Recently, investigators have used so-called implicit memory tests to investigate 'hearing' during general anaesthesia. Some of them were able to demonstrate auditory perception in anaesthetized patients, but others have failed to find any evidence for intraoperative stimulus registration. We argue that differences in anaesthetic cocktails and memory tests can explain why some researchers have produced positive results, supportive of information-processing under anaesthesia, and others have not. It appears that stimulus registration can occur during surgery, regardless of anaesthetic technique. However, processing of complex information may only be possible during nitrous oxide anaesthesia.