads to a drop in the ability to maintain attention (a similar meal given in the evening does not produce these effects) (17). A second aim of the present study was to examine whether consumption of lunch has different effects on the performance and mood of sleep deprived and nondeprived subjects. A third aim of the study was to examine individual differences in the effects of sleep deprivation and lunch. Both sleep deprivation and the effects of lunch show strong effects of personality. In the case of sleep deprivation, the effects of sleep deprivation are greater in extraverts (5) than introverts. Subjects with low levels of anxiety typically show greater postlunch decrements than anxious subjects (20). The relationship between anxiety and postlunch changes in performance is only observed in subjects who have eaten lunch. Smith and Miles (20) have suggested that the lunch-anxiety relationship can be used as an indicator of whether an effect observed in the early afternoon (following consumption of lunch) reflects an effect of the meal or an endogenous effect. For example, if sleep deprivation interacted with the effect produced by the meal, one might expect it to also interact with anxiety. In contrast, if it interacted with endogenous circadian changes, one might expect a different effect in the early afternoon compared to the late morning, but one would not necessarily expect an interaction with anxiety. These individual differences in the effects of sleep deprivation and lunch were examined in the present experiment. The experiment was designed to provide a profile of the effects of sleep loss and lunch on mood, cardiovascular function, and two types of performance. One performance test, a cognitive vigilance task, was selected because it has been shown that performance of this type of task is frequently impaired by both sleep loss and lunch, and the other, a logical reasoning task, was chosen because performance of it is not generally altered by the factors considered here. EXPERIMENT

Method A between-subjects design was used with subjects being randomly assigned to the sleep deprivation or nondeprived conditions. All of the subjects were given extensive practice at the tests prior to the experiment and completed questionnaires measuring personality [Eysenck Personality Inventory (7)-measuring introversion, impulsivity, sociabilty, and neuroticlsm; Horne and Ostberg morningness questionnaire (10); Spielberger trait anxiety inventory (23); and the cognitive failures questionnaire (3)], recent minor psychoneurotic symptoms (a modified version of the Middlesex Hospital Questionnaire measunng anxiety, depression, and somatic symptoms over the last six weeks) (2), and a health-related behaviours questionnaire asking about usual eating, drinking, and sleeping patterns. The main aim of collecting most of this information was to examine whether there were any significant differences between the two groups prior to the sleep deprivation manipulation. METHOD

Subjects Twenty-one university students (12 males, 9 females, age range 18-24 years) took part in the study. Ten subjects were assigned to the sleep-deprived condition and 11 to the nondeprived condition. None of the subjects were taking prescribed or over-the-counter medication and all were nonsmokers. None

of the subjects regularly consumed large amounts of catl~inc (range 2-4 caffeinated drinks per day) or alcohol (range 0-16 units per week). The subjects in the sleep-deprived and nondeprived conditions were of comparable body weight (sleep deprived: mean = 66.7 kg, SD - I0.0: nondeprived: mean ~ 685) kg, SD 7.3).

Procedure Sleep deprivation. Subjects in the sleep-deprived condition arrived at the laboratory at 2200 h and during the night (until 0600 h) completed a series of performance tests at two hourly intervals. This meant that the study was not only examining the effects of sleep deprivation but also the after-effects of working at night. From 0600 h until the first testing session they remained in the laboratory where they were given breakfast and then allowed to read or watch a video, The nondeprived subjects followed their normal sleeping routine. Testing procedure. Subjects were tested in groups of up to six and each day two groups were tested with one group (the early group) being tested approximately 45 min before the other. The times for the early group are shown throughout the paper. The first (prelunch) test session took place at 1115 and subjects followed the procedure shown below. The subjects were then given lunch. Details of this are shown in Table 1. The second session took place at 1330 h and the third at 1445 h. Subjects remained in the laboratory between sessions. Measurement of Mood, Cardiovascular Function, and Performance At the start of each session blood pressure and pulse were measured. Following this, subjects rated their mood and carried out the performance tests described below. The mood rating and performance tests were presented on an IBM compatible computer. Mood was assessed using 18 bipolar visual analogue scales (e.g., drowsy-alert, tense-calm, etc.) (8). Logical reasoning task. This test was developed by Baddeley (1) and the subjects were shown statements about the order of the letters A and B followed by the letters AB or BA (e.g., A follows B: BA). The subjects had to read the statement and decide whether the sentence was a true description of the order of the letters. If it was, the subject pressed the T key on the keyboard, if it wasn't, they pressed the F key. The sentences ranged in TABLE 1 DETAILS OF THE LUNCH

