Skip to main content
eScholarship
Open Access Publications from the University of California

Scholarly Works (3 results)

  • Thesis
  • Peer Reviewed
UCLA Electronic Theses and Dissertations (2014)

Autism Spectrum Disorders (ASD) are a collection of neurodevelopmental disorders with features of impairments in two domains: social communication, and restrictive repetitive behaviors/interests. Quantitative electroencephalography (QEEG) holds promise as a translational method for investigating abnormal neural oscillatory activity in ASD. Use of QEEG in ASD research is increasing, however discrepancies exist among the types of methods, artifact handling, and analyses researchers use. To facilitate comparing results across researchers, a solid foundation for use of QEEG in ASD research is needed. This dissertation builds a framework for using QEEG in future studies on children with ASD through a detailed description of data processing, artifact effects, statistical analysis methods, and differences in QEEG measures between children with ASD and typically developing controls. Chapter 2 presents a systematic approach to establishing clean, artifact-free data, and investigates the effect of artifacts on QEEG band power measures in children. We demonstrate that: (1) data quality is similar between a diverse group of children with ASD and a control group of typically developing children (TD), and (2) group differences in QEEG measures are not confounded by group differences in artifacts. Chapter 3 focuses on incorporating individual diversity observed in ASD into frequency band power analyses. It first demonstrates the need for including phenotypic information in QEEG analyses, and then describes frequency band power differences between typically developing children and children with ASD.

    Cover page: Quantitative electrophysiology as a biomarker of Autism Spectrum Disorder
    • Article
    • Peer Reviewed
    UCLA Previously Published Works (2015)
    We quantified the potential effects of physiologic artifact on the estimation of EEG band power in a cohort of typically developing children in order to guide artifact rejection methods in quantitative EEG data analysis in developmental populations. High density EEG was recorded for 2 min while children, ages 2-6, watched a video of bubbles. Segments of data were categorized as blinks, saccades, EMG or artifact-free, and both absolute and relative power in the theta (4-7 Hz), alpha (8-12 Hz), beta (13-30 Hz) and gamma (35-45 Hz) bands were calculated in 9 regions for each category. Using a linear mixed model approach with artifact type, region and their interaction as predictors, we compared mean band power between clean data and each type of artifact. We found significant differences in mean relative and absolute power between artifacts and artifact-free segments in all frequency bands. The magnitude and direction of the differences varied based on power type, region, and frequency band. The most significant differences in mean band power were found in the gamma band for EMG artifact and the theta band for ocular artifacts. Artifact detection strategies need to be sensitive to the oscillations of interest for a given analysis, with the most conservative approach being the removal of all EMG and ocular artifact from EEG data. Quantitative EEG holds considerable promise as a clinical biomarker of both typical and atypical development. However, there needs to be transparency in the choice of power type, regions of interest, and frequency band, as each of these variables are differentially vulnerable to noise, and therefore, their interpretation depends on the methods used to identify and remove artifacts.
      Cover page: Physiologic artifacts in resting state oscillations in young children: methodological considerations for noisy data
      • Article
      • Peer Reviewed
      Research Reports (2015)
        Cover page: Identifying Longitudinal Trends within EEGExperiments
        Top