Elsevier

Clinical Neurophysiology

Volume 131, Issue 2, February 2020, Pages 437-445
Clinical Neurophysiology

Complexity changes in preclinical Alzheimer’s disease: An MEG study of subjective cognitive decline and mild cognitive impairment

https://doi.org/10.1016/j.clinph.2019.11.023Get rights and content

Highlights

  • Brain complexity is decreased in the early stages of Alzheimer's disease (AD), even before dementia onset.

  • First source-level complexity analysis revealed impairment in key AD areas such as precuneus.

  • Complexity alterations were related to hippocampal atrophy, a hallmark of AD neurodegeneration.

Abstract

Objective

To analyse magnetoencephalogram (MEG) signals with Lempel-Ziv Complexity (LZC) to identify the regions of the brain showing changes related to cognitive decline and Alzheimeŕs Disease (AD).

Methods

LZC was used to study MEG signals in the source space from 99 participants (36 male, 63 female, average age: 71.82 ± 4.06) in three groups (33 subjects per group): healthy (control) older adults, older adults with subjective cognitive decline (SCD), and adults with mild cognitive impairment (MCI). Analyses were performed in broadband (2–45 Hz) and in classic narrow bands (theta (4–8 Hz), alpha (8–12 Hz), low beta (12–20 Hz), high beta (20–30 Hz), and, gamma (30–45 Hz)).

Results

LZC was significantly lower in subjects with MCI than in those with SCD. Moreover, subjects with MCI had significantly lower MEG complexity than controls and SCD subjects in the beta frequency band. Lower complexity was correlated with smaller hippocampal volumes.

Conclusions

Brain complexity – measured with LZC – decreases in MCI patients when compared to SCD and healthy controls. This decrease is associated with a decrease in hippocampal volume, a key feature in AD progression.

Significance

This is the first study to date characterising the changes of brain activity complexity showing the specific spatial pattern of the alterations as well as the morphological correlations throughout preclinical stages of AD.

Introduction

Cognitive health, and its evolution with age, is an area of intense research, with one of the main drivers of this research being the increase in life expectancy in our modern day. Brain health is a difficult term to define as it covers a wide range of aspects, nevertheless, the Healthy Aging Research Network which is part of the Center for Disease Prevention and Control (CDC) in the United States has defined a healthy brain with preserved cognition as “one that can perform all the mental processes that are collectively known as cognition, including the ability to learn new things, intuition, judgement, language and remembering” (Centers for Disease Control and Prevention and Healthy Aging, 2007).

Ageing is typically associated with a decline in cognitive ability (Ardila et al., 2000). For some of the older adults this decline could be accompanied with a strong feeling of a more accentuated cognitive worsening compared to their age counterparts. This phenomenon is known as Subjective Cognitive Decline (SCD), and it could represent one of the earliest indicators of early neurodegeneration (Jessen et al., 2015). Recent studies into the development of dementia, particularly Alzheimer’s Disease (AD), have suggested that there may be a link between SCD and AD (Rabin et al., 2017). This is partly due to the deviations from healthy ageing that have been observed in SCD patients as well as the similarities of the symptoms of SCD and AD. Nevertheless, although some studies such as those performed by López-Sanz et al. (2017b), and Rabin et al. (2017) have highlighted the subtle presence of AD biomarkers in SCD, at this stage SCD is currently largely undetectable by standard neuropsychological tests as its effects can be compensated (Erk et al., 2011).

The continued decline in cognitive ability can result in the development of MCI. MCI is a state characterised by a significant reduction in cognitive abilities affecting one or more domains. However, MCI patients cannot be classified as having dementia, as their independent functioning is still preserved (Petersen, 2004). As cognitive abilities continue to decline, studies have shown that MCI individuals are at a higher risk of suffering from AD than healthy individuals (Petersen and Negash, 2008, Landau et al., 2010, Petersen, 2011). Therefore, due to the numerous similarities between MCI and AD (Fischer et al., 2007, Vos et al., 2013), MCI and SCD (López-Sanz et al., 2017a), as well as SCD and AD (Rabin et al., 2017), it becomes important to investigate the progression and early onset of AD as it is the most prevalent form of dementia (accounting for 60–80% of reported cases (Barnes and Yaffe, 2011)). Additionally, there is a growing socioeconomic demand to provide care for fully demented patients. Thus, it is necessary to understand the development of cognitive decline in its early stages to develop a long-lasting solution to successfully combating dementia. Overtime these solutions can result in robust, efficient and successful early diagnosis and, therefore, will ease the burden on the socioeconomic, healthcare and scientific communities. Hence, it is with this in mind that this investigation was performed.

