Discovering EEG resting state alterations of semantic dementia

https://doi.org/10.1016/j.clinph.2016.01.025Get rights and content

Highlights

  • First evidence of altered EEG resting state in semantic dementia.

  • Topographical comparison of resting state EEG microstates between semantic dementia, Alzheimer’s disease, and healthy elderlies.

  • Altered microstate class found in semantic dementia is related with decreased MMSE scores.

Abstract

Objective

Diagnosis of semantic dementia relies on cost-intensive MRI or PET, although resting EEG markers of other dementias have been reported. Yet the view still holds that resting EEG in patients with semantic dementia is normal. However, studies using increasingly sophisticated EEG analysis methods have demonstrated that slightest alterations of functional brain states can be detected.

Methods

We analyzed the common four resting EEG microstates (A, B, C, and D) of 8 patients with semantic dementia in comparison with 8 healthy controls and 8 patients with Alzheimer’s disease.

Results

Topographical differences between the groups were found in microstate classes B and C, while microstate classes A and D were comparable. The data showed that the semantic dementia group had a peculiar microstate E, but the commonly found microstate C was lacking. Furthermore, the presence of microstate E was significantly correlated with lower MMSE and language scores.

Conclusion

Alterations in resting EEG can be found in semantic dementia. Topographical shifts in microstate C might be related to semantic memory deficits.

Significance

This is the first study that discovered resting state EEG abnormality in semantic dementia. The notion that resting EEG in this dementia subtype is normal has to be revised.

Introduction

Semantic dementia (SD) is a variant of the Frontotemporal Lobar Degeneration (FTLD), a spectrum of non-Alzheimer’s dementias (Neary et al., 1998). It is characterized by a progressive language disorder with fluent, empty spontaneous speech, and loss of word meaning, commonly manifested by impaired naming and semantic paraphrases. Following this, SD has also been referred to as the semantic variant of primary progressive aphasia (Gorno-Tempini et al., 2011). According to the diagnostic guidelines of this clinical syndrome, it also encompasses a variety of symptoms such as loss of sympathy and empathy, and narrowed preoccupations (Snowden et al., 2001, Rankin et al., 2005).

Brain imaging methods, such as magnetic resonance imaging (MRI) and positron emission tomography (PET) show brain tissue atrophy and hypoperfusion in orbitofrontal and temporal lobes (Galton et al., 2001, Chan et al., 2002, Rosen et al., 2002, Diehl et al., 2004). Approximately 60% of the patients with SD show left lateralized atrophy and one third asymmetrical atrophy to the right hemisphere (Brambati et al., 2009). These regionally limited pathological changes circumscribe the unique character of this clinical syndrome. In particular, patients with left side atrophy show the typical semantic dysfunction, whereas right atrophied patients show dysfunctions of object recognition, or even the inability to recognize their family members’ faces (Gorno-Tempini et al., 2004).

Although MRI and PET provide invaluable insight in the neuropathology of SD, they are costly and invasive in the case of PET. In contrast, electroencephalography (EEG) is a non-invasive technique highly sensitive to changes in the functional state of the human brain. Due to its high temporal resolution, EEG is advantageous in the detection of rapidly and subtle changing brain activations as compared with other methods such as functional MRI (fMRI). Commonly, EEG recordings can be divided into two approaches. One approach is used to assess spontaneous brain activity where subjects are at awake rest, also referred as resting state. The other approach involves stimuli or tasks to evoke brain activity patterns, so-called event-related potentials (ERP). Contrasting varying task or stimulus conditions allows the investigation of specific perceptive or cognitive functions. Despite these advantages, resting EEG has neither been used to investigate the electrophysiological mechanisms of SD nor validated for a clinical purpose. Instead, the guidelines declare only that resting state EEG is normal in SD (Neary et al., 1998). This notion has not been changed in a more recent consensus on a classification of PPA including the semantic variant, where no EEG markers are reported (Gorno-Tempini et al., 2011). Although a couple of EEG studies were conducted with FTLD patients (Chan et al., 2002, Pijnenburg et al., 2008), little attention has been given to characterize resting EEG markers of SD. The opposite is the case for other dementias, especially for Alzheimer’s disease (AD). In particular, numerous studies have found changes in frequency resting state EEG power as compared to other dementias and healthy controls (Rice et al., 1990, Dierks et al., 1993, Ihl et al., 1993, Baiano et al., 2007, Schreiter Gasser et al., 2008).

