|
Brown, Dr. Glen D. Why Dendrites? A New Null Hypothesis Studies of dendritic function typically start with the null hypothesis that the dendrite's raison d'etre is to increase membrane surface area (e.g. Mel, 1999). However, this simple increased-surface-area hypothesis is not supported by results showing that there is space on dendrites not occupied by synapses at maximum packing density (e.g. Colonnier, 1968). Another proposal is that dendrites exist primarily to increase the computational power of the brain (Mel, 1994,1999; Koch, 1999). Experimental evidence is scarce, but multiple synaptic integrators can be found in a single neuron (Spencer and Kandel, 1961). In principle, dendrites could act to compartmentalize multiple types of information processing. For example, neurons seem to be wired to do selective AND-NOT operations in different dendrites. Here we consider the hypothesis that dendrites evolved to communicate over distance. Distance between neurons is widely regarded as the driving force for the evolution of axons and action potentials. However, this is less widely appreciated for dendrites. This hypothesis predicts a decreasing inter-synaptic interval from soma to distal dendrites, which is consistent with anatomical data (e.g. Colonnier, 1968). The communication-over-distance hypothesis also makes some predictions about the spatial distribution of active conductances that are similar to those made by the increased-surface-area hypothesis (see Mel, 1999). Like macroscopic features of nervous system organization (e.g. encephalization, topographic maps), dendritic arbors should conserve space (Mitchison, 1991). Thus, sub-optimal wiring may be taken as evidence against the communication-over-distance hypothesis. However, some deviation from optimal wiring may also be expected. Conductance load limits the number of inputs that can be summed on a single dendritic branch (Segev and Rall, 1998). Even making branches thicker cannot compensate for limitations imposed by thresholded active conductances needed to propagate signals over distance. Thus, extra dendrites may be necessary to preserve the neuron's ability to do coincidence detection even when fewer dendrites would result in shorter wire length. In summary, why dendrites? To communicate over distance. Why so many? To preserve neurons' ability to integrate large numbers of inputs in a short time window. Apparent exceptions to these rules will also be considered.
|