All subjects ate the followinglunch: --Tomato, orange and herb soup with croutons --Chicken a la King --Boiled rice --Strawberries and raspberrieswith cream. Nutritional Composition Protein Fat Carbohydrate Energy Weight of meal

39 g 41g ll8g 970 kcal 870 g

All subjects had one cup of decaffeinated coffi~e after lunch.

EFFECTS OF SLEEP DEPRIVATION

969

syntactic complexity from simple active to passive negative (e.g., A is not followed by B). Subjects carried out the task for 3 min. Repeated digits vigilance task. This task was developed by Smith (14) and a modified version with a fast presentation rate was used here. This version shows a very rapid vigilance decrement (performance begins to decline after only l or 2 min). The subjects were shown three-digit numbers on the screen at the rate of 100 per min. Each number was normally different to the preceeding one, but occasionally (eight times a min) the same number was presented on successive trials. Subjects had to detect these repeats and respond as quickly as possible. The number of hits, reaction times for hits, and false alarms were recorded. The task lasted for 8 min.

90

•

NON-SLEEP DEPRIVED

•

SLEEP DEPRIVED

80 to v-7

70 60

ta o ~

50

t.) etI.tl Q.

RESULTS

40

Statistical Analysis Analyses of variance were carried out to determine whether sleep-deprived and nondeprived groups differed in personality, MHQ scores, or eating/drinking/sleeping habits. Analyses of variance were then performed on the cardiovascular, mood, and performance data with sleep condition as a between-subject factor and test sessions as a within-subject factor. Task parameters were also included as within-subject factors in the analyses of the performance tasks (e.g., logical reasoning: sentence type. Repeated digits: time on task).

Comparability of the Two Groups There were no significant differences between the two groups for the following measures: introversion, impulsivity, sociability, neuroticism, morningness, cognitive failures, MHQ depression, MHQ anxiety, MHQ somatic symptoms, or sleeping/ eating/drinking habits. This makes it difficult to attribute any effects obtained to differences present before the sleep manipulation.

Alert

50

•

non-sleep deprived

[]

sleep deprived

40

FIG. 2. Effectsof sleep deprivation on accuracy of performing the repeated digits task (group means are shown, SDs as bars).

Cardiovascular Measures No significant differences were found between the nondeprived and sleep-deprived subjects for either pulse or blood pressure.

Mood Significant effects of sleep deprivation were found for the mood scales relating to alertness, with the sleep-deprived subjects reporting that they were felt less alert [drowsy-alert; F(1, 19) = 9.34,p < 0.01; strong-feeble: F(1, 19) = 5.98,p < 0.05; muzzyclearheaded: F(1, 19) = 7.24, p < 0.01; lethargic-energetic: F(1, 19) = 11.77, p < 0.005; mentally slow-quick witted: F ( l , 19) = 7.33, p < 0.01; attentive-dreamy: F(I, 19) = 17.64, p < 0.001; incompetent-proficient: F( l, 19) = 6.01, p < 0.05; interestedbored: F(I, 19) = 5.78, p < 0.05]. The general trend is shown in Fig. 1 for the drowsy-alert ratings. It can be seen that the sleep-deprived subjects showed a linear decline in alertness, whereas the nondeprived subjects reported a slight drop in alertness after lunch but, generally, showed comparable levels over the sessions. The analyses revealed no significant interactions between session and sleep conditions.

Performance Tests

30

20

10

Drowsy

3O

0 11:15

1:30

3:00

Time of Day

FIG. 1. Effects of sleep deprivation on subjective ratings of alertness (scores are the distance in arbitrary units, 0-50, from the drowsy end of the scale. Group means are shown, SDs as bars).

