Using the Golgi impregnation technique, Ramón y Cajal carefully documented the morphological diversity of cortical neurons. Almost a century later, the introduction of the patch clamp technique by Neher and Sakmann provided the neuroscience community with beautiful tools to probe the functional properties of the various cortical cell types.

In today's talk, I will illustrate how we recorded functional properties of cortical neurons during sensory processing in vivo and subsequently reconstructed recorded neurons to obtain dendritic architecture. Next, we used the structure-function information of individual neurons to build a 3D model of the cortical column.

This approach allowed us to assess — with single cell resolution — how functional characteristics in vivo emerge from structural properties of the cortical column.

We have applied different Golgi techniques and developed 3D metrical [Uylings & Van Pelt, 2002] and topological [Van Pelt & Uylings, 2002] methods for the study of dendritic trees. With these methods we showed different growth phases in the dendritic patterns in human frontal cortex during lifespan [Koenderink et al., 1994, 1995; Petanjek et al., 2008, 2011]. In addition we found another pattern of aging changes in normal aging than described before [De Brabander et al., 1998; Petanjek et al., 2008].

Dendritic alterations are also studied in Alzheimer\'s disease in the human dentate gyrus of the hippocampus [De Ruiter & Uylings, 1987] and the human frontal cortex [Uylings & De Brabander, 2002].

Modern computer tools make it possible to model neurons in high detail, including their fine dendritic structure. By combining various neuronal simulation tools (Python, NEURON, LFPy) and Finite-Element methods (FEM), both intracellular and extracellular responses of neurons can be modeled, and electrical experimental measurements emulated. Here, we show results from a modeling study of local-field potentials (LFPs) measured by MEAs in an acute brain slice setting when current is injected into somas of individual pyramidal neurons. The results are compared with recent example experiments from our collaborating group at the Radboud University [Bakker et al, Neural Networks 22, 1159 (2009)].