"Placebo sleep" can boost your mental performance

Believing that you've had a good night's sleep can influence your mental performance, regardless of how much sleep you actually had. That's according to a new paper, by Christina Draganich and Kristi Erdal, who tricked students into thinking there's a medical technique that can establish objectively how well you slept the previous night.

Fifty students first said how well they'd slept. Next, they were wired up to measures of their brain waves, pulse and heart-rate, and half of them were told the fiction that in fact they'd had just 16.2 per cent REM sleep the previous night (below average sleep quality); the other half were told they'd had an above average night of sleep, with 28.7 per cent REM sleep. Confronted with a difficult mental arithmetic task, the students told they'd had a good night's sleep then outperformed those who were told they'd had a poor night's sleep. In contrast, their initial subjective sense of their previous night's sleep quality was not related to their performance.

It might be tempting to take from this first result the idea that we can boost our mental performance if we convince ourselves we slept well last night. However, bear in mind that the students told they'd had a good night's sleep scored 34.81 on average on the arithmetic test, whereas the average score on this test for an adult is 36. It's a shame there wasn't a baseline control condition to see how students would have performed without receiving any information about their sleep quality. Those students told they had a bad night's sleep scored 22.13 on the test. If anything then, this first result looks like a "nocebo effect": belief that last night's sleep was bad undermined performance, but being told last night's sleep was good made little difference.

A second study was similar but this time, after receiving fictitious positive or negative feedback on their previous night's sleep, dozens of students completed a range of tests: the same arithmetic task used earlier, a word association task, a measure of visual-motor processing speed, and a digit-span test of short-term memory. Another improvement from the first study is that the researcher who interacted with the participants did not know which condition they'd been allocated to. There were also two control groups – they answered questions about their sleep the previous night and then performed the tests. So they didn't receive the false "objective" feedback on their sleep quality.

For those who received it, feedback on sleep quality was correlated with performance on the arithmetic task and the word association task, with those told they'd slept well scoring higher than those told they'd had a disturbed night. The students' initial sense of how well they'd slept, before they received the objective sleep measure, was not correlated with their performance on any of the tests. This was also true for the students in the two control conditions.

This time there was some evidence of a beneficial placebo effect. Students told they'd had a good night's sleep scored an average of 51 on the word association task, whereas the adult average score on this test is 43.51. It's shame the researchers didn't provide the word association scores for the students in the control conditions for comparison. Their scores could have acted as a baseline rather than referring to published adult norms for the tests.

"We have shown that decrements in performance can be elicited when verbal instruction and technological displays convey poor sleep quality to the individual," the researchers said. "We have also shown that increments in performance can be elicited when verbal instruction and technological displays convey high-quality sleep."

These results build on a study published ten year's ago that found people with insomnia experienced less sleep-related anxiety when a real objective measure of sleep (an actigraph) showed that they tended to overestimate how long it took them to get to sleep.

Further reading

Draganich C, and Erdal K (2014). Placebo Sleep Affects Cognitive Functioning. Journal of experimental psychology. Learning, memory, and cognition PMID: 24417326