In recent years, the methods to analyze electrophysiological signals have been profoundly extended. For example, the use of whole scalp electrode montages and thus increasing numbers of sensors has led to what can be referred to as topographical analyses. Topographic relates to its spatial characteristics, that is, changes of the amplitude configuration across electrodes over time. Concretely, a change of topography reflects a transformation of the activation of the neuronal network. In contrast, a shift in amplitude alone indicates that the number of involved neurons firing coherently alters (Koenig et al., 2002). Although topography changes occur in a time range of milliseconds (ms), it has been shown that momentary stable spatial patterns lasting approximately 100 ms can be observed (Lehmann, 1990, Koenig et al., 1999). One possibility to make such discrete, repeatedly appearing topographies or microstates more intelligible are pattern recognition algorithms (Lehmann, 1987). Microstate analyses have been used increasingly to quantify modulations of cognitive processes such as language and memory. For example, in ERP studies, subtle changes between stimuli conditions of certain tasks (e.g. correct vs. incorrect sentences, related vs. unrelated word pairs) were expressed as differences in microstate onset, offset, duration, number of presence, and topography (Brandeis et al., 1995, Wirth et al., 2007, Grieder et al., 2012). Moreover, microstate differences were found between patient and control groups in a study that investigated alterations of brain activity in patients with AD and SD (Grieder et al., 2013). These results thus demonstrate that EEG microstate analysis is a powerful tool for disentangling the brain dynamics that are related to external stimuli as well as cognitive disorders.

Furthermore, microstate analyses have served to characterize topographical patterns of the resting state EEG. The majority of the studies have clustered the resting EEG into four microstate classes, which has been found to be the optimal number according to the cross-validation criterion (Pascual-Marqui et al., 1995, Koenig et al., 2002, Britz et al., 2010). With this approach, Van de Ville et al. (2010) for instance succeeded to link the microstate classes to the resting state networks obtained with fMRI. Furthermore, Yuan et al. (2012) studied the relationship of simultaneous fMRI and EEG microstate using independent component analysis. Moreover, it has been shown that depending on the sleep stage, the topographies of the microstate classes differ slightly (Brodbeck et al., 2012). Alterations of resting state microstate duration as well as microstate transitions were found in patients with frontotemporal dementia, AD, and schizophrenia as compared to controls (Dierks et al., 1997, Kikuchi et al., 2007, Nishida et al., 2013). However, findings of resting EEG microstates in SD is missing and the notion that the resting EEG in SD is normal has not been investigated further. On the other hand, the literature strongly supports the assumption that if a cognitive state is altered, so should be the microstates. Consequently, one could expect different microstate parameters (e.g. topography, duration etc.) of patients with SD than of healthy controls, as an indication of semantic memory deficits seen in SD.

In accordance with this rationale, the aim of this study was to investigate resting state EEG microstates in patients with SD. One group of healthy elderly (HC) participants and one group of patients with AD served as control groups. We hypothesized that differences in microstates should be expected between SD and HC as well as AD. This assumption is based on the fact firstly that brain atrophy in SD differs from AD, which should lead to altered neuronal generator configurations. Secondly, patients with SD show a more severe semantic memory deficit as patients with AD, which has been found to alter the topography of ERP microstates in relation to a semantic task (Grieder et al., 2013). However, no specific hypotheses were made on which parameter was expected to be altered in SD compared with the control groups, as no data was available in the literature for this purpose.

Section snippets

Subjects

In the current study, 8 patients with semantic dementia (SD), 8 healthy elderly controls (HC), and 8 patients with Alzheimer’s disease (AD) were included, while one half of each group was assessed in Osaka, Japan and the other half in Stockholm, Sweden. The main reason for conflating the two-center data was the common difficulty of recruitment of patients with SD due to the low prevalence, difficult diagnosis, and high dropout rate as a consequence of severe cognitive impairment. In order to

Microstate analysis

The microstate k-means clustering analysis yielded four individual microstate classes as illustrated in Fig. 1 on the left side. Moreover, group mean microstate maps are depicted on the right side of Fig. 1.

Microstate statistics

The TANOVA including all three participant groups yielded significant group effects for microstate class B and C (see Fig. 1 on the right side). Topographies of microstate classes A and D did not differ between the groups. Post-hoc TANOVAs for microstate class B and C showed that SD differed

Discussion

The current study is the first that revealed deviations in resting EEG in semantic dementia. Consequently, the notion that resting EEG is normal in semantic dementia has to be revised. In particular, patients with semantic dementia showed altered resting state microstate topographies in class B and C, as compared with a healthy and an Alzheimer’s disease control group.

