People who take more risks with their health - including smoking and drinking - are less likely to develop Parkinson's disease, a study suggests.
The study links low risking taking behaviour to Parkinson's disease
A team led by University College London quizzed 212 people - half of whom had Parkinson's - and identified a risk-averse "Parkinsonian Personality".
Smokers and drinkers seemed to be protected - maybe because these habits denoted a risk-taking personality.
Parkinson's disease experts said the results should be treated with caution.
The study appeared in the Journal of Neurology, Neurosurgery and Psychiatry.
Parkinson's is a degenerative brain disorder which affects movement and becomes more common in older age.
Its causes remain unclear, but it is often associated with a shortage of brain chemical dopamine which controls connections between nerve cells.
Earlier studies suggest a link between smoking and caffeine intake and lower rates of Parkinson's.
And it has been suggested the dopamine system is protected by nicotine and caffeine.
Professor Andrew Lees, who led the research, said: "If you have never smoked you have double the risk of Parkinson's disease but we do not know the reasons for that.
"It may be that there's something protective within cigarette smoking or it may be another factor.
"Certain personality characteristics of individuals destined to develop Parkinson's disease may make them less prone to start smoking."
The study found patients with Parkinson's disease had smoked less, drank less alcohol and caffeine.
They also scored lower on sensation-seeking and risk taking behaviour, and higher on anxiety and depression than the comparison group.
This prompted the authors to suggest their may be a link between low sensation-seeking behaviour and what might be termed a "Parkinsonian personality".
They draw on earlier studies which suggest Parkinson's disease patients tend to reject hedonistic behaviour, to be scrupulous, socially withdrawn and disinclined to take risks.
They suggested it may be the risk-averse nature of the Parkinsonian personality, rather than the physical effects of smoking, caffeine and alcohol on the brain that influences disease rates.
Professor Lees acknowledged that behaviour may be being affected by something going wrong in the dopamine system long before symptoms of the disease materialised.
However, he added that there was evidence that certain personality types were more likely to develop certain diseases.
Kieran Breen, research director at the Parkinson's Disease Society, said determining cause and effect in retrospective studies such as this is extremely difficult and highly unreliable.
He said: "As dopamine levels can start to fall many years before difficulties with movement become noticeable and before a diagnosis of Parkinson's disease is made, an assumption that the personality traits outlined in the study cause the condition is flawed.
"A more likely explanation would be that declining dopamine levels prior to the onset of the motor symptoms is associated with these behaviours."
He added that research which implies that people with Parkinson's have set characteristics or lifestyles must be treated with caution.
Dr Paul Wicks, research psychologist at the Institute of Psychiatry, said the theory was interesting but that it needed to be tested in a long-term study.
He added: "Needless to say for all the pleasure they bring there are a multitude of other health risks, associated with nicotine, alcohol, and caffeine.
"I hope we won't be seeing headlines suggesting everyone start binge-drinking, chain-smoking, and knocking back espressos to ward off Parkinson's disease."