M.Sc. thesis topics

Working memory is a key function in humans. It allows us to temporarily maintain small amounts of information (i.e., generally less than 3 or 4). If we draw a parallel with computing, we could compare it to RAM memory. This cognitive function is one of the most studied in the world of cognitive psychology. Nevertheless, some of its aspects still remain misunderstood. Working memory abilities have been shown to increase as a function of linguistic characteristics of stimuli. For example, the words “paint, table, brush” will be easier to memorize and manipulate than the words “tire, biped, demography”. Working memory is therefore not an isolated function, but seems to interact in a complex way with the linguistic system.

The goal of this project is to better understand the interactions between working memory and the linguistic system. To do this, the student will have to adapt a so-called interactive activation model. This type of model makes it possible to explain a large number of phenomena occurring within our memory. The problem with these models is that they are particularly unstable and difficult to handle. The challenge of this project will be to adapt one of these models, in order to make it more stable and able to model the behavior observed in humans.

Understanding how specific ion channel conductances affect the input–output behaviors of a neuron remains a challenging task. In particular, the link between ion channel degeneracy (which refers to multiple ‘different’ mechanisms conveying equivalent function) and the stimulus encoding properties of these ion channels is not well-understood.

In this project, we will develop a pipeline that combines numerical simulations of biophysical models (conductance-based models) and estimation of statistical models (point process generalized linear models) to explore the link between variations in ion channel conductances and stimulus encoding. We will illustrate this pipeline on published biophysical models of thalamic neurons and/or spinal cord neurons.

A fundamental question in neuroscience is how to link observed neural activity to the unobserved biophysical mechanisms that generate this activity. Therefore, there is a critical need for methods to incorporate the partial and noisy data that we observe with detailed, mechanistic models of neural activity. 

In this project, we will explore how to estimate the parameters and the hidden variables of neuronal models from neuronal spike train responses. In particular, we will compare modern simulation-based inference methods to more traditional methods like particle filters. Depending on the progress, we will also investigate how to actively collect new data in closed-loop experiments to improve the inference.