Dear All, The next talk in our Friday seminar series will be given by: *Ma'te' Lengyel* (Computational and Biological Learning Lab, Department of Engieering, University of Cambridge) *INTERNAL MODELS OF THE BRAIN: FROM BEHAVIOUR TO NEURAL REPRESENTATIONS* Date: April 6, 13:30 Location: Institute of Cognitive Neuroscience and Psychology, MTA-TTK, Szondi u. 83-85, 1st floor Abstract: Our percepts rely on an internal model of the environment, relating physical processes of the world to inputs received by our senses, and thus their veracity critically hinges upon how well this internal model is adapted to the statistical properties of the environment. However, the level of statistical sophistication with which such internal models are adapted or learned has not been investigated systematically, and their neural underpinning has also remained unknown. We conducted behavioural and electrophysiological studies to address these issues. Using a visual statistical learning paradigm, we have shown that humans learn about the statistics of visual stimuli well beyond the level of pair-wise associations, approximating the performance of a Bayesian ideal learner (Orban et al, PNAS 2008). Using multielectrode recordings in awake ferrets, we have argued that spontaneous activity in V1 represents a key component of Bayesian inference: the prior distribution (Fiser et al, TICS 2010), and showed that this prior becomes gradually adapted to the statistic of natural visual scenes over the course of development (Berkes et al, Science 2011). [This is joint work with Gergo Orban, Pietro Berkes, and Jozsef Fiser.] Everyone is welcome to attend. Kindest regards, Tama's Bo"hm