Tải bản đầy đủ - 0 (trang)
2 Falling Asleep Versus Staying Awake

2 Falling Asleep Versus Staying Awake

Tải bản đầy đủ - 0trang




(according to the 15 second criterion) within the 20-minute limit of a

MSLT session, although, of course, this likelihood is higher during early

afternoon sessions.

If these sessions are extended to 30 minutes, then most people will

indeed fall asleep, which shows just how many of us can fairly easily fall

asleep during the daytime, given the right circumstances. But whether

this is still illustrative of hidden ‘sleep debt’ is another matter. One of

the best examples comes from a large study [4] of 100 volunteers, equal

numbers of men and women, aged 25–65 years. All were screened to

exclude sleep complaints and any subjective signs of even minor daytime

sleepiness, and who had maintained their usual sleep habits for two weeks

prior to undergoing a day of five MSLTs. However, these sessions were

extended to 30 minutes, and of the total of 500 sessions, sleep onset

occurred on 413 (83 %) occasions, and only two participants never fell

asleep. The average MSLT scores were rather longer, as one might expect.

Nevertheless, one might ask why so many people can fall asleep at

times of the day when they would not normally do so, and whether this

indicates sleep debt having reached epidemic proportions. Or, on the

other hand, maybe there is a voluntary ability to fall asleep under low levels of sleepiness, given the conducive circumstances to do so. One might

call this ‘high sleepability’, which is distinct from needing sleep due to

obviously insufficient sleep. When healthy participants, free of sleep disorders, come to a sleep laboratory they might have an expectancy to be

somewhat sleepy, which can be likened to someone well satiated, neither

hungry nor looking for food, then taken to an attractive café and confronted with aromas of fresh bread, croissants and so on, and presented

with a choice of these items. The likelihood is that an enticing item will

be selected and soon eaten with relish. Was there an unmasked, hidden

hunger after all, or was it just eating for pleasure [4]? Besides, we can eat

simply out of boredom. Here, I see high sleepability and the decision

actually to go to sleep as a ‘conscious and voluntary’ act following appetitive sleepiness.

Thus, a normal, healthy sleeper in a sleep laboratory, who is relaxed,

maybe rather bored, with few interesting distractions and no waking pressures, will probably show that their most likely behaviour is a propensity

to fall asleep, albeit after longer than the 12-minute criterion of ‘excessive



daytime sleepiness' but, nevertheless, even within 20 minutes. My point

is that as sleepiness presumably appeared before this sleep onset, then it

was likely to have been generated by the circumstances and one’s volition

rather than be indicative of any real need for sleep.

Not only is the MSLT very sensitive to sleepiness, even to very low

levels that maybe only exist under exacting laboratory conditions, but

the same can apply to the monotonous reaction time tests I will come to.

Such levels of sleepiness can go unnoticed by and be of little concern to

people during their everyday situations, who are not prepared to change

their waking habits for more daily sleep, which would only provide minimal benefits in terms of greater waking alertness and productivity. Hence,

it could be argued that people do not really know how sleepy they are,

whereas these objective tests will indicate this.

There is a variant of the MSLT, the ‘Maintenance of Wakeful Test’

(MWT [3]), which is more real-world, where the participant sits in a

reclining chair, wired up with an EEG, as with the MSLT, but in a quiet

room, under brighter lighting, and this time is asked to ‘try and stay

awake’. The time taken to fall asleep is again measured, with the test ending if and when the participant falls asleep. Up to 40 minutes is allowed

for each session, and like the MSLT, this test is also repeated four to five

times every two hours from 10 a.m. The average time to fall asleep, in

minutes, is the overall MWT score. Nil sleep is scored as 40 minutes. As

might be expected, MWT scores are two to three times longer than those

for the MSLT.

Despite these latter scores being longer, one might expect there to be

a high correlation between the outcomes of the MSLT and MWT; that

is, people who have low scores on one will have relatively low scores on

the other, but this is not the case as the association is weak [4]. These two

tests seemingly measure two semi-independent abilities—falling asleep

and staying awake [5, 6].

One extreme example of this difference concerns those patients with

obstructive sleep apnoea syndrome, causing them to have excessive daytime sleepiness (EDS, see Sect. 9.2), and unwanted (involuntary) episodes of falling asleep in the daytime. When successfully treated their

nighttime sleep usually returns to normal, with no complaint or signs of

daytime EDS. Nevertheless, they seem to be very sleepy when returning




to the sleep clinic for a check-up, as their MSLT scores can remain very

low [6, 7], as they fall asleep quickly when instructed to ‘relax and try

and go to sleep’, as required by this test. But there is no other evidence

of sleepiness and their MWT scores are usually normal [7]. So it seems

that their many years of suffering from the disorder has left them with the

ability to fall asleep rapidly if they so wish, but with little real sleepiness.

