Five-step wavelet approach to detect and characterize changes in the dynamics of wheel running and food intake behaviors in mice subjected to feed restriction

KEMBRO, J.M.; FLESIA, A. G.; NIETO, P.S.

La Plata

Congreso; XIX Congreso Regional de Física Estadística y Aplicaciones a la Materia Condensada (TREFEMAC 2022); 2022

Comisión organizadora del XIX Congreso Regional de Física Estadística y Aplicaciones a la Materia Condensada (TREFEMAC 2022)

The temporal dynamics of behavior is complex because it reflects processes occurring at different temporaland spatial scales, susceptible to perturbations by external factors. Locomotor activity in mammals presentscircadian (approx. 24h) rhythms and other dynamical patterns, specifically, the ultradian rhythms (URs, i.e.rhythms with periods shorter than 24h) and long-range auto-correlation. These patterns can be differentiallyaffected by environmental perturbations. Identifying changes in such patterns is important to understand behavioraladaptation to the environment. Specifically, how animals are fed can have an important impact not only on theanimal’s behavioral dynamics but also in their physiology and health. Yet, there are no clear tools to characterizeand quantify these dynamical perturbations at different behavioral scales.Thus, herein, we analyze behavioral timeseries of food intake and wheel running behaviors derived from mice that are either fed ad-libitum (i.e. withoutrestrictions) or exposed to temporal restriction of feeding (TR) or caloric restriction (CR) throughout a 42-dayexperiment. Since mice are mostly nocturnal, feed restrictions were either applied during the daytime or nighttimeto evaluate the impact of time at which such restriction paradigms were applied. Specifically, we applied a 5-stepwavelet approach to detect and characterize behavioral rhythms and detrended fluctuation analysis (DFA) toassess changes in auto-correlation properties. We found that TR and CR differentially modulates the acrophaseand strength (power) of circadian rhythms in both running wheel and food intake activity. As expected, daytimeTR and CR induced an abrupt change of phase and a transitory decrease in power of the circadian food-intakerhythm, consistent with a transitional state associated with the perturbation induced by the change in feedingparadigme. Importantly, the wavelet approach we propose allows us to detect ultradian rhythms in feed-intake andwheel running time series, and we were able to show, for first time, how to characterize the persistence, prevalenceand robustness of URs. which depend on the behavior studied, feeding paradigm and individual genetic/metabolicbackgrounds. In addition, we detect long-range correlations in the wheel-running, but not in feed-intake for scalesshorter than 100 min. Surprisingly, the correlation properties were susceptible to be modulated by the feedingparadigm. We conclude that the feed paradigm affects the temporal dynamics of behavior over a broad rangeof temporal scale, highlighting the diversity of behavioral patterns beyond the circadian activity, building up thebehavioral architecture that composes these complex systems.