Saslab pro

We conducted the recording and analysis of USVs as previously described [28 (link),35 (link)]. In brief, USVs were recorded using an UltraSoundGate 116H audio device with a CM16/CMA microphone (Avisoft Bioacoustics, Berlin, Germany). For acoustic analyses, the recordings were transferred to SASLab Pro (version 5.2, Avisoft Bioacoustics). 50-kHz USVs were analyzed manually according to Wright et al.’s classification [26 (link)] and our previous study [28 (link),35 (link)].

This experiment used three sound types: wolf howl as predator sound, deer rut call as interspecific call representing loud, potentially frightening sound, however, produced by non-dangerous herbivores, and white noise as control. To minimise the risk of pseudoreplication (McGregor 2000 (link)), it was decided to use six variants for the wolves’ howls and two for the deer calls. The wolf howls come from captive wolves (Canis lupus baileyi) at the Wolf Conservation Centre, New York. The deer calls are rut calls from red deer (Cervus elaphus elaphus and Cervus elaphus hippelaphus) obtained from bioacoustica.org. Based on the availability and suitability of the calls (without disturbing elements), the Mexican wolves instead of the European were chosen; however, as howling is a social communication process that is significant for all canid species (Kershenbaum et al. 2016 (link)), it is relevant to use the non-native wolf subspecies for the experiment. The white noise is a static sound generated through the software Avisoft-SASLab Pro (Avisoft Bioacoustics, Berlin, Germany). The sounds were equalised in intensity and length (30 s). The Python random.choice command randomly selected sounds for every test.
The sounds were played with a Lamax PartyBoomBox 500 (Lamax, Prague, Czech Republic) at a sound pressure of 100 DB at 1 m, at a distance of 75 m [40DB] measured with a Nikon Monarch 3000 rangefinder (Nikon, Tokyo, Japan). The speaker was covered by dark cloth to decrease its visibility by the horses. This setting avoided overstimulation with an unrealistic sound level while still being hearable by the ponies. At longer ranges, the sound would quickly drop below 30DB and could easily be covered by the sound of the wind.
As the conditions allowed it, it was decided to favour the number of herds over the number of repetitions to avoid any risk of habituation. Every herd was tested only once per sound, and the observations at each herd were carried out with at least one day in between two playbacks, following similar previous studies (Christensen and Rundgren 2008 (link); Janczarek et al. 2020a (link), b (link), 2021 (link); Watts et al. 2020 (link)). The testing was performed by two people, one playing the sounds and one video recording the herd. The procedure was as follows: the herd was recorded for 5 min without disturbance, then the sound was played for 30 s, and finally, the herd was recorded for another 15 min (Watts et al. 2020 (link)). The recordings were performed with a Panasonic HC-VX1 4 K camera (Panasonic, Osaka, Japan). Since remaining hidden was not feasible in most cases, the observers chose not to conceal themselves. The ponies were, however, habituated to human presence and showed no detectable reaction to it.
Other recorded parameters were the herd size, sex and age distribution, with ponies under three 3-years-old considered immature (Rogers et al. 2021 (link)). External factors were also noted: the time of the day (hours and minutes), the weather conditions (wind and rain scored as a binary option for the presence or not), which deeply affect horses´ behaviour (Bernátková et al. 2022 (link)), the environment type (open or closed habitat) and the presence of other ungulates in the enclosure.

Litters chosen for testing contained more than seven pups for BTBR (7.8±1.05), B6 (7.1±0.58) and FVB/NJ (7.5±0.34), and more than five pups for 129X1 (5.5±0.40; a strain known for small litters). One female and one male from each litter of BTBR, B6, FVB/NJ and 129X1 mice (n = 10 litters each strain) were used for baseline measurements of the ultrasonic vocalizations from pnd 2 to 12. Body weights and body temperatures of pups were measured after the ultrasonic vocalization test on pnd 2, 4, 6, 8 and 12. On each day of testing, each pup was placed into an empty plastic container (diameter, 5 cm; height 10 cm), located inside a sound-attenuating styrofoam box, and assessed for USVs during a five minute test. At the end of the five minute recording session, each pup was weighed and its axillary temperature measured by gentle insertion of the thermal probe in the skin pocket between upper foreleg and chest of the animal for about 30 seconds (Microprobe digital thermometer with mouse probe, Stoelting Co., Illinois, USA). No differences in patterns of calling were detected in a comparison of male and female pups, therefore data were collapsed across sex.
An Ultrasound Microphone (Avisoft UltraSoundGate condenser microphone capsule CM16, Avisoft Bioacoustics, Berlin, Germany) sensitive to frequencies of 10–180 kHz, recorded the pup vocalizations in the sound-attenuating chamber. The microphone was placed through a hole in the middle of the cover of the styrofoam sound-attenuating box, about 20 cm above the pup in its plastic container. The temperature of the room was maintained at 22±1°C. Vocalizations were recorded using Avisoft Recorder software (Version 3.2). Settings included sampling rate at 250 kHz; format 16 bit. For acoustical analysis, recordings were transferred to Avisoft SASLab Pro (Version 4.40) and a fast Fourier transformation (FFT) was conducted. Spectrograms were generated with an FFT-length of 1024 points and a time window overlap of 75% (100% Frame, Hamming window). The spectrogram was produced at a frequency resolution of 488 Hz and a time resolution of 1 ms. A lower cut-off frequency of 15 kHz was used to reduce background noise outside the relevant frequency band to 0 dB. Call detection was provided by an automatic threshold-based algorithm and a hold-time mechanism (hold time: 0.01 s). An experienced user checked the accuracy of call detection, and obtained a 100% concordance between automated and observational detection. Parameters analyzed for each test day included number of calls, duration of calls, qualitative and quantitative analyses of sound frequencies measured in terms of frequency and amplitude at the maximum of the spectrum.
Waveform patterns of calls were examined in depth in twenty sonograms collected from every strain, one from each of the pups tested. The sonograms were one minute in length and selected from recordings at postnatal day 8. We classified 3633 BTBR calls, 2333 B6 calls, 1806 129X1 calls and 2575 FVB/NJ calls. Each call was identified as one of 10 distinct categories, based on internal pitch changes, lengths and shapes, using previously published categorizations [21] (link), [22] (link), [24] (link). Classification of USVs included ten waveform patterns described below, and illustrated visually in Figure 2 and S1 and as audiofiles (Sounds S1, S2, S3, S4, S5, S6, S7, S8, S9, S10) in Supporting Information.
Inter-rater reliability in scoring the call categories was 98%. Call category data were subjected to two different analyses: a) strain-dependent effects on the frequency and duration of the vocalizations emitted by each subject at pnd 8 b) strain-dependent effects on the probability of producing calls from each of the ten categories of USV, as described below under Statistical analysis.