Abstract : GABAergic internereurons are crucial components of the neocortical network, and the functional characterization of the neocortex has been greatly hindered by the lack of consensus regarding the way they should be classified. Interneurons differentiate from an electrophysiological, and a morphological point of view, as well as by the expression of molecular markers, and it remains debated if the combination of these features delineates separate classes, or if it defines a phenotypical continuum. During my PhD, I proposed myself to study the diversity of neocortical interneurons, while taking all these criterions into account. Patch-clamp recordings coupled to single-cell RT-PCR have been performed in mouse, on a sample of more than 300 neocortical interneurons, and the arborization of nearly 200 of them has been reconstructed in 3 dimensions. The electrophysiological, morphological and molecular phenotypes of sampled neurons have been quantified through a set of 56 parameters. Using this sample, we first undertook to characterize interneurons in layer VI, a region where they had only been seldomly described. Using an unsupervised approach relying on electrophysiological, morphological and molecular parameters, four classes of layer VI GABAergic interneurons could be identified. In addition, using immunohistochemistry on Knock-IN GAD67:GFP mice, we mapped the distributions of GABAergic interneurons expressing characteristic markers across the radial extent of layer VI, showing that distinct population accumulate in particular sublaminar regions. These analyses have been submitted for publication. We have then undertaken to characterize the diversity of GABAergic interneurons across all neocortical layers. Using unsupervised methods relying on electrophysiological and molecular properties, we showed that the classes of interneurons which were characterized in layer VI could be found again in our complete sample. However, by analyzing the separation of these groups with analytical methods, we found that, if certain classes appeared to form clearly distinct archetypes, a substantial fraction of our sample presented phenotypes which were intermediary to particular classes. Therefore, this work suggest that neocortical interneurons do segregate into several populations, but that these groups are not separate classes but are better conceived as phenotypical archetypes. These results supports a new way of looking at the the diversity of GABAergic interneurons and will be reported in a second publication.