Discussion
Summary
Main (expected) results:
-
ERF waveforms are explained on the systems level by the weight matrices, that is, the interactions across the whole of AC. This dynamical-system element of the waveform is modified by the topography of the AC and its connections (the K matrix)
-
Variations across subjects are explained not only by subject-specific cortical topographies but also by subject-specific dynamics, however, P randomizations have a much stronger effect on the waveform than K randomizations.
Novel approach for ERP generation
-
Similarities and differences between our network approach for the generation of ERPs and other, already existing models (equivalent current dipoles, phase reset model).
-
In our approach ERPs are characterised by two physical terms (decay constant \(\gamma\), damping frequency \(\delta\)) which are not available from discrete source models; ERPs emerge from the network dynamics and can be explained in terms of normal modes.
-
Analytical solutions enable powerful parametric studies to obtain a further understanding of the dynamics of the signal processing in auditory cortex. They provide (computationally) fast solutions. One of the most intriguing practical implication is the potential of fitting the model to experimental data and, thus, obtain subject-specific parameters.
-
Tradition of linearizing non-linear systems.
Origin of subject specificity
-
Two sources of subject specificity (P, K-matrix). We observe different impact between changes of oscillation dynamics and changes in topography.
-
K-matrix and its impact on the waveform (P50, M100, etc.); dependence of peak amplitudes on feedforward and feedback connections.
-
Separating the effects of P and K; investigating whether the effects of P and K on ERPs do not differ between linear and non-linear case.