Waveform shape better explains the relationship between respiration and neural oscillations than cross-frequency coupling

Poster Session F - Tuesday, April 16, 2024, 8:00 – 10:00 am EDT, Sheraton Hall ABC

Lee Eena Kosik¹ (ekosik@ucsd.edu), Bradley Voytek^1,2,3; ¹University of California, San Diego, ²Halıcıoğlu Data Science Institute, ³Neurosciences Graduate Program

The breath cycle is a periodic rhythm consisting of asymmetric inhalation and exhalation phases. There is emerging evidence that respiratory rhythms entrain neural oscillations in various brain regions. More specifically, it is hypothesized that respiration influences neural processing via synchronization of coherent rhythms, as well as phase-amplitude coupling (PAC). These mechanisms of action do not account for a fundamental attribute of these physiological rhythms: waveform shape. Respiratory and neural rhythms are both nonsinusoidal in nature: they wax and wane for various durations, and have sharper or smoother peaks and troughs. Here, we test the hypothesis that the classically reported PAC measures in respiration-mediated neural oscillations may conflate the effects of nonsinusoidal waveform shape with multiple oscillations. To do this, we analyzed a dataset of iEEG patients at rest with simultaneously recorded respiration and electrodes in medial temporal lobe (MTL; hippocampus and amygdala). To extract the true waveform shape of the respiration-entrained neural oscillations, we used spatio-spectral decomposition (SSD) to isolate MTL oscillations at the respiration frequency. From this signal, we used bycycle to extract cycle-by-cycle waveform shape features of the SSD sources. To assess the relationship between these rhythms, we examined their cross correlation and Granger causality. Additionally, we computed classically measured metrics of PAC to test if there were spuriously generated coupling with these nonsinusoidal rhythms. We found that even in the respiration rhythm itself, there is significant PAC indicating either people breathe with nested oscillations (which is unlikely) or PAC of these nonsinusoidal waveforms is being misinterpreted.

Topic Area: METHODS: Electrophysiology