Mobile phone use does not raise the risk of brain tumours, a Japanese study suggests.
The research is the first to look at the effects of hand set radiation levels on different parts of the brain.
Tokyo Women's Medical University found no increased risk of the three main types of brain cancer among regular mobile phone users.
The study, comparing 322 brain cancer patients and 683 healthy people, appears in British Journal of Cancer.
The cancer patients had one of the three most common types of brain tumour - glioma, meningioma or pituitary adenoma.
The researchers rated each subject according to how many years they had been using a mobile phone, and how long they spent talking on it each day.
They studied the radiation emitted from various types of mobile phone, and placed them into one of four categories relating to radiation strength.
And they also analysed how each phone was likely to affect different areas of the brain.
Lead researcher Professor Naohito Yamaguchi said: "Using our newly developed and more accurate techniques, we found no association between mobile phone use and cancer, providing more evidence to suggest they don't cause brain cancer."
Previous research on the safety of mobile phones has produced contradictory results.
However, most have suggested no association with an increased risk of cancer.
The largest study to date, involving 420,000 people, failed to find any evidence of a cancer trend even after 10 years of use.
Dr Lesley Walker, of the charity Cancer Research UK, said: "So far, studies have shown no evidence that mobile use is harmful, but we can't be completely sure about their long-term effects.
A spokesperson for the campaign group Mast Sanity said: "The results of this new study need to be added to the body of evidence. No single study can stand alone.
"Interestingly, in this research they noted an increase in glioma on the side of the head where the phone is used but put it down to reporting bias.
"Industry and government funded studies tend to do this but it is not scientific to dismiss results that are inconvenient."