Comparison of alternative designs for reducing complex neurons to equivalent cables

Reduction of the morphological complexity of actual neurons into accurate, computationally efficient surrogate models is an important problem in compu tational neuroscience. The present work explores the use of two morphoelect rotonic transformations, somatofugal voltage attenuation (AT cables) and si gnal propagation delay (DL cables), as bases for construction of electroton ically equivalent cable models of neurons. In theory, the AT and DL cables should provide more accurate lumping of membrane regions that have the same transmembrane potential than the familiar equivalent cables that are based only on somatofugal electrotonic distance (LM cables). In practice, AT and DL cables indeed provided more accurate simulations of the somatic transie nt responses produced by fully branched neuron models than LM cables. This was the case in the presence of a somatic shunt as well as when membrane re sistivity was uniform.