Analog SFF, May 2010
Page 6
* * * *
Remote-Control Medicine: Waldo Lives Again
Surgeons can only do their job if they have both sensory input and close control of what is being done in an operation. Interfacing with surgical machines has become more natural, with controls that follow their hand movements, increasingly resembling the “waldoes” in Robert Heinlein's story. Initially there were problems such as depth perception, which in humans depends on having two eyes and a computational system at the back of your brain that triangulates an object's position based on the difference in images from each eye. Both internal robots and systems with multiple optic sensors offer more than one viewpoint and enable stereoscopic imaging rather than a “one-eyed” view of the situation.
A more difficult problem is that touch is sometimes as important as sight. Not only is tactile feedback crucial to avoid damaging delicate tissue, it is an essential part of diagnosis and safety. A rough or hard polyp is more strongly indicative of cancer than a soft, smooth one, and a nearby artery that would make tissue sampling a risky idea may be noticeable only if an exploring finger feels a suspicious pulsation. Until recently, surgeons had to learn to use their eyes as a substitute for tactile feedback, seeing how easily tissue compresses with an instrument or watching for pulsations transmitted from one instrument to another.
* * * *
Fig. 3. University of Nebraska mini-robot. Its body is held to the upper abdominal wall by internal magnets attracted to an external magnetic handle that can be moved along the skin to reposition the robot, allowing the surgeon different points of view and access to other parts of the abdominal cavity. (Reprinted from Surgical Clinics of North America 88 (2008) p. 1128, with permission from Elsevier Limited and author Dmitri Oleynikov MD)
* * * *
The answer to this problem lies in “haptic feedback,” transmitting sensations at the end of a probe or endoscope to the hands of the controlling surgeon. At UCLA, a piezoelectric force sensor on robotic instruments transmits changes in force to changes in voltage, which are then converted to inflation pressures in an array of 3mm balloons mounted on the surgeon's hand controls (fig. 4). Working together with the da Vinci Surgical System, researchers are beginning to combine tactile feedback with surgical instruments, and they hope that in the near future, it may be possible to develop enough touch sensation to discriminate between velvet and leather . . . or between bone and delicate blood vessels. Further miniaturization is still needed: The closely packed nerve endings on your fingertips allow you to tell whether one or two pinpoints are touching them, even if the separation is only one or two millimeters, so 3mm balloons are still too large to duplicate normal sensation.
Haptic feedback can be used not only in surgery, but also to train surgeons. Medical schools have used virtual-reality programs for several years to teach surgical anatomy and familiarize surgeons-in-training with procedures before they try them on actual patients. These have two drawbacks: visual-only input, and the fact that there is no way for the trainee to feel the virtual patient.
* * * *
Fig. 4. UCLA haptic system. The two plates are designed to interact with the surgeon's index finger and thumb, with pneumatic tubing connecting to control surfaces attached to his fingers by Velcro straps. (From a presentation at the 30th Annual International IEEE Conference, Vancouver, Canada, August 20-24, 2008, by M. Culjat et al., permission from IEEE)
* * * *
Some current training programs combine physical models (either cadavers or anesthetized animals) with virtual-reality images of the operating field. “Augmented reality” simulations, however, can only train surgeons to use currently available instruments on whatever can be modeled in the lab. Adding tactile feedback to computer-generated visual simulations is making physical models unnecessary and greatly expanding the number of situations that can be simulated. Haptic devices marketed by SensAble Technologies, for example, are used in “TraumaVision,” an orthopedic training simulator that creates a virtual patient; the company proudly boasts of its ability to simulate the “scraping sensation when moving the drill alongside the bone."
Tactile feedback is also making possible another concept dear to the hearts of science fiction writers: cyborgs. Amputees fitted with present-day artificial legs control them using their remaining muscles, so learning to walk means hours of practice, with vision and balance substituting for the foot and ankle sensations that the rest of us take for granted. Artificial hands are even more difficult to guide for the same reason, and adjusting grip pressure to pick up a peach instead of a baseball will be extremely difficult . . . at least until UCLA's artificial limb system is ready to be marketed.
Robotic exoskeletons are being developed for military and construction uses; for paraplegics, who lack sensation as well as strength, haptic feedback will be essential to allow them to respond to changes in position and to avoid accidents.
* * * *
The Brain-to-Robot Interface: Making it “Natural"
Surgeons using robotic equipment find their job is easier every year, as more user-friendly waldoes and headsets are developed. But their interaction unavoidably takes place at a human-machine interface, and while the actions and sensations may be complex, they are being monitored and controlled by hands connected to a computational system located inside the surgeon's head. Designing a computer system to direct the many small hand and arm muscles is as complex as the human arm itself, but those muscles and nerves were fully mapped by the late twentieth century. A more challenging problem is getting the system to respond to changes in position and to react to stimuli without computational delays and a lengthy training program for the user.
