The cerebral cortex (neocortex) is the site where all sensory information is integrated to generate complex behaviors and sophisticated cognitive functions. This is accomplished through the concerted, synchronous, and often rhythmic activity of intertwined cortical networks formed by highly heterogeneous neuronal populations. In particular, locally-projecting, inhibitory interneurons (using GABA as neurotransmitter) encompass a vast number of cell subclasses, and this rich cellular diversity results in an efficient division of labor dictating virtually all cortical activities. Importantly, some interneurons are specialized in targeting dendrites, whereas others, known as basket cells, innervate the perisomatic region of cortical principal cells. Whereas dendrite-targeting interneurons are believed to control the integration of glutamatergic synapses onto principal cells, inhibitory cells innervating the perisomatic region of pyramidal neurons are responsible for their precise spike timing. Basket cells act therefore as precise ‘metronomes’: by providing pyramidal neurons with a temporal code, they end up synchronizing large neuronal ensembles. Importantly, we have previously found that a particular basket cell subtype, the fast-spiking (FS) interneuron, is self-connected by powerful functional autapses (i.e. synapses that a neuron makes with itself). Our preliminary results support the hypothesis that these basket cells connect more strongly with themselves via autaptic contacts than with other elements of the cortical microcircuit. This proposal aims at addressing unanswered questions, such as: Why do FS interneurons auto-connect so strongly? What is their role during synchronous rhythmic activities? How do GABAergic autaptic connections integrate FS cell input-output properties?In addition we have found that GABAergic neurotransmission onto principal pyramidal neurons undergoes non-hebbian long-term plasticity, which is bi-directional (i.e. potentiating or depressing), depending on pyramidal neuron cell type and cortical layer. In the last decades, the plasticity of excitatory glutamatergic synapses has been extensively studied and proposed to be the synaptic correlate of learning and memory. In contrast, mechanisms and function of plasticity of GABAergic synapses are still poorly understood. Therefore, in a second set of experiments, we will elucidate the mechanisms underlying and the cellular players involved in bi-directional long-term plasticity of GABAergic synapses onto pyramidal neurons, and we will identify its functional relevance during cortical network activity.Results of these experiments will lead to a better understanding of the physiological properties and function of neocortical interneurons, fundamentally advancing the general knowledge of neocortical physiology underlying both normal behaviors and pathological activities.