StriatalPlas
Multiscale Striatal Models for Neuronal Plasticity: Integration to the Brain Simulation Platform
Project description
The dorsal striatum is crucially involved in reward learning and addiction, including the habit learning involved in compulsive behaviour and the rapid relearning involved in relapse.
Striatal synaptic plasticity is a mechanism of memory storage, but it is unclear how the changes in molecular mechanisms observed in withdrawal produce the rapid re-learning of relapse.
Numerous studies on ex vivo synaptic plasticity have revealed a complex and diverse set of required molecular mechanisms which differ depending on the pattern of cortical stimulation (e.g. Fino et al., 2010; Hawes et al., 2013; Wu et al., 2015). Since in vivo cortical activity resembles none of the patterns used ex vivo and in vivo synaptic plasticity experiments in the striatum are prohibitively difficult, it is unclear which molecular mechanisms are critical for in vivo synaptic plasticity. Furthermore, in vivo cortical activity varies from trial to trial and spike trains have variable inter-spike intervals that can change plasticity development (Graupner et al., 2016).
Our prior signalling pathway models (Blackwell et al. 2019, Kim et al. 2010, 2011) represent state-of-the-art and have successfully made predictions of in vitro synaptic plasticity and underlying mechanisms. Our multi-compartmental models of neuron electrical activity (Dorman et al. 2019, Bhalla 2017), represent state-of-the art and address questions of dendritic integration. However, multi-scale models are required for delineating mechanism of in vivo synaptic plasticity, memory storage, withdrawal and relapse.
Objectives
The objective is to integrate knowledge across multiple levels to create striatal models for investigating mechanisms underlying habit learning in both normal and drug or alcohol withdrawn individuals. For this we like to use both the workflows for building subcellular level models and to reuse and enhance already existing SP6 subcellular level models.
First, we will create multi-scale models by connecting models of intracellular signalling pathways underlying synaptic plasticity (e.g. Blackwell et al., 2019; SP6 models) with models of neuronal electrical activity (Dorman et al. 2019; SP6 striatal models, e.g. see Lindroos et al 2018 and https://f1000research.com/posters/7-1336 . Our approach is to start with populating the FindSim database with striatal experimental data. Use case: we will create and optimize SBML specified striatal signaling pathway models in Moose, and interface them with our striatal neuron models to create a multiscale model specified using standard declarative formats.
If time permits, we will couple our parameter optimization algorithm (Jȩdrzejewski-Szmek et al., 2018) with our data-harvesting and model validation pipeline (Viswan et al., 2018), and use that software to validate the multi-scale model. Here we like to compare the results with parameter search approaches already used in SP6 (see e.g. Eriksson et al. 2019). We believe the outcome will strengthen the simulation services provided at the subcellular scale.
Key facts
Time frame: 2020-2022
Origin: HBP Voucher Programme
