What the Sleeping Brain Does with Memories

(1) For much of the twentieth century, sleep was widely regarded as a passive state in which the brain simply rested. It was known that people who slept after learning tended to remember more than those who stayed awake, but the prevailing explanation was essentially negative: sleep, it was argued, merely shielded new memories from the interference of fresh experiences. Over the past few decades, however, a substantial body of research has suggested that the sleeping brain actively reorganizes what it has learned.

(2) Some of the most striking evidence comes from studies of rats. Neurons in the hippocampus, a structure essential for forming new memories, fire in distinctive sequences as an animal explores a maze. When researchers recorded these neurons during subsequent sleep, they found that the same sequences recurred, often in compressed form, as though the experience were being replayed. Many scientists now believe that such replay, which occurs chiefly during deep, slow-wave sleep, helps transfer memories from the hippocampus, where they are initially fragile, to the cortex, where they can be stored over the long term.

(3) Sleep may do more than preserve information; it may also transform it. In one experiment, participants solved a series of number problems that, unknown to them, contained a hidden shortcut. Those who slept between training and testing were considerably more likely to discover the shortcut than those who remained awake for an equivalent period. Findings of this kind suggest that sleep can help the brain extract general rules from specific experiences.

(4) Not all researchers emphasize strengthening. According to an influential hypothesis proposed by Giulio Tononi and Chiara Cirelli, connections between neurons grow progressively stronger during waking hours, and one function of sleep is to scale them back down. This process, they argue, conserves energy and frees capacity for new learning, while the connections that matter most remain relatively strong. The two accounts are not necessarily incompatible, and a central task for the field is to determine how they fit together.

(1) For much of the twentieth century, sleep was widely regarded as a passive state in which the brain simply rested. It was known that people who slept after learning tended to remember more than those who stayed awake, but the prevailing explanation was essentially negative: sleep, it was argued, merely shielded new memories from the interference of fresh experiences. Over the past few decades, however, a substantial body of research has suggested that the sleeping brain actively reorganizes what it has learned.

(2) Some of the most striking evidence comes from studies of rats. Neurons in the hippocampus, a structure essential for forming new memories, fire in distinctive sequences as an animal explores a maze. When researchers recorded these neurons during subsequent sleep, they found that the same sequences recurred, often in compressed form, as though the experience were being replayed. Many scientists now believe that such replay, which occurs chiefly during deep, slow-wave sleep, helps transfer memories from the hippocampus, where they are initially fragile, to the cortex, where they can be stored over the long term.

(3) Sleep may do more than preserve information; it may also transform it. In one experiment, participants solved a series of number problems that, unknown to them, contained a hidden shortcut. Those who slept between training and testing were considerably more likely to discover the shortcut than those who remained awake for an equivalent period. Findings of this kind suggest that sleep can help the brain extract general rules from specific experiences.

(4) Not all researchers emphasize strengthening. According to an influential hypothesis proposed by Giulio Tononi and Chiara Cirelli, connections between neurons grow progressively stronger during waking hours, and one function of sleep is to scale them back down. This process, they argue, conserves energy and frees capacity for new learning, while the connections that matter most remain relatively strong. The two accounts are not necessarily incompatible, and a central task for the field is to determine how they fit together.

(1) For much of the twentieth century, sleep was widely regarded as a passive state in which the brain simply rested. It was known that people who slept after learning tended to remember more than those who stayed awake, but the prevailing explanation was essentially negative: sleep, it was argued, merely shielded new memories from the interference of fresh experiences. Over the past few decades, however, a substantial body of research has suggested that the sleeping brain actively reorganizes what it has learned.

(2) Some of the most striking evidence comes from studies of rats. Neurons in the hippocampus, a structure essential for forming new memories, fire in distinctive sequences as an animal explores a maze. When researchers recorded these neurons during subsequent sleep, they found that the same sequences recurred, often in compressed form, as though the experience were being replayed. Many scientists now believe that such replay, which occurs chiefly during deep, slow-wave sleep, helps transfer memories from the hippocampus, where they are initially fragile, to the cortex, where they can be stored over the long term.

(3) Sleep may do more than preserve information; it may also transform it. In one experiment, participants solved a series of number problems that, unknown to them, contained a hidden shortcut. Those who slept between training and testing were considerably more likely to discover the shortcut than those who remained awake for an equivalent period. Findings of this kind suggest that sleep can help the brain extract general rules from specific experiences.

(4) Not all researchers emphasize strengthening. According to an influential hypothesis proposed by Giulio Tononi and Chiara Cirelli, connections between neurons grow progressively stronger during waking hours, and one function of sleep is to scale them back down. This process, they argue, conserves energy and frees capacity for new learning, while the connections that matter most remain relatively strong. The two accounts are not necessarily incompatible, and a central task for the field is to determine how they fit together.