Here, it seems that the process of instigating sleep is somewhat separate

from sleepiness itself. For them, falling asleep can be a pre-meditated

voluntary decision under much conscious control, presumably enabling

them easily to fall asleep when asked to do so during the MSLT. That is,

they seem to have ‘high sleepability’, in easily being able to fall asleep

seemingly with minimal sleepiness if they so wish, whereas they are quite

able to remain awake under ‘awake-promoting’ circumstances. Of course,

at bedtime there is the usual sleepiness, which itself also promotes sleep.

On the other hand, it seems that for many of those other people with

insomnia, they are at the other extreme, in having ‘poor sleepability’,

with their apparent inability to fall asleep.



Arguably the other most commonly used test of sleepiness is the

Psychomotor Vigilance Test (PVT) [1, 8]. It is a reaction time test

whereby participants sit facing a computer screen, in a sound-dampened,

small cubicle, with minimal visual distractions. The index finger of one’s

dominant hand rests on a button, ready to depress it in response to a digital millisecond clock that appears on the screen in random intervals of

between 5 and 12 seconds. The response stops the clock, which remains

in view for 1–2 seconds, to give the participant some feedback on their

performance. Typically, the task lasts 10 minutes. Two main scores are

usually obtained: the average reaction time of all responses less than 500

milliseconds, and the number of responses greater than 500 milliseconds,

called ‘lapses’. The longer the reaction time and the more the lapses, the

greater is the sleepiness.

The PVT is particularly sensitive to ‘lapses’, that can last up to a few

seconds, which not only consist of ‘microsleeps’, but also distractions



from the screen, despite there being little around by way of distraction.

Microsleeps are those gradually droopy eyelids and loss of responsiveness,

typically followed by a shake of the head and a return to alertness for a

minute or so until the process repeats itself. This drifting in and out of

wakefulness is also what naturally happens when we close our eyes at

bedtime to go to sleep, or rather ‘drift in and out of sleep’, when microsleeps rapidly become longer and longer, eventually coalescing to become

sustained sleep.

Under these mundane testing conditions, when people are bored or

sleepy and try to stay awake, the brain and behaviour need and seek

stimulation. In looking away from the computer screen people can miss a

signal which is recorded as a lapse. Under PVT scoring conditions maybe

this does not really matter, as a distraction is probably as good as a microsleep in terms of a lapse. However, we [9, 10] found that this ‘distractability’ is itself an important measure, as it can reveal three different aspects

of behaviour, as a result of: (1) otherwise alert boredom, merely looking

away from the screen and missing a signal; (2) an actual microsleep; or

(3) distraction due to more severe sleep loss, including attempts to seek

some stimulation as a means of trying to stay awake; this latter aspect is

covered in Sect. 11.2. All this points to the advisability of recording the

participant’s EEG as well as view their face via a camera, which together

can distinguish between these three aspects. Simply measuring reaction

times alone can miss these other important clues as to what is really going


Interestingly, reaction times do not generally slow down much with

sleepiness, as is usually thought, as between lapses individual reactions

are near normal, typically slowing by less than 100 milliseconds, whereas

there is no reaction during a lapse, of course. However, if all the responses

are averaged together, including those within lapses that might only

occur every couple of minutes or so, and compared with a dozen normal

responses, then the overall reaction time will appear significantly slower.

Although the MSLT and PVT are deemed to be objective tests of

sleepiness, again, findings between them do not always concur, especially as the PVT is ‘experimenter-paced’, unlike the ‘subject-paced’

MSLT. That is, the sudden appearance of a stimulus on a screen within

the PVT setting is computer-generated and unpredictable (within limits),




with the participant having no control over this. In contrast, if and when

the participant falls asleep during the MSLT, this is at their own pace.

Generally, experimenter-paced tasks are more sensitive to sleepiness than

are subject-paced ones. Other influential differences between the PVT

and MSLT are that the PVT involves sitting in a well-lit, small room with

the participant encouraged to do their best at responding to the stimuli

when these appear, whereas the MSLT entails the participant lying on

a comfortable bed in a darkened room and encouraged to go to sleep.

Thus, it is not surprising that these two tests produce different outcomes.


