University of Utah researchers are among collaborators at numerous universities, hospitals and companies who will work together to develop an artificial arm that looks, feels and works like a natural arm — controlled by thought and able to move naturally, clear to the fingertips, while sensing its environment.

The Department of Defense, which is funding the project for up to $55 million, wants a device with the properties of a natural arm, ready for clinical testing in 3 1/2 years. The impetus is replacing amputated arms of soldiers wounded in battle.

The U.'s contribution centers on communication between an artificial limb and what's left of the original arm. They're developing a peripheral nerve interface, a key task that could bring up to $10.3 million of the grant to the U. and its own subcontractors.

They're not starting from scratch.

"If we didn't have a head start on these issues, we wouldn't be able to have a dream of this," says Greg Clark, principal investigator for the U.'s piece of the project and a bioengineering associate professor.

"The teams tasked with different parts of the project were all selected in part for their previous advances in various areas that can be exploited to build the dream prosthetic arm. Current prosthetics have not yet managed articulated fingers."

The U. team is modifying an existing interface device, the Utah Electrode Array, a pill-size device developed and refined over the past 15 years by its inventor, Richard Normann, bioengineering professor, and colleagues. University researchers will use a "slant" electrode array that has tiny electrodes of varying lengths, since they must reach nerve fibers at different depths in the user's residual arm, says Clark.

Only because so much relevant work, like creation of the electrode array, has been done is the ambitious project feasible with its tight timetable, Clark says.

The electrode array goes directly into the nerves in the residual limb, where each electrode either talks or listens to specific nerve fibers. The human body determines the language any prosthetic will have to use.

"I can wire tap and extract information from the nervous system, but I can't change it," says Clark.

For the arm to work on neural signals, it will have to be built based on, and responding to, those signals.

More than $30 million of the funding will be used for phase one, says the project sponsor, the Defense Advanced Research Projects Agency.