Cross-Frequency Modulation, Network Information Integration and Cognitive Performance in Complex Systems

Abstract

The higher cognitive processes cannot be explained by processes of sequential neural activation and new systemic approaches to the study of electroencephalographic correlates of cognition are presently proposed. As the mental activity appears to be the result of a systemic integration of different neural network activities (brain networks, BNet), the associated mutual interaction between different frequency bands generates the phenomenon of cross-frequency modulation (CfM) that supports the synchronization and the coordination of the high-frequency activity of distant brain areas. Within this paper we present two kinds of evidence supporting the key role of CfM in allowing the above mentioned integration, one coming from experimental EEG data concerning human subjects, and the other coming from computer simulations of models of brain network activity. As concerns the experimental evidence in 8 subjects, aged between 20 and 51 years old, not suffering of any psychopathology or neurological disorders, EEG activity was recorded during the execution of a cognitive task in which they were asked to identify the non-previously known criteria used by the computer to classify some presented images. The study of the Event Related Potentials (ERPs) showed a longer latency in responses following an error. Independent EEG components with a strong fit with a dipole model were identified and CfM of these components was computed either independently from the stimulus and related to the stimulus and to the subject response. The obtained results were consistent with the role of CfM in the functional activation of BNets and with the subject cognitive performance. Some computer simulations of the activity of networks of Integrate-and-fire neurons were also carried out. The latter showed that the presence of synchronization between neuron activities was facilitated by the presence of long-range synaptic interactions, periodicity of the stimulus inputs, adaptation of the neuronal threshold and presence of inhibitory synapses. We found also evidence of the key role played by network connectivity structure. Namely, both theoretical arguments and simulation results agreed in showing that, in the case of external discontinuous stimuli the scale-free networks are more likely to manifest a high frequency resonance phenomena coupled with persistent low frequency oscillations. All these data are supporting the role of CfM in the functional integration of complex networks and suggest that this phenomenon could be typical of a large number of other complex systems such as, e.g., the social networks.