Evidence that learning is consolidated during sleep

There is new scientific evidence to support the time-honored advice to students cramming for exams to get themselves a good night’s sleep after studying. Researchers who analyzed brain activity in sleeping volunteers who had learned to navigate through a computer-generated virtual town have discovered evidence that spatial memories are consolidated during deep sleep.

Also, the researchers say that they have shown for the first time that the activity level in the brain’s learning center, the hippocampus, correlates with the improvement in memory performance when the subjects are tested the next day. According to Philippe Peigneux and his colleagues, “A growing body of experimental evidence shows the influence of sleep on the consolidation of recent memory traces. The underlying hypothesis posits that the information that is acquired during wakefulness is actively altered, restructured, and strengthened during sleep.”

However, they said, exploring this consolidation process was difficult because of the complexities of both sleep and memory. For example, sleep consists of two major stages — rapid eye movement (REM) sleep and non rapid eye movement (NREM) sleep. Evidence from animal studies of learning and sleep indicated that spatial memories seem to be replayed in the hippocampus during the deep “slow wave sleep” (SWS) during the NREM sleep stage.

Peigneux and his colleagues explored the memory consolidation process by asking volunteers to learn the layout of a virtual town that the researchers adapted from a popular computer game. They then tested the subjects on their knowledge of the town by challenging them to quickly find routes through the town to specific locations.

The researchers divided the subjects into three experimental groups:

• one that underwent training in the virtual town and whose brains were scanned during testing while they were awake;
• one that underwent training and testing and whose brains were then scanned during sleep;
• one that did not undergo any training and whose brains were scanned.

The researchers measured the subjects’ brain activity using positron emission tomography (PET) to measure blood flow in the subjects’ brain regions. In PET, test subjects receive a harmless dose of radioactive tracer, and their brains are scanned as the tracer infuses through the brain. Blood flow through specific brain structures constitutes a measure of activity in that structure.

The researchers found that the first group — compared to the non-trained group — showed greater activity in their hippocampus and an adjacent learning-related region as they took the route tests, with greater activity correlated with better performance. The group scanned during sleep after testing also showed greater hippocampal brain activity during sleep, compared to the non-trained group. Importantly, when the researchers compared the hippocampal activity during SWS sleep in the trained group and the non-trained group, they found that the first group showed higher activity.

Next, the researchers tested the trained group after their sleep session and compared their performance with the brain activity measured during sleep. They found that the higher the gain in post-sleep performance, the higher had been their NREM brain activity during sleep. No such correlation was found in REM brain activity.

To ensure that they were, indeed, measuring brain activity due to spatial processing, Peigneux and his colleagues also compared brain activity data from the spatially trained group with data from a fourth group that had only taken a reaction time test. In that test, the subjects’ brains were scanned as they pressed a key as fast and accurately as possible corresponding to the location of a dot on a computer screen.

Thus, the data from the fourth group allowed the researchers to distinguish brain activity changes due to spatial processing from those due to general mental processing. The researchers’ comparisons of the brain activity in the groups confirmed that the hippocampal activity they detected was due to spatial processing.

“Our results provide critical evidence that spatial memory traces are processed during NREM sleep in humans,” wrote the scientists. “Moreover, the hippocampal activity during sleep is shown to correlate with the improvement in memory performance on the next day. To the extent of our knowledge, this effect has not yet been reported in the animal hippocampus.”