Characterizing Central-Autonomic Dynamics During an Episodic Memory Task Using Multi-Modal Neural and Cardiomechanical Signals.

Abstract

Memory function underlies mental and behavioral health. While the role of the central nervous system (CNS) during episodic memory encoding and retrieval is well researched, the interplay between the CNS and the autonomic nervous system (ANS) during these processes is not. We addressed this gap by analyzing fluctuations in CNS-ANS network coupling during an episodic memory task (EMT). Sixty-seven healthy adults completed an EMT consisting of three stages: Encoding, Retrieval 1, and Retrieval 2. From electroencephalogram (EEG) recordings, we estimated time-varying CNS indices via alpha and theta power. Simultaneously, from electrocardiogram (ECG) and seismocardiogram (SCG) recordings, we estimated time-varying parasympathetic and sympathetic indices via cardiac vagal index (CVI) and pre-ejection period, respectively (PEP). Using time delay stability and surrogate data analysis, for each protocol condition, we assessed coverage, the percentage of significant coupling links, over the CVI-Neural and PEP-Neural networks. Median coverage of CVI-Neural and PEP-Neural increased from Baseline to Encoding (+13.2%, $p$<0.001 and +7.9%, $p$<0.001), decreased from Encoding to Retrieval 1 (-5.3%, $p$<0.05 and -5.3%, $p$<0.001), and increased from Retrieval 1 to Retrieval 2 (+5.3%, $p$<0.001 and +5.3%, $p$<0.05). This is the first work to integrate EEG, ECG, and SCG time series to elucidate CNS-ANS interactions during an EMT. We demonstrated that healthy adults experienced parallel changes in CNS-parasympathetic and CNS-sympathetic network coverage during the task. In the future, measuring neural and cardiomechanical signals to estimate CNS-ANS coupling while probing memory function in clinical populations may help derive new mental health screening biomarkers.