Abstract
Oscillatory neuronal activity is implicated in many cognitive functions, and its phase coupling between sensors may reflect networks of communicating neuronal populations. Oscillatory activity is often studied using extracranial recordings and compared between experimental conditions. This is challenging, because there is overlap between sensor-level activity generated by different sources, and this can obscure differential experimental modulations of these sources. Additionally, in extracranial data, sensor-level phase coupling not only reflects communicating populations, but can also be generated by a current dipole, whose sensor-level phase coupling does not reflect source-level interactions. We present a novel method, which is capable of separating and characterizing sources on the basis of their phase coupling patterns as a function of space, frequency and time (trials). Importantly, this method depends on a plausible model of a neurobiological rhythm. We present this model and an accompanying analysis pipeline. Next, we demonstrate our approach, using magnetoencephalographic (MEG) recordings during a cued tactile detection task as a case study. We show that the extracted components have overlapping spatial maps and frequency content, which are difficult to resolve using conventional pairwise measures. Because our decomposition also provides trial loadings, components can be readily contrasted between experimental conditions. Strikingly, we observed heterogeneity in alpha and beta sources with respect to whether their activity was suppressed or enhanced as a function of attention and performance, and this happened both in task relevant and irrelevant regions. This heterogeneity contrasts with the common view that alpha and beta amplitude over sensory areas are always negatively related to attention and performance.
Highlights
Neuronal signals contain oscillations at many frequencies [1], and these have been shown to be implicated in many cognitive functions
Three key properties of our proposed approach for analyzing oscillatory neuronal activity In this paper, we present a novel approach for analyzing oscillatory neuronal activity, which uses a model-based method that separates and characterizes sources by their patterns of between-sensor phase coupling
We start with electrophysiological measurements Vjl(t) measured over time t, obtained from sensor j and trial l. Oscillatory activity in these recordings is described by Fourier coefficients, which we obtain from a spectral analysis involving multitapering (e.g. Welch [22] or Slepian [23] tapering; multitapering is optimal because it provides control over the frequency resolution, but it is not required for our method)
Summary
Neuronal signals contain oscillations at many frequencies [1], and these have been shown to be implicated in many cognitive functions (for a review, see [2]). Interpreting sensor-level activity is strongly hindered by the fact that the underlying sources generate overlapping sensor-level spatial patterns. This is especially problematic when neuronal activity is investigated in the context of a task, and the question of interest is whether activity differs as a function of experimental variables or behavior. This typically involves an analysis of some measure of neuronal activity (e.g. power, coherence) and an independent variable (e.g. conditions, reaction times) at every sensor or sensor-pair. It may appear that sensor-level activity is either suppressed or enhanced by an experimental manipulation, whereas different underlying sources are each modulated differently (e.g. with some sources being suppressed, while others are enhanced, as we will show later)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
