Elsevier

NeuroImage

Volume 191, 1 May 2019, Pages 225-233
NeuroImage

Desynchronizing to be faster? Perceptual- and attentional-modulation of brain rhythms at the sub-millisecond scale

https://doi.org/10.1016/j.neuroimage.2019.02.027Get rights and content

Highlights

  • We measured inter-peak intervals (IPIs) of alpha/beta rhythms over the visual cortex.

  • Mean and SD of IPIs were reduced when a stimulus appeared in contralateral hemifield.

  • The alpha/beta rhythms thus became faster and more regular during visual stimulation.

  • The same changes in IPIs were seen when subjects attended to contralateral hemifield.

Abstract

Neural oscillatory signals has been associated with many high-level functions (e.g. attention and working memory), because they reflect correlated behaviors of neural population that would facilitate the information transfer in the brain. On the other hand, a decreased power of oscillation (event-related desynchronization, ERD) has been associated with an irregular state in which many neurons behave in an uncorrelated manner. In contrast to this view, here we show that the human ERD is linked to the increased regularity of oscillatory signals. Using magnetoencephalography, we found that presenting a visual stimulus not only induced a decrease in power of alpha (8–12 Hz) to beta (13–30 Hz) rhythms in the contralateral visual cortex but also reduced the mean and variance of their inter-peak intervals (IPIs). This indicates that the suppressed alpha/beta rhythms became faster (reduced mean) and more regular (reduced variance) during visual stimulation. The same changes in IPIs, especially those of beta rhythm, were observed when subjects allocated their attention to a contralateral visual field. Those results revealed a new role of the event-related decrease in alpha/beta power and further suggested that our brain regulates and accelerates a clock for neural computations by actively suppressing the oscillation amplitude in task-relevant regions.

Introduction

Many studies suggest that neural oscillatory signals could serve various important functions in the brain. Unit-cell recordings on animals showed that changes in amplitude and coherence of gamma-band oscillation are strongly linked to cognitive processes such as feature integration and attention (Engel et al., 2001; Fries, 2009), while oscillations in theta-band plays a key role in memory (Buzsaki and Moser, 2013; Knyazev, 2007). The close relationships between oscillatory signals and cognitive functions have been also demonstrated in the human brain (Jensen et al., 2007; Palva and Palva, 2011). In electroencephalography (EEG) and magnetoencephalography (MEG), an increase in power of a given frequency band is called the event-related synchronization (ERS), which reflects cooperative or synchronized behaviors of a large number of neurons (Pfurtscheller and Lopes da Silva, 1999) that would facilitate the information transfer across neuronal groups (Fries, 2005). Consistent with this view, a spatial allocation of attention induced the gamma-band ERS in the sensory cortex of a contralateral hemisphere (Bauer et al., 2006), and stronger theta-band ERS in a memory encoding stage predicted better performances in later retrieval stage (Osipova et al., 2006).

In contrast to the ERS, a decrease in power of oscillatory signals is called the event-related desynchronization or ERD (Pfurtscheller and Lopes da Silva, 1999), being associated with uncorrelated (irregular) behaviors of neural population. Paradoxically, many studies have reported the ERD in brain regions that would play a central role in a given task (task-relevant regions). A well-known example is the alpha ERD over the occipital cortex in response to visual inputs (Berger, 1929). Recent studies showed that not only the alpha rhythm (8–12 Hz) but also the beta rhythm (13–30 Hz) showed prominent ERD to visual (Kulashekhar et al., 2016; Minami et al., 2014; Wyart and Tallon-Baudry, 2009), auditory (Cirelli et al., 2014; de Pesters et al., 2016; Fujioka et al., 2015), and somatosensory (Bauer et al., 2006; Fransen et al., 2016) stimuli. The alpha/beta ERD in the sensory areas was also induced by attention (van Ede et al., 2012). A voluntary allocation of attention to a visual hemifield reduced alpha/beta power in the contralateral visual cortex (Bauer et al., 2014; Worden et al., 2000; Wyart and Tallon-Baudry, 2009).

Why was the ERD, a neural signature of uncorrelated activities, observed in task-relevant regions? Previous studies resolved this issue by assuming an inhibitory role of alpha oscillation on neural processing (Jensen and Mazaheri, 2010; Klimesch, 2012). For example, Jensen and Mazaheri (2010) argued that alpha activity provides pulsed inhibition reducing the processing capabilities of a given area. A recent study also reported the inhibitory role of beta rhythm on neural processing (Shin et al., 2017). The alpha/beta ERD observed in task-relevant regions thus would reflect a release from this periodic (regular) inhibition, representing the active processing of sensory information in a contra-stimulus and contra-attention hemisphere.