Magnetoencephalography is a non-invasive analysis technique used to record the magnetic fields generated by electrical activity in the human brain in a reference free manner (Gomez et al., 2008, Escudero et al., 2009, Jafarpour et al., 2013). The use of magnetoencephalograms (MEGs) to study the background activity of the brain has increased over the years due its advantages, including good temporal resolution, an adequate spatial resolution, particularly when combined with an individual MRI, good signal-to-noise ratio, and relatively robust against volume conduction when compared to other measures such as electroencephalography (Stam et al., 2007, Stam et al., 2016. Moreover, due to these advantages, the use of magnetoencephalography to investigate the early stages of AD has been recommended by the International Working Group (IWG) and the American Alzheimer's Association (Dubois et al., 2016). Therefore, in this study MEG recordings were used to investigate the changes in brain activity associated with cognitive decline.

Non-linear signal processing methods have successfully been used to investigate changes in the electrical and magnetic activity of the brain associated with various aspects such as age (Lutzenberger et al., 1995, Meyer-Lindenber, 1996, Shumbayawonda et al., 2017), sex (Li et al., 2014), and pathology (Li et al., 2014). One family of non-linear analysis methods focusses on the characterisation of complexity (Tononi et al., 1998). Complexity is a concept that can be used to assess changes in MEG brain activity to identify the direct/indirect effects of cortical functional organization of the brain (Li et al., 2014). Thus, complexity could be a robust indicator of the variances in cortical neuronal interaction as changes in complexity of the MEG signal may be coarsely related to the changes in the regularity of the interactions occurring within the neuronal network. It is also worth noting that, among the numerous complexity estimation techniques, those based on symbolic dynamic analysis are arguably the most computationally efficient. Lempel-Ziv Complexity (LZC) is a robust symbolic dynamic complexity technique that has been used to successfully investigate changes in electrophysiological activity associated with age (Li et al., 2014), AD (Gómez et al., 2006), schizophrenia (Fernández et al., 2011) or changes associated with cognitive decline (Fernández et al., 2012). LZC is a non-parametric method that can be used for the robust estimation of complexity of short time series and reflects the number of distinct substrings along a given sequence as well as their rate of recurrence (Fernandez et al., 2012). The LZC algorithm reflects the magnitude of the signal points and is frequency sensitive (Lempel and Ziv, 1976). Therefore, more complex or irregular time series would have greater LZC values than those featuring more regular patterns (Fernandez et al., 2012). In addition to this, when compared to other symbolic dynamic measures, LZC does not require the time series being analysed to be stationary. This is, therefore, a significant advantage when studying physiological signals (Shumbayawonda et al., 2018) and represents a major advantage of this metric. Nevertheless, previous MEG studies using LZC have analysed signals in the sensor space, but not in source space. The current study tries to address this limitation of previous work by extending the analysis of the complexity of MEG signals to the source space. It was hypothesised that LZC analysis in source space would reveal more specific information about the changes in MEG activity associated with cognitive decline in the preclinical stages of AD.

Section snippets

Subjects and diagnosis criteria

Subjects were volunteers from three centres, namely the: Department of Neurology from Hospital Universitario San Carlos, Centre for Prevention of Cognitive Impairment, and Seniors Centre of Chamartin District in Madrid (Spain). A total of 99 subjects MEG recordings (36 males, 63 female) with an average age of 71.82 ± 4.06 made up the dataset used in this study as shown in Table 1. All subjects were age-matched, with no significant differences between the genders. All the subjects signed an

Results

Statistical analysis using whole-brain broadband mean group LZC values for all the sources showed that only the differences between SCD and MCI, (SCD-MCI) were statistically significant (p = 0.02), while the differences between controls and MCI groups (CN-MCI) and controls and SCD groups (CN-SD) were not (p = 0.109 and p = 0.468, respectively),. When the average complexity values for each group were plotted on the same graph (see Fig. 1), it was found that there was an inverted U relationship

Discussion

The aim of this study was to evaluate changes between the brain activity from healthy controls, SCD and MCI recorded in MEG signals with LZC, a non-linear algorithm that quantifies complexity in time series. Our study found, for the first time in source space, significant alterations in non-linear brain dynamics in MCI patients. Particularly, MCI patients exhibited decreased broadband LZC with respect to subjects with SCD, who interestingly did not show significant alterations in these metrics.

Conclusions

In conclusion, results showed that MCI was associated with a significant decrease in complexity (measured by means of LZC) when compared to controls and subjects with SCD. Moreover, these changes in LZC values were significantly correlated to decreases in right hippocampal volumes. Therefore, the outcomes from this study show, for the first time, the relationship between source space complexity and changes in prodromal stages of Alzheimer’s disease.

Funding

This study was supported by a project from the Spanish Ministry of Economy and Competitiveness, PSI2009-14415-C03-01 and a project by Comunidad de Madrid, B2017/BMD-3760 (NEUROCENTRO-CM). Elizabeth Shumbayawonda was supported by a Santander Universities Postgraduate Research Award.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (74)

  • C.R. Jack et al.

    Biomarker modeling of Alzheimer’s disease

    Neuron

    (2013)
  • S. Liu et al.

    Abnormal EEG complexity and functional connectivity of brain in patients with acute thalamic ischemic stroke

    Comput Math Methods M

    (2016)
  • E. Maris et al.

