Research report
Long-range EEG synchronization during word encoding correlates with successful memory performance

https://doi.org/10.1016/S0926-6410(00)00011-2Get rights and content

Abstract

Distinct cortical activity during memory encoding of words, which were either recalled or not, was reported by a number of studies. This activity was mainly found at frontal and temporal/parietal brain regions. However, it was not clear if these regions interact with each other or work independently. In order to get a functional measure of the degree of neuronal large-scale cooperation, we calculated EEG coherence, which provides a statistical measure of synchronization between two EEG signals per frequency band. Therefore, coherence enables us to assess the functional interaction between cell assemblies of distant brain regions. The purpose of our study was to investigate if successfully recalled words show enhanced cortical synchronization compared with not recalled ones. Additionally, the influence of stimulus modality and the way different EEG frequencies participate in this process was examined. The EEG of 25 participants was recorded during memory encoding of concrete German nouns, either presented auditorily or visually and stimuli were separated according to the participant's memory performance. Recalled nouns exhibited overall enhanced synchronization but showed typical patterns, especially between anterior and posterior brain regions in all frequency bands except the alpha-1 band (8–10 Hz). Recalling nouns was accompanied by increased synchronization between more distant electrodes in relation to an increase of synchronization between adjacent electrodes. Moreover, the degree of intrahemispheric synchronization was higher for recalled nouns. The pattern of EEG coherence and amplitude changes during verbal memory encoding allowed us to assess the probability that nouns would be recalled or not.

Introduction

Functional differences in brain activity accompanying the memory encoding of words, which either could have been recalled or not, have often been reported. Brain activity seems to differ in many aspects between these two groups of words as was assessed with several different neurophysiological techniques. In a recent functional neuroimaging study Wagner et al. [78]examined the neural activation for subsequently recalled and subsequently forgotten verbal experiences. Similar regions within the left prefrontal and medial temporal cortices were engaged during the encoding of both recalled and forgotten words. However, subsequently recalled words demonstrated significantly greater brain activation within these anatomically specific brain regions. Wagner et al. [78]hypothesized that both left prefrontal and temporal regions jointly promote memory formation for words. Whether these regions operate together, as suggested by Wagner et al., or if they act independently during effective memory encoding, is still not known [61].

At least partly comparable results were obtained with PET studies. Kapur et al. [26]found a common activation of left prefrontal and left medial temporal cortex during deeper (semantic) encoding of word-pairs. Therefore, they hypothesized the common activation of these two systems to lead to a neurophysiological memory trace, which can be used to guide subsequent memory retrieval. Left prefrontal brain regions are substantially more involved during deeper encoding of verbal stimuli, which is correlated with an increased involvement of the semantic memory system and simultaneous encoding into the episodic memory 18, 77.

Beside these studies, which demonstrate the spatial extend of metabolic activity during memory encoding of words, there also exist a series of EEG studies that dealt with the difference in electrical brain activity between subsequently recalled and not recalled words. Early ERP studies investigated the kind of cognitive operation during memory encoding, which enhances the memorability of items. Generally, ERPs between 300 and 800 ms during the initial presentation of words were more positive going over frontal and parietal regions for words that were recalled than for words that were not recalled [for a review, see Refs. 34, 46]. Mecklinger and Friederici [38]found an early parieto-occipital and a late fronto-central ERP difference between recalled and not recalled nouns for concrete words.

Another aspect of EEG differences between recalled and not recalled words was addressed by Klimesch and collaborators, which studied frequency band-related band power differences between these word groups 29, 30, 31. Klimesch et al. [31]studied event-related synchronization (ERS) and desynchronization (ERD) in the theta and alpha bands during the encoding of visually presented words in an incidental learning paradigm. They reported a significant increase in band power in the theta band (4.49–6.69 Hz), especially at frontal and occipital recording sites, and a decrease in the alpha bands (6.69–12.69 Hz), mainly at occipital sites for recalled vs. not recalled words. Moreover, recalled nouns produced a larger ERD in the lower alpha band (8.2–10.7 Hz) and a smaller ERD in the upper alpha band (10.7–13.2 Hz) in contrast to not recalled nouns [30]. In summary, the results of all studies described above hint at various anatomical and physiological differences between the processing of recalled and not recalled nouns.

