Figures and data in Maternal behavior influences vocal practice and learning processes in the greater sac-winged bat

Figures
Tables
Additional files

2 figures, 12 tables and 1 additional file

Figures

Figure 1

Download asset Open asset

Behavioral displays and interactions during babbling and maternal influence on the amount of vocal practice.

(**A-C**) shows the typical social behaviors and interactions of mother-pup pairs during a babbling bout. Note: all these behaviors could occur singly or in an interactive way (i.e. a behavioral display was followed by another in less than one second). (A) In on average 80% of cases, the interaction was initiated by the pup, normally by crawling toward the mother while babbling. A mother initiated interactions by hovering in front of or landing next to her pup. (B) After the start of an interaction, mother and pup engaged in a number of different behaviors, which were performed equally by both mother and pup (Appendix 1—table 1), with the exception of ‘wing poking’, which was mostly observed in pups (Appendix 1—table 1). The different behaviors could variably succeed after one another (large arrows); however, ‘crawl towards’ and ‘crawl away’, as well as ‘short flights’ respectively ‘hover flights’ and a temporarily ‘change in the roosting spot’ were often observed in repetitive sequences (bold black arrows). Hover flights in front of the interactive partner were the most conspicuous behavior, followed by the short flights either next to the interaction partner or within the day roost. With increasing pup age, short flights became progressively longer, for example pups even briefly left the day roost and landed on an adjacent tree. (C) Interactions between mother and pup could be terminated in two ways: Either the pup continued to babble alone or the mother allowed nursing. Maternal behavior did not differ with regard to pup sex (SI). Drawing credit: C. A. S. Mumm. (D) shows the positive influence of the maternal behaviors displayed during babbling bouts on babbling bout duration [s] (N=19 pups, N=186 babbling bouts, GLMM, family Gamma with log link). (E) shows that higher maternal activity scores led to a longer babbling phase duration [days] (N=19 pups, LMER). The figure legends represent the raw data, whereas the fitted lines (with lower (2.5th percentile) and upper (97.5th percentile) confidence intervals) are based on the calculated model. Neither bout duration nor babbling phase duration differed between pup sex (SI).

Figure 2 with 5 supplements

Download asset Open asset

Maternal influence on learned syllables in babbling bouts.

(A) shows a typical multisyllabic territorial song produced by an adult male. The most characteristic syllable type is called ‘B2’ (i.e. the last six syllables of this song), composed of a buzzed part connected to a tonal part. This is also the most common song syllable type in pup babbling (Appendix 1—table 3). The spectrogram was created in Avisoft SasLab Pro (Hamming window, 1024 FFT, 50% overlap resulting in 317 Hz frequency and 2.048ms time resolution). (B) shows the positive significant influence of maternal behavioral displays on the total number of learned syllables within babbling bouts (N=10 pups, N=90 babbling bouts, GLMM, family negative binomial with log link). Territorial songs can be flexibly composed of different buzz syllable types (Figure 2—figure supplement 1). (C) shows that maternal behavior positively influences the versatility (i.e. number of different learned syllable types, Figure 2—figure supplement 2) of babbling bouts (N=10 pups, N=90 babbling bouts, LMER). (D) The most common song syllable type B2 (Appendix 1—table 3) was categorized into precursor and mature syllables (Figure 2—figure supplement 3). Maternal behavioral displays positively influenced the percentage of mature song syllables (Table 3) (N=10 pups, N=85 babbling bouts, GLMM family binomial with logit link). The figure legends represent the raw data, whereas the fitted lines (with lower (2.5th percentile) and upper (97.5th percentile) confidence intervals) are based on the calculated model. (E) shows the acoustic movement of pups during their vocal ontogeny based on acoustic parameters of the syllable type B2, the most common song syllable type (N=8 pups, N=220 syllable trains, Figure 2—figure supplement 4). The larger the value on the y-axis, the larger the acoustic movement during the pup’s vocal ontogeny (Figure 2—figure supplement 5). The acoustic movement was positively correlated with the maternal activity score.

Figure 2—figure supplement 1

Download asset Open asset

Five different song syllable types in babbling bouts.

