Elsevier

Neuroscience Letters

Volume 696, 23 March 2019, Pages 28-32
Neuroscience Letters

Research article
Visibility graph analysis of speech evoked auditory brainstem response in persistent developmental stuttering

https://doi.org/10.1016/j.neulet.2018.12.015Get rights and content

Highlights

  • Visibility graph (VG) and fractality are used to evaluate the complexity of speech-evoked auditory brainstem response (s-ABR).

  • VG of the s-ABR is proposed to differentiate subjects with persistent developmental stuttering (PDS) from the normal group.

  • Differential complexities between normal and PDS subjects is quantified using Graph index complexity (GIC).

  • The model is applied to 14 individuals with PDS and 15 normal subjects.

  • The results reveal promising ability of GIC for assessment of abnormal activation of brainstem level in PDS.

Abstract

In this research, the concept of fractality based on nonlinear science and chaos theory is explored to study and evaluate the complexity of speech-evoked auditory brainstem response (s-ABR) time series in order to capture its intrinsic multiscale dynamics. The visibility graph of the s-ABR series is proposed as a quantitative method to differentiate subjects with persistent developmental stuttering (PDS) from the normal group. Differential complexities between normal and PDS subjects is quantified using Graph index complexity (GIC). The model is applied to 14 individuals with PDS and 15 normal subjects. The results reveal the promising ability of GIC for assessment of abnormal activation of brainstem level in PDS group. It is observed that all s-ABR series have visibility graphs with a power-law topology and fractality in the s-ABR series is dictated by a mechanism associated with long-term memory of the auditory system dynamics at the brainstem level.

Introduction

Stuttering is a disruption in speech categorized by verbal dysfluency such as interjection, repetition and prolongation of syllables and words during communication [41,62,79]. Recent neurological studies have focused on the underlying mechanisms of stuttering. Subjects with persistent developmental stuttering (PDS) have shown structural and functional abnormalities in the cortex based on Magnetoencephalography (MEG) and imaging studies such as an abnormal planum temporale asymmetry and different hemispheric speech-motor processing [36,44]. PDS is associated with the abnormal neural activities in subcortical lesions [58,69]. Thus, auditory brainstem response (ABR) is a suitable clinical tool to recognize brainstem timing deficits.

Graña et al. employ resting state effective connectivity to discriminate auditory hallucination [37]. The scalp-recorded ABR is an auditory evoked potential extracted from the electrical activity generated by a large population of neurons along the auditory brainstem and recorded by electrodes placed on the scalp usually in response to simple auditory sounds (e.g., tones, click) [20,22]. The ABR is a noninvasive clinical test used to evaluate the auditory brainstem function in various clinical applications such as developmental and learning disorders [13], determining hearing loss [74], hearing screening of newborn [60], intraoperative monitoring [40], auditory threshold estimation [85] and hearing aid fittings [54].

Recently, the response of the brainstem to more complex stimuli such as speech or music has been studied [19,45]. The s-ABR includes transient and steady-state responses. Transient and non-periodic features of the stimulus are reflected in the transient responses, while steady-state time-locked responses are observed in the sustained responses [47,48,75]. The s-ABR is an objective tool suitable for evaluation of developmental and learning disorders, auditory processing problems and auditory neuropathy [19,45].

In previous studies, different linear time and frequency methods have been applied to examine s-ABR time series including amplitude and peak latency, cross-correlation, Fourier analysis, and cross-phaseogram [72,75,76]. These methods, however, cannot extract the intricate details of the s-ABR signals adequately. Furthermore, most existing quantitative methods used to evaluate the dynamics of s-ABR series assume the stationarity of the s-ABR time series where the mean and variance of signal do not change with time [67]. Recent studies, however, have brought to attention the non-linear and non-stationary nature of s-ABR series, thus limiting the utility of the existing methods for examining complexity [63,71].

Using peak latency and Fourier analysis, Tahaei et al. [77] report significant differences for the onset and offset transient peaks in PDS compared with the control group. These methods, however, fail to reveal the exact temporal dynamic features of the underlying signals, thus motivating the authors to explore the application of nonlinear methods [56]. In recent years, researchers have employed signal processing, nonlinear science and chaos theory, and machine learning to develop powerful models for computer-aided diagnosis of various neurological and psychiatric disorders such as epilepsy [24,29,33,73], Autism Spectrum Disorder (ASD) [7,9,16,17], Attention Deficit Hyperactivity Disorder (ADHD) [8,12], Mild Cognitive Impairment (MCI) [3,53,55], the Alzheimer’s disease (AD) [6,15,18,39,42], Parkinson’s disease (PD) [84], Creutzfeld-Jakob disease [61], depression including Major Depressive Disorder (MDD) [2,5,11], and sleep studies [1,21,27], seizure detection [26], and seizure control [65]. Mozaffarilegha et al. used Ensemble Empirical Mode Decomposition (EEMD) to evaluate the chaotic features of s-ABR signals [63].

