Striatum, the main input station of basal ganglia, is a network of inhibitory medium spiny neurons (MSNs) receiving excitatory cortical and thalamic input and playing a crucial role in motor and cognitive functions. Striatal MSNs express D1 or D2 dopamine receptors and form a two-population mutually inhibitory network. To understand the role of striatum in brain function and dysfunction it is important to characterize the differences in cortical and thalamic inputs to the two types of MSNs, and in their integrative properties.

D1- and D2-MSNs recorded under ketamine anesthesia exhibit transitions between depolarized and hyperpolarized membrane potentials, known as up- and down-states, respectively. It is presumed that activity during down-states is determined by intracellular processes, whereas large membrane voltage fluctuations during up-states are a product of increased synaptic input.

We measured statistics of MSN membrane potentials in both states and used it to estimate the effective membrane time constant (τ_eff) of neurons. The significant difference in τ_eff between up- and down-states is consistent with the assumption that MSNs in up-states operate in a synaptically driven high-conductance regime. By comparing D1- and D2-MSN statistics we found that on average D1-MSNs receive stronger input.

We found that the means of membrane potentials during up-states and their variances in high-gamma band varied in a correlated manner (ρ). Using a point neuron model of MSN and a simplified representation of cortico-striatal network we found that this effect could be explained by assuming that MSNs receive correlated effective inputs.

Finally, across different MSNs we observed high variability in ρ itself. Using a simple model of MSN we show that the main determinant of variability in ρ is the diversity of synaptic weights and input correlations and that most of the intrinsic properties of MSNs have little effect. This suggests that neuronal heterogeneity among MSNs could be obscured by statistics of synaptic inputs and synaptic weights.

In summary, by analyzing in vivo recorded data we show that MSNs operate in a high-conductance regime, and that D1 cells receive either stronger or more excitatory input than D2 MSNs. Furthermore, we show evidence that this input is correlated.