Elsevier

Brain and Language

Volume 96, Issue 1, January 2006, Pages 90-105
Brain and Language

EEG theta and gamma responses to semantic violations in online sentence processing

https://doi.org/10.1016/j.bandl.2005.06.007Get rights and content

Abstract

We explore the nature of the oscillatory dynamics in the EEG of subjects reading sentences that contain a semantic violation. More specifically, we examine whether increases in theta (≈3–7 Hz) and gamma (around 40 Hz) band power occur in response to sentences that were either semantically correct or contained a semantically incongruent word (semantic violation). ERP results indicated a classical N400 effect. A wavelet-based time-frequency analysis revealed a theta band power increase during an interval of 300–800 ms after critical word onset, at temporal electrodes bilaterally for both sentence conditions, and over midfrontal areas for the semantic violations only. In the gamma frequency band, a predominantly frontal power increase was observed during the processing of correct sentences. This effect was absent following semantic violations. These results provide a characterization of the oscillatory brain dynamics, and notably of both theta and gamma oscillations, that occur during language comprehension.

Introduction

Currently, there has been a growing interest in the oscillatory dynamics that can be observed in the electrophysiological activity of the brain. Unlike Event-Related brain Potentials or fields (ERPs/ERFs; the so-called evoked activity) that can be extracted from the EEG or MEG by straightforward averaging of single trials, event-related changes in oscillatory EEG/MEG activity (termed induced activity) that speak on the synchronization and desynchronization aspect of neural activity have received little attention. The goal of the present study is to explore the oscillatory dynamics that coincide with the well-studied semantic ERP effect, the N400 effect. We will focus on the modulation of oscillatory neuronal activity recorded in the EEG during the reading of sentences that contain either a semantically incongruent word (which elicits an N400 in the ERP analysis) or that are semantically congruent. More specifically, we want to establish whether semantic violations induce oscillatory activity that is qualitatively different from those induced by syntactic violations (which elicit a P600 in the ERP analysis) that have been found in a previous study (Bastiaansen, van Berkum, & Hagoort, 2002b).

Much of the research on language comprehension has focused on the integration of semantic and syntactic information. In particular, a large body of research has centered on assessing the influence of semantic information on the syntactic analysis of sentences. This has spawned a debate about whether semantic and syntactic processing are subserved by neuronally distinct networks [i.e., as proposed by Hagoort and Brown (1999)] or whether semantics and syntax are processed by a single network [i.e., as proposed by McClelland, StJohn, and Taraban (1989)]. Within the domain of event-related potential (ERP) research, this debate has resulted in a search for ERP components that are differentially sensitive to semantic and syntactic processing in language comprehension. Several different ERP profiles appear to be related to either aspects of semantic or syntactic processing. Two in particular are the N400, which is related to semantic processing, and the P600, which is related to syntactic processing.

The P600 component is a broadly distributed positive shift that begins around 500 ms after the occurrence of a syntactic violation. The P600 has been observed with various syntactical violations, such as phrase structure violations, verb argument violations, and morpho-syntactic violations such as breaches of number and gender agreement (Friederici et al., 1993, Hagoort et al., 1993, Neville et al., 1991, Osterhout and Mobley, 1995).

A seminal study by Kutas and Hillyard (1980) demonstrated that if participants are presented with sentences ending with a semantically inappropriate word (such as “He spread the warm bread with socks”) vs. a contextually appropriate word, (for example, “He spread the warm bread with butter”), a significantly larger negativity in the ERP waveform occurs from 250 to 600 ms after the onset of the inappropriate word, with a peak amplitude around 400 ms. Since Kutas and Hillyard (1980), a large body of research has utilized the N400 to investigate semantic processing. As a result of this research the N400 has been shown to occur to each and every word, but the amplitude and latency can be affected by a number of factors including word class, semantic relatedness to the context, cloze probability,1 word frequency, presentation modality, and verbal working memory load (Brown and Hagoort, 1999, Kutas and Van Petten, 1994; see Chao et al., 1995, Neville et al., 1986, Smith et al., 1986, Stuss et al., 1986; for examples of N400 effects related to working memory). In general, a larger amplitude N400 will be found to a word that is not related to the context, has a low cloze probability or a low frequency. The N400 effect has been interpreted as reflecting some aspect(s) of the processes that integrate the meaning of a particular word into a higher-order semantic representation (Brown and Hagoort, 1999, Osterhout and Holcomb, 1992, Rugg, 1990; however see Kutas & Van Petten, 1994; for an alternative suggestion).