Although this release-from-inhibition (disinhibition) hypothesis of alpha/beta ERD is widely accepted, here we explore a new role of ERD other than the disinhibition. Specifically, we focused on changes in frequency (Cohen, 2014) and regularity (Fransen et al., 2015) of neural rhythms caused by the ERD. Regularity (or periodicity) is one of the most fundamental information in oscillatory signals. Indeed, previous studies reported altered regularity of neural oscillations in patients with mental disorders such as Alzheimer's disease (Gomez et al., 2007; Poza et al., 2012). We hypothesized that, if the alpha/beta ERD represents weakened (less-regulated) pulses of inhibition, this would be associated with decreased regularity of oscillations in the same frequency band.

Basic procedures of our MEG experiment are shown in Fig. 1. Participants performed the standard Posner task in which their attention was directed to either a left or right visual field (Fig. 1A). A central cue (arrow) and a peripheral task stimulus (Gabor patch) would induce a decrease in alpha/beta power in a contralateral (contra-attention and contra-stimulus) hemisphere. To quantify the regularity, we measured inter-peak intervals (IPIs) of oscillatory signals (Fig. 1B), depicting a distribution of their occurrences (Fig. 1C). For example, an IPI distribution of alpha rhythm (8–12 Hz) would be centered on 100 ms, because its central frequency is 10 Hz. When the alpha rhythm is highly regular, most IPIs would be located around the mean (100 ms), which results in a smaller variance of the distribution (Fig. 1C, top). In contrast, an irregular alpha wave would be indexed by a large variance because such irregular rhythm produces IPIs distant from 100 ms (Fig. 1C, bottom). One advantage of our method is that it can measure various types of changes in oscillatory signals simultaneously. For example, if the decreased alpha/beta power causes an overall increase in an oscillation frequency, this would be detected as a decrease in a mean (rather than variance) of IPIs.

Section snippets

Participants

Twenty-two healthy human subjects (13 females, age: 20–54) participated in the present study. Informed consent was received from each participant after the nature of the study had been explained. All experiments were carried out in accordance with guidelines and regulations approved by the ethics committee of Kobe University, Japan.

Stimuli and task

We used the Matlab Psychophysics Toolbox (Brainard, 1997; Pelli, 1997) to generate visual stimuli (refresh rate: 60 Hz) and to record button-press responses by

Behavioral data

Hit and false-alarm (FA) rates in all types of trials (mean ± SE across participants) are shown in Fig. 2. Participants correctly pressed a button when a grating with a target orientation (target grating) appeared in an attended visual field (hit rates: 95.2% in LL target and 94.8% in RR target trials). Reactions times measured from an onset of a grating were 665.9 ± 23.6 ms (LL) and 665.7 ± 22.2 ms (RR). No difference in hit rates (t(21) = 0.30, p = 0.77, Cohen's d = 0.029) or reaction times (t

Discussion

In contrast to the prediction in Introduction, we presently found that the decreased alpha-to-beta power in task-relevant regions was associated with the increased regularity of oscillatory signals. Although regularity (periodicity) of neural oscillation has been measured using autocorrelation analysis (Red'ka and Mayorov, 2015), spectral entropy (Gomez-Pilar et al., 2016; Poza et al., 2012), sample entropy (Gomez and Hornero, 2010), auto-mutual information analysis (Gomez et al., 2007), Q

Acknowledgments

This work was supported by KAKENHI Grants Number 22680022 and 26700011 from the Japan Society for the Promotion of Science for Young Scientists to Y.N. We thank Mr. Y. Takeshima (National Institute for Physiological Sciences, Japan) for his technical supports. The authors declare no competing financial interest. All data supporting the findings of this study are available from Y.N. upon reasonable request.

References (57)

  • G.G. Knyazev

    Motivation, emotion, and their inhibitory control mirrored in brain oscillations

    Neurosci. Biobehav. Rev.

    (2007)
  • C.E. Lemercier et al.

    Aberrant alpha and gamma oscillations ex vivo after single application of the NMDA receptor antagonist MK-801

    Schizophr. Res.

    (2017)
  • A. Mazaheri et al.

