MEMBERS
Alexandre Pouget
Alexandre Pouget heads the Computational Cognitive Neuroscience Laboratory at the UNIGE. His research focuses on probabilistic theories, computation and learning in neural circuits. He is currently applying this framework to a wide range of topics, including olfactory processing, spatial representations, sensory-motor transformations, multisensory integration, attention control, decision-making, causal reasoning and simple arithmetic. According to his approach, cerebral knowledge takes the form of probability distributions, and learning is acquired through probabilistic inferences.
Alexandre Pouget humorously describes himself as a "failed experimenter". This led him to think that if he wanted to enter the field of neuroscience as a biologist, he needed to develop theory! So he developed solid skills in mathematics and physics. He set up his own laboratory in 1996 and has since devoted himself to his passion for neuroscience, and by extension, mental health research.
Probabilistic theories: an asset for neuroscience!
Probabilistic theories are rather counter-intuitive. Everything we know is subject to uncertainty. We know very little with certainty. Every time we observe something that interests us, we only get partial information about what we want to know. For example, when we want to cross the road, but see a car approaching, the crucial information is how far we are from the car. It tells you whether you have enough time to cross the road before being hit. But it's not enough to look at the car to know how far away it is. At best, we only get an estimate. 10 metres? More, less? Everything we perceive, everything we know, is subject to uncertainty. In science, we observe. Then we explain the observations using scientific theories. But we never have absolute certainty.
Alexandre Pouget's laboratory is trying to understand how neurons take uncertainty into account, how circuits represent uncertainty and how uncertainty is manipulated to make robust calculations. His team works with all the animals that perform such calculations. Starting with humans, but also monkeys, and currently rodents and even fly larvae and maggots.
Understanding the addictive properties of cocaine
In collaboration with Christian Lüscher's laboratory at the UNIGE, Alexandre Pouget and his team are studying how animals learn to obtain certain rewards, whether natural, such as food, or artificial, such as cocaine or direct stimulation of the ventral tegmental area (VTA).
The research team is trying to detect the presence of possible indicators that can discriminate between animals that will become dependent and those that will not. They use learning models to determine which ones best predict the phenomenon of dependence. Once these predictive parameters have been identified, the laboratory will highlight their neural correlates in order to understand where they occur. Their function will then be determined, potentially leading to the development of targeted therapies to combat addiction.
Thanks to the probabilistic theories of Alexandre Pouget's laboratory, the differences between individuals who become addicted and those who don't have been identified and better understood, an important step towards understanding addiction.