We defined five different syllable types based on spectro-temporal characteristics (i.e. best visualized by the combination of the spectrogram and the oscillogram). B1 has no tonal part. B2 has a clear pulsed part connected to one tonal part. B2 trill has a pulsed part and a tonal part with rippled frequency modulations. B3 has at least two tonal parts connected to a pulsed part. B4 has no discernible pulsed part but only a smeared noisy part. These syllables can be flexibly used for the composition of territorial songs; their composition is usually an indication of the current aggression level of the singer. The spectrogram was created in Avisoft SasLab Pro: Hamming window with 1024-point Fourier transformation and 87.5% overlap (frequency resolution: 244 Hz, time resolution: 0.512ms).

Figure 2—figure supplement 2

Download asset Open asset

Babbling excerpt with syllable labeling.

This excerpt of a babbling bout (pup ID 13) illustrates the acoustic analysis which was performed in order to analyse (a) the total number of song syllables (=amount of all song syllables; in this example the number of B2, B2 trill and B3), (b) the versatility per bout (=amount of all different types per bout, subsequent calculation of the diversity index using the R package vegan), and, (c) the number of precursor and mature B2 syllables within a babbling bout (=amount of mature (M) and precursor (P) B2 syllables). B2, B2 trill, and B3 belong to the adult-like syllables, being part of the later adult vocal repertoire. Besides the learned song syllables, babbling includes other adult-like syllable types (IC: Isolation call syllables, C1: chatter syllables) and syllables exclusively produced during babbling by pups (UP: undifferentiated proto-syllables). The spectrogram was created in Avisoft SASLab Pro: Hamming window with 1024-point Fourier transformation and 87.5% overlap (frequency resolution: 244 Hz, time resolution: 0.512ms).

Figure 2—figure supplement 3

Download asset Open asset

Precursor versus mature syllables.

This spectrogram illustrates the difference between precursor versus mature syllables (for the most common type B2 which was used in this analysis) from three different pups. Precursor syllables have distinct gaps, usually in the pulsed part, between the pulsed and the tonal part or even both. The spectrogram was created in Avisoft SASLab Pro: Hamming window with 1024-point Fourier transformation and 87.5% overlap (frequency resolution: 244 Hz, time resolution: 0.512ms).

Figure 2—figure supplement 4

Download asset Open asset

B2 syllable measurements for acoustic change tracking during ontogeny.

This figure depicts three B2 syllables which were measured to assess acoustic change of pups during ontogeny. If possible, three syllables from three different syllable trains (i.e. sequence of at least five syllables) were selected and acoustic parameters were extracted with the software Avisoft SASLab Pro. Acoustic parameters were extracted for the pulsed part (BZ1, BZ2, BZ3 in the figure) and for the tonal part (T1, T2, T3 in the figure). Measurements were subsequently averaged (separately for pulsed and tonal part) for each train to minimize temporal dependence among syllables produced within close succession. The box illustrates that for each syllable, acoustic parameters were taken from only one harmonic containing most energy (to render comparison across syllables and syllable parts possible). As described in the method section, we used derived parameters (principal components), original parameters (e.g. duration) and LFCCs (see Materials and methods). The spectrogram was created in Avisoft SASLab Pro: Hamming window with 1024-point Fourier transformation and 87.5% overlap (frequency resolution: 244 Hz, time resolution: 0.512ms).

Figure 2—figure supplement 5

Download asset Open asset

Acoustic movement during ontogeny.

This figure shows the DFA space for one of the colonies with three exemplary pups, showing the acoustic change between the early (square) and the late (triangle) babbling phase. For each pup, we calculated individual centroids for each phase based on the DFAs. The numbers next to the arrows depict the calculated Euclidean distances. The figure includes a pup showing almost no change (violet colors), some change (blue colors), and large change (green colors).

Tables

Table 1

The influence of the social environment on the amount of vocal practice.

Response variables (first column): A: Babbling bout duration [s]: the babbling bout duration measured in seconds (N=19 pups, N=186 babbling bouts, GLMM, family Gamma with log link, random factor pup ID: repeated measurements over time of the same focal pups). B: Babbling phase duration: The number of days pups spent babbling (N=19 pups, LMER, random factor colony ID: measurements of pups from the same colony). The predictor variables for both models were z-transformed, a standardization procedure that facilitates convergence of the model. The second column depicts the fixed and random effects for both models. The third column depicts the estimate and the standard error as well as the standard deviation for the random effect, and the last column the p-value (bold if influence is significant). Abbreviations: ‘Mat. behav.’=maternal behaviors.

