Transcriptional Regulation of the Human 5-HT1A Receptor Gene by Lithium: Role of Deaf1 and GSK3ß.

Serotonin 1A (5-HT1A) autoreceptors located on serotonin neurons inhibit their activity, and their upregulation has been implicated in depression, suicide and resistance to antidepressant treatment. Conversely, post-synaptic 5-HT1A heteroreceptors are important for antidepressant response. The transcription factor deformed epidermal autoregulatory factor 1 (Deaf1) acts as a presynaptic repressor and postsynaptic enhancer of 5-HT1A transcription, but the mechanism is unclear. Because Deaf1 interacts with and is phosphorylated by glycogen synthase kinase 3ß (GSK3ß)-a constitutively active protein kinase that is inhibited by the mood stabilizer lithium at therapeutic concentrations-we investigated the role of GSK3ß in Deaf1 regulation of human 5-HT1A transcription. In 5-HT1A promoter-reporter assays, human HEK293 kidney and 5-HT1A-expressing SKN-SH neuroblastoma cells, transfection of Deaf1 reduced 5-HT1A promoter activity by ~45%. To identify potential GSK3ß site(s) on Deaf1, point mutations of known and predicted phosphorylation sites on Deaf1 were tested. Deaf1 repressor function was not affected by any of the mutants tested except the Y300F mutant, which augmented Deaf1 repression. Both lithium and the selective GSK3 inhibitors CHIR-99021 and AR-014418 attenuated and reversed Deaf1 repression compared to vector. This inhibition was at concentrations that maximally inhibit GSK3ß activity as detected by the GSK3ß-sensitive TCF/LEF reporter construct. Our results support the hypothesis that GSK3ß regulates the activity of Deaf1 to repress 5-HT1A transcription and provide a potential mechanism for actions of GSK3 inhibitors on behavior.

A User's Guide to Generalized Integrate-and-Fire Models.

The generalized integrate-and-fire (GIF) neuron model accounts for some of the most fundamental behaviours of neurons within a compact and extensible mathematical framework. Here, we introduce the main concepts behind the design of the GIF model in terms that will be familiar to electrophysiologists, and show why its simple design makes this model particularly well suited to mimicking behaviours observed in experimental data. Along the way, we will build an intuition for how specific neuronal behaviours, such as spike-frequency adaptation, or electrical properties, such as ionic currents, can be formulated mathematically and used to extend integrate-and-fire models to overcome their limitations. This chapter will provide readers with no previous exposure to modelling a clear understanding of the strengths and limitations of GIF models, along with the mathematical intuitions required to digest more detailed and technical treatments of this topic.

Temporal derivative computation in the dorsal raphe network revealed by an experimentally driven augmented integrate-and-fire modeling framework.

By means of an expansive innervation, the serotonin (5-HT) neurons of the dorsal raphe nucleus (DRN) are positioned to enact coordinated modulation of circuits distributed across the entire brain in order to adaptively regulate behavior. Yet the network computations that emerge from the excitability and connectivity features of the DRN are still poorly understood. To gain insight into these computations, we began by carrying out a detailed electrophysiological characterization of genetically identified mouse 5-HT and somatostatin (SOM) neurons. We next developed a single-neuron modeling framework that combines the realism of Hodgkin-Huxley models with the simplicity and predictive power of generalized integrate-and-fire models. We found that feedforward inhibition of 5-HT neurons by heterogeneous SOM neurons implemented divisive inhibition, while endocannabinoid-mediated modulation of excitatory drive to the DRN increased the gain of 5-HT output. Our most striking finding was that the output of the DRN encodes a mixture of the intensity and temporal derivative of its input, and that the temporal derivative component dominates this mixture precisely when the input is increasing rapidly. This network computation primarily emerged from prominent adaptation mechanisms found in 5-HT neurons, including a previously undescribed dynamic threshold. By applying a bottom-up neural network modeling approach, our results suggest that the DRN is particularly apt to encode input changes over short timescales, reflecting one of the salient emerging computations that dominate its output to regulate behavior.

Cell-type-specific responses to associative learning in the primary motor cortex.

The primary motor cortex (M1) is known to be a critical site for movement initiation and motor learning. Surprisingly, it has also been shown to possess reward-related activity, presumably to facilitate reward-based learning of new movements. However, whether reward-related signals are represented among different cell types in M1, and whether their response properties change after cue-reward conditioning remains unclear. Here, we performed longitudinal in vivo two-photon Ca2+ imaging to monitor the activity of different neuronal cell types in M1 while mice engaged in a classical conditioning task. Our results demonstrate that most of the major neuronal cell types in M1 showed robust but differential responses to both the conditioned cue stimulus (CS) and reward, and their response properties undergo cell-type-specific modifications after associative learning. PV-INs' responses became more reliable to the CS, while VIP-INs' responses became more reliable to reward. Pyramidal neurons only showed robust responses to novel reward, and they habituated to it after associative learning. Lastly, SOM-INs' responses emerged and became more reliable to both the CS and reward after conditioning. These observations suggest that cue- and reward-related signals are preferentially represented among different neuronal cell types in M1, and the distinct modifications they undergo during associative learning could be essential in triggering different aspects of local circuit reorganization in M1 during reward-based motor skill learning.

Parallel and Recurrent Cascade Models as a Unifying Force for Understanding Subcellular Computation.

Neurons are very complicated computational devices, incorporating numerous non-linear processes, particularly in their dendrites. Biophysical models capture these processes directly by explicitly modelling physiological variables, such as ion channels, current flow, membrane capacitance, etc. However, another option for capturing the complexities of real neural computation is to use cascade models, which treat individual neurons as a cascade of linear and non-linear operations, akin to a multi-layer artificial neural network. Recent research has shown that cascade models can capture single-cell computation well, but there are still a number of sub-cellular, regenerative dendritic phenomena that they cannot capture, such as the interaction between sodium, calcium, and NMDA spikes in different compartments. Here, we propose that it is possible to capture these additional phenomena using parallel, recurrent cascade models, wherein an individual neuron is modelled as a cascade of parallel linear and non-linear operations that can be connected recurrently, akin to a multi-layer, recurrent, artificial neural network. Given their tractable mathematical structure, we show that neuron models expressed in terms of parallel recurrent cascades can themselves be integrated into multi-layered artificial neural networks and trained to perform complex tasks. We go on to discuss potential implications and uses of these models for artificial intelligence. Overall, we argue that parallel, recurrent cascade models provide an important, unifying tool for capturing single-cell computation and exploring the algorithmic implications of physiological phenomena.