Elsevier

NeuroImage

Volume 56, Issue 2, 15 May 2011, Pages 837-842
NeuroImage

Causal influence of gamma oscillations on the sensorimotor rhythm

https://doi.org/10.1016/j.neuroimage.2010.04.265Get rights and content

Abstract

Gamma oscillations of the electromagnetic field of the brain are known to be involved in a variety of cognitive processes, and are believed to be fundamental for information processing within the brain. While gamma oscillations have been shown to be correlated with brain rhythms at different frequencies, to date no empirical evidence has been presented that supports a causal influence of gamma oscillations on other brain rhythms. In this work, we study the relation of gamma oscillations and the sensorimotor rhythm (SMR) in healthy human subjects using electroencephalography. We first demonstrate that modulation of the SMR, induced by motor imagery of either the left or right hand, is positively correlated with the power of frontal and occipital gamma oscillations, and negatively correlated with the power of centro-parietal gamma oscillations. We then demonstrate that the most simple causal structure, capable of explaining the observed correlation of gamma oscillations and the SMR, entails a causal influence of gamma oscillations on the SMR. This finding supports the fundamental role attributed to gamma oscillations for information processing within the brain, and is of particular importance for brain–computer interfaces (BCIs). As modulation of the SMR is typically used in BCIs to infer a subject's intention, our findings entail that gamma oscillations have a causal influence on a subject's capability to utilize a BCI for means of communication.

Introduction

Higher-frequency oscillations of the electromagnetic field of the brain, known as γ oscillations, have been associated with a diversity of cognitive processes including attention (Bauer et al., 2006, Gruber et al., 1999, Sokolov et al., 2004), short-term memory (Tallon-Baudry et al., 1998, Tallon-Baudry et al., 1999, Jokisch and Jensen, 2007), motor control (Crone et al., 1998, Pfurtscheller et al., 2003), and the integration of different object features into a coherent percept (Engel et al., 2001, Tallon-Baudry and Bertrand, 1999). One explanation for the ubiquity of γ rhythms in cognitive processes is provided by the hypothesis that γ rhythms constitute a fundamental mechanism of cortical information processing, dynamically routing signals within a fixed anatomical network (Fries et al., 2007). This hypothesis has gained further support by recent evidence that γ oscillations are indeed linked to performance in behavioral paradigms, providing information on successful encoding of new verbal- (Sederberg et al., 2007) and declarative memories (Osipova et al., 2006). It remains unclear, however, how γ rhythms interact with other brain rhythms. While recent studies have demonstrated cross-frequency correlations of γ rhythms with electromagnetic oscillations at different frequencies (Osipova et al., 2008, Canolty et al., 2006, Darvas et al., 2009, de Lange et al., 2008), to date there is no direct evidence for a causal influence of γ oscillations on other brain rhythms.

In this work, we study the relation of γ oscillations and the sensorimotor rhythm (SMR) in healthy human subjects using electroencephalography (EEG). The SMR is of particular importance for research on brain–computer interfaces (BCIs), as modulation of the SMR, typically induced by motor imagery (Pfurtscheller and Neuper, 2001), constitutes the most frequently used paradigm in research on BCIs (Mason et al., 2007). We first demonstrate that modulation of the SMR, induced by motor imagery of either the left or right hand, is positively correlated with the power of frontal and occipital γ oscillations, and negatively correlated with the power of centro-parietal γ oscillations. We then proceed to investigate the causal relation of the SMR and γ oscillations, based on the framework for causal inference developed by Pearl, Spirtes, and others (Pearl, 2000, Spirtes et al., 2000). Specifically, we investigate the dependency structure of our empirical observations, and present strong evidence that the observed dependence of γ oscillations and the SMR has been generated by a causal structure in which γ oscillations exert a causal influence on the SMR. We thereby present the first empirical evidence for a cross-frequency causal influence of γ oscillations on other brain rhythms, which supports the fundamental role attributed to γ oscillations for information processing within the brain.

Our findings are furthermore of particular significance for research on BCIs based on motor imagery. Since in BCIs modulation of the SMR is typically used to infer a subject's intention, our findings entail that γ oscillations have a causal influence on a subject's capability to utilize a BCI for means of communication. We thereby provide the first non-trivial neurophysiological explanation for the large variation of BCI-performance across subjects, which we consider to be crucial for addressing the problem of “BCI-illiteracy”, i.e., the incapability of about 20% of subjects to reliably communicate by means of a BCI (Popescu et al., 2008). Specifically, our results indicate that subjects should be trained to volitionally shift γ power from centro-parietal to frontal and occipital regions in order to learn how to communicate by means of a BCI.

Section snippets

Experimental paradigm

Subjects participated in an experiment consisting of right- and left-hand motor imagery. Subjects were placed in a dimly-lit and shielded room, approximately 1.5 m in front of a screen. Each trial started with a centrally-displayed fixation cross. After three seconds, the fixation cross was superimposed by an arrow pointing either to the left or to the right, instructing subjects to initiate haptic motor imagery of the left (class label − 1) or right hand (class label + 1), respectively. The arrow

Results

In order to assess the subjects' capability to modulate their SMR, we computed group-level as well as single-subject classification accuracies (percentage of trials in which the sign of the SMR lateralization score corresponds to the intended movement imagination) and associated p-values for rejecting the null-hypothesis that classification accuracy equals chance level (second column of Fig. 1). With the exception of subject S7, all subjects performed significantly above chance level, with p

Discussion

In this work, we have provided empirical evidence that γ rhythms have a causal influence on SMR modulation during motor imagery. As such, this study presents the first empirical evidence for a causal influence of γ oscillations on other brain rhythms, supporting the fundamental role attributed to γ rhythms for information processing within the brain.

To provide a more intuitive understanding of the discovered relation between γ oscillations and the SMR, Fig. 3 shows the actual values of the SMR

References (27)

  • F. de Lange et al.

    Interactions between posterior gamma and frontal alpha/beta oscillations during imagined actions

    Front. Hum. Neurosci.

    (2008)
  • A. Engel et al.

    Dynamic predictions: oscillations and synchrony in top-down processing

    Nat. Rev. Neurosci.

    (2001)
  • K. Fukumizu et al.

    Kernel measures of conditional dependence

  • Cited by (96)

    • Application of self-adaptive multiple-kernel extreme learning machine to improve MI-BCI performance of subjects with BCI illiteracy

      2023, Biomedical Signal Processing and Control
      Citation Excerpt :

      Previous research results proved that BCI based on evoked potentials have no BCI illiteracy when there are sufficient training and an adequate stimulation time period [3,4,5,6], Therefore, this study focused on motor imagery-BCI illiteracy (MI-BCI illiteracy), which is based on spontaneous EEG. Many researchers have explored the following three aspects: (1) Identification of the characteristics that can predict who are indeed BCI illiterate, such as questionnaire, gamma rhythm and neurophysiological predictor [7,8,9,10,11]; (2) Enhancement of BCI performance based on experimental methods and protocols, such as adding feedback during motor imagery, adding vibratory stimulation to the corresponding wrist [12,13,14,15]; (3) Enhancement of BCI performance via data processing. Based on the existing research on MI-BCI, this paper focuses on the application of data processing to reduce the BCI illiteracy based on MI, so as to improve the performance of BCI.

    View all citing articles on Scopus
    View full text