The bitter wrangle over the fate of Terri Schiavo, the severely brain-damaged woman at the centre of a political and emotional storm in the United States, is highly distressing. Her husband has fought to allow her to die; her parents have argued that she might yet recover. As Nature went to press, Schiavo's parents seemed to have exhausted their legal options, and their daughter has had her feeding tube removed.

It is not our intention to pass judgment on Schiavo's fate. But for patients with related conditions, and their families and doctors, science may have a role to play that is not being fully explored.

In the case of Schiavo, whose brain was temporarily starved of oxygen 15 years ago, medical opinion seems clear. Neurologists say that she is in a persistent vegetative state and is unable to respond to instructions. Her chances of recovery are close to zero.

But when the prognosis is less certain, research may be able to help. Some severely brain-damaged patients are described as being in a minimally conscious state. These patients are occasionally able to respond to commands and are thought to have a slightly better chance of some recovery.

In a study published last month (N. D. Schiff et al. Neurology 64, 514–523; 2005), researchers probed the brains of two such patients using functional magnetic resonance imaging while relatives read them personal stories. The researchers found activity in language networks that was similar to that of healthy individuals. Other scans have shown the brain regions that are physically damaged, but not which circuits can respond to stimuli.

The results are only preliminary: it is not clear, for example, what the patients actually experienced. But researchers could potentially distinguish a persistent vegetative state from a minimally conscious state, or even identify conditions in between. They might even find signature patterns that can help predict the likelihood of recovery.

Perhaps the biggest obstacle is the perception in the medical and research community that severely brain-damaged patients are a lost cause. This is reflected in the number of research groups scanning the brains of such patients; they can be counted on one hand.

Some researchers have to battle for permission to work with brain-damaged patients. Patients who cannot consent to research must clearly be safeguarded. But the situation could be simplified through living wills, which allow people to specify the treatment they want if they become incapacitated. These could include requests about participation in research, as some already do about organ donation.

The ethical and medical issues involved in these cases will always be excruciatingly difficult, even when the diagnoses are as clear-cut as Schiavo's. An increased motivation to tackle research in this area won't make the problems go away, but it could help to inform the difficult decisions that families and doctors are forced to take.