    EEG spectral dynamics during discrimination of auditory and visual targets

    Brain Res Cogn Brain Res

    (2005)
  • A. Mierau et al.

    State-dependent alpha peak frequency shifts: experimental evidence, potential mechanisms and functional implications

    Neuroscience

    (2017)
  • T. Minami et al.

    Decreased beta-band activity is correlated with disambiguation of hidden figures

    Neuropsychologia

    (2014)
  • A. Mueller et al.

    Linking ADHD to the neural circuitry of attention

    Trends Cognit. Sci.

    (2017)
  • N. Nishitani et al.

    Viewing lip forms: cortical dynamics

    Neuron

    (2002)
  • Y. Noguchi et al.

    Direct behavioral and neural evidence for an offset-triggered conscious perception

    Cortex

    (2015)
  • G. Pfurtscheller et al.

    Event-related EEG/MEG synchronization and desynchronization: basic principles

    Clin. Neurophysiol.

    (1999)
  • M. Suzuki et al.

    Temporal dynamics of neural activity underlying unconscious processing of manipulable objects

    Cortex

    (2014)
  • M. Bauer et al.

    Tactile spatial attention enhances gamma-band activity in somatosensory cortex and reduces low-frequency activity in parieto-occipital areas

    J. Neurosci.

    (2006)
  • M. Bauer et al.

    Attentional modulation of alpha/beta and gamma oscillations reflect functionally distinct processes

    J. Neurosci.

    (2014)
  • C.B. Beaman et al.

    Sensory coding accuracy and perceptual performance are improved during the desynchronized cortical state

    Nat. Commun.

    (2017)
  • Y. Benjamini et al.

    Controlling the false discovery rate - a practical and powerful approach to multiple testing

    J. Roy. Stat. Soc. B

    (1995)
  • H. Berger

    Uber das Elektrenkephalogramm des Menschen

    Arch Psychiat Nervenkr

    (1929)
  • J.W. Bisley

    The neural basis of visual attention

    J. Physiol.

    (2011)
  • D.H. Brainard

    The psychophysics toolbox

    Spatial Vis.

    (1997)
  • G. Buzsaki et al.

    Memory, navigation and theta rhythm in the hippocampal-entorhinal system

    Nat. Neurosci.

    (2013)
  • Cited by (6)

    • Disrupted relationship between intrinsic neural timescales and alpha peak frequency during unconscious states – A high-density EEG study

      2023, NeuroImage
      Citation Excerpt :

      This temporal property of APF can be measured by an analysis method first developed in (Cohen, 2014) called “frequency sliding”, which involves the computation of the first temporal derivative of the range-restricted phase time series of the neural signal: as a result, one can obtain a time series of instantaneous frequencies in the selected range, which is useful to capture its variability at a fine-grained temporal resolution. Recent studies report a role for alpha frequency sliding in temporal processing of incoming inputs, similarly to what has been put forward for INTs: the speed of alpha frequency sliding predicts the temporal resolution of visual perception (Samaha and Postle, 2015), regulates event-related desynchronization during a visual perception task (Noguchi et al., 2019), tracks the insurgence of on- and off-thoughts (Hua et al., 2022) and predicts more general temporal integration mechanisms (Shen et al., 2019). These results support the hypothesis that the duration of alpha cycles organizes the gating of incoming inputs.

    • Temporal codes of visual working memory in the human cerebral cortex: Brain rhythms associated with high memory capacity

      2020, NeuroImage
      Citation Excerpt :

      Most studies listed above have investigated changes in power (amplitude) of alpha/beta rhythm related to vWM. Here we address this issue by measuring a speed of those waves, using the inter-peak interval (IPI) analysis of magnetoencephalography (MEG) (Noguchi et al., 2019). In this analysis, raw waveforms of MEG are filtered with a pass-band of interest (e.g. 8–30 Hz for alpha-to-beta band).

    • Changes in latency of brain rhythms in response to affective information of visual stimuli

      2020, Biological Psychology
      Citation Excerpt :

      Moreover, the IPI analysis enables us to evaluate the regularity of a brain rhythm (see 2.6 for details). Using this analysis, we previously found a significant reduction of IPIs caused by a spatial allocation of attention (Noguchi et al., 2019); the mean IPI over the left/right visual cortex became shorter when participants attended to a right/left visual field. This attentional modulation of IPIs in a contralateral hemisphere was most clearly observed when we focused on oscillatory signals in alpha-to-beta range (8–30 Hz).

    View full text