Response variable	Parameter	Estimate (SE)	Significance
A Babbling bout duration [s]	Fixed effects
	Intercept	6.038 (0.074)	p<0.001
	Mat. behav. during babbling (standardized)	0.537 (0.044)	p<0.001
	Pup age (standardized)	0.100 (0.044)	p=0.022
	No. of tutors (standardized)	–0.041 (0.077)	p=0.6
	Random effect (standard deviation)
	Pup ID	0.19
B Babbling phase duration [days]	Fixed effects
	Intercept	44.322 (5.222)	p=0.001
	Maternal activity score (standardized)	7.317 (2.824)	p=0.02
	No. of tutors (standardized)	3.560 (5.167)	p=0.530
	Random effect (standard deviation)
	Colony ID	11.20

Table 2

The influence of the social environment on learned syllables.

Response variables: A: Total number of song syllables: the learned territorial song syllables which were present in a babbling bout, (N=10 pups, N=90 babbling bouts, GLMM, family negative binomial with log link, random factor ID: repeated measurements over time of the same focal pups). B: Song syllable versatility: how many of the five different song syllable types (Figure 2—figure supplement 1) were present in a babbling bout, (N=10 pups, N=90 babbling bouts, LMER, random factor ID: repeated measurements over time of the same focal pups). Neither the total number of song syllables nor song syllable versatility differed between pup sex (SI). C: Percentage of mature song syllables: We investigated the influence of different predictor variables on the percentage of mature song syllables of the most common syllable type, B2 (Appendix 1—table 3 and N=10 pup, N=85 babbling bouts, GLMM family binomial with logit link, random factor ID: repeated measurements over time of the same focal pups, random factor observation-level: to avoid overdispersion). The predictor variables for both models were z-transformed, a standardization procedure that facilitates convergence of the model. Fixed and random effects are depicted in the second column. The third column depicts the estimate with standard error, and the last column the p-value (significant influence indicated in bold). Abbreviation: ‘Mat. behav.’=maternal behaviors.

Response variable	Parameter	Estimate (SE)	Significance
A Total number of song syllables	Fixed effects
	Intercept	4.188 (0.194)	p<0.001
	Maternal behavior during babbling (standardized)	0.460 (0.098)	p<0.001
	Pup age (standardized)	0.446 (0.096)	p<0.001
	No. of tutors (standardized)	0.036 (0.193)	p=0.9
	Random effects (standard deviation)
	Pup ID	0.55
B Song syllable versatility	Fixed effects
	Intercept	0.889 (0.030)	p<0.001
	Maternal behavior during babbling (standardized)	0.105 (0.031)	p<0.001
	Pup age (standardized)	0.125 (0.030)	p<0.001
	No. of tutors (standardized)	–0.022 (0.030)	p=0.479
	Random effects (standard deviation)
	Pup ID	0.04
C Percentage of mature song syllables	Fixed effects
	Intercept	–0.698 (0.355)	p=0.049
	Maternal behavior during babbling (standardized)	0.376 (0.188)	p=0.045
	Pup age (standardized)	0.185 (0.177)	p=0.296
	No. of tutors (standardized)	0.107 (0.341)	p=0.754
	Random effects (standard deviation)
	Pup ID	0.98
	Observation-level	1.24

Table 3

Terminology.

Term	Description
Syllable type	Sound unit with characteristic spectro-temporal features
Undifferentiated proto-syllables	Syllables only produced by pups, occurring in babbling and vanish after weaning.
Adult-like syllable	Syllables which are present in the later adult vocal repertoire (after weaning).
Buzz syllable types of song	Territorial song syllables with discernible pulsed part; 5 distinct syllable types: B1, B2, B2 trill, B3, B4 (only syllable with smeared buzz part) (Figure 2—figure supplement 1).
Mature syllable	Mature syllables show no gap (e.g. between the pulsed and tonal part, within the buzzed part) (Figure 2—figure supplement 3).
Precursor syllable	Distinct gaps within the pulsed part and/or between pulsed and tonal part (Figure 2—figure supplement 3).
Early babbling phase	First half of entire babbling phase
Late babbling phase	Second half of entire babbling phase
Song syllables	Syllables of which the territorial song is composed of (Figure 2A)

Key resources table

Reagent type (species) or resource	Designation	Source or reference	Identifiers	Additional information
Biological sample (Saccopteryx bilineata)	Saccopteryx bilineata; S. bilineata; greater sac-winged bat	other	N/A	Wild animals: two different populations (Costa Rica and Panamá), both male and female pups (N=19 mother-pup pairs) ‘see Material and methods, sections Study species, study sites, acoustic and behavioral data collection’
Software, algorithm	SPSS	IBM SPSS Statistics	RRID:SCR_002865
Software, algorithm	MATLAB	MATLAB	RRID:SCR_001622
Software, algorithm	R	R Studio	RRID:SCR_000432
Software, algorithm	Avisoft SASLab-Pro	Avisoft Bioacoustics	RRID:SCR_014438
Software, algorithm	VLC	VLC media player	N/A

Appendix 1—table 1

Ethogram of behaviors during babbling bouts.

