How practice makes
In the August 8 Science, Reza Shadmehr reports that for several hours after a person first learns a new motor skill such as playing the violin or riding a bike, the brain is hard at work consolidating and storing the new memory.
"We tend to think that practicing a task is the most important aspect," says Shadmehr. But this "offline learning system," he says, appears to allow for the brain's short-term storage region to teach new motor skills to the long-term memory system.
Exactly how and where the brain stores motor skills has been a big gap in neuroscientists' understanding of memory, notes Shadmehr. Brain injury can impair motor memories, a condition called apraxia. But studies of such patients have not pointed to a single location for motor skills.
Shadmehr, a professor of biomedical engineering with an expertise in building robots, comes at this research indirectly. To learn how to build robot control systems, he turned to studying the brain, specifically asking how the brain learns to control the arm.
"Soon we realized something that shocked us," says Shadmehr: "While it was true that the brain learned a lot during a practice session, quite important things were happening when our subjects were not practicing a task."
For the recent studies, 16 volunteers practiced using a robotic arm to move an object. The arm was programmed with a particular force so that participants felt they were dragging a very light object through a viscous medium. The volunteers practiced first in the morning and then returned to the lab six hours later to repeat the task.
During each session, Shadmehr, along with Henry Holcomb, a radiologist at Hopkins and the University of Maryland, imaged the brains of the volunteers using positron emission tomography (PET). PET shows cerebral blood flow, a measure of neural activity.
The researchers found that during the first learning phase, volunteers' PET scans showed activity in the dorsolateral prefrontal cortex, a region toward the front of the brain responsible for temporary storage of information. During the repeat of the task six hours later, however, the memory had shifted to regions farther back in the brain that are associated with motor skills: the motor cortex, parietal cortex, and cerebellum.
"During the six hours, while the participants were going about their daily routines, their brains had reorganized the skill learned in the morning," says Shadmehr. During this process, he adds, the brain appears to transform a fragile memory into a stable one.
The time period appears to be a critical one. In research published last year in Nature, Shadmehr found that volunteers could disrupt the long-term memory of a new motor skill if, during the hours between initial learning and memory consolidation, they attempted to learn a second, similar task, such as using the robotic arm to manipulate a new object.
Do Shadmehr and Holcomb's results mean that practicing a skill once is enough, and then the brain takes over? No such luck, says Shadmehr. Even though an operation is preserved within your long-term motor memory banks, you can still fortify that neural circuitry, which will enable you to ride a bike with greater balance, dance more gracefully, play the violin more fluidly. So the old joke still holds. If you still really want to get to Carnegie Hall... --MH
Blood cells blast off
Tracing groundwater's aquatic
Chuck Sarabun and Dan Ondercin get a lot of funny looks when they go boating. Their makeshift craft consists of a 22-foot catamaran platform that is covered by computer equipment. Plastic tubing drags from the back of the boat. Glass electrodes dot port and starboard.
Sarabun, an oceanographer, and Ondercin, a physical chemist, at the Johns Hopkins Applied Physics Laboratory (APL), are leading a team of scientists who are studying where groundwater goes when it seeps into tidal rivers and estuaries. Solvents and other chemical wastes can penetrate the soil and contaminate the groundwater, which potentially can then carry these contaminants into rivers and bays, say the researchers.
The team ultimately hopes to determine whether groundwater rapidly dilutes, or if chemicals it carries remain concentrated, thereby posing a bigger threat to the health of plant and animal life. But to do that, they first must establish that they can trace groundwater, by telling it apart from the freshwater it seeps into.
During a recent early morning field test, the team rolled their research craft into Lauderick Creek, adjacent to the U.S. Army's Aberdeen Proving Ground. The creek flows into the Bush River, a tributary of the northern Chesapeake Bay.
The operation is obviously low-budget. The electrodes are the best money can buy, but milkcrates serve as chairs and computer tables. The scientists boot up their computers. Jim Velky, an electrical engineer, turns on a centrifugal force pump, which sucks up water through the tubing hanging out the back of the boat at a rate of one liter per second.
Groundwater enters a creek such as the Lauderick either by diffusing slowly along the shoreline, or by entering through gravel lenses in the soil underlying the creek, says Sarabun. But beyond that, scientists do not know much more. "People have measured groundwater [in the soil], but we don't know how it acts once it reaches an estuary," he says.
In soil, groundwater has characteristic features such as a low pH and low level of chlorophyll. By using their electrodes (which are sensors), the researchers hypothesize that they can see and map distinct regions within tributaries that bear this distinct profile.
Water sucked up by the pump now begins to flow across the electrodes. The sensors detect and measure pH, temperature, conductivity, lead, uranium, and chlorophyll count--data that appear on the computer monitor.
Today's test and field tests done earlier in the summer appear to support the team's hypothesis. They show a plume of groundwater first hugging the shore, and then gradually fanning out into the larger body of water. The scientists plan to use a mass spectrometer based at APL to screen water samples collected during the tests for noxious organic molecules. Their presence would indicate that groundwater has brought along contaminants.