Repeated digits vigilance task. A main effect of sleep deprivation was found in the analysis of the number of hits, F(1, 19) = 18.0, p < 0.0005, with the sleep-deprived subjects detecting fewer targets. This result is shown in Fig. 2. The effect of sleep deprivation on this task confirms that short, paced tasks with a high memory load and rapid presentation rate are impaired by sleep loss. No significant differences were found between the sleep conditions for the number of false alarms or response times. Similarly, no interactions between sleep conditions and sessions were significant. A clear vigilance decrement was observed with time on task, but sleep deprivation did not modify the extent of this. Logical reasoning task. No significant effects of sleep deprivation (or interactions between sleep deprivation and sessions) were found in the analyses of the total number completed or

970

SMITH AND MABEN

percentage correct. A second series of analyses included sentence type (active vs. passive; simple vs. negative) as a factor and the response time analysis revealed an interaction between sleep deprivation and active/passive and simple/negative sentences, F(l, 19) - 4.54, p < 0.05. This is shown in Fig. 3. It can be seen that there is little difference between the sleep-deprived and nondeprived groups apart from their responses to the simple, passive sentences (which were slower in the sleep-deprived group).

ALERT

50

•

NSD/LA

[]

SD/LA

• []

NSD/HA SD/HA

40

Individual Differences The questionnaire measures of individual differences collected in the present study have been widely used in our laboratory. A principal component analysis on data from 200 subjects yielded three factors--anxiety, extraversion, and morningness/obsessionality. As already mentioned in the Introduction section, there is evidence that sleep deprivation may have a greater effect on extraverts than introverts, and that the effects of lunch may be related to anxiety. Morningness is also a dimension that one might expect to be important in studies involving sleep deprivation. Trait anxiety, extraversion, and morningness were, therefore, included as factors in the analyses. Subjects were subdivided into high and low scorers on the basis of a median split. Separate analyses were carried out for each personality dimension. Mood. Interactions between sleep, session, and trait anxiety were found in the analyses of the drowsy/alert, F(2, 34) = 4.69, p < 0.05, muzzy/clearheaded, F(2, 34) = 4.69, p < 0.05, wellcoordinated/clumsy, F(2, 34) = 6.54, p < 0.005, and incompetent/proficient, F(2, 34) = 5.96, p < 0.01, ratings. The general trend of these effects is illustrated in Fig. 4 for the drowsy/alert ratings, If one considers the ratings given by the nondeprived subjects first, one can see that the high anxious group show a postlunch dip in alertness, whereas those with low trait anxiety do

m

[] [] [] []

0

NSD/SIM PLF-JACTIVE SD/SIMPLE/ACTIVE NSD/SIM PLFJPASSIVE SD/SIMPLE/PASSIVE NSD/NEGATIVE/ACTWE SD/NEGATIVE/ACTIVE NSD/NEGATIV F-JPASSIVE SD/NEGATIVE/PASSIVE

l

w

Z 0 0 ILl

30

20

10

DROWSY 0 1

2

3

SESSION

FIG. 4. Effects of sleep deprivation, trait anxiety, and test session on subjective ratings of alertness (scores are in arbitrary units, 0-50, from the drowsy end of the scale. Group means are shown, SDs as bars. SD = sleep deprived, NSD = non sleep deprived. LA = low trait anxiety, HA = high trait anxiety).

not. The sleep-deprived subjects show a different pattern, with the low anxious subjects reporting the lowest level of alertness at session 1, then becoming less alert at session 2, and staying at that level for the third session. In contrast to this, sleepdeprived subjects with high levels of trait anxiety reported comparable levels of alertness to their nondeprived counterparts until the third session, when they reported their lowest level of alertness. No interactions between morningness, sleep, and session were significant for any of the mood scales, and similar negative results were obtained in the analyses including extraversion. Logical reasoning. A significant interaction between extraversion, sleep conditions, and sentence type was obtained in the response time analysis, F(1, 17) = 9.27, p < 0.01. This effect is shown in Fig. 5. It can be seen that the sleep × sentence type interaction reported earlier was entirely due to the sleep-deprived extraverts. No interactions between either trait anxiety or morningness and sleep conditions, sessions, and task parameters were significant. Repeated digits vigilance task. The results from this task also showed an interaction between extraversion, sleep, and session. This was obtained in the analysis of the speed of responses to targets, F(2, 24) = 9.54, p < 0.001, and is shown in Fig. 6. All of the groups except the sleep-deprived extraverts had fairly consistent response times across sessions. The sleep-deprived extraverts became slower and slower as the experiment progressed and on the final session were about 200 ms slower than the other groups. No interactions between sleep conditions and morningness or trait anxiety were significant. DISCUSSION

FIG. 3. Effects of sleep deprivation on response time for the different sentence types in the logical reasoning task (group means are shown, SDs as bars. SD = sleep deprived, NSD = non sleep deprived).

