THIS WEEK
Dreams of Doom help gamers learn Ewen Callaway, Chicago
SEAN GALLUP/GETTY
FREUD thought that dreams reveal unfulfilled sexual desires – but if slumbering video game players are anything to go by, their job is to help us learn new skills. “It really looks like if you’re not dreaming about it, you’re not getting better,” says Robert Stickgold, a sleep researcher at Harvard Medical School. He led one of two gaming studies that strongly suggest that dreaming and learning are intertwined. That sleep aids learning and memory is well known. Whether the specific content of dreams plays a role in this process was not. To investigate, Sidarta Ribeiro and André Pantoja of the Edmond and Lily Safra International Institute of Neuroscience of Natal in Brazil turned to the visceral, monster-filled, first-person shoot’em-up game Doom. They persuaded 22 volunteers to spend two nights in their
20 | NewScientist | 7 November 2009
recordings of play to tot up player deaths, enemies killed and secret passages discovered, as well as shooting accuracy. The seven volunteers who made the biggest gains had dream intrusion scores far higher than the seven who dreamed least or not at all about Doom. However, the eight who dreamed most about the game improved only modestly. The same pattern has been seen in studies of how a stimulus such as stress or caffeine affects learning: a little can help, too much can hinder. “If you’re too obsessed about something, you’re dreaming about blood and monsters, you can’t do well,” says Ribeiro, who presented the work at the Society for Neuroscience meeting in
laboratory wearing scalp electrodes that measure brain activity. On the first night, volunteers simply got used to the lab. On the second, before bed, they played Doom for an hour. The next morning, Ribeiro’s team roused them during rapid eye movement (REM) sleep – the stage of most dreaming – to be questioned on their dreams, and to play Doom again. To estimate how much each person was dreaming about the video game, the researchers asked volunteers to list things they “If you’re too obsessed, associated with Doom and with you’re dreaming about the sleep lab. Blood, monsters, blood and monsters, you chainsaws and shotguns topped can’t do well” the Doom list, while the lab brought beds, pillows and electrodes to mind. Ribeiro’s team Chicago last month. A further then scoured dream reports for sign that dreams may be linked these words and used them to to learning came in brain activity measure the extent that the game measurements that suggested “intruded” into the dreams. that volunteers were replaying They also rated improvement the game during REM sleep. at Doom after sleep, using It’s not just REM sleep that seems to accompany learning. At the same conference, Stickgold and his colleague Erin Wamsley reported initial results from a study in which volunteers played a virtual maze game, before and after taking a nap that was too short to involve REM. The work is still preliminary, but the pair report that four volunteers who dreamed about the game solved the maze faster after their siesta than before. Those whose dreams didn’t incorporate the maze, as well as volunteers not allowed a nap, improved little or not at all. Neither study proves that dreams cause improvement. Ribeiro admits that the people who dreamed about Doom most could simply have been those most motivated to improve. But he says dreaming and learning –I sleep to better myself– must be intimately linked. ■
Old drugs reveal unexpected new tricks USING a database of molecular shapes may help to prevent drugs’ side effects by predicting how they will interact with the body’s receptors. It might even suggest new target diseases for existing drugs. Most drugs work by binding to a receptor or other molecule on cells, triggering a biochemical change. But many bind to other receptors too, producing side effects that only show up after tests in animals or people. Thinking about how to predict these unintended targets, Brian Shoichet, a computational chemist at the University of California, San Francisco, and colleagues reasoned that if a drug’s shape resembles that of another molecule known to bind to a certain target, then the drug may bind to that target too. Using published databases of interactions between small molecules, or “ligands”, and potential targets, they used software to search for similarities in chemical structure between these ligands and 3665 drugs either in use or under development. This yielded a list of nearly 4000 previously unknown potential targets (Nature, DOI: 10.1038/nature08506). Some of these new targets might explain known side effects. The software predicted that the antinausea drug domperidone should also bind to a receptor that contributes to cardiac arrhythmia – one of the drug’s side effects. Others point to new uses: the antihistamine mebhydrolin should block a receptor involved in Alzheimer’s disease. Experiments are still needed to prove that these drugs really do bind to the predicted targets, and that this has a biological effect. However, the approach provides a way to identify potential targets, and might help drug designers find drugs that hit several targets at once, says Andrew Hopkins, a pharmacologist at the University of Dundee, UK. “It is often by not being selective that drugs exert their effects.” Bob Holmes ■