To form a full understanding of language input all of the different parts of information that are processed in different brain areas must be integrated. How this binding between the distributed nodes of the language network is implemented still an open question, but it has been suggested that it may be propagated by synchronization and desynchronization of oscillatory neural activity. In this view, synchronization and desynchronization link spatially distributed brain areas together to form transient functional networks (Singer, 1993, Singer, 1999).

It is generally agreed upon that analyzing event-related changes in either amplitude, or (phase) coherence of EEG oscillations provide a window onto the processes of synchronization and desynchronization of neuronal populations (Tallon-Baudry and Bertrand, 1999, Varela et al., 2001). Two types of increase are studied: an increase in amplitude of oscillatory EEG activity is taken to at least partially reflect an increase in synchrony of the underlying local neuronal population. On the other hand, an increase in (phase) coherence between EEG sensors is thought to reflect increased synchrony between spatially distributed neuronal populations. These event-related increases in synchrony in turn are considered to reflect the transient formation of local and spatially distributed functional networks (for review, see Varela et al., 2001).

Oscillatory dynamics cannot be studied with standard ERP methodology. The reason for this is that although event-related changes in amplitude of a given oscillation may be time-locked to an experimental event, the phase of the activity at a given point in time will differ from trial to trial. As a result, oscillations largely cancel out during the averaging process that is used in ERP analysis (see Bastiaansen and Hagoort, 2003, Tallon-Baudry et al., 1996; for a more detailed account of the signal-analytic distinction between ERPs and event-related oscillatory activity). For a proper analysis of oscillatory dynamics, different analytic tools have to be used such as wavelet-based time-frequency analysis (for quantifying amplitude changes) or event-related coherence analysis (for quantifying changes in phase coherence between electrodes; for a further review, see Bastiaansen & Hagoort, 2003). Recent studies employing such techniques have clearly demonstrated that synchronous oscillations have functional significance during the execution of tasks engaging a variety of cognitive operations, such as memory encoding and retrieval (e.g., Burgess and Ali, 2002, Fell et al., 2001, Klimesch, 1999), working memory (e.g., Jensen & Tesche, 2002; Kahana et al., 1999, Tesche and Karhu, 2000), face perception (Rodriguez et al., 1999), object detection (Tallon-Baudry & Bertrand, 1999), and attentional processes (e.g., Bastiaansen and Brunia, 2001, Foxe et al., 1998; Fries et al., 2001, Klimesch, 1999). Recently, such studies are also being performed in the domain of language comprehension (e.g., Bastiaansen et al., 2002a, Bastiaansen et al., 2002b, Bastiaansen et al., 2005, Pulvermüller et al., 1999, Schack et al., 2003, Weiss and Rappelsberger, 1996, Weiss et al., 2001, Weiss et al., 2000; see also the review in Weiss & Müller, 2003). Still, relatively little is known about synchronous oscillations and their possible functions during language comprehension.

In our previous studies we characterized the time course of event-related amplitude modulations in oscillatory activity in the theta (4–7 Hz), alpha (8–12 Hz), and beta (>13–30 Hz) frequency bands during the processing of correct sentences (Bastiaansen et al., 2002a), of sentences containing syntactic violations, which elicited a P600 in the ERP analysis (Bastiaansen et al., 2002b), and of open-class words vs. closed-class words occurring within a sentence context (Bastiaansen et al., 2005). In all three studies, event-related power changes in the theta frequency range showed a marked reactivity to the various linguistic manipulations, suggesting a functional role for theta band activity during language processing. The general pattern of results that has emerged from these studies is that theta increases can be observed over left occipital areas, bilaterally over the temporal cortex, and over frontocentral regions. The left occipital theta response has been related to the encoding of the visual word form. The (left) temporal theta response has been associated either with lexico-semantic retrieval (Bastiaansen et al., 2005) or with increased verbal working memory load (Bastiaansen et al., 2002a). The frontocentral theta increase is more difficult to characterize in functional terms: it increases with working memory load (Bastiaansen et al., 2002a) is present after both open-class and closed-class words (Bastiaansen et al., 2005), and was found to increase following syntactic violations. In addition, different types of syntactic violations led to a different lateralization of the frontal theta increase, a result that we will return to in the discussion (Bastiaansen et al., 2002b).

