Neural circuits as computational dynamical systems

• Published in: Current opinion in neurobiology Vol. 25; pp. 156 - 163
• Main Author: Sussillo, David
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.04.2014
• Subjects: Animals; Humans; Models, Neurological; Nerve Net - physiology; Neural Networks (Computer); Neurology; Prefrontal Cortex - physiology; Psychiatry
• ISSN: 0959-4388; 1873-6882; 1873-6882
• DOI: 10.1016/j.conb.2014.01.008

•Many cortical circuits can be viewed as computational dynamical systems.•A new tool to help us understand cortical dynamics is the optimized recurrent neural network (RNN).•RNNs are useful because they have fundamental similarities to biological neural systems.•RNNs are optimized to perform tasks analogous to those given to subjects in experimental settings.•RNNs can generate novel ideas and hypotheses about the mechanisms of computation in biological neural circuits. Many recent studies of neurons recorded from cortex reveal complex temporal dynamics. How such dynamics embody the computations that ultimately lead to behavior remains a mystery. Approaching this issue requires developing plausible hypotheses couched in terms of neural dynamics. A tool ideally suited to aid in this question is the recurrent neural network (RNN). RNNs straddle the fields of nonlinear dynamical systems and machine learning and have recently seen great advances in both theory and application. I summarize recent theoretical and technological advances and highlight an example of how RNNs helped to explain perplexing high-dimensional neurophysiological data in the prefrontal cortex.