Behavior	Description	Observed in:
crawl away*	One individual crawls a certain distance (>2 cm) away from the other.	P / M
crawl towards*	One individual crawls towards another individual (approaches up to 2 cm).	P / M
hover*	Hover flight in front of an individual (for a few seconds).	P / M
wing-poke*	Poking a conspecific with wrist. Pups normally repeatedly poke their mothers. Mothers usually poke the pup only once to terminate the babbling bout. Includes wing pokes with and without physical contact.	P / M
short flight*	Individual flies to another spot in the day roost or individual flies towards interaction partner.	P / M
rock	Rocking with entire body from side to side; also used when mother wants pup to detach from teat.	P / M
wing stretch	Wing is stretched out completely.	P / M
head stretch	Head and neck are conspicuously bent and stretched to either left or right side.	P
push up	Push wrists against wall and push body up.	P / M
dribble	Pound with wrists against wall, very fast and repetitive.	P
wrist lift	Pound with wrists against wall once, not hitting other individual.	P
babble	Vocal practice.	P

The behaviors labeled with a * either occurred as solitary behavior or as behavioral sequence (i.e. interaction). These interactive behaviors were used to assess the maternal behavioral influence on pup babbling. The behaviors without a * (with the exception of babble) were defined as comfort behaviors and not included as behavioral feedback. The third column depicts if the behavior was observed in the pup (P) or both pup and mother (P/M).

Appendix 1—table 2

Maternal display rates: duration of babbling is not a confounding factor.

Response variable	Parameter	Estimate (SE)	Significance
A Babbling bout duration [s]	Fixed effects
	Intercept	6.132 (0.091)	p<0.001
	Mat. behav. rates during babbling (standardized)	–0.146 (0.053)	p=0.006
	Pup age (standardized)	0.418 (0.050)	p<0.001
	No. of tutors (standardized)	–0.063 (0.095)	p=0.51
	Random effect (standard deviation)
	Pup ID	0.24

The number of displays in longer bouts could just reflect that more displays are possible in a longer period. To assess this potential confounding factor, we investigated the effect of display rates on vocal bout duration (N=19 pups, N=186 babbling bouts, GLMM family Gamma with log link).

Appendix 1—table 3

The average proportion of syllable type production.

ID	B1	B2	B2-trill	B3	B4
11	0	47.8	6.6	24.7	12.6
12	25.8	56.1	6.6	9.7	1.8
13	0.13	50.2	12.7	16.1	6.7
14	0	54.9	25.8	27.8	19.2
15	0.1	54.9	16.2	9.0	19.7
16	4.9	67.2	1.6	20.6	5.7
17	2.3	73.5	3.4	20.4	0.4
18	0	54.2	14.5	24.1	7.2
19	0	53.7	7.8	24.8	13.8
20	0	50.9	24.5	11.6	5.3

For each pup, the average proportion of each learned syllable type produced during all analysed babbling bouts is displayed. Averaged over all pups B2 is the most commonly produced syllable type within babbling bouts (53.8%), followed by B3 (18.6%), B2 trill (11.4%), B4 (8.9%), and B1 (2.8%).

Author response table 1

AIC	BIC	loglik	deviance	df.resid
2594.9	2614.2	-1291.4	2582.9	180

Author response table 2

Scaled residuals.

Min	1Q	Median	3Q	Max
-1.4604	-0.7379	-0.1924	0.5238	2.6564

Author response table 3

Random effects.

Groups Name	Variance	Std.Dev.
ID (Intercept)	0.0569	0.2385
Residual	0.3664	0.6053

Number of obs: 186, groups: ID, 19.

Author response table 4

Fixed effects.

	Estimate	Std. Error	t value	Pr(> \|z\|)
(Intercept)	6.13150	0.09114	67.278	< 2e-16 ***
age.z	0.41835	0.05048	8.287	< 2e-16 ***
behavioural_quotient.log.z	-0.14627	0.05310	-2.754	0.00588 **
nomales.z	-0.06251	0.09473	-0.660	0.50934