Font Size: 

Experimental data from the hippocampus and the neocortex haveestablished a link between Phase-Amplitude Cross-Frequency Coupling, where theamplitude of faster gamma oscillations is synchronized with the phase of the slowertheta/alpha activity, and cognitive processes such as attention and learning.Traditionally, it is hypothesized that the phase of the slow wave dictates gammawave amplitude. Recently however, indexes such as the Cross-FrequencyDirectionality (CFD) have revealed that interactions occur in both directions: fromslow to fast (CFD>0) and from fast to slow (CFD<0). In this computational study, wedemonstrate that the connectivity of simple yet common neural motifs isassociated with the direction of these interactions. Specifically, feedforwardinhibition modeled through a theta-modulated ING (Interneuron Network Gamma)model exhibits fast-to-slow interactions while feedback inhibition modeled througha theta-modulated PING (Pyramidal Interneuron Network Gamma) model exhibits slow-to-fast interactions. In the former case, the negative directionality arises dueto the rapid dynamics of basket cell interneurons, which respond in gammasufficiently fast to anticipate theta locally. Notably, this negative directionality ismaintained across plausible synaptic strength parameters and modeled timedelays. Furthermore, we report that each motif is optimized to integrate distinct
types of inputs, with the fast-to-slow (slow-to-fast) motif better integrating non-theta (theta inputs). Our results suggest that directionality measurements reflect
the activation of different underlying circuit motifs that serve distinct computationalneeds.