Human outperform mouse Purkinje cells in dendritic complexity and computational capacity

Abstract

Purkinje cells (PC) of the cerebellum are amongst the largest neurons of the brain and have been extensively investigated in rodents. However, their morphological and physiological properties in humans are still poorly understood. Here, we have taken advantage of high-resolution morphological reconstructions and of unique electrophysiological recordings of human PCs ex vivo to generate computational models and estimate computational capacity. An inter-species comparison showed that human PCs had similar fractal structure but were bigger than mouse PCs. Consequently, given a similar spine density (2/m), human PCs hosted about 5 times more dendritic spines. Moreover, human had higher dendritic complexity than mouse PCs and usually emitted 2-3 main dendritic trunks instead than 1. Intrinsic electroresponsiveness was similar in the two species but model simulations revealed that the dendrites generated ~6.5 times (n=51 vs. n=8) more combinations of independent input patterns in human than mouse PCs leading to an exponential 2n increase in Shannon information. Thus, while during evolution human PCs maintained similar patterns of spike discharge as in rodents, they developed more complex dendrites enhancing computational capacity up to the limit of 10 billion times.