A robotic arm breakthrough?

by Edyta Zielinska

We at The Scientist are always on the lookout for overhyped research, and we suspect we've run across an example in an Associated Press story about a brain-controlled prosthetic arm published yesterday (December 2). The article stated that researchers in Italy had used a robotic arm, controlled via electrodes implanted in the nerves of an amputee's arm, that allowed him "to feel sensations in the artificial limb and control it with his thoughts." Patient Pierpaolo Petruzziello next to the robotic arm Image: Courtesy of Universita Campus Bio-Medico di Roma and Scuola Superiore Sant'Anna di Pisa. The feeling sensations bit would be a major breakthrough in the field, said Chet Moritz from University of Washington School of Medicine, with the potential to give patients the ability to control a robotic arm by sensation alone, rather than adjusting their movements based on where they see their arm move. But the story isn't based on a published study. The researchers, led by Paolo Maria Rossini from the University Campus Bio-Medico of Rome, say they've submitted their results to several journals, including PNAS and Science Translational Medicine, but the research was announced at a press conference held by the university before publication of the paper. The AP's account of the research is "poorly described so it's hard to have an opinion," said Moritz who was not involved in the research. Still, based on the article and the 12 page press document prepared by the university, "I don't think this is a breakthrough," he added. In the article, the amputee implanted with the electrodes reported saying he received sensations that were as good as what he could sense with his intact right arm. "I find that a little bit hard to believe," said Moritz. Actually, Rossini told The Scientist, the researchers hadn't really tested the ability of the robotic arm to transmit sensory information to the patient. Instead, they had directly stimulated the electrodes leading into the nerve, which the patientsaid he could feel. The stimulation only worked for 10 days, after which the patient was no longer able to sense, said Rossini. Indeed, he said in regard to their work on sensation, "there is not much new in respect to the previous experiments." Instead, Rossini said, the group's novel contribution was the use of signals from the brain, recorded by EEG, to augment the signals received from electrodes implanted in the patient's arm. The additional information allowed the patient to control the robotic arm with a higher accuracy, said Rossini. Moritz pointed out, though, that "without seeing the data it is hard to know if this is indeed a new combination of EEG and nerve recording." While it's highly unusual for a research team to announce its results to the press ahead of publication in a peer-reviewed journal, and most journals forbid early publicity, Rossini explained that "we could not wait anymore, there was a lot of competition and the results are so clear." Several groups, funded by Defense Advanced Research Projects Agency (DARPA) are working on integrating sensation into the control of a robotic hand, said Moritz, and according to their grant deadline, they're scheduled to announce their progress by the end of 2009. Still, he added, "the first group to publish a real scientific paper will get the credit in the long term." However, with regard to demonstrating a patient's ability to use sensory information to adjust the movements of a prosthetic arm, the research seems to have come up short. "They shouldn't be criticized for trying," said Moritz. "They just haven't gotten there yet."