When in doubt, go to bed

Latest research suggests that sleeping on a problem and trusting to gut instinct both have a sound scientific basis. Geoff Watts investigates.

When the choice is hard or the problem defies reasoned solution, sleep on it. Leave the brain alone; leave it, so to speak, on autopilot. That's what folk wisdom has long dictated - and so now does science. A smattering of research evidence suggests that unguided cerebral activity may do a better job than the conscious reasoning of which we're all so proud.

An example. From time to time I serve on judging panels, often for science writing awards. I like to rank entries according to the overall impression they make: gut feeling, if you like. But some award organisers, attempting as they see it to be more systematic and so fairer, insist on entries being judged by marks out of ten for each of several set criteria. Adding up your scores should waft you painlessly to your personal preference. Simple. Except that I find it doesn't.

When forced to use this approach I often feel that the entry coming out on top is the "wrong" one. If the set criteria have been poorly chosen, this is no surprise. But even when the criteria are fine, the result that emerges can still leave me feeling uncomfortable. So uncomfortable that, to achieve the "right" result without upsetting the judging arrangements, I've sometimes put the process into reverse. I've ranked the entries, then allocated the points in a way that yields my intuitive pecking order. Ridiculous, of course.

I have never felt entirely happy about doing this: about elevating gut instinct over reason. But since coming across some studies by a handful of European researchers I no longer worry. I am now pretty sure I'm doing the right thing.

Two of the researchers who have led me to this happy conclusion are Ap Dijksterhuis and Loran Nordgren at the University of Amsterdam. They base their argument on the premise that there are two modes of thought: conscious and unconscious. In the former, the more familiar variety, we contemplate a particular issue or question, mentally list all the relevant information, weigh up the pros and cons, then come to a reasoned decision. Unconscious thought, by contrast, requires no focused deliberation or even attention. The answer "just comes". Because this mode of thinking requires no wilful effort, and generates conclusions that may be tricky if not impossible to explain, it tends to have the status of poor relation - even though we all rely on it.

Dijksterhuis and Nordgren argue that, far from representing an inferior form of thought, this is actually a better method of reaching certain types of decision. Conscious thought, they say, has a limited capacity for dealing simultaneously with more than a small quantity of information and is poor at weighing the relevant importance of different factors. When dealing with complicated issues it does not necessarily lead to sound choices.

In recent years they've conducted various laboratory and real-life experiments to investigate the relative performance of the two thought modes. One study was of people who'd bought things from one or other of two shops: Ikea, specialising in relatively large and complex products, mainly furniture; and Bijenkorf, a department store selling simpler articles including clothing and kitchen accessories. As shoppers left the stores they were questioned about what they'd bought, whether they'd known in advance what they wanted, and how much consideration they'd given their purchase before seeing it and buying it.

The researchers categorised their shoppers as "conscious thinkers" or "unconscious thinkers" according to the amount of conscious thought they claimed to have devoted to their choice. A few weeks later the shoppers were contacted to find out how satisfied they were with what they'd bought. As Dijksterhuis and Nordgren had predicted, with the simpler articles bought at Bijenkorf, conscious thinkers were the more satisfied; with the more complex purchases bought at Ikea, it was the unconscious thinkers who were happiest with their choices. As experiments go, this one is hardly without weaknesses. But its outcome does find support in a number of similar findings from the lab and the field.

Dijksterhuis characterises conscious thought as rule-based and precise: the sort of processing required to perform an arithmetic calculation. Unconscious thought - although it too follows rules in that it can, for example, detect recurring patterns - is less precise. But it can handle, evaluate and integrate larger amounts of sometimes conflicting data. Conscious thought makes good choices in simple matters; unconscious thought leads to better choices when the circumstances are more multifaceted.

Jan Born of the University of Lubeck has a rather different take on the importance of unconscious thought. He and his colleagues set out to test the notion that sleeping on a problem - literally - can trigger new insights. For this purpose he employed a simple but rather tedious arithmetic task. Via a computer screen, he presented subjects with successive strings of eight digits. Using a couple of simple rules they had to transform this string into a new one - only the last digit of which was of interest to the testers. Because each new digit had to be entered into the computer keyboard, the testers were able to measure their subjects' speed at the task, and its gradual increase as they became more practised.

What subjects were not told was that the last digit of any new string would always turn out to be identical to that in the second position. Anyone noticing this could save themselves a deal of work - because, as they'd been told, the testers were interested only in the identity of the last digit. The moment at which subjects achieved this insight was apparent from a step reduction in the time taken to process each string.

All subjects went through an initial training period that was long enough for them to understand the task, but not for the penny to drop about the labour saving ploy. Eight hours later they were given a further series of number strings to process. Some had stayed awake during the intervening period; others had slept. Those who had slept were twice as likely as those who had stayed awake to notice the second digit trick.

Robert Stickgold of Harvard Medical School and Center for Sleep and Cognition in Boston is familiar with both pieces of work. He sees no conflict between them, pointing out that while unconscious thought processes are clearly involved in both, whether sleep is or is not essential may depend on the type of problem to be solved or decision to be made.

He offers what he describes as a classic example: whether or not to accept a really good job in a really bad place. "If you're anal-compulsive you get out your sheet of paper and you write 'job 1' and 'job 2' and start listing the pluses and the minuses. And it never helps. Then what you do is sleep on it. You wake up in the morning and you know in your gut what the decision is."

But why should sleep - as opposed merely to a period of time for unconscious reflection - be helpful? Stickgold suspects that it is all to do with memory. Certain phases of sleep seem to provide an opportunity for the brain to decide what experiences it needs to store, and what it can reject. This involves reviewing information acquired during the day and making new associations between different portions of it. So the capacity for better decisions and fresh insights that follows sleep might be the fortuitous by-product of a process carried out by the brain for a quite different purpose.

Isn't it faintly unnerving to be so reliant on a system of mental processing of which we are completely unaware and over which we have no sway? "Yes, I suppose it can seem disconcerting," Nordgren concedes. "But not when you think of all the other things that you trust your brain to handle for you."

Stickgold agrees. "We should use our conscious effort for the things that need it. The things it does much better." There's plenty of evidence, he adds, that the rationality of many of our decisions is actually a post-hoc explanation. Although we feel we should always be able to explain ourselves, often we cannot. All rather damaging to our intellectual egos - and one reason, perhaps, why there is so little research in this area.

Either way, requests for more time or an opportunity to sleep on something can no longer be dismissed as excuses intended to put off the process of making a decision. They are that process. Thank you, brain.