<<Back to Note Summary Page

<<Back to Cognition Class Home Page

---------------------------------------------------------------------------------------

SCOTT DORAN - Ph.D. candidate in cognitive neuroscience

What I will discuss is how poor sleep effects cognitive performance and how good sleep improves a couple of specific cognitive capacities. This is a difficult intersection of research fields so I need to provide some background about the basics of normal sleep.

During sleep, the brain changes its neuronal patterns of activity globally, all parts of the brain move to synchronized firing. Non-REM sleep (NREM) comprises about 75% of all our sleep while REM sleep tends to be around 20% of an average night's sleep. Sleep stages occur in a precise order unless you are narcoleptic. NREM precedes REM sleep and deep sleep occurs before light NREM sleep. REM sleep becomes longer and contains more phasic events as the night goes on. Phasic sleep will be important later but suffice to say that it is the more active portion of REM sleep.

The key here is that sleep is dynamic and can change according to your biological and or mental state. For example:

effects of staying up late, = less REM and less delta sleep

drugs, alcohol, stress on sleep states. = less REM and delta

How do these dynamic changes in your nightly sleep affect your day? Both too little sleep and too much sleep can make you moody but only too little sleep produces cognitive impairment, as measured in the laboratory. In fact, lets see just how sleepy you are.

What about our nation's communal sleepiness? How sleepy are most people?

* Thirty-one percent of all drivers have fallen asleep at the wheel at least once in their lifetime.

The National Sleep Foundation reports a conservative figure of at least 100,000 accidents and 1,500 fatalities (the equivalent of four fully loaded Boeing 747 airplanes) are due to failing asleep at the wheel. However, most states do not keep specific statistics on falling asleep at the wheel. The actual annual figures might be as high as 200,000 accidents and 5,000 fatalities (the equivalent of twelve fully loaded 747s).

* In the PBS television documentary "Sleep Alert," a Boeing 747 captain noted: "It is not unusual for me to fall asleep in the cockpit, wake up twenty minutes later and find the other two crew members totally asleep." In another report, "A Boeing 757 captain told how his forehead hit the control column on his approach to New York's Kennedy airport as the need for sleep became overwhelming."

* What might our communal sleepiness be costing us?

The direct costs of sleep disorders and sleep deprivation for 1990 alone were estimated at $15.9 billion. Indirect costs, in terms of productivity are thought to be $150 billion but this is really undocumentable. Neither of these figures takes into account the incalculable costs of suffering, family dysfunction, and the loss of human life.

James Maas is a sleep educator from Cornell University and his survey of the research finds that most people get only 6.5 - 7 hours sleep a night. 60% sleep less than 7 hours and one third of the population sleeps less than 6 hours per night. Working mothers appear to be the most severely sleep deprived.

* Optimal amounts of sleep = enough to not feel sleepy during the day.

* Not enough sleep = sleeping more on weekends, need an alarm clock every day, fall asleep (even partially) in a boring lecture (how about if I turn off the lights?).

MOST PEOPLE NEED 8 hours of sleep. Some estimates claim that before the invention of the light bulb (1879) most people slept up to 10 hours each night during the winter. You sleep enough if you can be completely awake all day without need to sleep at all, even in a boring lecture, in a dark room, or as a passenger in a car.

What are the real-world problems invoked by inadequate sleep? For the purposes of cognitive psychology we want to know how individual cognitive processes, such as those you are discussing in this class, are altered by a lack of sleep.

Lets look at some data showing how sleep deprivation effects optimum responses during a simple reaction time experiment. This data shows how two days sleep effects the only the FASTEST responses. Focusing only on fast responses is useful because sleepiness contributes to lapses and totally missing responses; lapse trials greatly increase average RTs. Variance stays about the same for these fastest of responses but you can see how the circadian effects have a bigger influence as sleep deprivation increases. What is worse is how sleep deprivation affects tasks that require sustained attention to the task and sustained performance, a more real-world situation than are the changes imposed by simple reaction times.

