A Common Cortical State Underlying Neuronal Population Coding


Zhiqin John Xu,New York University Abu Dhabi and Courant Institute


2018.06.12 14:00-15:00


601, Pao Yue-Kong Library


To understand how large-scale neuronal networks function, it is important to identify their common dynamical operating states. The probabilistic characteristics of these operating states will underlie network functions such as its coding schemes. Here, directly from multi-electrode data from three separate experiments, we quantitatively identify a cortical operating state (the “probability polling” or “p-polling”), common across different species (mouse and monkey) and different behaviors. Regarding this state’s functional impact, in the three experiments we confirm that it establishes the accuracy of low-order maximum entropy representations of the distribution of neuronal firing patterns. Our simulations also show that the “balanced state”, common in large networks, is also a p-polling state; and that the p-polling state is closely related to weakly correlated networks. However, the p-polling state is more general than these two concepts. These results provide evidence for the p-polling state’s commonality and its potential importance for neuronal coding.

*This is a joint work with David Cai, David W. McLaughlin, Douglas Zhou, et al.