The researchers say they can adapt their instrumentation to solve a number of environmental problems. In October, they planned to take their craft to Michigan, to study a heavily industrialized stretch of the St. Clair River. --MH
Jason, the sequel
While plying the Mediterranean Sea 2,000 years ago, a wooden trading ship laden with olive oil, wine, and fish sauce sank beneath 2,500 feet of water. Clay pots and vessels settled to the muddy bottom, where they remained for centuries, cloaked by murky water and unscathed except for the occasional barnacle.
This past summer, a closet-size contraption built from titanium, aluminum, and stainless steel motored over to the pottery-strewn field, reached out a shiny arm, and began scooping up the ancient vessels.
The contraption was the remote-controlled robot named Jason, which has traveled on dozens of undersea expeditions where human divers cannot survive, including a trip to the mid-Atlantic in the summer of 1996 (See the September 1996 Magazine).
Hopkins mechanical engineer Louis Whitcomb, along with Dana Yoerger and Hanumant Singh, of the Woods Hole Oceanographic Institution, designed Jason's navigation and control system--the "brains" that enable the robot to sense and maneuver around its environment. With funding from the Office of Naval Research, during the past year, the engineers boosted Jason's technical prowess.
Until recently, the robot could only be steered manually, by a pilot using a joystick in a control room. (Jason is tethered to a surface ship by fiber-optic cables.) Even the best piloting is far from precise, notes Whitcomb, and Jason's path was like a "drunken sailor's."
The robot had another limitation. Pilots had to land it before commanding it to unfold its robotic arm. Whitcomb knew that would not do for the Mediterranean expedition. Landing a two-ton robot on a field of priceless ancient pottery would not put Jason in the good graces of archaeologists.
So Whitcomb and his fellow engineers affixed a bottom-lock Doppler sonar device to the robot's underside. The coffee-can-size instrument measures the distance to the ocean bottom and records Jason's Doppler shift, which tells how fast Jason is moving toward or away from an object. This information allows Jason's control system to keep the robot on a fixed path, or to extend its arm while hovering.
The researchers put their new technology to the test this past
summer, when they traveled with Jason and a group of
archaeologists and oceanographers to explore a 20-square-mile
region of the Mediterranean. The area lay beneath what was once a
trade route between Carthage and Rome's port city of Ostia, and
turned out to contain a treasure trove of ancient wrecks, the
largest ever discovered in deep water. The oldest of the eight
ships dates from the late second or early first century B.C.|
During the expedition, which was organized by Robert Ballard, president of the Institute for Exploration, a Navy nuclear research submarine scoured the bottom for wrecks. Whenever it spotted one, it signaled back to scientists aboard a surface research ship. Navigating from a control room on the mother ship, Whitcomb and his colleagues then steered the robot in for a closer look.
During the six-week expedition, barnacle-encrusted amphorae and other artifacts that were retrieved from the sea floor almost daily won the "oohs" and "aahs" of the archaeologists, but the highlight for Whitcomb was of a higher tech variety.
"The most exciting part of the trip was one morning at 2 a.m. when we took our hands off the controls," says Whitcomb. As he and his colleagues observed on a control room monitor, they saw that their 2,200-pound "drunken sailor" had sobered up. Under its own steering system, it sailed smoothly and perfectly over a pre-programmed course.
Using its new Doppler sonar and with other instruments including cameras and a side-scanning sonar, Jason meticulously photographed the site and recorded the topography for every two centimeters of ocean bottom. "The topographic map is more detailed than any previous map in deep ocean," says Whitcomb.
Jason's journeys are not over. Next year, the scientists plan to take the robot to a field of hydrothermal vents in the Gulf of California. --MH
Toothpaste manufacturers are not about to start mining the skies, however. The newly discovered fluoride is 20,000 light-years away, in an interstellar gas cloud called Sagittarius B2. It is also in the form of a gas, hydrogen fluoride, not the element fluorine, which is used in toothpaste.
The discovery, the first sighting of fluoride in the interstellar medium, or space between stars, now gives astronomers the opportunity to see how hydrogen fluoride is produced in the interstellar gas, and thereby understand what this unusual medium is like, says Neufeld. "All stars and everything we're made of, and all the fluoride and everything else in a tube of toothpaste was at one time this interstellar gas."
The gas is a sparse, frigid medium, he explains. Millions of years ago, a slightly denser pocket of this gas collapsed under the force of gravity to form the sun and a spinning disk of material that eventually became the planets. Some of the fluorine in the interstellar medium became incorporated into these solid bodies. Thus, today, fluorine is found in soil, plants, and other materials throughout the Earth. But other fluorine atoms reacted with hydrogen molecules in the interstellar gas, to form hydrogen fluoride.
Neufeld and colleagues at Cal Tech and the Max-Planck Institute for Radio Astronomy, in Bonn, hunted for fluoride using the European Space Agency's Infrared Space Observatory satellite. Before the satellite's launch two years ago, astronomers had no means of observing hydrogen fluoride. That's because it absorbs light at a wavelength of about one two-hundredth of an inch, within the far infrared range of the electromagnetic spectrum. Such observations are not possible from Earth because our planet's atmosphere absorbs radiation at this wavelength.
The astronomers report their finding in the October 20 Astrophysical Journal Letters. --MH
Written by Melissa Hendricks
RETURN TO NOVEMBER 1997 TABLE OF CONTENTS.