In order to examine the fractality and complexity in signals, a number of nonlinear methods have been proposed [51,80,83]. Among them is detrended fluctuation analysis (DFA) used to compute the Hurst exponent in order to identify the fractality and the long-range correlation behavior in different signals [68]. Long-range correlation or long-range dependence is a phenomenon arising in the analysis of time series. It shows how the statistical correlation between two points decreases when the time intervals increase. On the other hand, a visibility graph (VG) is an alternative approach to calculating the Hurst exponent [49].

In recent years, graph theory has been used to model and study brain connectivity [34,59,82]. The VG can convert fractal signals into scale-free graphs [50]. The degree of fractality of the signals can be identified using the degree of scale-freeness. It has been demonstrated that different biological time series including electrocardiography (ECG) [14,35] and electroencephalography (EEG) [4,32,78,80] behave as scale-invariant processes. Ahmadlou and Adeli present visibility graph similarity as a new way of measuring complexity in coupled dynamic systems and to evaluate brain activities of neurodegenerative disorders such as autism [4].

In this research, VG and the concept of fractality based on nonlinear science and chaos theory is explored to study and evaluate the complexity of the s-ABR time series in order to capture its intrinsic multiscale dynamics. To the best of our knowledge, the VG has not been used on the s-ABR series in the literature. The VG of the s-ABR series is proposed to differentiate adults with PDS from the control group. Differential complexity between normal and PDS subjects is quantified using Graph index complexity (GIC). The model is applied to 14 individuals with PDS and 15 normal subjects.

Section snippets

Participants

This experiment was performed and data were collected by the first author under a protocol approved by the review board of the Iran University of Medical Sciences. All participants in this study signed the informed consent. The data were collected from 14 right-handed individuals with PDS (6 women and 8 men), aged 15–33 years (mean ± SD = 26.50 ± 1.04). A speech-language pathologist made clinical diagnoses of developmental stuttering. To evaluate the severity of stuttering, we used the Persian

Results

The research was performed on an Intel(R) Core(TM) i5 CPU, M 450 @ 2.40 GHz with microcode of 0 × 1, 2400.000 cpu MHz and 3072 KB cache size. The MATLAB R2016b was used to implement the methodology. The CPU computation time was 0.09 s.

Fig. 5 shows the Log–Log plot of the cumulative degree distribution p(k) of the series of s-ABR for healthy and PDS subjects, where k indicates the order of a node, i.e. the number of edges connected to a node, P is the probability distribution of edges

Discussions

In this study, we have examined the large-scale dynamics of neural oscillations in the auditory system in the brainstem level of normal and subjects with PDS. The long-range power-law correlations have been found in a wide range of complex biological systems, including heart rate [14], medullary sympathetic neurons in neurophysiology [70] and long memory in human coordination [30] exhibiting the fractal dynamics of biological systems. On the other hand, recent studies have discovered long-range

Conclusion

In this paper, the VG algorithm is used to quantify the long-range correlations in s-ABR signals. The s-ABR series of PDS group were compared with normal subjects. The results reveal the promising ability of GIC for assessment of abnormal activation of brainstem level in PDS group. The GIC values of subjects with stuttering are higher than the normal group confirmed the evidence of insufficiency of the cortical activities in PDS subjects. It is also observed that the VGs of all s-ABR series

References (86)

  • J. Kim et al.

    What is a complex graph?

    Phys. A: Stat. Mech. Appl.

    (2008)
  • A. Koravand et al.

    Speech-evoked auditory brainstem responses in children with hearing loss

    Int. J. Pediatr. Otorhinolaryngol.

    (2017)
  • N. Kraus et al.

    Brainstem origins for cortical ‘what’ and ‘where’ pathways in the auditory system

    Trends Neurosci.

    (2005)
  • A. Krishnan

    Human frequency-following responses: representation of steady-state synthetic vowels

    Hear. Res.

    (2002)
  • L. Parish et al.

    Long-range temporal correlations in epileptogenic and non-epileptogenic human hippocampus

    Neuroscience

    (2004)
  • N. Russo et al.

    Brainstem responses to speech syllables

    Clin. Neurophysiol.

    (2004)
  • E. Skoe et al.

    Cross-phaseogram: objective neural index of speech sound differentiation

    J. Neurosci. Methods

    (2011)
  • L. Wang et al.

    EEG analysis of seizure patterns using visibility graphs for detection of generalized seizures

    J. Neurosci. Methods

    (2017)
  • R. Yuvaraj et al.

    Brain functional connectivity patterns for emotional state classification in Parkinson’s disease patients without dementia

    Behav. Brain Res.

    (2016)
  • U.R. Acharya et al.

    Nonlinear dynamics measures for automated EEG-based sleep stage detection

    Eur. Neurol.

    (2015)
  • U.R. Acharya et al.

    A novel depression diagnosis index using nonlinear features in EEG signals

    Eur. Neurol.

    (2015)
  • M. Ahmadlou et al.

