06-Mar-2000
For individual e-mail subscriptions go to Subscriptions.
Previous issue 2000.09 | Next issue 2000.11
There have been a number of pills that were to reset our internal clock (for instance melatonin) suggesting a chemical basis for our circadian rhythms. There have been also reports that melatonin works best if one faces bright sunlight and people even used neon lights directly on the skin to speed up the clock resetting process.
Now it seems that there might actually be a scientific basis for this modern folk-wisdom. Although scientists don't quite understand how there is now strong evidence that light can directly influence expression of the clock gene in heart and kidney cells of zebra fish without any influence from the eyes.
Similar light synchronized oscillations in Clock messenger RNA concentrations have been reported from fruit flies and even mammals. The miracle genes that are suspected to be responsible for this mysterious phenomenon are called "cryptochromes". They are known to participate in light entrainment of plants (!) and fruit flies but their role as light receptors in mice is still controversial. On the other hand the evidence that they are involved somehow is quite compelling: Mutant mice with defects in the cryptochrome genes don't exhibit circadian rhythms when left in the dark. But since neither retinal rods nor cones (the cells that are responsible for our ordinary vision) are required for the light synchronization to work, it is still a mystery how the organism senses the lighting changes.
- Light Acts Directly On Organs And Cells In Culture To Set The Vertebrate Circadian Clock, David Whitmore, Nicholas S. Foulkes & Paolo Sassone-Corsi, Nature, 404, 87-91 (2000)
Their finding, that the brain resonates with its dimensional complexity to those chaotic sounds could be confirmed in the paper by Patel and Balaban: They performed state of the art measurements of the MEG (magnetic fields produced by the brain) of volunteers who listened to synthesized tone sequences ranging in complexity from pure noise and 1/f sequences to simple scales. But instead of playing just the tone sequences at different pitch values they additionally modulated the sound amplitudes with the brain-physiologically notorious 40 Hz frequency. (This frequency is currently thought to be associated with cognitive binding processes.) The researchers could show that magnetic brain wave patterns from certain brain areas would follow the pitch contour of the tone sequences. The accuracy of tracking increased as the tone sequences become more predictable. This monotonic increase of pattern matching could be interpreted as a low level recognition process that is not dependent on the characteristic features of music.
If one looks, however, at the degree to which different brain regions interact, one can see a resonance behavior in agreement with the discovery of Birbaumer et al. : The coherence is low (corresponding to a high dimensional complexity) for both noisy and monotonous tone sequences and it was high in coherence (low in dimension) for music-like sequences with 1/f like characteristics.
One might wonder if these insights into the physiological foundations will help to understand the role of music in an evolutionary context.
- Temporal Patterns Of Human Cortical Activity Reflect Tone Sequence Structure, Aniruddh D. Patel, Evan Balaban, Nature, 2 March 2000, Volume 403 No. 6773
- See also: Perception of Music and Dimensional Complexity of Brain Activity, International Journal of Bifurcations and Chaos, N. Birbaumer, W. Lutzenberger, H. Rau, G. Mayer-Kress, C. Braun, 6(2): 267-278, 1996
Baker et al. used an autoimmune model (CREAE) of multiple sclerosis on mice that induced spasms and tremor very similar to the symptoms of human patients. Then they applied "cannabinoid (CB) receptor agonism", one of the active ingredients of cannabis and found a significant amelioration of both tremor and spasticity in diseased mice.
These results are very interesting especially in connection with earlier results that indicated that high doses of the major psychoactive ingredient of cannabis can inhibit the development of CREAE in rodents. Those results have not been convincing since their effects could find an alternative explanation besides the therapeutic effect of cannabis ingredients.
The authors suggest the consideration of cannabinoids for treatment purposes also because their hydrophobic nature allows their rapid access to the central nervous system. The authors also express hope that a more selective use of the CB agonist can reduce the psychoactive side effects without reducing the therapeutic benefits.
- Cannabinoids Control Spasticity And Tremor In A Multiple Sclerosis Model, David Baker, Gareth Pryce, J. Ludovic Croxford, Peter Brown, Roger G. Pertwee, John W. Huffman & Lorna Layward, Nature, 2 March 2000, Volume 403 No. 6773
The changes in gene expression caused by the chemicals are read and displayed by computer, showing up as dots of color on the computer's screen. Potentially, the new NIEHS Microarray Center could provide safety information better and faster than do animal tests -- and would replace, augment or improve on many of them. (…)
The new NIEHS Microarray Center makes use of the ToxChip, developed at NIEHS, which contains copies, or clones, of about 2,000 of the 80,000 genes in the human body. Millions of cloned copies of each gene form a nearly invisible dot that is "arrayed" -- hence the name -- in a grid pattern on the glass slide. The center also uses an even newer microarray, called the Human ToxChip, containing clusters of each of 12,000 different cloned genes.
