Elsevier

Neurocomputing

Volume 347, 28 June 2019, Pages 230-239
Neurocomputing

Brain network analysis for auditory disease: A twofold study

https://doi.org/10.1016/j.neucom.2019.04.013Get rights and content

Abstract

Tinnitus is the phantom perception of sound and can have negative effect on the quality of life. Sudden deafness is generally defined as sensorineural hearing loss of 30 dB or greater over at least three contiguous audiometric frequencies and within a three-day period, which is usually unilateral and can be associated with tinnitus. The exact pathogenesis of such auditory diseases is still unknown, and alterations in the functional connectivity of the brain network are suspected to involve one possible pathogenesis. Scalp electroencephalogram (EEG) analysis is a common approach to capture brain neural activity. This paper conducts a twofold study in terms of brain network analysis for auditory diseases based on EEG data. Specifically, the brain network is constructed in view of the phase lag index (PLI) between signals. The unsupervised study is first conducted via the unsupervised network community detection algorithm. The goal is to reveal the underlying characteristics of functional connectivity concerning subjects with and without auditory diseases from the brain network perspective. The second study is the supervised learning. A new strategy termed FBA (Features from Brain Areas) is proposed to abstract each brain network as a feature vector and the supervised classification method, i.e., Support Vector Machine (SVM) is utilized to perform the three-class classification concerning the regular control group, tinnitus patient group and sudden deafness patient group. Extensive experiments have been conducted on totally 146 subjects including 41 regular controls, 51 sudden deafness patients and 54 tinnitus patients. The results show the necessary of the unsupervised study that may shed light on the pathogenesis and the classification performed in the supervised study may serve as a portable detection method for auditory diseases.

Introduction

Tinnitus is the phantom perception of sound [1] often associated with emotional symptoms such as low-mood, annoyance, anger, anxiety, depression, reduced attention, and poor sleep [2], [3], [4]. The tinnitus signal may arise from compensatory mechanisms following sensory deafferentation, with cochlear hair cell loss resulting in reduced spontaneous firing rate of the auditory nerve fibers [5]. Reduction in afferent activity may lead to hyperactivity in the auditory cortex through reductions in cortical inhibition [6], [7]. Emotion and other residuals have been linked to the various cortical and subcortical networks of the brain [2]. On the other hand, sudden deafness or sudden sensorineural hearing loss is generally defined as sensorineural hearing loss of 30 dB or greater over at least three contiguous audiometric frequencies and within a three-day period [8]. The exact pathogenesis of such auditory diseases is still unknown. The alterations in the functional connectivity of the brain network are suspected to involve one possible pathogenesis [9]. Thus, brain network analysis can be helpful to reveal the underlying mechanisms of such auditory diseases. This is because examining the human brain as an integrative network of functionally interacting brain regions can provide new insights about neuronal communication in the human brain. It also provides a platform to examine how functional connectivity and information integration relate to human behavior and how this organization may be altered in some auditory diseases caused by neurodegeneration [10], [11]. In the meanwhile, electroencephalogram (EEG) is one of the most effective methods to observe brain neural activities [12], [13], [14], [15], [16].

A lot of studies on auditory diseases have been made in the last decade. The most common denominator for these studies is the goal of elucidating the underlying neural mechanisms of auditory diseases with the ultimate purpose of finding a cure [2], [9], [17], [18], [19], [20], [21]. Transcranial direct current stimulation (tDCS) was identified to be an effective research tool for transient tinnitus neuromodulation [2]. Henry et al. made a review of the underlying mechanisms of tinnitus [17]. A conclusion has been drawn that long-term maintenance of tinnitus is likely a function of a complex network of structures involving central auditory and nonauditory systems. Behavioral, electrophysiological, functional magnetic resonance imaging (fMRI) techniques were used to identify the tinnitus-hyperacusis network. In [18], Chen et al. proposed a testable model accounting for distress, arousal, and gating of tinnitus and hyperacusis. The graph theoretical network analysis method was used to detect brain connectome alterations in unilateral sudden sensorineural hearing loss (SSNHL) [9]. It suggested that the alteration of network organization already exists in unilateral sudden sensorineural hearing loss patients within the acute period. It also suggested that the functional connectome of unilateral SSNHL patients is characterized by a shift toward small-worldization. Chen et al. concluded that both cochlear and vestibular endorgans/afferents were identified to be severely affected bilaterally by the vestibular test battery and resulted in poor hearing outcome [19]. However, the above mentioned studies have neither taken into account the functional connectivity of the brain from the viewpoint of network nor analyzed the brain network through the general structural measurements such as modularity.

In this paper, we first construct the brain network based on the EEG data in view of the phase lag index (PLI) between signals. And then a twofold study is conducted from the perspectives of unsupervised study and supervised study. Specifically, the unsupervised study is conducted by means of the unsupervised network community detection algorithm, where the structural measurement, i.e., modularity of the network is calculated and compared among brain networks. This unsupervised study may help to reveal the underlying characteristics of brain networks of the subjects who suffer from auditory diseases. And then the supervised study is conducted, where each brain network is abstracted as a feature vector through the proposed strategy termed FBA (Features from Brain Areas). And one typical supervised classification method, i.e., Support Vector Machine (SVM) is utilized to perform the three-class classification that makes a distinction among subjects with sudden deafness, tinnitus and regular controls. Extensive experiments have been conducted on totally 146 subjects including 41 regular controls, 51 sudden deafness patients and 54 tinnitus patients. The results show the necessary of the unsupervised study that may shed light on the pathogenesis. And the classification performed in the supervised study may serve as a portable detection method for auditory diseases.

