Dr Jon Simons and Dr Paul Burgess led the study at the UCL Institute of Cognitive Neuroscience. Dr Burgess said: "In our tests volunteers either thought they had imagined words which they had actually been shown or said they had seen words which in fact they had just imagined - in over 20 per cent of cases. That is quite a lot of mistakes to be making, and shows how fallible our memory is - or perhaps, how slim our grip on reality is!
"Our work has implications for the validity of witness statements and agrees with other studies that show that our mind sometimes fills in memory gaps for us, and we confuse what we imagined occurred in a situation - which is related to what we expect to happen or what usually happens - with what actually happened.
"Most of us, though, have a critical reality monitoring function so that we are able to distinguish well enough between what is real and what is imagined and our imagination does not have too great an impact on our lives - unless the reality check system breaks down such as after stroke or in cases of schizophrenia."
The study found that the areas that were activated while remembering whether an event really happened or was imagined in healthy subjects are the very same areas that are dysfunctional in people who experience hallucinations.
Dr Burgess said: "We believe that hallucinations are caused by a difficulty in discriminating information present in the outside world from information that is imagined. In schizophrenia the difficulty you have in separating reality from imagined events becomes exaggerated so some people have hallucinations and hear voices that simply aren't there." These results indicate a link between the brain areas implicated in schizophrenia and the regions that support the ability to discriminate between perceived and imagined information.
In the tests, healthy subjects were shown 96 well-known word pairs from pop culture such as 'Laurel and Hardy', 'bacon and eggs', and 'rock and roll'. The participants were asked to count the number of letters in the second word of the pair. Often the second word wasn't actually shown and the subject had to imagine the word - such as 'Laurel and ?'.
Participants were then asked which of the second words they had actually seen on screen and which ones they had only imagined. The subjects' brain activity was observed using fMRI scans while they remembered whether words had been imagined or seen on screen.
When people accurately remembered whether they had actually seen a word or just imagined it brain activity in the key areas increased - many of which are found in brain area 10, which is involved in imagination and reality checking, develops last in the brain and is twice as big in humans as in other animals. In the people who did not remember correctly, activation in brain area 10 was reduced.