, 1999). In some experiments, we also revealed neighboring GPe neurons by immunoreactivity for human neuronal protein HuC/HuD (HuCD). Neurochemical verification was performed by assessing Selleck CB-839 immunofluorescence in single-plane confocal images. A neuron was classified as not expressing the tested molecular marker only when positive immunoreactivity could be observed in other cells on the same focal plane as the tested neuron. To visualize the somatodendritic and axonal architecture of identified neurons using brightfield microscopy, we then revealed the neurobiotin tracer with a permanent reaction product, Ni-DAB (Magill et al., 2001 and Sadek et al., 2007). When targets of GPe

neurons were to be identified, sections not containing Ni-DAB-labeled somata were further processed by the “peroxidase-anti-peroxidase” method to reveal other neurons expressing PV, nitric oxide synthase, or ChAT with a DAB reaction product (Bevan et al., 1998). Reconstructions were performed blind to electrophysiological phenotype. Five identified GP-TI neurons and five GP-TA neurons (cells #1–10; see Figures learn more 3 and 4) were traced in three dimensions using Neurolucida software (MBF Bioscience) (Sadek et al., 2007). Morphometric analyses were carried out using Neurolucida Explorer

software (MBF Bioscience). Electron microscopy was carried out according to standard protocols (Sadek et al., 2007), and was only

performed if just one GP-TA neuron was juxtacellularly labeled in the brain. After examination in the light microscope, pieces of striatal tissue containing the axonal arborizations of GP-TA neurons were dissected out. Serial ultrathin sections (∼50 nm) were cut and, for labeled axon terminals forming synapses, images of serial sections (up to 10) were recorded. The striatal structures postsynaptic to GPe axon terminals (i.e., dendritic shafts or spines) were characterized. Spines were identified on the basis of their emergence from a dendritic shaft, their Sodium butyrate relatively small size, the absence of mitochondria, and/or the presence of spine apparatus. The classification of GPe units recorded during slow-wave activity as either “GP-TI” or “GP-TA” was performed by computing the “activity histogram” of single-unit activity with respect to the cortical slow (∼1 Hz) oscillation (Mallet et al., 2008a). The coefficient of variation of the interspike interval (CVisi) was calculated as an indicator of firing regularity. Linear phase histograms were used to examine the temporal relationships between the firing of identified GPe neurons and cortical beta oscillations (Mallet et al., 2008a). Modulations of unit activity in time with cortical beta oscillations were tested for significance using Rayleigh’s Uniformity Test (Oriana; Kovach Computing).