Brain-computer interfacing (BCI) is the next logical step, not only for control of remote devices, but also for bypassing blocked pathways, such as the spinal nerves of a paraplegic or the failing nerve-muscle connection in neurologic diseases like ALS (Lou Gehrig's disease, in which progressive muscle weakness takes away first mobility and later a patient's ability to communicate). At several institutions here and in Europe, devices are being developed to convert brain impulses into signals that control computer-driven devices.
A number of different approaches are in use for the “receiver.” The conventional means of monitoring brain activity is electroencephalography (EEG), which records brain electrical activity at different sites. The second uses infrared spectroscopy, measuring tiny changes in temperature that reflect the increase in blood flow when different areas of the motor cortex become active. Both modalities are being fine-tuned to allow high-resolution scans to monitor the areas of the brain controlling individual muscles. At New York's Wadsworth Center, BCI researchers have used electrical monitoring to allow humans to control a two-dimensional computer cursor; in Switzerland an EEG-based system has been used to steer a wheelchair along complex paths.
Amputees could operate artificial limbs using BCI, not by painfully learned techniques, but by transmitting the brain's normal control signals to a prosthesis. An intermediate option being developed in Chicago by Dr. Todd Kuiken's team involves using the remaining nerves in an amputated limb and amplifying their signals to operate artificial muscles in a prosthesis. If tactile-feedback technology can be added to send sensations back to the brain, it may revolutionize artificial limbs and return people to the workforce far more rapidly than is possible at present.
While BCI offers the possibility of independence to handicapped people, unfortunately most of these handicaps are the result of illness or injuries that make it more difficult for them to use the equipment than healthy volunteers. Nevertheless, they could expand what could be done safely by surgeons, emergency medical technicians, or rescue workers in dangerous environments, where the speed of direct brain-machine connections could be critical in stabilizing or recovering accident or disaster victims.
Funding for UCLA's tactile feedback project comes in part from the Department of Defense, as does that for most of the robotic medical systems in this country. Since t
he Vietnam era, military medicine has been at the forefront of areas like emergency care and trauma surgery, including emergency medical transport. Robotic surgery may make it possible to remove the age-old logistical problem of getting the doctor and patient together. If a surgeon can control an operation remotely from a seat in the same room, why can't he operate on a patient in another room? Or miles away?
Surgeons are expensive to train and replace. Telesurgery can let a surgeon in a field hospital operate on a wounded soldier in a MASH unit; “telementoring” allows a more experienced surgeon to guide another surgeon, or even a military medic, in the care of an injured patient.
* * * *
Fig. 5. Robot amphibian snake developed by Japan's Hirose Fukushima Lab. It moves both in water and on the ground using paddles and passive wheels; each joint is able to operate independently or under control of an operator, so that individual units can be detached and modified depending on the task. (Courtesy of Tokyo Institute of Technology and Hirose Fukushima
* * * *
At the annual “Medicine Meets Virtual Reality” conference, the focus is shifting from visualization-only systems that were originally used for training to robotics, feedback and control devices, and human-computer interfaces. The last few years have seen robotics become a standard part of the program, with da Vinci's surgical systems dominating the field and offering robotic surgery for more and more areas—most recently eye surgery and guidance systems for heart procedures. Haptic feedback was included in a number of papers presented this year, both as a way to improve surgical procedures and for training healthcare professionals from dentists to battlefield medics.
Telesurgery has been a reality since 2000, when Johns Hopkins and the National University Hospital in Singapore collaborated on transoceanic gallbladder surgery. In some remote areas of Canada local healthcare workers can get not only advice, but also surgical help from a doctor miles away. Military research is somewhat shorter-range, using unmanned aerial vehicles to take surgical instruments to where they are needed and perform surgery miles away from the surgeon. These have been tested in California and will have obvious applications on the battlefield.
* * * *
Robot Autonomy: Will Medicine Be Its Last Frontier?
What is still missing in medical robotics is the robot that makes its own decisions. Presently, robot surgery is confined to what most writers bluntly describe as “master-slave” systems. While advances in miniaturization mean that robots in the operating room will look and act very different from Jack Williamson's humanoids, there remains the question of how much autonomy medical robots may be given in the future.
A recent review of robotic surgery called for “more automation and more computer interface"—a logical approach that other industries already use in robotic systems. Both the manufacturing and domestic service industries use robots able to adapt to changes in situation, and automated monitoring systems are beginning to change the way that people with chronic medical problems are followed outside of the hospital or doctor's office. Japanese and Korean companies are developing and marketing adaptable robots for dozens of uses, with the field of “evolutionary robotics” expanding to develop algorithms and computational systems that analyze and respond to changes in orders, tasks, or conditions.
A very clear reason to change the relationship arises when the “master's” judgment is impaired. At the University of British Columbia, researchers are experimenting with “intelligent wheelchairs” programmed to avoid collisions, help navigate new areas, and remind a confused or forgetful person of social and hygienic needs. A camera-equipped computer compares the user's location and direction with a map of the facility. Speech-recognition technology allows the wheelchair to follow instructions, while a task processor edits out unsafe or incomprehensible directions from the user and can give instructions or warnings as needed.