Subjective Sleepiness

Of the various sleepiness scales given to participants to rate their own

sleepiness, the most coherent and unambiguous is the 9-point Karolinska

Sleepiness Scale (KSS) [11], seen below. Earlier (Sect. 1.13), I described

the often used Stanford Sleepiness Scale (SSS) that contains too many

ambiguities. There are also alternative ‘analogue’ sleepiness scales, having a 10 cm line ‘anchored’ at either end by the terms ‘very alert’ or ‘very

sleepy’, requiring a cross to be placed at an appropriate point along the

scale, with the distance in millimetres along the scale indicating the level

of perceived sleepiness.

Karolinska Sleepiness Scale (KSS)

1. Extremely Alert

2. Very Alert

3. Alert

4. Rather Alert

5. Neither alert, nor sleepy

6. Some signs of sleepiness

7. Sleepy, but no effort to keep awake

8. Sleepy, some effort to keep awake

9. Very sleepy, great effort to keep awake—fighting sleep.



As the KSS takes less than a minute to complete, this contrasts with

the much lengthier objective measures of sleepiness, seen with the PVT

and MSLT, requiring, respectively, 10 minutes and up to 20 minutes

duration, and thus able to detect a sleepiness that might be missed by the

short duration subjective scales, or because of claims that we are apparently poor at assessing our own sleepiness, with which I disagree, for the

following reasons.

Both the MSLT and PVT are insensitive to sleepiness if they were also

to be given for only a minute, that is, for a duration comparable with

that allowed for the KSS and similar scales. Moreover, detecting sleepiness using any of these methods depends not only on the test duration,

but also on the degree of tedium that develops after one’s self-motivated

initial alertness wanes. Furthermore, the MSLT and PVT have the advantage of being given under quiet, non-distracting and relaxing settings,

whereas this is often not the case with subjective measures, when the participant might have just sat down to complete the scale, and then moved

to these other test settings. With these subjective measures there is usually

insufficient time to settle down and relax before completing the scale, and

certainly not enough time to let one’s true feelings of sleepiness become

apparent, as with the PVT for example. Remember, sleepiness feeds on

boredom and monotony. We [12] have clearly shown that people are

quite able to detect their own level of sleepiness if they are allowed five

minutes or so to settle down and relax, which is a comparable period by

which time the PVT usually begins to indicate sleepiness.


Mind over Matter

What people do and think about before and during all these tests is also

critical to the outcome, as we and others have also shown on many occasions, with the PVT also being prone to these factors. In one of our studies [13] on the effects of boredom, healthy, non-sleep-deprived volunteers

were quietly ushered into a lounge to wait alone for half an hour prior to

undergoing the PVT. There were two conditions, one having the lounge

containing dull and outdated technical magazines, with a TV showing a

tedious video of how to grade potatoes. In the other condition there were




interesting, topical magazines and amusing TV cartoons. Subsequently,

there was a marked difference in the reaction time scores between the

conditions, with many more lapses following the dull wait.

In another of our studies [14], using a more prolonged 30-minute

PVT session during the mid-afternoon and with sleepy participants, two

comparable groups were both given decaffeinated coffee. One was told

that it was regular coffee with caffeine to keep them awake, and the other

group was told that it was ‘only decaffeinated coffee’. Performance was

markedly better under the former condition, despite being caffeine-free.

On the other hand, if moderately sleepy people are told that they

will receive an attractive monetary reward for avoiding lapses during a

10-minute PVT, then there are no lapses, unlike under the normal conditions [15]. However, by extending the PVT, here, to 30 minutes even the

determination of the rewarded group to remain alert, fails towards the

end, despite their best efforts to perform well.

Money can also influence the MSLT [16], as we have also seen with

healthy, only moderately sleepy participants undergoing the usual MSLT,

with half of them unexpectedly told that they will receive a financial

reward for falling asleep as soon as they can. Accordingly, they get more

comfortable, adopt their more idiosyncratic sleeping position, settle

down more rapidly, fluff up the pillows and fall asleep around five minutes faster than when told nothing beyond the standard instruction of

‘relax and try and go to sleep’.

Thus the attitude of the participant to the testing procedure is so very

important, especially for the MSLT, as this test is often seen to be a ‘physiological’ and ‘clinical’ index of sleepiness, and especially as the EEG is

also being measured, unlike the usual, seemingly more ‘psychological’

PVT.  Yet, as the PVT is often also referred to as a ‘neurobehavioural’

measure of sleepiness, this also seems to make it more physiological and

clinical. Nevertheless, apart from sleepiness, and as we have seen, both

measures clearly show the influence of emotions, alertness and general

‘state of mind’.

