Hybrid Brains

Abstract

Computational modeling of neurons and circuits is a growing component of neuroscience research and can fruitfully complement experimental investigations, ideally in a mutually informative feedback loop between experiments and modeling studies. The dynamic clamp technique allows for an even more direct interaction between experimentation and modeling by interfacing living neurons and circuits with computational models in real-time and at multiple levels, ranging from models of cellular components and synapses to models of individual neurons or entire circuits. The dynamic clamp thus creates hybrid in vivo – in silico systems in which living brains and computer models directly ‘talk to each other’. This takes advantage of both worlds, combining the ground truth of experimental investigation of living neural systems with the complete control over neural, synaptic, and circuit parameters provided by computational models. This chapter explains how the dynamic clamp operates, describes various dynamic clamp applications, and gives examples of dynamic clamp studies that have furthered our understanding of circuit operation. I also discuss caveats of the technique, including technical issues and limitations, and the inherently embedded question of what we can learn from computational models and hybrid systems, and to what extent they can be ‘trusted’.