Imprimir Republish


The alphabet of vision

Brazilian team identifies pattern of the electrical signals that conduct the information from the eyes to the brain of flies

FABIO COLOMBINIWho has ever tried to smash one of those green flies that insist in being a bother at barbecues know that it is not easy. It is because the reaction time of these insects is far shorter than it is in human beings: the fly prepares to dodge a swipe in 30 thousandths of a second, while we take at least four times more to redirect out hand and to try to hit it. Measuring the transmission of electrical impulses emitted by a neuron associated with the vision of blow-flies, a trio of physicists is beginning to unveil how information captured by the insect’s eyes manage to arrive so quickly – and preserved to the point of concerning the essential details of the environment – to the central nervous system, responsible for the command to change the inclination of the wings and to alter the direction of flight, escaping from the flyswatter.

In the University of São Paulo in São Carlos, in the interior of the state, physicist Roland Köberle submitted oriental latrine flies of the Chrysomya megacephala species to 40 minute sessions of simulated flight. In what was almost the work of a craftsman, he catches the fly in a plastic tube by the part of the body that would correspond to the shoulders and inserts microscopic electrodes into a special pair of neurons located in the insect’s head: the H1 neurons, sensitive to movements that occur in the horizontal direction and let the fly know whether it is doing a curve to the right or to the left. If the fly is flying in a straight line, these neurons trigger off electrical signals (nerve pulses) with the same frequency. When it diverts its flight to the right-hand side, for example, the frequency of the pulses emitted by the right-handed neuron increases. The same occurs with the neuron on the left-hand side if it moves in the other direction.

In each session, the fly watches a sort of ultrarapid video clip in a special monitor, in which vertical bars varying to the right or to the left every 2 thousandths of a second recreate for the insect the feeling of being in full flight. At the same time, a computer records the electrical signals that the H1 neurons trigger off as a reaction to over 1 million visual stimuli that the fly is given during the experiment. The analysis of these electrical pulses showed that they all have similar characteristics. What changes is the interval that separates one pulse from another, a sort of neural silence, which varied from 2 to 200 milliseconds. “The short intervals suggest the need for a rapid response to the visual stimuli, whereas the long ones appear when the neuron is not being stimulated”, Köberle says. Comparing this raw data, though, did not bring the researcher much information, because the variation in these intervals was very large: the longest neural silence lasted a hundred times more than the short one.

Letters and numbers
Köberle then decided to regroup these intervals, not by the unit of duration (millisecond), but by bands of duration, which generally comprised several milliseconds. To simplify, he attributed a letter of the alphabet to each one of these bands and now called (a), for example, the intervals with up to 4 milliseconds of duration, (b) those between 5 and 20 milliseconds, and so on successively. By means of calculations made in partnership with two other Brazilian physicists – Murilo Baptista, currently at the University of Potsdam, Germany, and Celso Grebogi, a professor at the University of Aberdeen, in Scotland -, Köberle found that 15 letters or less would already represent all that variety of neural silences. Better still: four different letters (a, b, c, d) would suffice, provided that they were associated with words of up to ten letters, understandable only to the neurons.

When translating the sequences of electrical pulses and silences emitted by the fly’s H1 neurons to this neural alphabet, Baptista, Grebogi and Köberle finally found some order behind the apparent confusion. The pulses and intervals repeated themselves according to patterns that would come back and appear on an increasingly smaller scale – for example, aaabbbccc, aabbcc, abc. Known as multifractal, this pattern is similar to the pattern observed in the design formed by the foam of a cream coffee disturbed by a spoon in movement, and can be described by the mathematical formulas of the Dynamic Systems Theory, better known as the Chaos Theory. The physicists also evaluated the probability of the different possible sequences grouping themselves together in words recognized only by the neurons, and they confirmed that, for reasons as yet unknown, specific sets of letters appear more than others.

“We began to identify a language that, in future, may make it possible to understand how these sequences of electrical signals and pauses are interpreted in the fly’s brain”, says Köberle, the coordinator of the study that presented these results in the Physical Review Letters of October 2006. In short, it is the first step towards understanding how the fly sees. But not just flies, but also animals with a more complex nervous system. Besides vision, this language that is beginning to be decoded may explain how the stimuli captured by the sense organs that need a rapid and reliable response reach and are interpreted by the central nervous system, like the feeling of pain when one steps on a nail, which in less than a second covers 2 meters of nerve cells to the brain, where it is interpreted. “We believe that these properties are universal”, Köberle explains. “If a network works well at a given level, nature generally reproduces it at higher levels, sometimes with adaptations.”

The Project
Exploring the fly’s neural code (nº 02/03565-4); Modality: Thematic Project; Coordinator: Roland Köberle – São Carlos Physics Institute/USP; Investment: R$ 179,742.12