Crossover between industries happens all the time. A welding technique developed for mining equipment finds its way to the automotive industry, chemical plant reactor insulation is adapted for space, remote-control devices can monitor and operate oil-well production systems or smelting operations. Medical technology has taken what it needed from other industries in the past, so what directions are future medical robots likely to take?
The technology transfer is likely to go both ways. Biomimetic robots now mimic snakes, fish, and four-footed mammals using mobility systems copied from biologic organisms (Boston Dynamics's BigDog robot displays its rough-ground ambulation programming on its own YouTube video). An amphibian snake robot developed in Japan (fig. 5) is able to fit into tight spaces to look for earthquake victims; why not further miniaturize it to operate in large blood vessels? Battlefields are by no means the only dangerous environments, and robots developed for military applications could make it possible to bring medical care to accident victims instead of waiting for transportation to a hospital.
Robot autonomy and adaptability are probably going to be the biggest hurdles, and not just for technical reasons. Human research in the United States faces far more regulatory obstacles than it does in other countries; more than a few projects, such as one version of the artificial liver, have been moved overseas due to the paperwork burden that comes with testing new medical technology. Ethical and legal questions involve not only patient safety, but whether decisions made according to an algorithm programmed into a computer are better or worse than those made by a surgeon with more experience. And surgeons themselves may resist the idea that part of their job can be replaced by a robot, however sophisticated the computer controlling it. (The argument that robots don't get tired is generally countered by the widely held belief that surgeons don't either.)
The advantages of smaller size, greater speed and accuracy, and faster patient recovery from surgery have made robotics a part of medical care already. But the technology being developed now may lead to medical applications that even science fiction writers would not have predicted.
Copyright © 2010 Stella Fitzgibbons, MD
* * * *
About the Author: Stella Fitzgibbons holds a master's degree in chemical engineering from MIT and worked for four years in the petrochemical industry before starting medical school. She is board certified in internal medicine and has practiced in both primary care and hospital medicine. When not writing or practicing medicine, she stays busy with piano and handbells.
[Back to Table of Contents]
Novelette: HANGING BY A THREAD by Lee Goodloe
* * * *
Illustrated by Vincent Di Fate
* * * *
Problems that are relatively simple in one environment may become exceedingly tricky in another.
* * * *
It sure looks homelike,” Amy commented to no one in particular, staring at a viewport. Fleecy clouds, Coriolis-spun into cyclonic spirals . . . one large whirl of white cloud even had a distinct eye. Elsewhere, patches of unshaded ocean glittered, an occasional point of light winking as the water surface flashed light directly toward the viewer. Aside from the mellow yellowish illumination from Gomez's Star, the misnamed “red” giant, it could be a view of Earth.
And except for the other difference, obvious on closer inspection: the complete absence of land.
"But we know better, don't we,” a man's voice said. Amy turned slightly. He looked like a spacer, one of the engineering crew that had put the entire infrastructure in place before the majority of the scientific staff was awakened. He seemed about thirty-five or so—but didn't everybody who was out of adolescence? Appearances were no clue to age.
"Yeah, I still wonder what I've gotten myself into,” she said. “And to be living down there for the next few years.” She shook her head. “The tales of that first expedition..."
"Well, we're a lot better prepared. It's a wonder they survived at all, with all they had to improvise.” He paused for a moment, then continued. “Hi, I'm Matt Simkins. Engineering and planetary support. I'm in charge of Waterstation."
"You're just up? One of the students?” Matt made them questions. She nodded. “Yes, I have to find a thesis. If I survive, I guess.” She looked again at the monitor.
Matt looked at her. “Well, caution will keep you alive. But there's a fine line between caution and fatalism. If you don't think you'll survive, you won't."
A chime interrupted them. “Ah, here we go. Elevator to Waterstation I. After you, ma'am.” He gestured with an exaggerated sweeping bow at the doors that had just opened.
Amy smiled at him, in some confusion, and entered the airlock that led to the waiting cab. Matt followed, and then the rest of those who'd been waiting. She was quiet for a moment; his comment had given her pause. She tried again with the conversation. He'd taken the seat by hers.
"Why an airlock? We don't need it here!"
"Ah, but we might need it somewhere. Suppose someone—or something—needs to get out to do maintenance on the stalk? It would be nice not to have to open the whole elevator cab to vacuum."
Amy paused. “Oh. Dumb question, huh?"
"There are no dumb questions. Only dumb mistakes.” He said the old cliché so good-naturedly that she had to laugh.
"So you've been dealing with this—this environment for a while,” she said. It was a statement.
"I've been awake for six months. Look what we built for you!” He gestured at the elevator cab around them. “And wait'll you see the station. All the comforts of home.” He grinned self-deprecatingly. “Well, I must admit. We had just a little help from the nanofabs."