A paralyzed Massachusetts woman picked up a bottle of coffee and sipped from it by moving a robotic arm with her thoughts, researchers reported Wednesday — the latest advance in the race to restore movement to people who have lost control of their muscles.

The moment marked the first time in 15 years the 58-year-old, who had suffered a stroke, was able to pick up anything of her own volition.

Researchers called the advance the first demonstration of reaching and grasping by a brain-controlled prosthetic arm. In recent years, other paralyzed patients have high-fived with a different robotic arm and moved a cursor around a computer screen simply by thinking about it.

The scientists involved said it will be years before such devices are widely available, but they hailed the advance as a milestone.

"Things in this field are exploding right now," said Andrew Schwartz, who is developing another thought-controlled robotic arm at the University of Pittsburgh, and who was not involved in the work reported Wednesday.

In the study, researchers implanted an electrode chip — the size of a baby aspirin — into the brains of two patients. Both had suffered strokes in their brainstems that left them in a "locked-in" state. While their brains worked normally, connections to the muscles below had been severed, leaving them quadriplegic and unable to speak.

Placed onto a sliver of brain that controls movement — the motor cortex — the chip listened to signals generated by the patients' brain cells as they thought about moving their own arms. A computer read that signal, interpreted it and sent movement messages to the robotic arm.

"I just imagined moving my own arm, and the (robotic) arm moved where I wanted it to go," the second patient, a 66-year-old man, told the researchers in response to questions submitted earlier by journalists. He can slowly communicate by moving his eyes as an assistant points to letters on a board.

A cable attached to the skull transmitted the signal.

"They're basically plugged in," said John Donoghue, a Brown University neuroscientist involved in the new work, which was reported in the journal Nature.

The trials reported Wednesday took place in 2011. In them, the patients touched and grabbed foam balls on a table with two different robotic arms. The woman succeeded about 50 percent of the time with an arm made by the company DEKA as a prosthetic for amputees. With a second arm, made by the German Aerospace Center, she succeeded about 70 percent of the time. The male patient did even better.

A team of brain scientists, engineers, mathematicians and computer scientists at Brown University, Harvard Medical School and elsewhere worked on the brain-computer interface — called BrainGate — for more than a decade. The National Institutes of Health and the Department of Veterans Affairs funded much of the work.

The key breakthrough: programming computers to interpret the brain signals as arm and hand motion.

The researchers did so by training the computers — and the patients. With the brain interface plugged in, the researchers moved the robotic arm and told the patients to imagine moving their own arm in the same way. The computer recorded the brain signals generated. Later, when the patients were asked to move the arm, the computer responded by tapping into the databank of brain signals it had stored.

Despite being paralyzed for 15 years, the woman's brain still generated signals corresponding to hand and arm movement.

"This means the motor cortex continues to operate in what appears to be a normal way," Donoghue said. "That allows us to tap into them even years after some event like a stroke or spinal cord injury."