Computational Neuroscience

Computational Neuroscience is a field of neuroscience that investigates the biophysical principles governing the electrochemical activity of the neurons, along with the computational processes involved in cognitive functions. The brain possesses a highly intricate architecture across multiple spatial scales. It is still not fully understood how biophysical mechanisms, which can be studied in isolation, function when integrated into larger networks. To that end, the central nervous system has been extensively studied across a wide variety of animal species (including humans), in many contexts involving more or less complex tasks, using increasingly elaborate instrumentation. Advancing our understanding of healthy cognition contributes to a better characterization of neuropathological alterations.

To understand the brain, a wide range of biophysical and computational models have been developed, along with numerical simulation methods aimed at explaining experimental data and guiding the design of new experiments. Computational neuroscience regroups methods and approaches from multiple fields, including physics, mathematics, numerical science and psychology in addition to biology.

Historically, one of the earliest models in computational neuroscience is the “integrate-and-pull” model introduced by Louis Lapicque in 1907. This influential model is still used nowadays to build neuronal networks that account for key biophysical mechanisms. For many years, the biophysics of neurons has been the focus of scientists to understand the heterogeneity of their activity. In parallel, models for cognition have interpreted neuronal activity in terms of algorithms and information processing. Interactions between neuroscience and neighboring fields have been bidirectional with some preferred directions depending on the period, partly reflecting hypes in science. Recently, the development of modern artificial intelligence has enabled the processing of complex data, providing powerful tools for data analysis, but also offering a model for high-level human cognition.