Grand Challenge 5: Mind and Brain Architecture Bottom Up Brain Architecture Development based on Biologically Plausible Neurons Christian R. Huyck, Soodamani Ramalingam, and Usama Hasan Middlesex University c.huyck, s.ramalingam, and u.hasan@mdx.ac.uk 1 Introduction At some level, the basis of the Architecture of the Brain and thus of the Mind is clearly Neurons. There is a significant understanding of Neurons, their computa- tional properties, and though this understanding is incomplete, we can currently use biologically plausible neural models for a range of computational tasks. We propose using simulated biologically plausible neurons to develop a brain architecture. We have made initial explorations of such an architecture and have found them promising. One main advantage of this approach is that we can consistently use human neural and cognitive functioning as guidance. This work could also direct research in Neuroscience. 2 Plausible Neurons Neurons are leaky integrators [6]. They learn via Hebbian learning. Our current model uses a simplified neural model including discrete time steps. Neural fatigue is an additional feature that is biologically motivated. Together these neural features enable the creation of Cell Assemblies based on Hebb’s hypothesis [9]. Cell Assemblies are reverberating neural circuits, and these circuits enable a great deal of computational power. 3 Computational Properties of Plausible Neurons We have used our current model as a categoriser (e.g. [10]). This is consistent with a range of work in attractor networks [2]. The network is presented with a 1