The rest of the paper is organized as follows. In Section 2, we describe in detail the subjects involved in the experiment, the EEG data collection, the brain network construction and the methods used in the twofold study. Comprehensive results and the corresponding analysis are presented in Section 3. Section 4 makes a conclusion of this paper.

Section snippets

Subjects

146 subjects participate in our experiments, including 41 regular controls, 51 sudden deafness patients and 54 tinnitus patients. The subjects are selected by both taking psychological questionnaires and being diagnosed by doctors from Sun Yat-sen Memorial Hospital. None of the subjects use medications that are expected to influence EEG signals. Informed written consents are obtained from all the subjects prior to the recordings.

EEG data collection

Resting state EEG data are collected in the experiments. All the

Results and analysis

Extensive experiments are conducted on the EEG data collected from Sun Yat-sen Memorial Hospital. Detailed results and analysis in terms of both unsupervised study and supervised study will be reported in this section.

Conclusion

In this paper, we perform a twofold study in terms of the brain network analysis for auditory diseases. Specifically, the brain network is constructed based on EEG data in view of the phase lag index (PLI) between signals. The unsupervised study is first conducted, where the unsupervised network community detection algorithm is utilized and the analysis is given at both subject level and group level. This unsupervised study aims at finding the underlying characteristics of brain networks of

Acknowledgments

This project was supported by NSFC (61876193), Guangdong Natural Science Funds for Distinguished Young Scholar (2016A030306014), and Tip-top Scientific and Technical Innovative Youth Talents of Guangdong special support program (2016TQ03X542). The authors would like to thank Department of Otolaryngology of Sun Yat-sen Memorial Hospital, Sun Yat-sen University for help collecting EEG data.

Pei-Zhen Li received her bachelor’s degree in 2017 from Sun Yat-sen University. She is a graduate student at Sun Yat-sen University from Sept. 2017. Her current research interests include graph clustering and medical data mining.

References (36)

  • L. Koessler et al.

    Automated cortical projection of EEG sensors: anatomical correlation via the international 10–10 system

    Neuroimage

    (2009)
  • L.R. Peraza et al.

    Volume conduction effects in brain network inference from electroencephalographic recordings using phase lag index

    J. Neurosci. Methods

    (2012)
  • M.P. Van Den Heuvel et al.

    Exploring the brain network: a review on resting-state fMRI functional connectivity

    Eur. Neuropsychopharmacol.

    (2010)
  • S. Maldonado et al.

    Profit-based feature selection using support vector machines – general framework and an application for customer retention

    Appl. Soft Comput.

    (2015)
  • G.S. Shekhawat et al.

    Modulation of perception or emotion? A scoping review of tinnitus neuromodulation using transcranial direct current stimulation

    Neurorehabil. Neural Repair

    (2015)
  • TsaiM.-C. et al.

    Tinnitus abnormal brain region detection based on dynamic causal modeling and exponential ranking

    Biomed. Res. Int.

    (2018)
  • SunZ.-R. et al.

    Multi-view intact space learning for tinnitus classification in resting state EEG

    Neural Process. Lett.

    (2019)
  • L. Rüttiger et al.

    The reduced cochlear output and the failure to adapt the central auditory response causes tinnitus in noise exposed rats

    PLoS One

    (2013)
  • Pei-Zhen Li received her bachelor’s degree in 2017 from Sun Yat-sen University. She is a graduate student at Sun Yat-sen University from Sept. 2017. Her current research interests include graph clustering and medical data mining.

    Ling Huang received her undergraduate and master degree in 2009 and 2013, respectively, from South China University of Technology. She is currently working toward the Ph.D. degree at Sun Yat-sen University. She has published several papers in international journals and conferences such as Pattern Recognition, Information Sciences, AAAI, IEEE ICDM, IEEE BIBM and DASFAA. Her research interest is data mining.

    Chang-dong Wang received the Ph.D. degree in computer science in 2013 from Sun Yat-sen University, Guangzhou, China. He is a visiting student at University of Illinois at Chicago from Jan. 2012 to Nov. 2012. He joined Sun Yat-sen University in 2013 as an assistant professor with School of Mobile Information Engineering and now he is currently an associate professor with School of Data and Computer Science. His current research interests include machine learning and data mining. He has published over 100 scientific papers in international journals and conferences such as IEEE TPAMI, IEEE TKDE, IEEE TCYB, IEEE TSMC-C, Pattern Recognition, Neurocomputing, AAAI, ICDM, CIKM, SDM, BIBM and DASFAA. His ICDM 2010 paper won the Honorable Mention for Best Research Paper Awards. He won 2012 Microsoft Research Fellowship Nomination Award. He was awarded 2015 Chinese Association for Artificial Intelligence (CAAI) Outstanding Dissertation.

    Chuan Li received his Ph.D. in data mining area in June 2006 from Sichuan University. He is currently an associate professor and vice dean of database and knowledge engineering institute of computer school, Sichuan University. His main research interests include the mining of information networks, neural networks, biological computing and data warehousing. He works actively in academic serving and events. He also works for local government of Guizhou Province focusing on the advisory and policy making of data mining industry.

    JIAN-HUANG LAI received his M.Sc. degree in applied mathematics in 1989 and his Ph.D. in mathematics in 1999 from Sun Yat-sen University, China. He joined Sun Yat-sen University in 1989 as an Assistant Professor, where currently, he is a Professor with School of Data and Computer Science. His current research interests are in the areas of pattern recognition, data mining, and computer vision. He has published over 200 scientific papers in the international journals and conferences on image processing and pattern recognition, e.g. IEEE TPAMI, IEEE TCYB, IEEE TKDE, IEEE TNN, IEEE TIP, Pattern Recognition, ICCV, CVPR and ICDM.

    View full text