The work of this thesis concerns how cortical memories are stored and retrieved. In particular, large-scale simulations are used to investigate the extent to which associative attractor theory is compliant with known physiology and in vivo dynamics.

The first question we ask is whether dynamical attractors can be stored in a network with realistic connectivity and activity levels. Using estimates of biological connectivity we demonstrated that attractor memories can be stored and retrieved in biologically realistic networks, operating on psychophysical timescales and displaying firing rate patterns similar to in vivo layer 2/3 cells. This was achieved in the presence of additional complexity such as synaptic depression and cellular adaptation.

Fast transitions into attractor memory states were related to the self-balancing inhibitory and excitatory currents in the network. In order to obtain realistic firing rates in the network, strong feedback inhibition was used, dynamically maintaining balance for a wide range of excitation levels. The balanced currents also led to high spike train variability commonly observed in vivo. The feedback inhibition in addition resulted in emergent gamma oscillations associated with attractor retrieval. This is congruent with the view of gamma as accompanying active cortical processing.

While dynamics during retrieval of attractor memories did not depend on the size of the simulated network, above a certain size the model displayed the presence of an emergent attractor state, not coding for any memory but active as a default state of the network. This default state was accompanied by oscillations in the alpha frequency band. Such alpha oscillations are correlated with idling and cortical inhibition in vivo and have similar functional correlates in the model. Both inhibitory and excitatory, as well as phase effects of ongoing alpha observed in vivo was reproduced in the model in a simulated threshold-stimulus detection task.