Now the time on task decrement- The dramatic thing here is that sustained performance shows an ACCELERATION of decreased performance as the sleep deprivation becomes greater. The graph shows how performance decreases as the task goes on (each line represents the reaction times for the first minute, the second minute, and so on. As you can see, people who are sleep deprived have slower overall reaction times and they get worse (as time on task increases) much faster.

So, these data show us that increasing sleepiness impairs performance generally in several cognitive domains (simple response speed, memory, vigilance, decision-making acuity). In these simple studies there is very little increase in variability just an overall trend to be slower.

Now we can ask, how do specific changes in sleep, that is the structure of a single night's sleep, effect cognitive performance? The paradigm I will focus on is a perceptual learning experiment called the Karni paradigm Karni developed a paradigm to test for improvements in perceptual processing. How much better can you get at quickly identifying and responding to a change in your visual environment?

TASK:

Subjects look at the fixation a mask screen comes on for 17 ms then the target screen comes up for a period between 0-400ms. Then a masking screen returns for 17ms. Subjects must report which letter occurred at fixation (to assure they keep their eyes there) AND which was the orientation of the peripheral target. The total stimulus exposure time is reduced in each of 8 experimental blocks (10ms at a time) and mastery is considered getting 80% correct.

Subjects then wait a period of time (or sleep) and are retested. Subjects tend to get faster but only after 6-8 hours and they improve their perception for targets only in the visual quadrant being tested. Look at the results comparing subjects who slept compared to those who did not.

If you look at the relationship between how long subjects slept and their improvement in performance you see that longer sleep is better. But what about the effect of each kind of sleep? Previous research for this task found that REM sleep deprivation wiped out the sleep benefit.

Stickgold looked at the correlation between performance and the amount of each sleep type found in each quartile of the night. Having more slow wave sleep (SWS) in the first quarter of the night was significantly correlated with improved performance on the Karni task as was having a lot of REM sleep in the last quarter of the night. Stickgold then multiplied these two effects to get a measure of "Throughput" Think of throughput as the amount of performance improvement that can be attributed to the different types of sleep the subject had, SWS, REM and both. SWS in the first quarter explains 49% of the subjects improvement, REM sleep in the last quarter explains 56% of the improvement. The product SWS1xREM4 explains 80% of his subjects' improvement.

This suggests that a sequential relationship of sleep types is critical for perceptual learning in this task. In the field of sleep and cognition we have hypothesis called "sequential hypothesis" which attempts to explain why the strict NREM - REM - NREM - REM pattern might be necessary for consolidating learning. It has been explored at the behavioral and the neuronal level and it appears to be critical for improved performance perhaps through neuronal plasticity. Next, I will show you a video in which Carlyle Smith explores the effects of REM sleep using experimental paradigms which are ecologically relevant.

Carlyle Smith sums up our current understanding very well. Declarative, item-related knowledge does not seem to depend so much on sleep. Full integration and usability of knowledge does require good sleep. So, today sleep scientists are asking the more precise question about sleep's role in cognition.

Historical Q: How does sleep affect learning?

Historical A: Sleep improves learning.

Modern Q: How does sleep affect specific cognitive processes?

Modern A: Sleep improves implicit process, perhaps by corticalizing new information.

IMPLICIT PROCESSING (NON-DECLARATIVE): wake may be sufficient to learn someone's name or a list of capitol cities but sleep appears essential for developing skill. Using knowledge in a context-specific way (like an exam) may benefit most from the integration of new information to a level that you cannot discuss but you "just know". For example, perceptual processing (Karni), Complex Logic task (Carlyle Smith) and motor learning appear to be impaired most by NREM sleep disruption The different sleep stages may participate in corticalizing or re-processing either different types of information meant for storage or the different sleep stages may each contribute something unique to all informational transformations.

---------------------------------------------------------------------------------------

<<Back to Note Summary Page

<<Back to Cognition Class Home Page