0961 Applying Fractal Analysis Towards the Understanding of Long-COVID-Induced Insomnia

Abstract

Abstract Introduction As part of the recent international COVID Sleep Study-II, long COVID syndrome was found to be associated with chronic insomnia. However, an article submitted for publication out of the Hôpital Hôtel-Dieu found no significant difference between long-COVID-related insomnia and standard insomnia with respect to routine polysomnographic (PSG) parameters. In light of these findings, we seek to better understand the difference between long COVID insomnia and standard chronic insomnia by conducting a fractal analysis of the raw PSG signals of our subjects. Methods 15 long COVID patients with complaints of chronic insomnia were included as subjects, as well as 34 matched controls without long COVID who had similar complaints of chronic insomnia. All participants underwent one night of polysomnographic recording. We perform our analysis on movements of the abdomen (as an approximation of deep breathing quality) and on EEG waves, considering that shortness of breath and cognitive dysfunction are commonly associated with long COVID. Building upon an existing repertoire of medical research using the fractal dimension (FD) of patients’ signal data to differentiate pathologic and physiologic processes, we make use of Detrended Fluctuation Analysis (DFA) to calculate the FD of the abdomen (FDA) and of an aggregated EEG wave (FDE) for each patient. Results We find a decrease in the FDA (p=0.036) and increase in the FDE (p=0.012) of long COVID insomniacs with respect to those of our control group. These results are shown to be significant after applying a Holm-Bonferroni correction for multiple hypotheses. Conclusion Despite our small data sample, an in-depth analysis of PSG signals shows significant differences in the fractal nature of long-COVID-related insomnia with respect to standard chronic insomnia. Our results furthermore match our expectations, given that related research outside of the COVID domain indicates a decrease in the FD of muscle movements and an increase in the FD of EEG waves in unhealthy individuals. Further research could employ more participants in order to extend our results to other data channels outside of EEG signals and the abdomen. Support (if any)