Toxic substances produce changes that express, or turn on and off, genes, the center scientists said, and the chips and the accompanying computer support used to read the slides, take advantage of that linkage.
Initially the new center is evaluating known toxins -- for example, chemicals that are known to cause cancer and/or mutations -- to build a library or database showing the typical genetic changes that these known poisons produce. Once they have "signature" profiles of how known toxins change genes, the scientists said, they can evaluate other chemicals for potential harm by comparing the gene changes they produce with those made by the known toxins.
A match between an expression signature or "on/off" pattern produced by an unknown compound and that from an established toxic compound would indicate a potential danger in the test compound. (…)
In addition to creating the new microarray center in its laboratories at Research Triangle Park, N.C., NIEHS plans to help other organizations -- including some of its 20 university-based environmental health centers -- establish microarray capabilities and skills. Already, visiting scientists, including several from abroad, have worked alongside the center's 12 scientists. A notice making grant support available can be found at http://www.niehs.nih.gov/dert/ma-supp.htm. (…)
If you looked carefully at a glass slide that has been turned into a ToxChip, you would see a grid of dots, each looking like a bit of dust. Each dot is a cluster of millions of cloned copies of a single gene and is precisely placed in a grid by a robot.
When genes are expressed, or turned on or off, gene messengers -- a template or mirror image of the DNA -- is made called messenger RNA to carry out the DNA's will. These can be marked to appear either red or green when lasers are shined on them. These RNA molecules search for and connect with their DNA match on the slide and are read by a scanner as a red or green dot.
Changes in the red and green dots following exposures to test chemicals and to known poisons are read and compared using computers -- a process similar to reading two scanner codes on a supermarket purchase but then comparing them, rather than adding them up.
When these match up -- when, say, there is a similarity in the pattern of green or red spots -- there will be a good chance the new substance is also a poison, or toxin, and should be studied further.
- Environmental Health Institute To Use Gene Chips To Evaluate Chemicals For Potential Harm To Humans, Science Daily 3/1/200
- Press Release, National Institute Of Environmental Health Sciences
Lobemeier describes in some detail how many enzyme chemists cannot give up their linear hammer and how that can really mess up the interpretation of their "Lineweaver-Burk" and "Scatchard" plots. They provide some nice and intuitive arguments why blind application of statistical packages needs to be replaced by some careful and context specific application of most often non-linear data analysis methods. They also provide a number of links to websites with useful commercial or free non-linear regression software.
- Linearization Plots, Time for Progress in Regression, BioMedNet, Martin L. Lobemeier, March 3, 2000 · Issue 73,
In the 80s a different conceptual approach to modeling was developed at places like the Santa Fe Institute: It is "bottom-up" or "agent-based" and has as conceptual guiding principle the rules that are followed by individual agents. One would expect that for describing realistic systems reliably one should eventually get the same results from both types of models describing the same system, unfortunately there is still a long ways to go to reach that goal.
One class of these agent based simulation models have been stimulated by the Artificial Life approach that started off as a small seminar organized by Chris Langton at the CNLS in Los Alamos. There have been a number of spin-off applications ranging from industrial sized market forecast applications (see ComDig 2000.3 for examples) and educational tools (see MIT MediaLab's Starlogo, http://www.media.mit.edu/starlogo) all the way to the best selling series of simulation computer games created by Will Wright and his team atMaxis. In his first game SimCity you can play mayor of an existing city or create a new city from scratch. You can collect taxes but you have to worry about crime rates and traffic or your population will decline as "people" move somewhere else. In SimEarth you were put in charge of running our planet and you have to figure out how to keep the oceans from evaporating and keeping a nicely balanced eco-system. In the early 90s when ants became one of the complexity showcases of academic researchers Will Wright came out with his own really funny version of SimAnt, making you the king of the anthill.
When I visited Will a few years ago he was showing me his latest project where objects themselves would interact with the agents appropriately. He was excited to demo one of his first objects, a toilet that would respond in its own toilet way when an agent approached it. Now this immensely complex artificial world is finally finished and you can buy it as The Sims. Now you cannot just watch other people's daily life in sitcom shows but you can create your own Seinfeld episodes interactively.
