Differences in brain networks during consecutive swallows detected using an optimized vertex–frequency algorithm

Abstract

Patients with dysphagia can have higher risks of aspiration after repetitive swallowing activity due to the “fatigue effect”. However, it is still unknown how consecutive swallows affect brain activity. Therefore, we sought to investigate differences in swallowing brain networks formed during consecutive swallows using a signal processing on graph approach. Data was collected from 55 healthy people using electroencephalography (EEG) signals. Participants performed dry swallows (i.e., saliva swallows) and wet swallows (i.e., water, nectar-thick, and honey thick swallows). After standard pre-processing of the EEG times series, brain networks were formed using the time-frequency based synchrony measure, while signals on graphs were formed as a line graph of the brain networks. For calculating the vertex frequency information from the signals on graphs, the proposed algorithm was based on the optimized window size for calculating the windowed graph Fourier transform and the graph S-transform. The proposed algorithms were tested using synthetic signals and showed improved energy concentration in comparison to the original algorithm. When applied to EEG swallowing data, the optimized windowed graph Fourier transform and the optimized graph S-transform showed that differences exist in brain activity between consecutive swallows. In addition, the results showed higher differences between consecutive swallows for thicker liquids.

Computational Deglutition: Using Signal- and Image-Processing Methods to Understand Swallowing and Associated Disorders [Life Sciences]

“…The network-based approach is also very interesting to the signal processing community, as it opens many interesting problems for the field of graph signal processing. In our own research, we use the vertex-frequency analysis (see a recent lecture note in this magazine [10]) to understand swallow-by-swallow changes in brain networks [11]. However, other graph signal processing approaches are anticipated to be suitable as well.…”

Studying the Swallow using Surface Electroencephalography: A Systematic Review

“…The most consistent trials were observed in 3 of the studies carried out by Jestrovic's team [15][16][17] with the swallowing of five different stimuli. All other studies used different tasks and stimuli to investigate swallowing under various conditions.…”

On Hilbert transform, analytic signal, and modulation analysis for signals over graphs

“…A fast algorithm for implementation of vertex-frequency representations for graphs was developed by Jestrovic et al in [53]. They also developed an optimized vertex-frequency representation for investigating brain characteristics during consecutive swallows [54]. Kotzagiannidis and Dragotti extended the notion of finiterate-of-innovations to circulant graphs [55] and also considered splines and wavelets for circulant graphs [56].…”