Variation and selection in neural function.

In a recent TINS article 1 , Purves et al. present a critique of darwinian theories of neural development. According to these authors, neural darwinists share the common belief that early in development the nervous system contains an initial excess of neural elements, from which those elements (neurons or synapses) that are less well-suited to the existence of the organism are subsequently eliminated. Purves et al. examine the evidence cited in favor of this notion. While some observations do suggest an initial excess of neurons or synapses, Purves et al. stress the importance of other evidence, pointing towards the progressive elaboration of dendritic and axonal processes as well as an overall increase in the complexity of neuronal morphology and connectivity as major developmental trends. They conclude that these tendencies are incompatible with a regressive (and therefore, in their view, darwinian) view of development. This portrayal and critique of selectionism is both ill-conceived and misleading. It is directed primarily against theoretical views that rely strongly but not entirely on elimi-native selection, but it either misinterprets or ignores other brain theories based on variation and selection 2–4. These selectionist theories emphasize three major components: (1) the generation of variability within populations of neurons, which manifests itself structurally through cell replication and cell death, and ongoing neurite extension, retraction and synaptic remodeling, and dynamically through continuous modifications of firing patterns 5 ; (2) the interaction of the variable circuitry and firing patterns with the organism's environment 6 ; and (3) differential amplification or attenuation of the contribution of neuronal or synaptic populations to neuronal function, either by local rules of plasticity based on correlated firing, or by global changes mediated by diffuse ascending (value) systems 7. The argument put forward by Purves et al. rests on the alleged temporal separation of initial overproduction and subsequent elimination of synaptic or neuronal elements. Overproduction and elimination thus appear in their argument as two consecutive and separate stages of development. It is important to realize that most available evidence suggests that the generation of variation , the interaction with the environment and differential amplification occur simultaneously and concurrently throughout development and adult life of the organism. In basic agreement with selectionist theories 2–4,8 , the article by Purves et al.) lists ample evidence for significant variability among the brains of individuals. These authors, however, regard such variability as an epi-phenomenon of trophic mechanisms controlling the orderly …