*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Title: Modeling Stress Dynamics: Non-Invasive Estimation and Neuromodulation of Latent Autonomic State
Committee:
Dr. Inan, Advisor
Dr. Rozell, Co-Advisor
Dr. Vela, Chair
Dr. Davenport
Abstract: The objective of the proposed research is to enable closed-loop stress regulation via inference and control of latent autonomic state, leveraging non-invasive physiological sensing and neuromodulation. Two synergistic components are necessary to accomplish this. The first involves estimation. Specifically, a multimodal set of physiological signals will be used to infer a lower dimensional latent autonomic state's dynamics. To do so requires biosignal processing and feature extraction; however, the foundational literature on respiration signal processing and feature extraction remained unsatisfactory. Our first completed contribution, therefore, was to engineer an end-to-end respiration processing and respiration pattern variability (RPV) extraction pipeline and establish RPV’s associations with neural and psychological measures. Our proposed next step for estimation is to design a novel latent autonomic state estimation for stress (LASES) framework and validate state correspondence with neural and psychological measures. The second necessary component of the proposed closed-loop system involves control, whether that be simple on-off control or advanced algorithms. As our control input, we propose the use of transcutaneous cervical vagus nerve stimulation (tcVNS). tcVNS's static effects studied in prior work indicate sympathetic arousal attenuation. However, biomarker dynamics remained poorly understood. Thus, our second completed contribution was to study these biomarker dynamics using state-space models of varying dead time. By characterizing the responses and time delays, the on-off effects of tcVNS were elucidated for preliminary control design. We propose furthering these control efforts by modeling the tcVNS-induced dynamics of latent autonomic state via system identification. This will enable in silico control design and evaluation.