The aim of the present study was to investigate the behavior of other functional EEG parameters during the encoding of nouns namely amplitude and especially coherence, which turned out to yield an important contribution to the understanding of brain functioning during mental activities 4, 33, 36, 41, 48, 49, 54, 58, 63, 69, 74, 80, 81.

EEG amplitude (square root of power) provides a quantitative measure of the synchrony of more local sources for each frequency within the range of an electrode. In contrast, the recorded coherence of EEG scalp is a large-scale measure, which depicts dynamic functional interactions between electrodes separated by longer distances. Thus, amplitude and coherence are mathematically and functionally independent measures of neocortical dynamics. Coherence between two EEG signals, which is the correlation coefficient in the frequency domain, may be understood as a measure of phase stability and is a measure for the degree of synchronization between brain signals of certain brain regions. High coherence between EEG signals recorded at different sites of the scalp may hint an increased functional interplay between the underlying neuronal networks.

The purpose of studying EEG coherence during certain mental activities, like memory processes, is not to localize the change of electrical brain activity within distinct brain regions, but to register specific patterns of large-scale decreasing and increasing synchronization of brain activity between the underlying neuronal networks. The functional interaction between cell assemblies distributed throughout the cerebral cortex as large-scale networks is a prerequisite for cognitive information processing [8]. Especially, in the light of recent neurobiological findings, coherence analysis of the human EEG got additional importance. Several studies, especially of the cat's and monkey's visual cortex have shown that neuronal synchronization processes between adjacent and also between distant brain regions correlate with the perception of objects or the successful performance of visuomotoric tasks 9, 14, 15, 20, 32, 60, 72.

Changes of synchronization are performed within different frequency bands, which may reflect functionally different components of information processing, e.g., mere sensory perception or memory encoding. Since most of the animal experiments mentioned above reported narrow-band synchronization within high frequencies (Gamma>30 Hz) to be correlated with perceptual binding, an increased number of studies on human EEG started to deal with the role of high frequencies during various cognitive processes [e.g., Refs. 24, 41, 52, 53, 59]. Astonishingly, some authors did only report findings in the higher gamma range, but no changes within low frequencies 41, 59. Apart from the importance of gamma activities for cognitive processing in humans, one may not fail to notice numerous studies, which described a lot of findings in lower frequencies, which are very well correlated with different kinds of cognitive information processing, like memory, language, concept retrieval and music processing [e.g., Refs. 2, 19, 23, 28, 33, 35, 48, 55, 56, 57, 63, 64, 65, 68, 71, 74, 80, 81, 82, 83]. It may be that high- and low-frequency synchronization play different roles in cognitive information processing [75]. Nevertheless, it is important to note that parts of the high (gamma) frequencies are harmonics of activities in the alpha range, which was discussed by Jürgens et al. [25]and that a close mathematical relationship may exist between high and low EEG frequencies [66]. Therefore, in the present study the investigation of lower frequency bands was presented, which does not suffer from well-known methodological difficulties like the investigation of gamma frequencies and yields reliable data concerning cognitive information processing since many years.

Up to now, only a few studies were performed on the investigation of EEG coherence during memory processes in healthy humans 5, 33, 63, 41, 79. Generally, during the use of distinct types of memory, these studies reported an increase of synchronization between brain regions involved in the respective task. Krause et al. [33]found an increase of coherence in the left parieto-temporal region for concept activation and in the prefrontal region for coordination during memory retrieval of a concept. Sarnthein et al. [63]reported a significant enhancement in coherence in the theta range (4–7 Hz) between prefrontal and posterior electrodes during a working memory task, and Miltner et al. [41]found that gamma-band coherence increased between regions of the brain that received two classes of stimuli involved in an associative-learning paradigm. Moreover, good memory performers showed higher coherence in lower frequencies than bad memory performers [79].