We hypothesized altered microstate classes in semantic dementia due to the deteriorated neurophysiology that leads to semantic

Acknowledgments

We thank Raffaella Crinelli for performing recruitment and neuropsychological assessment, Francisco Lacerda and Petter Kallioinen of the Department of Linguistics, Stockholm University, Stockholm, Sweden, for their support at the EEG lab. This study was supported by the Swiss Synapsis Foundation and the Swedish Alzheimerfonden.

Conflict of interest: The authors declare that there is no conflict of interest associated with this study.

References (47)

  • T. Koenig et al.

    Millisecond by millisecond, year by year: normative EEG microstates and developmental stages

    Neuroimage

    (2002)
  • K. Nishida et al.

    EEG microstates associated with salience and frontoparietal networks in frontotemporal dementia, schizophrenia and Alzheimer’s disease

    Clin Neurophysiol

    (2013)
  • Y.A. Pijnenburg et al.

    Investigation of resting-state EEG functional connectivity in frontotemporal lobar degeneration

    Clin Neurophysiol

    (2008)
  • U. Schreiter Gasser et al.

    Alzheimer disease versus mixed dementias: an EEG perspective

    Clin Neurophysiol

    (2008)
  • W.K. Strik et al.

    Three-dimensional tomography of event-related potentials during response inhibition: evidence for phasic frontal lobe activation

    Electroencephalogr Clin Neurophysiol

    (1998)
  • W.K. Strik et al.

    Data-determined window size and space-oriented segmentation of spontaneous EEG map series

    Electroencephalogr Clin Neurophysiol

    (1993)
  • J. Wackermann et al.

    Adaptive segmentation of spontaneous EEG map series into spatially defined microstates

    Int J Psychophysiol

    (1993)
  • H. Yuan et al.

    Spatiotemporal dynamics of the brain at rest–exploring EEG microstates as electrophysiological signatures of BOLD resting state networks

    NeuroImage

    (2012)
  • D. Brandeis et al.

    Mapping event-related brain potential microstates to sentence endings

    Brain Topogr

    (1995)
  • D. Chan et al.

    Differing patterns of temporal atrophy in Alzheimer’s disease and semantic dementia

    Neurology

    (2002)
  • Clinical and neuropathological criteria for frontotemporal dementia. The Lund and Manchester Groups

    J Neurol Neurosurg Psychiatry

    (1994)
  • T. Dierks et al.

    EEG-microstates in mild memory impairment and Alzheimer’s disease: possible association with disturbed information processing

    J Neural Transm

    (1997)
  • A. Du et al.

    Different regional patterns of cortical thinning in Alzheimer’s disease and frontotemporal dementia

    Brain

    (2007)
  • Cited by (38)

    • Event-related microstate dynamics represents working memory performance

      2022, NeuroImage
      Citation Excerpt :

      For example, several researchers have suggested that static microstate parameters illustrate the differences between groups or correlations based on cognitive function, diseases, or aging: mental arithmetic (Seitzman et al., 2017), major depressive disorder (Murphy et al., 2020), anxiety (Pan et al., 2020); ASD (D'Croz-Baron et al., 2019; Jia and Yu, 2019), schizophrenia (da Cruz et al., 2020), and Alzheimer's disease (AD; Grieder et al., 2016; Tait et al., 2020). Grieder et al. (2016) examined differences in the topographical map among healthy controls, patients with semantic dementia, and patients with AD. They confirmed significant cognitive functions (Boston Naming Test, Animal Fluency, and Verb Fluency) between healthy controls and patients and demonstrated significant differences in spatial patterns of classes B and C. Particularly for patients with semantic dementia, the topographical map of classes B and C were distributed from the right to left hemisphere symmetrically and from left frontal to right temporal regions, respectively.

    • EEG microstates as markers of major depressive disorder and predictors of response to SSRIs therapy

      2022, Progress in Neuro-Psychopharmacology and Biological Psychiatry
    • Altered EEG microstate dynamics in mild cognitive impairment and Alzheimer's disease

      2021, Clinical Neurophysiology
      Citation Excerpt :

      It has been found that temporal dynamics, especially in terms of duration, are the key characteristic of microstate (Van De Ville et al., 2010). Some studies showed no difference in microstate duration between AD and healthy controls (Nishida et al., 2013; Grieder et al. 2016), but most recent studies supported that the longer duration of microstate represents the more severe stage of cognitive impairment (Smailovic et al., 2019; Musaeus et al., 2020; Tait et al., 2020). Moreover, the sequence of microstates (‘microstate syntax’) proposed by Lehmann et al. (2005) opened a novel physiological comparison of mental operations between patients and controls.

    View all citing articles on Scopus
    View full text