Clustering Brain Signals: a Robust Approach Using Functional Data Ranking
- PDF / 3,858,249 Bytes
- 18 Pages / 439.642 x 666.49 pts Page_size
- 34 Downloads / 207 Views
Clustering Brain Signals: a Robust Approach Using Functional Data Ranking Tianbo Chen1 · Ying Sun1 · Carolina Euan1 · Hernando Ombao1 Accepted: 30 October 2020 © The Classification Society 2020
Abstract In this paper, we analyze electroencephalograms (EEGs) which are recordings of brain electrical activity. We develop new clustering methods for identifying synchronized brain regions, where the EEGs show similar oscillations or waveforms according to their spectral densities. We treat the estimated spectral densities from many epochs or trials as functional data and develop clustering algorithms based on functional data ranking. The two proposed clustering algorithms use different dissimilarity measures: distance of the functional medians and the area of the central region. The performance of the proposed algorithms is examined by simulation studies. We show that, when contaminations are present, the proposed methods for clustering spectral densities are more robust than the mean-based methods. The developed methods are applied to two stages of resting state EEG data from a male college student, corresponding to early exploration of functional connectivity in the human brain. Keywords Central region · Functional median · Robustness · Spectral analysis · Time series clustering
1 Introduction Most of the research on clustering of brain signals currently focuses on how populations of neurons respond to external stimuli or how they behave during the resting state. Brain Ying Sun
[email protected] Tianbo Chen [email protected] Carolina Euan [email protected] Hernando Ombao [email protected] 1
Statistics Program, King Abdullah University of Science and Technology (KAUST), Thuwal 23955, Saudi Arabia
Journal of Classification
activity following the presentation of a stimulus and even during resting state is the result of highly coordinated responses of large numbers of neurons both locally (within each region) and globally (across different brain regions) (Fingelkurts et al. 2005). The electroencephalogram (EEG) is a tool for monitoring the spontaneous electrical activity of the brain over a period of time. EEGs are typically recorded from multiple electrodes placed on the scalp, referred to as EEG channels. In practice, EEGs are often used to diagnose brain disorders, such as tumors, stroke, and coma, because the signals capture macroscopic oscillations caused by coordinated activities in the brain. Although the EEG has limited spatial resolution compared to functional magnetic resonance imaging, it remains a valuable tool due to its millisecond-range temporal resolution. The goal of this paper is to develop robust time series clustering algorithms that are resistant to outliers for the identification of similar EEG channels. The clustered EEG channels are useful for understanding the functional connectivity of the brain. Medical research and diagnostic applications generally focus on neural oscillations that are captured in EEG signals, or the spectral aspect of EEG. In this paper, we an
Data Loading...
