Intra-cortical propagation of EEG alpha oscillations
Introduction
Alpha oscillations can be recorded from mammalian neocortex, particularly during wakeful detachment from the environment. Although initially regarded as “idling rhythms”, recent observations of their ability to modulate sensory awareness and reaction times suggest a central role in perceptual and motor function (Dijk et al., 2008, Klimesch et al., 2007). In addition, recent experiments have provided evidence that alpha oscillations are intimately linked to early sensory-evoked potentials (Makeig et al., 2002, Mazaheri and Jensen, 2008, Nikulin and Brismar, 2006). Besides their involvement in perceptual processing, experiments in human and macaque have demonstrated their involvement in regulating attention, working memory processes, and provide selective inhibition of task-irrelevant networks (Jensen et al., 2002, Klimesch, 1999, Klimesch et al., 2007, Mo et al., 2011).
In contrast to the dynamics of localized alpha sources, which have been investigated in numerous studies, spatial propagation of alpha oscillations has received considerably less attention. Human EEG studies have shown however, that resting-state as well as stimulus-induced alpha oscillations travel over the scalp (Burkitt et al., 2000, Ito et al., 2005, Klimesch et al., 2007, Nunez et al., 2001) and that their wave properties correlate with reaction times (Fellinger et al., 2012, Patten et al., 2012), indicating their potential functional relevance. These observations however, pertain to scalp signals and the properties of the underlying waves on the cortical sheet are unclear at present. Because scalp velocities lie in the range 5–15 m/s, it has been suggested that alpha scalp waves are propagate through cortico-cortical axons, which have conduction velocities in this range (Kandel et al., 2000). The resulting long wavelengths, combined with predominant propagation along the medial axis have been proposed to endow alpha oscillations with suitable properties to mediate top-down and bottom-up interactions between occipital and frontal networks (Burkitt et al., 2000, Ito et al., 2005, Klimesch et al., 2007, Nunez et al., 2001, Patten et al., 2012).
The reported velocities obtained from scalp EEG recordings seem inconsistent with those obtained from invasive recordings, which are about an order of magnitude lower. Local field potential (LFP) recordings in dogs visual cortex lead to estimates of about 0.3 m/s (Lopes Da Silva et al., 1980). More recently, electrocorticographic (EcoG) recordings in human subjects have led to estimates of propagation velocities in the range 0.7–2.1 m/s (Bahramisharif et al., 2013). Intriguingly, the authors observed gamma bursts propagating with the same velocity and riding on the troughs of the propagating alpha oscillations. The slow propagation velocities obtained from invasive recordings are consistent with voltage-sensitive dye (VSD), optical imaging, and local field potential recordings of other kinds of neocortical activity, which lead to propagation velocities < 1 m/s suggesting mediation through intra-cortical instead of cortico-cortical axons (Ermentrout and Kleinfeld, 2001, Han et al., 2008, Wu et al., 2008, Zheng and Yao, 2012), Besides the inconsistency in reported traveling velocities between scalp EEG and invasive recordings, another difference is that in the latter, systematic propagation of single-cycle oscillations is reported, while the former report propagation of single-cycle oscillations only in a fraction of cycles (Ito et al., 2005, Patten et al., 2012).
In this study we resolve these issues by combining a method to extract wave-activity from scalp EEG recordings with current source density simulations. The proposed method shows that single-cycle alpha oscillations propagate over the scalp, consistent with the observations of VSD and LFP studies. Furthermore, the simulations show that intra-cortical propagation of alpha oscillations through cortical sulci or gyri can account for the observed scalp velocities. We further verify the intra-cortical hypothesis of alpha propagation by formulating and verifying a number of predictions on the behavior of alpha scalp waves. Importantly, the wavelengths implied by the intra-cortical hypothesis are estimated to be in the order of several centimeters, much smaller than predicted by the cortico-cortical hypothesis, provoking a re-examination of the notion of alpha-mediated long-range interactions.