    Nonparametric statistical testing of EEG- and MEG-data

    J Neurosci Methods

    (2007)
  • J. Peter et al.

    Gray matter atrophy pattern in elderly with subjective memory impairment

    Alzheimers Dement

    (2014)
  • O. Sporns et al.

    Organization, development and function of complex brain network

    Trends Cogn Sci

    (2004)
  • C.J. Stam et al.

    The relation between structural and functional connectivity patterns in complex brain networks

    Int J Psychophysiol

    (2016)
  • G. Tononi et al.

    Complexity and coherency: integrating information in the brain

    Trends Cogn Sci

    (1998)
  • J.A. Yesavage et al.

    Development and validation of a geriatric depression screening scale: A preliminary report

    J Psychiatr Res

    (1982)
  • D. Abasolo et al.

    Lempel-Ziv complexity of cortical activity during sleep and waking in rats

    J Neurophysiol

    (2015)
  • M. Aboy et al.

    Interpretation of Lempel ziv complexity measure in the context of biomedical signal analysis

    IEEE Trans Bio-Med Eng

    (2006)
  • L.G. Apostolova et al.

    Conversion of mild cognitive impairment to Alzheimer disease predicted by hippocampal atrophy maps

    Arch Neurol

    (2006)
  • V. Berti et al.

    Early detection of Alzheimer's disease with PET imaging

    Neurodegener Dis

    (2010)
  • R.L. Buckner et al.

    Cortical hubs revealed by intrinsic functional connectivity: mapping, assessment of stability, and relation to Alzheimer's disease

    J Neurosci

    (2009)
  • J.L. Cantero et al.

    Regional hippocampal atrophy and higher levels of plasma amyloid-beta are associated with subjective memory complaints in nondemented elderly subjects

    J Gerontol A Biol Sci Med Sci

    (2016)
  • Centers for Disease Control and Prevention and Healthy Aging, 2007. What is a Healthy Brain? New Research Explores...
  • Y. Cheng et al.

    From mild cognitive impairment to subjective cognitive decline: conceptual and methodological evolution

    Neuropsychiatr Dis Treat

    (2017)
  • J. Dauwels et al.

    Diagnosis of Alzheimer's disease from EEG signals: where are we standing?

    Curr Alzheimer Res

    (2010)
  • S.I. Dimitriadis

    Complexity of brain activity and connectivity in functional neuroimaging

    J Neurosci Res

    (2018)
  • S. Erk et al.

    Evidence of neuronal compensation during episodic memory in subjective memory impairment

    Arch Gen Psychiatry

    (2011)
  • A. Fernández et al.

    Complexity and schizophrenia

    Prog Neuro-Psychopharmacol Biol Psychiatry

    (2012)
  • P. Fischer et al.

    Conversion from subtypes of mild cognitive impairment to Alzheimer dementia

    Neurology

    (2007)
  • N.C. Fox et al.

    Presymptomatic hippocampal atrophy in Alzheimer's disease. A longitudinal MRI study

    Brain

    (1996)
  • P. Garcés et al.

    Choice of magnetometers and gradiometers after signal space separation

    Sensors

    (2017)
  • T. Gili et al.

    Regional brain atrophy and functional disconnection across Alzheimer's disease evolution

    J Neurol Neurosurg Psychiatry

    (2011)
  • C. Gomez et al.

    Nonlinear forecasting measurement of magnetoencephalogram recordings from alzheimers disease patients

    (2008)
  • M. Grundman et al.

    Mild cognitive impairment can be distinguished from Alzheimer disease and normal aging for clinical trials

    Arch Neurol

    (2004)
  • A. Husseen Al-Nuaimi et al.

    Complexity measures for quantifying changes in electroencephalogram in Alzheimer’s disease

    Complexity

    (2018)
  • Cited by (19)

    • Efficacy of preclinical pharmacological interventions against alterations of neuronal network oscillations in Alzheimer's disease: A systematic review

      2021, Experimental Neurology
      Citation Excerpt :

      As for such functional dynamics, there is substantial evidence demonstrating that neuronal oscillatory activity, particularly θ and γ brain rhythms, is essential for memory function (Buzsáki, 2002; Kim et al., 2020). Furthermore, compounding evidence indicates that changes in brain rhythm patterns may be a suitable early (functional) biomarker in AD (Ahnaou et al., 2017; Babiloni et al., 2013b; Goutagny et al., 2013; Hatz et al., 2015; Kitchigina, 2018; Montez et al., 2009; Nakamura et al., 2017; Shumbayawonda et al., 2020). In addition, there is a correlation between the ApoE phenotype, alterations in electroencephalographic connectivity, and the prediction of disease advancement from mild cognitive impairment to AD (Vecchio et al., 2018), indicating that the disruption of cerebral rhythms is a main characteristic of AD in addition to the canonical neuropathological biomarkers (Triggiani et al., 2017).

    View all citing articles on Scopus
    1

    These authors have contributed equally and thus share first authorship.

    View full text