Marr's three levels of analysis are useful as a framework for neuroscience.

Abstract

In September 2019, I had the pleasure of attending a symposium (Galliano et al., 2019) commemorating the legacy of David Marr, one of the pioneers of computational neuroscience – the field concerned with developing mathematical models and analysis tools for studying the nervous system. At the symposium, I also participated in a live debate about the merits of one particular contribution that Marr made to the field: the so-called three levels of analysis Marr (1982). This CrossTalk article is an extended summary of the arguments I made there for why I find this contribution still valuable, nearly 40 years after its publication.