The present study has shown that two short performance tests are sensitive to the effects of sleep deprivation. The cruieial

EFFECTS OF SLEEP DEPRIVATION

971

features of the detection of repeated numbers task are that it has a memory load and a fast presentation rate. This greatly increases the sensitivity of the test and gives it advantages over traditional vigilance tasks, which have to be performed for a long time before the effects of reduced levels of alertness can be demonstrated. The results from the logical reasoning test demonstrate the importance of examining the microstructure of responding rather than merely examining global measures such as the number done or percent correct. Similar results showing that certain aspects of the task are more sensitive than others have been demonstrated with other changes of state (e.g., effects of low doses of alcohol) (15). Computerised tests are necessary to examine such effects, which gives them distinct advantages over the paper and pencil measures used in earlier studies. While the performance measures and ratings of mood were sensitive to the effects of sleep deprivation, they showed no evidenee that sleep deprivation effects were greater for the group as a whole in the immediate postlunch period compared to either earlier or later in the day. This would appear to cast doubts on the view that sleep deprivation and postlunch effects influence the same mechanisms. However, the data demonstrated that there were personality differences in the effects of sleep deprivation and lunch. Generally, extraverts showed greater effects of sleep deprivation on performance than introverts, which is

• • • []

7.5

NSD/INTROVERT SD/INTROVERT NSD/EXTRAVERT SD/EXTRAVERT

0 I.t,.I O3 v U.I I--r 0 l,-13= I.U

SA

SP

NA

NP

SENTENCE TYPE

FIG. 5. Effects of sleep deprivation, extraversion, and sentence type on response times in the logical reasoning task (scores are group means, SDs shown as bars. SD = sleep deprived, NSD = non sleep deprived. Sentencetype: SA = simple,active; SP = simple,passive;NA = negative, active; NP = negative, passive).

• • • []

800

~o

NSD/INTROVERT SD/INTROVERT NSD/EXTRAVERT SD/EXTRAVERT

750

t.)

tu CO

700

m

650

~-

600

r.

r

o (D (E

m at-

550 500

450 400

1

2

3

SESSION

FIG. 6. Effects of sleep deprivation,extraversion,and sessionon response times in the repeated digits task (group means are shown, SDs as bars. SD = sleep deprived, NSD = non sleep deprived).

consistent with earlier findings. In contrast to this, trait anxiety appeared to be the important factor related to mood changes. High anxious subjects reported a drop in alertness after lunch, and this effect was observed for both sleep-deprived and nondeprived subjects. The subjects with low trait anxiety only reported a postlunch drop in alertness if they were sleep deprived. The present findings clearly need to be replicated and the personality results should be treated with caution as they are based on small numbers of subjects. Future studies should also include a no-meal condition, as this will enable one to determine whether any sleep deprivation/postlunch effects reflect the effect of the meal or endogenous variation. The range of physiological markers should also be extended, as this may provide a clearer indication of the mechanisms underlying the effects. Such studies should surely be undertaken as the present experiment has demonstrated that we need further information about the combined effects of factors that change the state of the person. This approach will add to the knowledge obtained from studies of single factors in isolation, and will allow us to assess the extent to which different changes in state involve similar or different mechanisms. ACKNOWLEDGEMENT This research was supported by a grant by the Agricultureand Food Research Council.

REFERENCES 1. Baddeley,A. D. A three-minutereasoningtest based on grammatical transformation. Psychonom. Sci. 10:341-342; 1968. 2. Broadbent, D. E.; Broadbent, M. H. P.; Philpotts, R. J.; Wallace,J. Some further studies on the prediction of experimental colds in volunteers by psychologicalfactors. Psychosom. Res. 28:511-523; 1984. 3. Broadbent, D. E.; Cooper, P. J.; FitzGerald, P. F.; Parkes, K. R. The cognitivefailuresquestionnaire (CFQ) and its correlates. Br. J. Clin. Psychol. 21:1-6; 1982.

