Despite rapid progress in deep learning, its inspiration from biology remains shallow. Conversely, systems neuroscience lacks tools for building predictive, functional models of cognition at scale.
We recommend reviewing:
1) Yamins & DiCarlo (2016) – Using goal-driven deep learning models to understand sensory cortex
https://www.nature.com/articles/nn.4244
2) Richards et al. (2019) – A deep learning framework for neuroscience
https://www.nature.com/articles/s41583-019-0187-4
3) Kriegeskorte & Douglas (2018) – Cognitive computational neuroscience
https://www.nature.com/articles/s41583-018-0029-7
Key idea: Rather than merely drawing metaphors from the brain, we should formalize brain function as a set of learnable, testable, and simulatable constraints.