Reading the minds of comatose patients.

Nature profiles Adrian Owens, a man who uses brain scans to communicate with patients in a persistent vegetative state:

Adrian Owen still gets animated when he talks about patient 23. The patient was only 24 years old when his life was devastated by a car accident. Alive but unresponsive, he had been languishing in what neurologists refer to as a vegetative state for five years, when Owen, a neuro-scientist then at the University of Cambridge, UK, and his colleagues at the University of Liège in Belgium, put him into a functional magnetic resonance imaging (fMRI) machine and started asking him questions.

Incredibly, he provided answers. A change in blood flow to certain parts of the man’s injured brain convinced Owen that patient 23 was conscious and able to communicate. It was the first time that anyone had exchanged information with someone in a vegetative state.

Patients in these states have emerged from a coma and seem awake. Some parts of their brains function, and they may be able to grind their teeth, grimace or make random eye movements. They also have sleep–wake cycles. But they show no awareness of their surroundings, and doctors have assumed that the parts of the brain needed for cognition, perception, memory and intention are fundamentally damaged. They are usually written off as lost.

It was June 2006. Wimbledon was on, and in a headline-stealing study, Owen took fMRI scans of a 23-year-old woman in a vegetative state while he asked her to imagine playing tennis and walking through the rooms of her house. When healthy, conscious adults imagine playing tennis, they consistently show activation in a region of the motor cortex called the supplementary motor area, and when they think about navigating through a house, they generate activity in the parahippocampal gyrus, right in the centre of the brain. The woman, who had been unresponsive for five months after a traffic accident, had strikingly similar brain activation patterns to healthy volunteers who were imagining these activities, proving, in Owen’s mind, that she was conscious. The result, published in a one-page article in Science5, evoked wonder and disbelief. “I got two types of e-mail. People either said ‘this is great’ or ‘how could you possibly say this woman is conscious?’,” Owen says.

Other researchers contended that the response was not a sign of consciousness, but something involuntary, like a knee-jerk reflex. Daniel Greenberg, a psychologist at the University of California, Los Angeles, suggested in a letter to Science that “the brain activity was unconsciously triggered by the last word of the instructions, which always referred to the item to be imagined”6.

But Owen went on to bolster his case. Working with neurologist and neuroscientist Steven Laureys from the University of Liège, Owen showed that of 54 patients in a vegetative or minimally conscious state, five responded in the same way as the first woman1. Four of them were in a vegetative state. After refining their methods, the researchers asked patient 23 to use that capability to answer yes-or-no questions: imagine playing tennis for yes, navigating the house for no. They then asked about things that the technicians scoring the brain scans couldn’t possibly know.

Is your father’s name Thomas? No. Is your father’s name Alexander? Yes. Do you have any brothers? Yes. Do you have any sisters? No. The experiment is no easy feat for the patient. Owen’s protocol demands patients maintain focus for 30 seconds then rest for 30 seconds, with lots of repetition.

In front of a computer screen showing the fMRI data, Owen traces a blue line indicating activity in the supplementary motor area — a ‘yes’ — as it rises during the ‘answer’ period. It dives during the rest periods. A red line — indicating activity in the parahippocampal gyrus — represents the ‘no’. The lines are sharp and clear, and Owen, who has a taste for puns, calls the implication “a no-brainer”. “You don’t need to be a functional-imaging expert to appreciate what this person is telling you,” he says. The patient answered five of six questions correctly. There was no discernible signal for the sixth.

More details from more experiments at the link.