Investigating the ability of post-stroke EEG measures of brain dysfuntion to inform early prediction of cognitive impairment or depression outcomes

Abstract

Background Stroke is a leading cause of death and disability worldwide. In addition cognitive impairment (and dementia) and depression are common sequelae. Additional assessment methods which would better inform early prognostication of these unfavourable outcomes would engender substantial clinical benefits. For example, more accurate prognostication would enhance clinical decisions regarding appropriate levels of rehabilitation or care. This thesis investigates the potential utility, in this context, of one such assessment methodology; quantitative EEG (QEEG), performed in the first days after stroke. Study 1 (Chapter 2) The primary aim was to investigate the relationship between post-stroke QEEG measures and cognitive outcomes. Resting-state EEG was recorded within 62–101h after onset of middle cerebral artery (MCA) territory, ischaemic stroke. QEEG indices sensitive to power of one or more frequency bands were computed. The functional independence measure and functional assessment measure (FIM–FAM) was administered at approximately 3-months post-stroke. Total (30 items) and cognition-specific (5 items) FIM–FAM scores were correlated with QEEG indices using Spearman's coefficient. Twenty cases (10 female; age range 38–84) were analysed. Two QEEG indices demonstrated highly-significant correlations with cognitive outcomes: relative alpha power across the scalp (ρ=0.67, p≤0.001) and frontal delta/alpha power ratio (ρ=−0.664, p≤0.001). These results demonstrate that QEEG measures are associated with cognitive outcomes, and may potentially inform prognostication of same. Study 2 (Chapter 3) The primary aim was to investigate the ability of QEEG indices to inform prognostication of cognitive outcomes, assessed using the Montreal Cognitive Assessment (MoCA). Resting-state EEG was recorded within 48-239h after stroke symptom onset. Various QEEG indices analysed across the scalp were calculated. The MoCA was administered at approximately 3-months post-stroke and scores correlated with QEEG indices using Spearman's coefficient. Binary logistic regression was employed, with cognitive impairment indicated by MoCA scores ≤25. Thirty-one participants (10 female, age range: 18-84) were included in the analyses. Two measures of alpha frequency slowing (relative power in the traditional upper theta frequency range in posterior regions and peak alpha frequency across frontal and posterior regions of the scalp) were significantly positively correlated with lower MoCA scores. Using regression analysis, relative power of slowed alpha activity from 2 posterior regions were selected as optimal predictors of cognitive impairment. The regression model utilising these QEEG measures was able to accurately prognosticate cognitive impairment outcomes in 81% of the participants (25/31). These results indicate that the power of slowed alpha activity, acquired within several days of stroke onset from several electrodes only, can inform early prognostication of post-stroke cognitive outcomes. This QEEG measure would be of clinical value given the substantial numbers of stroke patients whom are unable to complete cognitive assessments, such as the MoCA, prior to hospital discharge due to functional impairments (as was the case in 13% of the patients recruited in the study). Study 3 (Chapter 4) The aim of the final study was to investigate potential associations between pre-discharge QEEG indices and post-stroke depression (PSD) outcome measures as well as the ability of QEEG to inform prognostication of PSD. (Methodology is detailed in Study 2, above). Outcome was assessed using the Geriatric Depression Scale (GDS) at approximately 3-months post-stroke. GDS scores were correlated with QEEG indices using Spearman's coefficient. Binary logistic regression was used to build a QEEG model to classify depression, indicated by GDS scores ≤6 (out of 15). A significant correlation was obtained between PSD and an index of interhemispheric asymmetry of delta activity, patients with more symmetric delta activity tended to have higher GDS scores (r=-0.475, p=0.007, n=31). This variable was entered into the binary logistic regression model, however, it was not found to be a statistically significant predictor of subsequent PSD. Conclusions The results of the studies reported in Chapters 2 and 3 collectively indicate that QEEG markers sensitive to alpha activity – particularly alpha slowing – following ischaemic stroke can inform early prognoses regarding post-stroke cognitive impairment. Notably, pre-discharge measures of slowed alpha activity, from several electrodes only, was found to reliably predict cognitive outcomes in the majority of patients investigated. Importantly, virtually all patients can be assessed with EEG, including the substantial proportion of patients who cannot be assessed with more routine assessment scales such as the MoCA; hence QEEG would be of particular clinical value in such cases. It is postulated that the observed associations between post-stroke measures of alpha slowing and cognitive outcomes are reflective of attentional function which appears linked to both variables. These novel findings and interpretations constitute the principal contributions of this thesis to the knowledge base in this field. In summary this research has considerable potential to ultimately contribute towards enhanced care and benefits for stroke patients and the broader community.