Health Scan: Noisy as a mouse

Multiple algorithms can increase ‘hit’ rate in searching for archeological sites.

By JUDY SIEGEL-ITZKOVICH
For archeologists, finding a genuine buried site buried is often a hit-andmiss affair. There may be a better way, says Dr. Lev Epelbaum of the geophysics and planetary sciences department at Tel Aviv University’s Exact Sciences Faculty.
The new technique, which can provide a tool to learn more about buried cities, roads and other objects, was revealed recently by Epelbaum in the journal Advances of Geosciences. It is based on an algorithm that unites various bits of data while erasing “noise” (irrelevancies). Some data come from radio transmitters of the type used to communicate with nuclear submarines, while others use detailed magnetic field observations. As a result, decisions on how to proceed are more easily reached.
Seven different techniques are used, he said, so if one is neutralized, one can use any or all of the other six to find suitable locations. For example, to locate objects dozens of meters deep, bone densities can be compared with the densities of the environment (including basalt and sand) in which they are found.
Existing methods for locating sites of potential archeological and geological importance can produce significant background “noise,” he said. This tool helps archeologists “see” important artifacts, civilizations and objects and sort them out in a four-dimensional chart.
It’s a challenge, as the surface is not smooth, and irrelevant underground objects can cause much “noise.” Magnetic fields change with time, so variations complicate things, said Epelbaum.
Working with TAU scientists Dr. Leonid Alperovitz and Dr. Valery Zaldev, Epelbaum succeeded in reducing such “noise.” There are so many still-undiscovered objects in Israel that his techniques will be very practical, he said. He called his technology Multi-PAM, in which PAM stands for physical archeological models.
If a pilotless plane is flown a few meters over the surface to survey broad strips of land, that alone would provide important historic and archeological data. He estimates that there are over 20,000 undiscovered archeological sites in this country. While the tool is aimed at improving archeological investigations in the Fertile Crescent, Epelbaum suggests it could also help examine more recent sites, such as old Indian burial grounds in the Americas.
NOISY AS A MOUSE
What happened to being “quiet as a mouse”? Researchers have recently shown that, rather than being silent creatures, mice emit ultrasonic calls in a variety of social contexts like those of bats, and these calls have song-like characteristics.
So if mice sing, are they born with the songs fully formed in their heads, or do they learn them from peers? This question is of great interest to scientists as, while many organisms produce regulated vocalizations, only a few species (such as humans) can actually learn these vocalizations. If it turns out that mice can indeed learn new songs, it would provide a convenient mammalian model of vocal learning. Whether or not mouse song involves learning – either through auditory imitation or behavioral feedback such as from the mother, however – is a subject of hot debate. To highlight the difficulties, two studies published recently in the open-access journal PLoS ONE have come to differing conclusions about whether mouse vocalization patterns are innate or learned.
In the first study, researchers from America’s Northeastern Ohio Universities Colleges of Medicine and Pharmacy and the MRC Institute of Hearing Research conducted a study to understand developmental changes in mouse song that would allow rodent parents to distinguish older mice from younger mice.
They found that many features changed with age. For example, the pattern of syllables became more complex.
According to lead author Jasmine Grimsley, “We concluded that the increased complexity of song suggests that mice may are capable of vocal learning, but we also recognized that other factors besides learning, such as genetically controlled neuromuscular development, might explain the increased complexity.
We conducted our study in normal hearing, CBA/CaJ mice, and we intend to use the results to understand how the brain codes the meaning of these sounds.”
The second study, a collaboration among Azabu University, the RIKEN Brain Science Institute and the Okanoya Emotional Information Project in Japan, used an experiment to test whether the vocalization patterns were more strongly influenced by genetics or environment.
The researchers used males from two mouse strains that emit different vocalizations. Males from each strain were raised in litters of the opposite strain until weaning. Vocalization patterns were recorded at 10 to 20 weeks of age, and the researchers compared vocalizations of cross-fostered mice to control mice reared by genetic parents.
Which is it – nature or nurture? It appears too early to say for sure, and we don’t yet know whether the mating songs of mice, for example, are genetically determined or learned from their parents, the researchers concluded. What is certain, however, is that even careful scientific research does not always produce straightforward answers.