Repeated ultrasonic vocalizations during REM sleep in the rat

Abstract

Rats are highly social animals that are able to emit audible sounds, but mainly intercommunicate in an ultrasonic interval of frequency (Brudzynski, ILAR journal, 2009). Adult rats emit ultrasonic vocalizations in two different frequencies, 22KHz and 50KHz. Vocalizations at 22KHz are mostly used in situations with negative connotations and are used as warning and alert signals, while those emitted at 50 Hz express hedonic states (Brudzynski, Curr Opin Neurobiol, 2013). In an attempt to obtain a detailed evaluation of the animal behaviour in a study regarding sleep and obesity, we recorded the emission of vocalizations at 22Hz as monitor of the animals well being. Unexpectedly, we found vocalizations during REM sleep. Thus, we quantitatively analyzed this behavior, in an attempt to unravel its functional significance. 15 Sprague‐Dawley male rats, adapted to a standard Light‐Dark (LD) cycle of 12h‐12h (L: 09:00–21:00), were administered for eight weeks with a normocaloric diet (NC, n=7), or hypercaloric (35% of fat) (IC, n=8), with subsequent obesity development. Under general anaesthesia, animals were implanted with: i) electrodes for electroencephalography (EEG), and nuchal (nEMG) and diaphragmatic electromyography (dEMG); ii) a catheter in the femoral artery for telemetric arterial pressure recording; iii) a hypothalamic thermistor. After recovery from surgery, animals’ parameters in basal conditions were recorded for two consecutive days. The sound emission was recorded by means of a “Bat Detector” ‐‐‐‐(BatBox III D) set for acquisition at 22KHz. Similarly to what observed in the wake state, during REM sleep the vocalization occurs during what appears to be a prolonged expiratory event (diaphragmatic contraction). The event can manifest as a single, or, more often, in ‘series’ that appear during the subsequent ‘expiratory’ events. Thus, for each animal, separating L(resting), and D (active) phases, the following parameters were evaluated: a) number of series/hour of REM sleep; b) series duration, calculated as time from the beginning of the first event and the end of the last event every series; c) mean number of events for each series. A preliminary analysis shows that vocal series appear with similar probability (No./h of REM sleep) in lean and obese animals, both in the L and D phase (NCL=2.8±0.7; NCD=1.8±0.6; ICL=2.6±1; ICD=2.6±0.7). Similar results can also be observed for the series duration (NCL=7.2±1.2s; NCD=6.4±2.5s; ICL=8.4±3s; ICD=8.8±1.8s) and the mean number of events for each series (NCL=2.8±0.3; NCD=3.0±1.7; ICL=2.0±0.5; ICD=2.7±0.6). The statistical analysis (two‐way ANOVA for repeated measures), showed no significant effects induced by the LD cycle or the animals weight. The appearance of ultrasonic vocalizations during REM sleep is not affected by either the LD ciclicity or the animal weight, supporting the hypothesis that these events represent a REM sleep constitutive phenomenon rather than being the consequence of a respiratory pathophysiological process.