On the contrary, our brains can keep the gate slightly open. For example, we wake up more easily when we hear our own name or a particularly salient sound such as an alarm clock or a fire alarm compared to equally loud but less relevant sounds.
In research published in Current Biology, we went one step further to show that complex stimuli can not only be processed while we sleep but that this information can be used to make decisions, similarly as when we’re awake.
Our approach was simple: We built on knowledge about how the brain quickly automates complex chores. Driving a car, for example, requires integrating a lot of information at the same time, making rapid decisions and putting them into action through complex motor sequences. And you can drive all the way home without remembering anything, as we do when we say we’re on “automatic pilot.”
When we’re asleep, the brain regions critical for paying attention to or implementing instructions are deactivated, of course, which makes it impossible to start performing a task. But we wanted to see whether any processes continued in the brain after sleep onset if participants in an experiment were given an automatized task just before.
To do this, we carried out experiments in which we got participants to categorize spoken words that were separated into two categories: words that referred to animals or objects — for example “cat” or “hat,” in a first experiment; then real words like “hammer” vs. pseudo-words (words that can be pronounced but are found nowhere in the dictionary) like “fabu” in a second one.
Participants were asked to indicate the category of the word that they heard by pressing a left or right button. Once the task became more automatic, we asked them to continue to respond to the words, but they were also allowed to fall asleep. Since they were lying down in a dark room, most of them fell asleep while words were being played.
At the same time we monitored their state of vigilance thanks to EEG electrodes placed on their head. Once they were asleep, and without disturbing the flow of words they were hearing, we gave our participants new items from the same categories. The idea here was to force them to extract the meaning of the word (in the first experiment) or to check whether a word was part of the lexicon (in the second experiment) in order to be able to respond.
Of course, when asleep, participants stopped pressing buttons. So in order to check whether their brains were still responding to the words, we looked at the activity in the motor areas of the brain. Planning to press a button on your left involves your right hemisphere and vice-versa. By looking at the lateralization of brain activity in motor areas, it is possible to see whether someone is preparing a response and toward which side. Applying this method to our sleepers allowed us to show that even during sleep, their brains continued to routinely prepare for right and left responses according to the meaning of the words they were hearing.
Even more interesting, at the end of the experiment and after they woke up, participants had no memory of the words they heard during their sleep, though they recalled the words heard while they were awake very well. So not only did they process complex information while being completely asleep, but they did it unconsciously. Our work sheds new light about the brain’s ability to process information while asleep but also while being unconscious.
This study is just the beginning. Important questions have yet to be answered. If we are able to prepare for actions during sleep, why is it that we do not perform them? What kind of processing can or cannot be achieved by the sleeping brain? Can sentences or series of sentences be processed? What happens when we dream? Would these sounds be incorporated into the dream scenery?
But most importantly, our work revives that age-old fantasy of learning during our sleep. It is well known that sleep is important to consolidate previously learned information or that some basic form of learning like conditioning can take place while we are asleep. But can more complex forms of learning take place and what would be the cost in terms of what sacrifices the brain would make to do this?
Sleep is important for the brain and total sleep deprivation leads to death after about two to four weeks. Indeed, it should be borne in mind that sleep is a crucial phenomenon and universal to all animals. We proved here that sleep is not an all-or-none state, not that forcing our brain to learn and do things during the night would be ultimately beneficial in the long run.