Evolution by natural selection is the only known force to build up complex adaptations. Hence it makes sense to think that the nervous system is also the result of evolution, and that it fulfills many different roles. This is not to say, however, that every aspect of it is selectively significant, let alone perfect. Evolutionary optima do not necessarily coincide with engineering ones for at least three reasons: (a) there are several entrenched state characters, so-called hang-ups from former evolution; (b) the available genetic variation may be limited (so-called constraints), and (c) in the process of selection, only relative (rather than absolute) fitness matters. Thus, having a modern mammalian neocortex may render the hippocampus obsolete from an engineering point of view, but evolution cannot go back to the drawing board and start all over again. Evolution is a process of "irrational design" (Sydney Brenner), often reminiscent of tinkering (bricolage, François Jacob). It is not yet known to what extent various aspects of the nervous system are selectively neutral or maladaptive from an engineering point of view.
Cumulative selection can produce complex adaptations, as Darwin and Wallace already recognized, but this is not always obvious. In the case of the optical structure of the eye, Darwin had no choice but to assume that natural selection would have driven the populations through a series of incremental improvements, acting on appropriate genetic variation (Darwin 1859). Nilsson and Pelger (1994) did demonstrate about 130 years later that Darwin was essentially correct. They applied computer simulations to generate eyes, and selection was based on the quality of the imaging capacity of the structure. Fortunately, biophysics is advanced enough to make this calculation, which was then used later to calculate the relative fitness of each eye. It turned out that starting with only three cell layers (a transparent, a pigment, and a sensitive layer), they could reconstruct how the eye of, say, a fish could have arisen in a surprisingly small time in evolution. Of course, they did not consider the visual analyzer, but this had not been their task in the first place.
A relevant question is, therefore, to ask how the neural structures fulfilling certain (occasionally quite complicated) styles could have evolved. Ultimately, one would like to simulate such evolution. Leigh Van Valen said that evolution was the control of development by ecology. In a less telegraphic form, one can imagine the following algorithmic process (see Figure 19.6). What are the relevant traits for a simulated evolution of nervous systems? An essential ingredient is the genetic determination of neuronal types by the following traits: number of
Genes ^ Development ^ Learning ^ Behavior
Was this article helpful?