Phd Student In Modeling Horizontal Interactions In Primary Visual Cortex M/F

Universities and Institutes of France
November 04, 2022
Offerd Salary:Negotiation
Working address:N/A
Contract Type:Temporary
Working Time:Full time
Working type:N/A
Job Ref.:N/A
  • Organisation/Company: CNRS
  • Research Field: Neurosciences Psychological sciences › Cognitive science Psychological sciences › Psychology
  • Researcher Profile: First Stage Researcher (R1)
  • Application Deadline: 04/11/2022 23:59 - Europe/Brussels
  • Location: France › SACLAY
  • Type Of Contract: Temporary
  • Job Status: Full-time
  • Hours Per Week: 35
  • Offer Starting Date: 01/01/2023
  • Horizontal connections are extremely abundant in cerebral cortex, and with local connections, they constitute the majority of excitatory synapses received by a given cortical neuron, while external inputs (from thalamus) represent only a few percent of synapses. Despite this dominance of horizontal and local connections, their role is poorly understood.

    The goal of this thesis project is to understand, using computational models, the role of horizontal connections in the processing of visual information in primary visual cortex (V1). We will build computational models based on experiments realized in Diego Contreras lab (DCL) in U Penn, where the paradigm is to record in multiple columns in V1, while presenting different types of visual inputs to the animal. The experiments can control the "brain state" of the animal, ranging from asynchronous states to oscillatory states (gamma oscillations). By modulating independently the stimulus and the state of the network, the experiments will map the relation between network state and the propagation of visual information across columns in visual cortex, via horizontal connections.

    Computational models will be built based on these experimental data, and detailed anatomical information available for V1 from cats and ferrets. We will use network models of cortical columns, or multiple columns organized in 2D, and connected via horizontal connections. Because the data allow an independent control of network state and horizontal input, they will provide unprecedented information to build precise computational models. Models will be constrained based on measurements of postsynaptic potentials, spike output properties, neuronal correlation and the distribution of synaptic input in cortex under two network states. Models will allow to parametrically understand the role of cellular and network properties and generate experimentally testable predictions.

    The project consists of constructing neural networks to model interactions via horizontal connections in visual cortex. The experimental data will be provided by the laboratory of Diego Contreras (U Penn, USA), where the response to visual stimuli will be measured for different network states (asynchronous or oscillanting). The computational modeling approach (Alain Destexhe lab, NeuroPSI) will attempt to reproduce the experimental data and build a framework to understand the role of horizontal connections in vision in primary visual cortex.

    The modeling will be performed gradually in a few steps. In a first step, we will use existing network models of asynchronous and gamma-oscillation states, and adjust these models to the experiments. The goal is here to reproduce the network state as precisely as possible, so that it can be used as a reference cortical column later. In a second step, the model will be adjusted to evoked responses, also based on experiments. Reproduding the response evoked by the same input in different network states is a very strong evidence that the model provides correct predictions. In a third step, we will extend the model to multiple cortical columns. Here, we will use experiments where one column is recorded while other column receives the visual input, therefore the response to the recorded column is uniquely due to horizontal connections. Here again, the model will be carefully compared to experiments to reproduce the different responses observed for different network states and visual inputs. A mean-field model will be designed from the spiking model, in order to allow us to extend to large scales. At the end, we aim to obtain a multi- column model of V1 with correct horizontal interactions, and dependence on network state. We will also attempt to generalize these findings to extrapolate to other cortical areas, and obtain a general theory of information processing in cerebral cortex via horizontal connections.

    In this project, there will be a continuous exchange between experiments (in DCL) and modeling, the model will formulate predictions that will be directly tested in DCL, which will result in refinement of the model. This experiment- model loop will be used to obtain a model that finely matches the experiments. This exchange between the two labs will be supported by visits between experimentalists and computational modelers.

    Finally, if time permits, we will also identify critical parameters which can make the system go wrong. For example, an obvious choice is to alter inhibition, which may result in aberrant behavior which will be tested experimentally (by opto-genetically activating or silencing inhibitory neurons for example). Various drugs such as alcohol, propofol or barbiturates alter inhibition and also have strong effects on vision. The prediction of such effects can be done with the computational model, and later tested experimentally using the same drugs.

    Web site for additional job details

    https: //

    Required Research Experiences
  • Neurosciences

  • None

  • Psychological sciences › Cognitive science

  • None

  • Psychological sciences › Psychology

  • None

    Offer Requirements
  • Neurosciences: Master Degree or equivalent

    Psychological sciences: Master Degree or equivalent

  • FRENCH: Basic

    Contact Information
  • Organisation/Company: CNRS
  • Department: Institut des Neurosciences Paris-Saclay
  • Organisation Type: Public Research Institution
  • Country: France
  • City: SACLAY
  • From this employer

    Recent blogs

    Recent news