Slide Slam Q2
Resting-state EEG signal complexity correlates with quality of narrative discourse in healthy older adults
Roger Newman-Norlund1, Samaneh Nemati1, Nicholas Riccardi2, Sara Sayers1, Sarah Newman-Norlund1, Julius Fridriksson1; 1University of South Carolina, Department of Communication Sciences and Disorders, 2University of South Carolina, Department of Psychology
Introduction: Analysis of the non-linear complexity of electroencephalographic signals acquired during rest (rsEEG) can provide useful information concerning abnormalities in cortical and subcortical dynamics in the working brain. rsEEG complexity metrics have proven potent as biomarkers of symptom severity in multiple clinical models including mild cognitive impairment, Alzheimer’s disease, Parkinson’s disease, schizophrenia and depression. Acquiring rsEEG data is relatively inexpensive (compared to techniques such as MRI, PET and MEG) and is therefore may be a viable option for measuring brain health. The current study examines the relationship between rsEEG signal complexity (i.e. Lempel-Ziv complexity) and language production (i.e. quality of self-generated narrative discourse) in a sample of healthy older adults. Methods: As a part of the Aging Brain Cohort testing battery at the UofSC, a total of 47 healthy older adult participants produced narrative discourse and completed resting-state EEG testing. Narrative discourse was obtained using AphasiaBank’s Cat Rescue picture description task, which requires participants to generate a story about a visual scene, complete with beginning, middle and end. Discourse samples were transcribed and coded using the CHAT transcription guidelines and linguistic variables were extracted using the CLAN program. Our primary language measures were the fluency factor (maze index) and the semantic factor (percent noun, verbs, and pronoun index). Eyes-open EEG data (3 minutes) was acquired using a 128-channel actiCAP EEG cap and a Brain Vision ActiChamp EEG plus system housed in an electronically shielded, custom-wired TeleAcoustic sound-proof booth. Following artifact removal using the Harvard-based HAPPE pipeline, Lempel-Ziv complexity (LZC) was calculated using the formula described by Lempel-Ziv (1976). Specifically, the EEG signal recorded from each electrode was transformed into a binary sequence using its median value as a threshold after which complexity was calculated by dividing the number of observed distinct patterns by the maximum complexity of the observed sequence. For each participant, whole-brain rsEEG complexity was calculated by averaging this measure across all electrodes. Results: In order to test the relationship between rsEEG signal complexity and discourse variables, we computed Pearson correlations between our measure of Lempel-Ziv complexity and two discourse measures (fluency & semantic factors). We found a positive correlation between rsEEG-based Lempel-Ziv complexity and individual semantic factor scores, r(46) = 0.36, p = 0.006, one-tailed, and a non-significant trend in the same direction for the fluency factor, r(46) = 0.16, p = 0.13, one-tailed. Discussion: The current study demonstrates that complexity of rsEEG signals is associated with the quality of narrative discourse production (semantic but not fluency factor) in a population of healthy older adults. Decreased rsEEG complexity in individuals with lower semantic factor scores may reflect abnormal interactions between disparate information sources involved in discourse generation. Reductions in rsEEG complexity may be due to a number of factors including neuronal death or partial deactivation of, or alterations in neuronal synchronization of language related networks. Future studies should explore the basis of rsEEG complexity variability as well as its potential utility as a biomarker of impaired language function throughout life.