    Fractality and a wavelet-chaos-methodology for EEG-based diagnosis of Alzheimer disease

    Alzheimer Dis. Assoc. Disord.

    (2011)
  • M. Ahmadlou et al.

    Graph theoretical analysis of organization of functional brain networks in ADHD

    Clin. EEG Neurosci.

    (2012)
  • M. Ahmadlou et al.

    New diagnostic EEG markers of the Alzheimer’s disease using visibility graph

    J. Neural Transm.

    (2010)
  • M. Ahmadlou et al.

    Spatiotemporal analysis of relative convergence of EEGs reveals differences between brain dynamics of depressive women and men

    Clin. EEG Neurosci.

    (2013)
  • M.A. Ahmed et al.

    Learning disabilities in different types of attention deficit hyperactivity disorders and its relation to cortical and brainstem function

    J. Neurol. Res.

    (2014)
  • A. Bhaduri et al.

    Quantitative assessment of heart rate dynamics during meditation: an ECG based study with multi-fractality and visibility graph

    Front. Physiol.

    (2016)
  • S. Bhat et al.

    Autism: cause factors, early diagnosis and therapies

    Rev. Neurosci.

    (2014)
  • S. Bhat et al.

    Automated diagnosis of autism: in search of a mathematical marker

    Rev. Neurosci.

    (2014)
  • S. Bhat et al.

    Clinical neurophysiological and automated EEG-based diagnosis of the Alzheimer’s disease

    Eur. Neurol.

    (2015)
  • G. BinKhamis et al.

    Methodological approaches to recording speech auditory brainstem responses: effect of stimulus duration, background, consonant, and number of repetitions

    J. Hear. Sci.

    (2017)
  • J. Boston

    Noise cancellation for brainstem auditory evoked potentials

    IEEE Trans. Biomed. Eng.

    (1985)
  • J.C. Bruder et al.

    Physiological ripples associated with sleep spindles differ in waveform morphology from epileptic ripples

    Int. J. Neural Syst.

    (2017)
  • R.F. Burkard et al.

    Auditory Evoked Potentials: Basic Principles and Clinical Application

    (2007)
  • Y.J. Cha et al.

    Deep learning-based crack damage detection using convolutional neural networks

    Comput.-Aided Civ. Infrastruct. Eng.

    (2017)
  • D. Cogan et al.

    Multi-biosignal analysis for epileptic seizure monitoring

    Int. J. Neural Syst.

    (2017)
  • S. Corbera et al.

    Abnormal speech sound representation in persistent developmental stuttering

    Neurology

    (2005)
  • T. De Cooman et al.

    Online automated seizure detection in temporal lobe epilepsy patients using single-lead ecg

    Int. J. Neural Syst.

    (2017)
  • A. Dereymaeker et al.

    An automated quiet sleep detection approach in preterm infants as a gateway to assess brain maturation

    Int. J. Neural Syst.

    (2017)
  • B. Direito et al.

    A realistic seizure prediction study based on multiclass SVM

    Int. J. Neural Syst.

    (2017)
  • S.W. Ducharme et al.

    Fractal dynamics, variability, and coordination in human locomotion

    Kinesiol. Rev.

    (2018)
  • A. Fernández et al.

    A Pareto-based ensemble with feature and instance selection for learning from multi-class imbalanced datasets

    Int. J. Neural Syst.

    (2017)
  • Z.-K. Gao et al.

    Visibility graph from adaptive optimal kernel time-frequency representation for classification of epileptiform EEG

    Int. J. Neural Syst.

    (2017)
  • Cited by (25)

    • Modeling the efficacy of different anti-angiogenic drugs on treatment of solid tumors using 3D computational modeling and machine learning

      2022, Computers in Biology and Medicine
      Citation Excerpt :

      The current research focuses on the three-dimensional simulation of tumor growth, with and without a prescription, to identify the differential benefits of the anti-angiogenic drugs bevacizumab, ranibizumab, and brolucizumab and their efficacy in inhibiting tumor growth. Vilanova et al. [16] studied a combination of computational and imaging models of tumor-induced angiogenesis using graph theory [17–19] to overcome some challenges in tumor growth treatment. Additionally, Tang et al. [20] developed pressure-based 3D computational modeling of tumor growth and angiogenesis to evaluate chemotherapy drug responses.

    • Study of the brainstem auditory evoked potential with speech stimulus in the pediatric population with and without oral language disorders: a systematic review

      2020, Brazilian Journal of Otorhinolaryngology
      Citation Excerpt :

      Studies agree that these difficulties can compromise language development skills and interfere with the individual’s social communication. The application of BAEP-speech in the population with persistent developmental stuttering was identified only in the study of Mozaffarilegha et al.,19 which used graphic visibility and fractality to observe the complexity of the responses to the exam in individuals with oral language pathology and with typical development. The study group showed greater complexity in the graphic index, compared to the group without alterations.

    • Collaborative passenger flow control for an urban rail transit network

      2024, Computer-Aided Civil and Infrastructure Engineering
    View all citing articles on Scopus
    View full text