Cognitive neuroscience

General Objectives The course provides theoretical foundations for understanding cognitive processes through the practi-cal lens of computational modelling. Cognitive functions will be interpreted as computational problems and solved by information-processing architectures in particular connectionist (neural network) models. Students will acquire essential R programming skills to implement, simulate, analyse these models, and designing experiments to collect empirical data on key cognitive domains such as memory, perception and language. The course will endow students with concrete computer programming tools to build, run, analyse, and interpret computational models and learn how to relate their behaviour to empirical data. At the end of the course, students should have developed critical skills to design, evaluate and compare computational models that could predictively bridge theoretical principles and empirical evidence from behavioural and cognitive neuroscience. Specific Objectives Knowledge and understanding EN: Students will acquire core concepts in Computational Cognitive Science, including Marr's levels of analysis and the motivations for computational modelling. They will gain an understanding of the principles and commitments of the three major modelling paradigms: Symbolic processing (overview), Connec-tionism (Neural Networks - deep dive), and the Bayesian framework for inference and learning (only briefly). They will learn about fundamental algorithms relevant to these approaches, such as the concep-tual basis of neural network training (Backpropagation) and acquire knowledge of computational architectures (e.g., associative networks or Convolutional Neural Networks; CNNs) to tackle cognitive problems in perception (e.g., categorisation), or memory (e.g., association) as well as simulate effects of altered pro-cessing in neural networks, which is relevant to examining neurodegenerative disorders. They will also gain knowledge of computational techniques in natural language processing, including N-grams, Distribu-tional Semantics (Word Embeddings), and Recurrent Neural Networks (RNNs). They will understand how these different models attempt to capture and predict patterns of human responses. Applying knowledge and understanding During the laboratory sessions, students will acquire practical skills in R programming, covering syn-tax, data structures, control flow, functions, data handling, and visualization. They will learn to implement key neural network models in R (e.g., perceptron, MLPs, CNNs) using libraries like “keras” and apply it to simulate cognitive processes (e.g., memory through associative networks). When looking at language, students will apply classic computational text processing techniques (N-grams), explore word embeddings and understand core concepts such as distributional semantics. They will also acquire practical skills in using experiment builder software (e.g., OpenSesame) to run simple cognitive experiments to obtain relevant data to be modelled. Overall, students will learn to develop computational models in the R language and understand how apply them to data from cognitive experiments. Making judgments By actively participating in lectures, reading assigned papers, and engaging in practical laboratory ac-tivities, students will develop critical thinking skills applied to computational cognitive science. They will learn to critically evaluate the theoretical assumptions, computational mechanisms, strengths, and limita-tions of the different modelling paradigms (e.g., connectionist vs. Bayesian) in explaining specific cogni-tive phenomena across perception, memory, and language domains. Students will learn to interpret the results and behaviour of computational models and critically assess their ability to account for empirical data from behavioural experiments and findings from cognitive neuroscience. Communication skills Students will develop written and oral scientific communication skills relevant to computational cognitive science research throughout the course. They will practice summarising and critically discussing research papers that employ computational modelling techniques. In class discussions, students will learn to clearly and effectively present the rationale, computational methods, results, and critical evaluation (including theoretical implications, strengths, weaknesses, and open questions) of studies involving computational models (Connectionist, Bayesian, Symbolic) applied to Memory, Perception, and Language, and their relevance to cognitive neuroscience. Learning skills Besides core course materials, students must read, understand, and critically engage with key scientific papers from the computational cognitive science literature. This experience will foster their skills in autonomous learning and critical analysis of primary research. They will develop the ability to extract the core computational ideas from different modeling paradigms (Connectionist, Bayesian, Symbolic), evaluate the methods and conclusions in the context of empirical evidence from Memory, Perception, and Language research, and identify potential future directions for research on human cognition and its neural basis using computational approaches.