Abstract W P387: Neurological Correlates Of Brain Function After Acute Stroke--A Dense Array EEG Study

Abstract

Listen

Translate article

INTRODUCTION: Measures of brain function can complement assessment of injury to inform clinical decision-making after stroke, but the most useful metrics remain uncertain. An acute stroke alters brain function in widespread areas. We therefore reasoned that a whole brain measure of brain function would be better related to behavioral deficits than a regional measure of brain function. METHODS: In 24 patients hospitalized for acute stroke, resting EEG (256 leads) was recorded for 3 min at the bedside and analyzed offline. Two EEG measures of brain function were extracted: [1] whole brain connectivity, which found the EEG frequency (from 1-30 Hz) and seed point (from among the 256 leads) that best fit whole brain coherence with total NIHSS scores, using a partial least squares regression model; and [2] regional brain activity, which found the EEG frequency and lead where spectral power was most strongly correlated with total NIHSS scores. Analyses were repeated focused on NIHSS motor subscores (Q4-6). All models were validated using a leave-one-out approach. RESULTS: The 24 patients were age 60.9±13.1yr, 3.5 ± 2.9 d post-onset (range 3hr-12d), and were studied in settings that included ER, ICU, and stroke ward. Whole brain EEG connectivity explained a large fraction of the variance in total NIHSS scores (r^2=0.72); this was achieved in the 2-4 Hz range, with seed over ipsilesional motor cortex, and with model predicting higher NIHSS score when this seed had greater coherence with contralesional frontal/motor regions. Regional brain activity, by comparison, explained a smaller fraction of variance (r^2=0.51), with maximal correlation between total NIHSS and regional EEG power found using a lead over contralesional motor cortex, at 2 Hz. Similar results for whole brain EEG connectivity were obtained when modeling NIHSS motor subscores in the 14 subjects with motor deficits (validated r^2=0.71). CONCLUSIONS: Dense array EEG recordings could be obtained early after stroke, rapidly and reliably, and at the bedside in widespread hospital settings. Whole brain connectivity measures corresponded to behavioral state better than measures of regional brain activity do. Results support the utility of EEG as a bedside method for evaluating brain functional status after stroke.