Protracted withdrawal from alcohol impacts excitability and dynamical properties of neurons of the bed nucleus of stria terminalis

The juxta-capsular subdivision of bed nucleus of stria terminalis (jcBNST) is part of the extended amygdala and plays an important role in the regulation of stress and reward related behaviors. This compact nucleus contains at least 3 types of biophysically diverse populations of GABAergic neurons that play an important role in the regulation of downstream neural circuits of the central amygdala. While drug-related neuroadaptive changes in the synaptic properties of BNST and other extended amygdala neurons have been well documented, their integrative properties and circuit interactions are also to be studied in order to gain a better understanding of the neurophysiological aspects of addictive behavior. Here we performed a series of dynamic clamp experiments with synthetic presynaptic voltage waveforms that were transformed to simulated excitatory and inhibitory synaptic conductances and injected into jcBNST neurons. This methodological approach allowed us to characterize how the temporal structure of the firing responses and the precision of spike timing depended on the type of neuron and how these properties were modified after prolonged withdrawal in alcohol dependent animals. Additionally, we performed a comprehensive analysis of neuronal responses under standard current clamp stimulation and used up to 12 characteristics of the neurons voltage output to compare their physiological properties, including resting membrane potential, membrane resistance, rehobase, the occurrence and intensity of post inhibitory response spikes, among others. In order to compare this complex set of physiological and dynamical parameters of jcBNST neurons from rats with a history of prolonged withdrawal and controls, we used the fuzzy K-means clustering algorithm, an unsupervised learning technique derived from a standard K-means algorithm. The goal of cluster analysis is the classification of objects based on similar properties among them and the subdivision into groups or clusters. Among the advantages of the fuzzy clustering algorithm one is that it allows objects to belong to several clusters simultaneously, with different degrees of membership. Our results showed that jcBNST neurons from normal animals displayed a higher degree of physiological diversity than those from dependent animals, meaning that the former subdivided into a greater number of clusters than the latter. Our experiments also showed that long-term neuroadaptive changes induced by drugs of abuse can alter populations of extended amygdala neurons in a way that might remain undetected if evaluating only their gross physiological properties.