The principal aim of the present study is to explore the oscillatory dynamics of the EEG recorded during sentences that are either semantically coherent or have a semantic violation. Characterizing the oscillatory dynamics during semantic violations will yield additional information on the neural mechanisms underlying the processing of semantic information.

Section snippets

Participants

Thirty-nine native speakers of Dutch participated in the experiment, 30 of which were included in the final analysis (24 females, 6 males; aged 18–28). They were paid a small fee for their participation. All subjects had normal or corrected-to-normal vision, normal hearing, and all were right handed. Seven of the participants included in the analysis reported having left-handed relatives. None of the subjects had any neurological impairment nor had any of the subjects participated in the

ERP analysis

Grand-average ERPs for both conditions are presented in Fig. 2 for a selected number of electrodes. Furthermore, scalp topographies for the N400 are given in Fig. 3 for both conditions.

Discussion

An EEG experiment was performed in which subjects read either correct Dutch sentences, or sentences that contained a semantic violation. Event-related potentials (ERPs) and oscillatory brain dynamics were studied. The principal aim of the experiment was to characterize the nature, i.e., magnitude and scalp topography, of EEG power (i.e., squared amplitude) changes that were expected to occur as a result of the semantic violations.

The ERP analyses of the EEG data showed a typical pattern of

Acknowledgments

The authors thank Nicole Wicha, Dirk Janssen, Thomas Urbach, and two anonymous referees for their helpful comments on an earlier version of this paper, and Eric Maris for advice regarding the statistical analysis of the data. At the time of manuscript preparation, M.B. was supported by a VENI grant from the Dutch Organization for Scientific Research (NWO) and L.H. was supported by post-doctoral fellowship from the National Institute of Health (NIH 5-T32-DC00041-10).

References (56)

  • W. Klimesch et al.

    Episodic retrieval is reflected by a process-specific increase in human electroencephalographic theta activity

    Neuroscience Letters

    (2001)
  • W. Klimesch et al.

    A method for the calculation of induced band power: Implications for the significance of brain oscillations

    Electroencephalography and Clinical Neurophysiology

    (1998)
  • P. Luu et al.

    Regulating action: Alternating activation of midline frontal and motor cortical networks

    Clinical Neurophysiology

    (2001)
  • L.K. McEvoy et al.

    Neurophysiological signals of working memory in aging

    Cognitive Brain Research

    (2001)
  • H.J. Neville et al.

    Event-related brain potentials during initial encoding and recognition memory of congruous and incongruous words

    Journal of Memory & Language

    (1986)
  • L. Osterhout et al.

    Event-related potentials elicited by syntactic anomaly

    Journal of Memory & Language

    (1992)
  • L. Osterhout et al.

    Event-related brain potentials elicited by failure to agree

    Journal of Memory and Language

    (1995)
  • B. Schack et al.

    Phase-coupling of theta–gamma EEG rhythms during short-term memory processing

    International Journal of Psychophysiology

    (2002)
  • W. Singer

    Neuronal synchrony: A versatile code for the definition of relations?

    Neuron

    (1999)
  • M.E. Smith et al.

    Human medial temporal-lobe potentials-evoked in memory and language tasks

    Electroencephalography and Clinical Neurophysiology

    (1986)
  • C. Tallon-Baudry et al.

    Oscillatory gamma activity in humans and its role in object representation

    Trends in Cognitive Science

    (1999)
  • S. Weiss et al.

    EEG coherence within the 13–18 Hz band as a correlate of a distinct lexical organisation of concrete and abstract nouns in humans

    Neuroscience Letters

    (1996)
  • S. Weiss et al.

    The contribution of EEG coherence to the investigation of language

    Brain & Language

    (2003)
  • R.H. Baayen et al.

    The CELEX Lexical Database (CD-ROM)

    (1993)
  • M. Bastiaansen et al.

    Theta responses are involved in lexico-semantic retrieval during language processing

    Journal of Cognitive Neuroscience

    (2005)
  • C.M. Brown et al.

    On the electrophysiology of language comprehension: Implications for the human language system

  • T.H. Bullock

    Introduction to induced rhythms, a widespread, heterogeneous class of oscillations

  • J. Fell et al.

    Human memory formation is accompanied by rhinal-hippocampal coupling and decoupling

    Nature Neuroscience

    (2001)
  • Cited by (206)

    View all citing articles on Scopus
    View full text