Researchers hope the test will help determine which patients may recover
Patients with severe brain damage who do not appear to have signs of consciousness still seem able to learn, a Cambridge University study suggests.
Researchers tested for Pavlovian-like responses in 22 people in a persistent vegetative state by playing a noise prior to a puff of air to the eye.
Some subjects learnt to anticipate the puff of air causing the eye muscles to twitch, Nature Neuroscience reported.
The team hopes it may lead to tests to determine which patients could recover.
Study leader Dr Tristan Bekinschtein from the University of Cambridge said the consensus had been that learning to link one stimulus with another - in this case a noise and a puff of air - was dependent on explicit awareness of the association.
But the study, where up to 70 puffs of air were delivered over a 25 minute period, showed that this sort of conditioning is possible even in patients who, by all other measures, are not conscious.
Electrodes positioned by the eyes picked up whether the muscles began to respond or not.
And a control experiment doing the same tests on people under general anaesthesia did not produce the same responses, suggesting that the learning does not happen when truly unconscious, and that some patients may have some level of consciousness not apparent on traditional tests.
Sign of recovery
Dr Bekinschtein, who began the work as a PhD student in Argentina, said they also found that patients who started to anticipate the puff of air were more likely to show signs of recovery later on, in terms of increased ability to communicate.
Although the results are reported in 22 patients, some of whom were in what is termed a "minimally conscious state", the team has since used the test in many more and is planning a large clinical trial with colleagues in the US and Belgium.
"These were classic Pavlovian experiments and some patients, but not all of them, started to respond.
"They were clearly anticipating the stimulus would come, so there is some kind of perception and from the point of view of the patient who is allegedly unconscious this could have profound implications," he said.
He said an exciting aspect of the study was that the test was easy to do and could be carried out without specialist equipment.
"This is potentially a test that could be used to test the consciousness of the patient.
"Interestingly about 80% of those who showed learning had some improvement so it gives us some confidence that we can predict those who will show some level of recovery," he added.
Professor Christopher Yeo, an expert in behavioural neuroscience at University College London, said the results looked interesting although the effects seen were small.
"All the data is extremely subtle so a study in a bigger group will improve the reliability of the findings," he said.
"The only thing I would say in terms of practical relevance is the subjects who don't learn are the ones who don't improve their condition but among those who do learn there is even distribution between those who do and do not improve."