The current study was designed in order to investigate the different degree of synchronization during the memory encoding of concrete nouns, which either were recalled or not. Stimuli were presented as well in the auditory as in the visual modality to assess the influence of stimulus modality on the synchronization processes. EEG amplitude was also examined in order to compare the findings with previous studies, but main emphasis was laid on the calculation and interpretation of coherence changes. This study attempted at elucidating the functional interaction between the cortical regions serving efficient verbal memory encoding of concrete nouns and the way different frequencies of the EEG participate in this process. According to the literature reported above, we hypothesized, at first, significantly higher coherence (synchronization) for recalled vs. not recalled nouns. Secondly, recalled nouns should elicit a higher percentage of coherence changes than not recalled ones. The third question addresses the topography of coherence differences between these two groups of words. In view of the findings in literature, we postulated increased coherence between brain regions involved in memory processes like prefrontal and temporal/parietal regions. In the fourth place, the influence of stimulus modality on the changing synchronization patterns was investigated.

Section snippets

Participants

Twenty-five female monolingual students (mean age 23.7±2.7) with German as their native language participated in the experiment. All had normal or corrected-to-normal vision and their laterality quotients on a slightly modified version of the Edinburgh Handedness Inventory [44]ranged from 84 to 100, indicating strong right-handed preferences for all subjects. Due to technical artefacts only the data of 23 persons were used for further data analysis.

Stimuli

Subjects were presented with 75 disyllabic

Behavioral data

No predefined mnemonic strategy was demanded from the participants and they were asked for the performance strategies they had applied. Eighty two percent of the participants reported about elaborative rehearsal strategies during encoding of the concrete nouns (relating nouns to each other, imaging strategies, etc.). On average, 35% of the auditorily and visually presented nouns could be recalled in the free recall paradigm. No significant difference has been found between the auditory and the

Discussion

The results of the current experiment exhibited major EEG amplitude and coherence differences between recalled and not recalled nouns during the period of word encoding. Most of these findings occurred independently of the modality of stimulus presentation (either auditory or visual) and reflected different synchronization patterns of brain activity between both groups of words.

Acknowledgements

The authors which to thank Dr. Lorenz Sichelschmid for help with the statistical data analysis. We also like to thank Dr. Horst M. Müller for helpful discussions and the Experimental Neurolinguistics Group for technical support. We thank Thomas Schneider and Werner Chromecek for their assistance in data processing. This study was supported by the Austrian “Fonds zur Förderung der wissenschaftlichen Forschung” P13578-MED and the “Hochschuljubiläumsstiftung” Proj. H-168/98.

References (83)

  • T Harmony

    Do specific EEG frequencies indicate different processes during mental calculation?

    Neurosci. Lett.

    (1999)
  • S Kapur et al.

    The neural correlates of intentional learning of verbal materials: a PET study in humans

    Cognit. Brain Res.

    (1996)
  • W Klimesch

    Memory processes, brain oscillations and the EEG synchronization

    Int. J. Psychophysiol.

    (1996)
  • W Klimesch

    EEG alpha and theta oscillations reflect cognitive and memory performance: a review and analysis

    Brain Res. Rev.

    (1999)
  • W Klimesch et al.

    Event-related desynchronisation (ERD) and the Dm effect: does alpha desynchronization during encoding predict later recall performance?

    Int. J. Psychophysiol.

    (1996)
  • D Lehmann et al.

    Spatio-temporal dynamics of alpha brain electric fields, and cognitive modes

    Int. J. Psychophysiol.

    (1997)
  • R.C Oldfield

    The assessment and analysis of handedness: the Edinburgh inventory

    Neuropsychologia

    (1971)
  • H Petsche et al.

    Brain electrical mechanisms of bilingual speech management: an initial investigation

    Electroencephalogr. Clin. Neurophysiol.

    (1993)
  • H Petsche et al.

    The possible meaning of the upper and lower alpha frequency ranges for cognitive and creative tasks

    Int. J. Psychophysiol.

    (1997)
  • H Petsche et al.

    Thinking with images or thinking with language: a pilot EEG probability mapping study

    Int. J. Psychophysiol.