The power of this approach is a fascinating example of true "edutainment". One can think about using this game as a high level extension of role-playing games where students can create their own avatar characters with the most sophisticated behavioral repertoire available today.
I hope it will find also find acceptance in schools to use complex game simulations to simulate complex systems like schools.
- Will Wright: Playing God, ZDNet, Michelle C.I. Feldman
The exact nature of how a fire spreads in a building or in an airplane is highly complex and correspondingly the task to prevent or extinguish fires. For instance in the phenomenon of "sprinkler skip" the activation of one sprinkler can cool the next sprinkler to a degree that it is not activated which can cause the fire to spread faster.
The burning process itself can be highly non-uniform and can by itself create new, critical, meta stable states. The most dangerous of which is causing "flashovers", a rapid re-ignition of a fire that had died down due to lack of oxygen for instance in a room with closed windows and doors. Here two critical parameters can interact in a deadly manner (actually nearly 75% of fire fatalities occur in flashovers): The ignition point of inflammable material depends on both temperature and oxygen content of the air. Opening the door or window to a room were a fire had the flame extinguished is still smoldering can suddenly re-ignite the flames due to the sudden increase of the oxygen level.
A new "Fire Dynamic Simulator" of the Building and Fire Research Laboratory at the National Institute of Standards and Technology uses methods from computational fluid dynamics to simulate how different types of fire spread and evolve. With a spatial grid resolution of 6in the simulator is probably the most advanced and complex computational tool in this area available today (soon to be distributed via www on fire. nist.gov). As a consequence, detailed fire simulations can easily run on a fast computer for several days. In spite of this number-crunching power the simulator is still not quite there yet, to predict in a realistic situation when a flashover situation has developed. In order to move in that direction the simulator needs to be updated by sensors placed in a building that will provide real-time feedback from the spread of the fire.
Just like with the weather, burning fires display a "butterfly effect of sensitive dependence to initial conditions. Therefore the prediction accuracy will rapidly deteriorate with increased forecasting time.
- A Computer To Outsmart A Raging Fire, Mark Derr, NY Times, 3/7/00
Water vapor and clouds play a crucial part in the atmospheric branch of the earth's hydrological cycle, but they also play havoc with attempts to describe that process through computer-based mathematical models. David Randall of Colorado State joined a Feb. 20 symposium at the American Association for the Advancement of Science annual meeting in Washington, D.C., to talk about the problems and to seek solutions.
In a presentation from 3-6 p.m. with four other geophysical scientists, Randall presented no new, research-based information to an audience made up primarily of applied mathematicians. Rather, he provided some simple information on water vapor and clouds in the atmosphere and outlined some of the difficulties so-called Global Atmospheric Models have encountered.
His remarks, and questions he hoped the audience would raise, may spark a new approach to the modeling difficulties, he said. Randall likened the event to bringing together the great mathematician John von Neumann and popular television weatherman Willard Scot to discuss water vapor, clouds and how to represent them in numerical form.
"It's a very, very messy problem for a number of reasons," Randall said. "For example, water tends to be very lumpy. You have a wet spot (in the atmosphere) a few kilometers across, then a dry spot, then another wet spot, and so on.
"The motion of air in the vicinity of clouds is especially complicated," he said. "As soon as air enters a cloud its motion becomes turbulent, and that, too, is a real problem-causer for mathematical modeling."
Scientists know that water evaporates from oceans and land and may change phase (e.g., turn from water to vapor to water to ice) several times during the average eight to nine days it remains airborne before precipitating out as rain or snow. Water is crucial in energy absorption and transfer. Evaporating a kilogram of water takes about the same amount of energy as burning 25,000 100-watt light bulbs for one second.
Meanwhile, water condenses into clouds, warming up the atmosphere. Thunderstorms in particular, Randall said, move water and energy around in the troposphere, the five to 10 miles of air above the earth's surface in which weather takes place.
"Thunderstorms can carry air to the top of the troposphere in 30 minutes," he said. "A thunderstorm is like an express elevator, and thunderstorms have a tremendous effect on the flow of energy through the atmosphere."
Meanwhile, radiation--sunshine and infrared wavelengths reflected from earth--interact with clouds and water vapor. Clouds reflect sunshine back to space, cooling the planet off, but by trapping infrared, clouds can heat the planet.
These relatively simple atmospheric phenomena make modeling difficult, so much so that certain important mathematical methods used in Global Atmospheric Models have been scrapped. Randall, who enjoys the mathematics of his work, hopes the atmospheric scientists-applied mathematicians' interchange will spur new, fruitful ideas on both sides.
