School of Social and Political Science

Description

Abstract:

A central part of the Human Brain Project’s vision was to build a ‘unifying model’ of the human brain that would unify fragmented neuroscientific data and communities. Scientists and engineers in the Human Brain Project often referred to brains and models in the singular - ‘a brain model’, ‘a rat brain’, or ‘a mouse brain’ – but these models were never of singular brains because of the various types of data and models used to build them. What, then, can the ‘a’ stand for if it is not for the singular? This talk highlights the diversity and pluralism of methods, theories and data that computational brain models bring together. I develop the concept of ‘a-model’ to argue that computational models of brains are: 1) assemblages, in that they are composed of several models at different scales and levels of detail; 2) they draw on data that is a-contextual - it is purposely removed from the specific experimental contexts it is collected in; and 3) are 'organism agnostic' (Calvert and Szymanski 2020) since they use data and models from various animals and species. While this pluralism and diversity is essential to the construction of computational models, it raises questions about whether these kinds of large-scale models can deliver on their promise of ‘understanding brain disorders’.  

Speaker Bio: Tara Mahfoud, University of Essex and is co-convenor of the British Sociological Association's Science and Technology Studies Group.