The successful applicant will build a simulation model of the rodent visual cortex and use it to assess the role of dendritic nonlinearities on the connectivity and activity properties of the resulting memory engrams.
Selected model predictions will be tested in headfixed behaving animals performing a visual discrimination task.
Specifically, the successful applicant will build a reduced circuit model of L2 / 3 V1 neurons, that incorporate dendrites (modeled as linear, sub-linear or supra-linear activation functions) and synaptic plasticity rules (STC-based LTP / LTD, homeostatic plasticity, plasticity of intrinsic excitability etc.
Models will also incorporate key interneuron types such as PV, Somatostatin, calretinin and VIP-expressing cells, modeled as point neurons.
The basic biophysical properties of all models will be validated against existing experimental data provided by the literature and ongoing collaborations Stelios Smirnakis at Harvard University).
The V1 circuit model will be trained with a feature and an integrative learning paradigm. Network analysis will include the size of the neuronal population that responds to stimulus presentation, somatic and dendritic spiking, synaptic weight changes-connectivity, the spatial distribution of modified synapses and the sparsity of the resulting engram.
By systematically manipulating dendritic biophysics ionic and synaptic conductances) so as to generate linear, sublinear and supralinear dendrites as well as the various synaptic / intrinsic plasticity rules, and interneuron strengths, we will assess the effect of these manipulations on a) learning capacity, b) network connectivity and c) engram size / connectivity / recall stability etc.
We will validate our modeling predictions via optogenetic manipulations of pyramidal and / or GABAergic (SST, PV or VIP) neurons in V1, while mice execute a visual discrimination task.
Reduced, biologically relevant model of the L2 / 3 V1 network with active dendrites and plasticity rules
Implementation of realistic visual learning protocols in the model
Assessment of dendritic contributions to a) learning capacity, b) emergent network connectivity, c) engram size / recall stability etc.
Radboud University (NL), Fleur Zeldenrust, Tansu Celikel (months 30-32)
Mathworks (DE), Philip Laserstein (months 18-21)