Poster · Alzheimer's & Memory Disorders
Cardiac Autonomic Context Moderates the Association Between EEG Alpha Reactivity and Reaction Time in Mild Cognitive Impairment
Jayson Guo1,2, Victor Lee2,3, Edwin Arii2,4, Alana Wickham2,4, Kenji Aoki1,2, Hyeong Jun Ahn, PhD5, Michael Sonson, MD2, Barbara Pitts, PhD2, Enrique Carrazana, MD1, Kore Liow, MD1,2
- 1 John A. Burns School of Medicine, University of Hawaiʻi at Mānoa, Honolulu, HI
- 2 Alzheimer’s Neural Network EEG Research Laboratory & Memory Disorders Center, Hawaiʻi Pacific Neuroscience, Honolulu, HI
- 3 Columbia University, New York, NY
- 4 University of Hawaiʻi at Mānoa, Honolulu, HI
- 5 Biostatistics Core Facility, Department of Quantitative Health Sciences, John A. Burns School of Medicine, University of Hawaiʻi at Mānoa, Honolulu, HI
Background: Alpha reactivity—the reduction in alpha power from eyes-closed (EC) to eyes-open (EO) resting EEG—is reduced in mild cognitive impairment (MCI) and may reflect the capacity to transition from internally oriented rest to external engagement. Whether greater alpha reactivity is associated with more efficient task processing, and whether its functional significance varies across physiologic states, remains unclear.
Methods: This retrospective cross-sectional study included 182 patients with MCI who underwent Biomarker-Based Electrophysiology for Advanced Monitoring (BEAM), providing EC/EO resting EEG; heart-rate variability (HRV) metrics; and Auditory Oddball (AO), 3-Choice Vigilance (CVT), and Standard Image Recognition (SIR) task measures. Separate multivariable linear regression models examined associations of alpha reactivity with each reaction-time outcome, task-specific event-related potential (ERP) latency (namely AO P300, CVT P1, and SIR P2), and Mini-Mental State Examination (MMSE) score. All models adjusted for age and sex, with depressive symptoms additionally included in the MMSE model. Exploratory interaction models tested whether resting EEG spectral measures (peak alpha frequency and individualized theta power) or cardiac autonomic indices (RMSSD, SDNN, and LF/HF) modified associations between alpha reactivity and task outcomes, with false-discovery-rate (FDR) correction within outcome families.
Results: Greater alpha reactivity was independently associated with higher MMSE scores after adjustment for age, sex, and depressive symptoms (β=0.024 ± 0.009, p=.005), but no direct associations were observed with AO, CVT, and SIR reaction time or ERP latencies. In interaction models testing effect modification, EC LF/HF modified the association between alpha reactivity and SIR reaction time (N=99; interaction β=15.32 ± 4.63 ms; p=.001; FDR q=.031; ΔR²=.097). That is, greater alpha reactivity was associated with faster responding at lower EC LF/HF, whereas this association weakened and reversed at higher LF/HF. Nominal interactions were observed for EO LF/HF with SIR reaction time (p=.006) and CVT reaction time (p=.022), although these did not survive FDR correction.
Conclusion: Alpha reactivity was associated with global cognitive performance but showed no uniform association with task-processing efficiency, as indexed by reaction time and ERP latency. Moderation of the alpha reactivity-reaction time relationship by LF/HF suggests that the functional significance of alpha reactivity varies with resting cardiac autonomic context, supporting the potential of integrated EEG-HRV assessment to refine biomarker interpretation and characterize physiologic heterogeneity in MCI.