The field of computational neuroscience originates from Physics, where there is a strong tradition to model nature from first principles. For instance, Joao has modeled working memory using ring networks, in which their connectivity structure is designed so that the required attractor dynamics emerges. This approach is extremely successful in modeling simple tasks, but has been proven limited when modeling more complex tasks. This has motivated a new approach, originated instead in Machine Learning, where the parameters (e.g. connectivity) of very flexible neural networks are instead trained to perform arbitrarily complex tasks. We also rely on this approach, to understand how computations necessary to solve complex tasks can be distributed across segregated brain regions and to study how different neuromodulators shape network dynamics. Finally, we also try to explain complex behavior at the computational level, essentially explaining away the specific interactions between neurons or brain regions, and focusing instead on the computations performed by them (e.g. what variables are learned and at what speed during complex tasks).