Babies born as young as 25 weeks experience "true" pain, rather than merely displaying reflex reactions as previously thought, British scientists have confirmed for the first time.

Brain scans taken while premature babies were having blood tests registered a surge of blood and oxygen in the area of the brain known to be linked to pain sensation in adults, researchers at University College London have found.

The breakthrough should lead to doctors being able to determine appropriate levels of pain relief for premature babies, something which is not currently known.

At present, 80% of 255 specialist units have no guidelines for managing pain, with a large number giving no relief for procedures such as blood tests, which babies have several times a day in intensive care. Before now, many clinicians assumed that neonatal babies - those born before 40 weeks' gestation and up to a month old - are likely to experience pain, but had no way of knowing for sure or of measuring this. Behaviour such as crying, kicking legs or scrunching up a face could indicate pain but could also indicate cold or hunger.

The research, on babies in the neonatal intensive care unit and special care baby unit at Elizabeth Garrett Anderson hospital in London, provides the first objective proof that when babies are given procedures such as heel lancing for blood tests or a chest drain inserted they do feel pain.

"We have been able to objectively stand back and measure activity in the pre-term infant brain in response to a noxious or painful stimulus. That's never been done before," said Professor Maria Fitzgerald, from UCL's department of anatomy and developmental biology, who led the research. "Anyone working with pre-term babies would feel intuitively that they're likely to feel pain - but before now we didn't know."

Prof Fitzgerald and her team attached sensors to the heads of 18 babies aged 25-45 weeks, to measure blood levels and oxygenation in the brain. The results showed that pain was processed in the somatosensory cortex, a part of the brain known to be linked to pain sensation in adults. The older the baby, the bigger the response - due to nerve cells becoming more connected together - and the faster the response, due to nerves being coated with a fat called myelin which speeds conduction. Responses were reduced when babies were asleep, but even the very youngest infant, aged 25 weeks, "demonstrated a clear and significant response", the researchers, writing in the Journal of Neuroscience, found.

Prof Fitzgerald said scientists now needed to assess if pain-relieving drugs given at current levels were working. Pain management was crucial not only to improve the babies' quality of life, but to minimise the possibility of altering connections in the brain that might increase their experience of pain in the future. Prof Fitzgerald said: "The UK has the highest rate of low birth-weight babies in western Europe, with 2.5% needing neonatal intensive care - 17,000 each year. Estimates show that in intensive care, each baby is subjected to an average of 14 procedures a day, many of which are considered by clinical staff to be painful. Furthermore, there is evidence that these repeated painful procedures ... lead to increased sensitivity to other non-painful procedures. Since pain information is transmitted ... from 25 weeks, there is the potential for pain experience to influence brain development from an early age so the brain is highly malleable at this stage of development."

A survey by the premature baby charity Bliss has found that only 20% of 255 units in the UK specialising in neonatal care have a procedure for managing pain in premature babies. A spokeswoman said: "We strongly believe there is no justification for babies to be in pain and that more attention should be paid to providing comfort and relief when painful procedures are undertaken whilst they are in neonatal care."

Professor Neil McIntosh, professor of child life and health at the University of Edinburgh, who speaks for the Royal College of Paediatrics and Child Health on the subject, described the research as "a really important piece of information".

But he stressed pain management depended on the procedure. "A large number of units give nothing for heel pricks. But with a very painful procedure like a chest drains, virtually no one would not give analgesia."