Spatial Eigenmodes and SynchronousOscillation : Co { Incidence Detection inSimulated Cerebral

Zero{lag synchrony arises between two points on the cerebral cortex when these receive concurrent independent inputs and has generally been ascribed to nonlinear mechanisms. We report results obtained by Principal Component Analysis (PCA) applied to simulations of cerebral cortex which exhibit zero{lag synchrony and realistic spectral content, and show that synchrony can arise by distinct and separable linear and nonlinear mechanisms. For lower levels of cortical activation synchrony between the sites of input can be accounted for by the eigenmodes associated with the wave activity generated by the inputs. The rst spatial eigenmode arises from even 2 Clare L. Chapman et al. components in the independent input signals and the second spatial eigenmode arises from odd components in the inputs. Together these account for most of the signal variance, while the predominance of the rst mode over the second within the near{{eld of the inputs accounts for zero{lag synchrony in the neighbourhood of the inputs. Thus the simulated cortical surface acts as a coincidence detector with regard to concurrent, but uncorrelated, inputs. Synchronous response in the extended cortical eld to changes in the inputs occurs on a time{scale of a few milliseconds. At a critical point of increasing cortical activation, local damped oscillation in the gamma band undergoes a transition to highly non-linear undamped activity at 40 Hz dominant frequency. Such activity \locks" between active sites to create patterns of activity with segregated phases. The complementary character of the synchrony dependent on damped wave interactions, and that attributable to locked nonlinear oscillations may permit speedy organisation of synchronous patterns and partition of multiple patterns of activity within the same eld of neu-rons. Synchronous Oscillation 3 Key words synchronous oscillation { Principal Component Analysis { zerolag synchrony { cortical activation { spatial eigenmode { nonlinear phase locking { stimulus feature binding 1 Introduction and Background.