Section snippets
Cortico-cortical versus intra-cortical propagation
Fig. 1A illustrates the cortico-cortical and intra-cortical hypothesis on alpha propagation. According to the cortico-cortical hypothesis, electric potential waves on the scalp reflect cortical waves with similar velocity (5–15 m/s) that are mediated through cortico-cortical axons, which are known to have conduction velocities in this range (Kandel et al., 2000). According to the intra-cortical hypothesis, advocated in this study, the traveling scalp waves result from relatively localized and
Propagation of single-cycle oscillations
To assess if single alpha-cycles propagate over the scalp, we computed the single-cycle waves and computed the instantaneous velocities and averaged these over the waves and subsequently over subjects. The resulting velocity distribution is shown in Fig. 5A. The average velocity is 2.26 ± 0.25 m/s. The narrow concentration around non-zero values demonstrates that single-cycle waves indeed propagate over the scalp. To assess if the wave-templates are representative for single-cycle waves, we
Discussion
In this study we have shown that scalp EEG traveling waves can be accounted for by cortical oscillations that propagate through neocortical tissue with velocities in the range of propagation velocities of spikes in intra-cortical axons. Using different reference electrodes in the proposed analysis method biases the extraction of wave-activity to different cortical regions and it is expected that MEG recordings, which are less affected by the spatial blurring of electromagnetic fields (
Acknowledgments
GD was supported by the ERC Advanced Grant: DYSTRUCTURE (no. 295129), by the Spanish Research Project SAF2010-16085 and by the CONSOLIDER-INGENIO 2010 Program CSD2007-00012, and the FP7-ICT BrainScales. The authors declare no competing financial interests.
References (46)
- et al.
Steady-state visual evoked potentials and travelling waves
Clin. Neurophysiol.
(February 2000) - et al.
Traveling electrical waves in cortex: insights from phase dynamics and speculation on a computational role
Neuron
(January 2001) - et al.
Evoked traveling alpha waves predict visual-semantic categorization-speed
NeuroImage
(February 2012) - et al.
Reverberation of recent visual experience in spontaneous cortical waves
Neuron
(October 2008) - et al.
Human cortical oscillations: a neuromagnetic view through the skull
Trends Neurosci.
(1997) EEG alpha and theta oscillations reflect cognitive and memory performance: a review and analysis
Brain Res. Brain Res. Rev.
(April 1999)- et al.
Induced alpha band power changes in the human EEG and attention
Neurosci. Lett.
(March 1998) - et al.
EEG alpha oscillations: the inhibition-timing hypothesis
Brain Res. Rev.
(January 2007) EEG and MEG: relevance to neuroscience
Neuron
(2013)- et al.
Relative contributions of intracortial and thalamo-cortical processes in the generation of alpha rhythms, revealed by partial coherence analysis
Electroencephalography
(1980)
Phase synchronization between alpha and beta oscillations in the human electroencephalogram
Neuroscience
Unifying and interpreting the spectral wavenumber content of EEGs, ECoGs, and ERPs
J. Theor. Biol.
Nonlinear multivariate analysis of neurophysiological signals
Prog. Neurobiol.
White matter architecture rather than cortical surface area correlates with the EEG alpha rhythm
NeuroImage
A population-average, landmark- and surface-based (PALS) atlas of human cerebral cortex
NeuroImage
A brain symmetry index (BSI) for online EEG monitoring in carotid endarterectomy
Clin. Neurophysiol.
A model of the spatial-temporal characteristics of the alpha rhythm
Bull. Math. Biol.
Study of human alpha rhythm: the alphon hypothesis and alpha suppression
Int. J. Psychophysiol.
Alpha oscillations in brain functioning: an integrative theory
Int. J. Psychophysiol.
Propagating neocortical gamma bursts are coordinated by traveling alpha waves
J. Neurosci.
Investigating the electrophysiological basis of resting state networks using magnetoencephalography
Proc. Natl. Acad. Sci.
Complex brain networks: graph theoretical analysis of structural and functional systems
Nat. Rev. Neurosci.
MEG characterization of spontaneous alpha rhythm in the human brain
Brain
Cited by (45)
A methodological framework for inverse-modeling of propagating cortical activity using MEG/EEG
2020, NeuroImageCitation Excerpt :EEG sensor-data give the impression that they propagate between frontal and occipital regions with high speeds and with wavelengths in the order of tens of centimeters. This has given rise to the cortico-cortical propagation hypothesis, according to which propagation of alpha oscillations is mediated by white-matter fibers, which have conduction speeds of about 10 m/s. However, the EEG sensor-data are also consistent with slow propagation of alpha oscillations over small distances, mediated by intra-cortical axons Hindriks et al. (2014), which have conduction speeds of about 0.3 m/s. According to this hypothesis, the high speeds measured on the scalp are a consequence of the local curvature of the cortex, giving rise to high angular velocities of the electric field at large distances from the cortex Hindriks et al. (2014).
Latency analysis of resting-state BOLD-fMRI reveals traveling waves in visual cortex linking task-positive and task-negative networks
2019, NeuroImageCitation Excerpt :Recent modeling work, however, using large-scale dynamical models on the human connectome, suggests that cortical flow patterns are multistable and exhibit phase-transitions (Roberts et al., 2019). To study such phenomena, a dynamic analysis in which no assumptions about stationarity are made, is required (Majeed et al., 2009, 2011; Hindriks et al., 2014; Thompson et al., 2014; Matsui et al., 2016). One way of extending a stationary latency analyses is to estimate the cross-covariance functions using a sliding-window.