    (1992)
  • F Pulvermüller et al.

    High frequency brain activity: its possible role in attention, perception and language processing

    Prog. Neurobiol.

    (1997)
  • D.S Ruchkin et al.

    Modality-specific processing streams in verbal working memory: evidence from spatio-temporal patterns of brain activity

    Cognit. Brain Res.

    (1997)
  • B Schack et al.

    The sensitivity of instantaneous coherence for considering elementary comparison processing: Part I. the relationship between mental activities and instantaneous EEG coherence

    Int. J. Psychophysiol.

    (1999)
  • B Schack et al.

    Adaptive phase estimation and its application in EEG analysis of word processing

    J. Neurosci. Methods

    (1999)
  • T Schanze et al.

    Phase correlation among rhythms present at different frequencies: spectral methods, application to microelectrode recordings from visual cortex and functional implications

    Int. J. Psychophysiol.

    (1997)
  • M Schürmann et al.

    Alpha oscillations shed new light on relation between EEG and single neurons

    Neurosci. Res.

    (1999)
  • W.D Sheppard et al.

    Pretrial EEG coherence as a predictor of semantic priming effects

    Brain and Lang.

    (1990)
  • W Skrandies et al.

    Electroencephalographic cortical oscillations and saccadic eye movements in humans

    Neurosci. Lett.

    (1999)
  • W Singer

    Coherence as an organizing principle of cortical functions

    Int. Rev. Neurobiol.

    (1994)
  • E.P Sloan et al.

    Anticholinergic drug effects on quantitative electroencephalogram, visual evoked potential, and verbal memory

    Biol. Psychiatry

    (1992)
  • R.W Thatcher et al.

    Cortico-cortical associations and EEG coherence: a two-compartmental model

    Electroenceph. Clin. Neurophysiol.

    (1986)
  • 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

    Neurosci. Lett.

    (1996)
  • G.F Wilson et al.

    EEG power changes during a multiple level memory retention task

    Int. J. Psychophysiol.

    (1999)
  • I.L Baschek et al.

    Bestimmung der Bildhaftigkeit (I), Konkretheit, (C) und der Bedeutungshaltigkeit (m′) von 800 Substantiven

    Z. Exp. Psychol.

    (1977)
  • V. Braitenberg, Cortical architectonics: general and areal, in: M.A.B. Brazier, H. Petsche (Eds.), Architectonics of...
  • St.L Bressler et al.

    Episodic multiregional cortical coherence at multiple frequencies during visual task performance

    Nature

    (1993)
  • J.D Cohen et al.

    Temporal dynamics of brain activation during a working memory task

    Nature

    (1997)
  • M Doppelmayr et al.

    Individual differences in brain dynamics: important implications for the calculation of event-related band power

    Biol. Cybernet.

    (1998)
  • R Eckhorn et al.

    Coherent oscillations: a mechanism for feature linking in the visual cortex?

    Biol. Cybernet.

    (1988)
  • R Eckhorn et al.

    High frequency (60–90 Hz) oscillations in primary visual cortex of awake monkey

    NeuroReport

    (1993)
  • M Essl et al.

    EEG coherence and reference signals: experimental results and mathematical explanations

    Med. Biol. Eng. Comput.

    (1998)
  • Cited by (240)

    • The rhythm of learning: Theta oscillations as an index of active learning in infancy

      2020, Developmental Cognitive Neuroscience
      Citation Excerpt :

      The relationship found was specific to the theta-band oscillations (3−5 Hz) recorded over the frontal cortex and was not present in any other frequency band or scalp area (Begus et al., 2015). The location where the learning-predictive theta modulation was recorded in this study is consistent with adult studies, which showed increased frontal theta power for later recalled compared to later forgotten items in adults (Cavanagh et al., 2012; Sederberg et al., 2003; Weiss and Rappelsberger, 2000). Importantly, the predictive relationship between theta oscillations during exploration and subsequent memory was not mediated by the amount of time infants spent looking at or manually manipulating the objects (Begus et al., 2015).

    View all citing articles on Scopus
    View full text