## Frontmatter | | | | --- | --- | | Authors | [[Umut Güçlü]], [[Marcel van Gerven]] | | Date | 2017/09 | | Source | [[Computational Models of Brain and Behavior]] | | URL | https://doi.org/10.1002/9781119159193.ch30 | | Citation | Güçlü, U., & van Gerven, M. (2017). [[Probing human brain function with artificial neural networks]]. In _Computational Models of Brain and Behavior_. [[URL](https://doi.org/10.1002/9781119159193.ch30)]. #Chapter | ## Abstract Artificial neural networks (ANNs) were originally conceived of as an approach to model mental or behavioral phenomena. This chapter addresses the question how ANNs can be used to probe human brain function with a focus on the state-of-the-art results that emerged from this approach. First, it describes how one can model the mapping between stimuli and responses in the human brain through the development of encoding models. Next, the chapter focuses on how the mapping from static naturalistic stimuli to neural responses can be realized using ANNs. It considers a special kind of multi-layer perceptrons (MLPs) for learning word embeddings. A recent development is to train deep neural networks (DNNs) consisting of up to a thousand hidden layers. Then the chapter describes how brain responses induced by dynamically changing naturalistic environments can be modeled. It ends by exploring future developments in the use of ANNs for investigating human brain function. ## PDF ![[Probing human brain function with artificial neural networks.pdf]]