Of course, most good experimental studies of sleepiness fully appreciate all these extraneous variables, and are careful to adhere to strict,

standard procedures. But even then participants can ‘outwit’ their experimenters, as we have seen ourselves with the PVT where, despite being



undertaken in a small soundproof, visually sterile cubicle, designed to

avoid all unwanted distractions, participants will even contrive ‘games’ to

play in order to offset this tedium, such as making words out of the first

row of the keyboard, working out the volume of the cubicle by counting

the soundproof tiles, and by doing sums in their head. Such diversions,

resulting in missed signals and the appearance of lapses, can be seen with

non-sleep-deprived, alert but very bored individuals, who might thus be

viewed to have ‘hidden sleepiness’ of which they were seemingly unaware,

as they had reported subjectively, on the KSS, for example, that they were

quite alert.


Twice as Sleepy or Half Alert?

Overall, and as I mentioned, the average MSLT score for normal sleepers

is about 12–15 minutes. As those who do not fall asleep in a session are

usually assigned a score of 20 minutes, in effect, this assumes that they

might well have fallen asleep during the 21st minute had the test continued for another minute, thereby causing a ‘ceiling effect’ that skews these

data. Various methods have been introduced to try and deal with this

problem, but the real difficulty, here, is what exactly do these scores mean

in terms of level of sleepiness? For example, 5-minute improvements in

MSLT score from 5 to 10 minutes or from 10 to 15 minutes and from

15 to 20 minutes are numerically equal in terms of changes in sleepiness.

But, put differently, someone with a MSLT of 5 minutes is apparently

twice as sleepy as someone with a score of 10 minutes who in turn is twice

as sleepy as someone with a MSLT of 20 minutes. However, there is no

real basis on which to make this latter claim, and few would actually do

so, but, in effect, this is how this linear (equidistant) time scale is used in

statistical terms.

The same argument applies to the PVT, which also produces data on a

linear scale, reflecting what is probably not a linear process—does a doubling of lapses indicate a doubling of sleepiness? And few would suggest

that we could attribute ‘twice as sleepy’ in terms of a doubling of reaction

times. A related issue mentioned in Sect. 6.5, is whether two nights of

total sleep loss, compared to one night, will double sleepiness, irrespective




of PVT or MSLT findings? Obviously, conceptualising sleepiness in this

manner makes as little sense as would suggesting that hunger can be ‘doubled’ by going without two meals instead of one. Nevertheless, we do use

the language of quantity to describe sleepiness, which is another reason

why language based, subjective scales such as the KSS do have a key role

to play and, in their own way, are just as important and meaningful as

the objective measures.

There are some similar conundrums relating to sleepiness that can be

illustrated by the MSLT. The first is based on the widely acknowledged

average MSLT score being around 12–15 minutes for healthy, good

sleeping adults. It implies that about half the healthy population will fall

asleep faster than this, and suggests that a few minutes or so before falling

asleep they would have been sleepy—sufficiently so to fall asleep. If they

had felt alert prior to entering the MSLT bedroom, then this again shows

that sleepiness can be ‘unmasked’, or maybe even generated in ostensibly

alert people, if they so wish. Of course, it might be argued that this average MSLT latency includes a substantial proportion of the population

with sleep debt, as they fail to attain a potentially desirable much longer

MSLT score, indicative of what might be viewed as ‘full alertness’. But

there is probably little to be gained by asking the participant who has

fallen asleep during the MSLT how sleepy they were prior to dropping

off, as we know that falling asleep itself clouds such a judgement, as will

be seen in the next section.

Another aspect of sleepiness, relating to the normal distribution of

sleep and sleepiness, is as follows. If healthy, fully sleep-satiated people

have their night sleep restricted by a fixed amount, either in absolute or

proportional terms, then due to the normal distribution between people,

some will naturally have shorter MSLT scores than others, despite the

same sleep loss. Alternatively, if the normal sleep lengths for a variety

of healthy sleepers of similar age and sex are ranked so that their MSLT

scores are identical, then presumably there will be a normal distribution

in these sleep durations. A similar finding may well be seen with other

measures of sleepiness despite these constant MSLT scores. These various

normal distributions may well belie more interesting interpretations, not

only in terms of the relationships between sleep durations and MSLT

scores in healthy good sleepers, but have wider implications.

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

2 Falling Asleep Versus Staying Awake

Tải bản đầy đủ ngay(0 tr)