"Water's a nightmare," he said. "Within the atmospheric science community, there are some of us who are mathematically inclined and some who are not, and even most who are inclined to deal with modeling flee from the complications of water. It's a dirty mess, mathematically speaking, but some of us like it because it makes the problem of understanding the atmosphere much more interesting."
- Colorado State Researcher Hopes Mathematicians Can Help Solve Problems That Water Causes In Mathematical Models Of Atmosphere, Science Daily
- Press release, Colorado State University
It becomes increasingly evident that the global climate system is of a complexity that is still underestimated. New observations bring into question some textbook theories that have been taught at universities for decades. One of them is the explanation for the recurrence of ice ages, which was believed to be driven by the changing orientation and distance of the earth in relation to the sun. In the 1930s Milankovitch invented the cycles named after him and which predict an Earth-orbit driven recurrence of ice ages is some 100,000-year rhythm. Basically all of these theories were developed in Europe or the US with correspondingly data based on the climate on the Northern Hemisphere. Only recently data of comparable accuracy from the Southern Hemisphere became available. The surprising finding is that the climate variations on both hemispheres are connected in some not completely understood way.Recent findings on the nature of the "ocean conveyor belt" and its extension across the equator provided some potential mechanism for this coupling (see ComDig 2000.6.9.1). New methods of dating sediment samples with the help of the radioactive elements uranium and thorium allow a more accurate estimate of the timing of the last ice ages. According to Henderson and Slowey the estimated end point of the last big ice age was at 135,000 years ago with an error of only 2000 years. This high accuracy result has some interesting consequences for the standard theories: They imply that the end of the last ice-age happened about 1,000 to 2,500 years earlier in the Southern Hemisphere than in the Northern hemisphere. This time difference is large enough to test the theory that ice ages are caused by the variations in the Earth' relation to the sun. Now it seem that the data are consistent that orbital variations of the Earth triggered an ice age either in the Southern Hemisphere or in the tropics but not in the Northern Hemisphere.
These results seem to be more in agreement with the hypothesis that global climate change is triggered by changes in the enormous heat (and CO2 reservoir of the pacific ocean, the place that also brings the world the El Nino events every couple of years.
- Climatology: Of Ice And Elephants, Daniel P. Schrag, Nature, 404, 23 - 24 (2000)
Other answers had to do with learning and memory and there seems to be mounting evidence that there is a scientific basis for that sleep-memory connection. It also would be consistent with the longer hours of sleep of young children, since up to the age of about ten the number of neuronal connections grows in the human brain, most rapidly in the first three years. After that age some sort of pruning takes place and the neuronal connections actually decrease in number.
Physiologically, (Hebbian) learning is explained as strengthening synaptic connections between neurons that takes place every time these neurons are active. But if that were the only process then it would not take too long and our brain would be "full" like in the far-side cartoon. Therefore it is also important that we learn to forget. Some theoretical arguments, supported by artificial neuronal net simulations have suggested that sleep might indeed the time when "un-learning" of all the irrelevant bits of random information takes place that we are exposed to during our waking hours.
Recent work by Stickgold et al. confirms again that learning and memory (retention) is improved if one has a full night's sleep of more than six hours. They studied the performance of volunteers to recognize and respond to target patterns. It is interesting that their result seems to indicate that a combination of deep sleep and rem sleep ("rapid eye movement" sleep, during which most of the dreaming occurs) are required for an improved memory and learning performance. On the other hand many higher mammals (including dolphins) only sleep for brief periods at a time and there is plenty of anecdotal evidence that many geniuses were notorious for their short hours of sleep.
It would be interesting to test if there is some evidence that together with the learning of a task there is an increased un-learning of non-relevant facts associated. But that might be a difficult experiment since it is hard to define what are non-relevant facts: Such a study might trigger the formation of groups who are interested exactly in those trivial factoids. Or has this already happened?
- For Better Learning, Researchers Endorse 'Sleep on It' Adage, Sandra Blakeslee, New York Times, March 7, 2000
- Sweet Dreams Can Lead to Better Learning, Cassie Ferguson, Harvard University Gazette, 1998
Also available in: Simple HTML format | TXT format | TXT format with links | Print
© 1999 - 2011 Complexity Digest.
Copyright Statement
|
Privacy Policy
| 
| Editor-in-Chief: Carlos Gershenson
| Founding Editor: Gottfried Mayer