4. Christie,M. J.; McBrearty, E. M. Psychophysiologicalinvestigations of post lunch state in male and female subjects.Ergonomics22:307323; 1979. 5. Corcoran, D. W. J. Noise and loss of sleep. Q. J. Exp. Psychol. 14: 178-182; 1962. 6. Craig, A.; Baer, K.; Diekmann, A. The effects of lunch on sensoryperceptual functioning in man. Int. Arch. Occup. Environ. Health 49:105-114; 1981.

972

7. Eysenck, H. J.; Eysenck, S. B. G. Manual of the Eysenck Personality Inventory. London: University of London Press: 1964. 8. Herbert, M.; John, M. W.; Dore, C. Factor analysis of analogue scales measuring subjective feelings before and after sleep. Br. Med. Psychol. 49:373-379: 1976. 9. Home, J. "'Core" and "'Optional" sleepiness. In: Broughton, R. J.; Ogilvie, R. D., eds. Sleep, arousal and performance. Boston: Birkhauser; 1991:27-44. 10. Horne, J.: Ostberg, O. A self-assessment questionnaire to determine morningness-eveningness in human circadian rh~hms. Int. J. Chronobiol. 4:9%110; 1976. 11. Kleitman, N. Sleep and wakefulness. University of Chicago Press; 1963. 12. Millar, K.; Hammersley, R.; Finnigan, F. Reduction of alcohol-induced performance impairment by prior ingestion of food. Br. J. Psychol. 83:261-278; 1992. 13. Smith, A. P. Effects of meals on memory and attention. In: Gruneberg, M.; Morris, P.; Sykes, R., eds. Practical aspects of memory: Current research and issues, vol. 2. Chichester: Wiley; 1988:477482. 14. Smith, A. P. Acute effects of noise exposure; An experimental investigation of the effects of noise and task parameters on cognitive vigilance tasks. Int. Arch. Occup. Environ. Health 60:307-310; 1988. 15. Smith, A. P.; Kendrick, A.; Maben, A. Effects of caffeine, lunch and alcohol on human performance, mood and cardiovascular function. Proc. Nutr. Soc. 51:325-333; 1992.

S M I T H A N D MABEN

16. Smith, A. P.; Leekam, S.: Ralph, A.: McNeiU, G. -lhe intluence of meal composition on post-lunch changes in performance efficiency and mood. Appetite 10:195-203: 1988. 17. Smith, A. P.; Maben, A.; Brockman, P. The effects of evening meals and caffeine on performance, mood and cardiovascular functioning. Appetite (in press). 18. Smith, A. P.: Miles, C. Effects of lunch on cognitive vigilance tasks. Ergonomics 29:125 l - 1261 ; 1986. 19. Smith, A. P.; Miles, C. Effects of lunch on selective and sustained attention. Neuropsychobiology 16:117-120; 1987. 20. Smith, A. P.: Miles, C. Acute effects of meals, noise and nightwork. Br. J. Psychol. 77:377-389; 1986. 21. Smith, A.; Ralph, A.; McNeill, G. Influences of meal size on postlunch changes in performance efficiency, mood and cardiovascular function. Appetite 16:85-91; 1991. 22. Smith, A.; Rusted, J. M.: Eaton-Williams, P.: Savory, M.; Leathwood, P. Effects of caffeine given before and after lunch on sustained attention. Neuropsychobiology 23:160-163; 1990. 23. Spielberger, C. D.: Gorsuch, R. L.; kushene, R. E. STAI Manual for the State-Trait Anxiety Inventory. Palo Alto: Consulting Psychologists Press: 1970. 24. Tilley, A.; Brown, S. Sleep deprivation. In: Smith, A. P.; Jones, D. M., eds. Handbook of human performance, vol. 2. London: Academic Press; 1992:237-260. 25. Wilkinson, R. T. Sleep deprivation. In: Edholm, O, G.; Bacharach, A. L., eds. The physiology of human survival. New York: Academic Press: 1965:399-430.