Machine learning reveals cryptic dialects that explain mate choice in a songbird

Maternal immune activation alters bout structure of rat 50-kHz ultrasonic vocalizations

Dysfunctional sequencing of behaviour and cognition is observed in schizophrenia across multiple domains, including during communication. We examined whether maternal immune activation (MIA), a risk factor for schizophrenia, disrupted the sequential organization of ultrasonic vocalizations (USVs) in a rat model. We analysed the structure of bursts of 50-kHz USVs (bouts) in two independent datasets (paired-rat: 19 control, 18 MIA; reward paradigm: 18 control, 20 MIA), using a Damerau-Levenshtein analysis with a k-fold cross-validation procedure. MIA animals showed greater variability in their bout sequences in both datasets, with lower Levenshtein similarity index (LSI) scores compared to control animals. Notably, MIA set median sequences were more similar to control bout sequences than to their own group's sequences, suggesting a breakdown in sequential organization. Additionally, we found an alteration to 50-kHz USV transitional preferences in MIA in a reward context. While sequence structure was altered, basic call production and call-type distribution remained largely intact across groups. These findings demonstrate that MIA specifically appears to affect the organization of vocal sequences at the bout level, while largely preserving basic vocalization patterns. This work extends our understanding of the effects of maternal infection during pregnancy, and how this can lead to altered communication sequences that are relevant to schizophrenia risk.

[Meaning and Mechanisms of Birdsong: Inspiration for Pneumology]

In contrast to humans, the location where sound is produced in birds is not the larynx, but rather the so-called "vocal box" (scientific term "Syrinx"). In some species the syrinx is located at the bifurcation point of the trachea into the two main bronchi (tracheal vocal head), while in some in the main bronchi (bronchial vocal head). During inspiration, part of the air flows into the lungs, and the part needed for singing flows into the air sacs adjacent to the lungs. During expiration, air leaves the air sacs and flows through the syrinx, where the song is created. When birds sing in two voices at the same time, individual sequences are formed simultaneously in the right and left parts of the syrinx.The song analysis is based on spectrograms (so-called sonagrams), which graphically represent the frequency spectrum of bird song.The song consists of one or more verses, which in turn consist of the variable or constant sequence of motives or syllables. Some songbirds have an enormous repertoire of syllables and verses (max. up to 7000 verses per day). In addition to singing, most bird species also have much simpler begging, contact, threatening, flight, alarm and copulation calls.Male birds sing primarily for two reasons: 1. They use song to woo a potential partner. This song provides the females with important information about the applicant's performance and health. 2. Singing serves to defend the territory.In around 40 % of songbird species, females also sing. Pairs of some species sing in perfect synchronization.A number of songbirds imitate both the voices of other songbirds and ambient noises, and many songbirds have regional dialects.Song development depends on genetics and other factors such as the environment, metabolism and hormonal influences. It proceeds step by step and initially includes relatively primitive sequences (so-called "subsongs"), then leads through more complex intermediate forms ("plastic songs") and finally to the completed singing pattern ("full songs").Young birds learn the song of their species at a time when they are not yet singing themselves, often as nestlings aged 10 to 50 days from older members of the species, usually from their fathers.The song of young birds develops, based on the template of adult song, in a network of sensory-motor neurons in the forebrain.Songbirds, especially the zebra finch, currently offer the best model for the neural basis of human language learning. In birds, the so-called "High Vocal Center" orchestrates all brain regions relevant to songs, with the neural control of song being sensitive to sex hormones.

Responses in adult pied flycatcher males depend on playback song similarity to local population

Song divergence driven by social learning has been proposed to be a key factor driving allopatric speciation in oscine birds. Songbirds often respond more to songs deriving from their local population, suggesting the potential for acoustic divergence across populations to shape both intra- and intersexual interactions. However, many of these studies were conducted on species with simple songs and, as a result, we know comparatively little about the emergence of population differences and song discrimination in species with complex songs. We addressed this question in the pied flycatcher (Ficedula hypoleuca) by calculating the dissimilarity of songs from 2 foreign populations as well as from our study site to the local centroid. We then conducted a paired-design playback experiment where both local and foreign songs were played simultaneously. We found that pied flycatcher males showed significantly stronger responses to those songs that sounded more similar to the local population. This suggests that despite the high complexity of the pied flycatcher song, individuals are still able to discriminate across populations. Our results support the hypothesis that learned song divergence can act as a mechanism for assortative mating and allopatric speciation.

The use of artificial songs to assess song recognition in imprinted female songbirds: a concept proposal

We propose an experimental paradigm to examine acoustic features responsible for song preference and recognition in songbirds. Song preference in female songbirds is often influenced by early song experience. That is why several Estrildid species, including our subject species, the Java sparrow (Padda oryzivora), are known to show an imprinted preference for their father's songs. After confirming that Java sparrow females preferred their father's song compared to non-imprinted through song playbacks (first step), we repeated the playback tests in the same subjects using synthesized stimuli (second step). To create synthesized stimuli, we removed all the complex frequency modulations and subharmonics from song notes that we used for the first step playback tests to see the effect of spectrometric features on song recognition. The results indicated that females showed higher rate of calling towards synthesized father song stimuli, suggesting that the macroscopic patterns would play more important roles in song recognition than the microscopic acoustic features. Although we looked at spectrometric features and father-imprinted song preference in this study, similar testing can be applied in many ways to test preference for local dialects or subspecies-specific songs.

Response of forest Turtur doves to conspecific and congeneric songs in sympatry and allopatry

Birds have a diverse acoustic communication system, and the ability to recognise their own species' song from a distance facilitates complex behaviours related to mate attraction and rival deterrence. However, certain species, including doves, do not learn songs and their vocal repertoires are much simpler than those of better-studied songbirds. In these so-called non-learning birds, relatively little is known about the role that bird song plays in intra- and interspecific interactions, and how such behaviours might be acquired (inherited or learned from experience). To investigate this question, we focused on two species of African wood doves whose long-range songs are used in a territorial context. Specifically, we examined the responses of sympatric and allopatric populations of male blue-headed wood-doves (Turtur brehmeri) and tambourine doves (Turtur tympanistria) to different types of simulated territorial intrusions, i.e. playback of conspecific, congeneric, and control songs. We aimed to assess (i) whether these species, which have similar songs, respond only to their own species' song or exhibit interspecific territoriality, and (ii) if the response pattern is affected by the presence or absence of congeners in the general area. We found that both species responded strongly to playback of their own species in both sympatric and allopatric populations. In allopatry, though, male tambourine doves misdirected their response and also approached the playback of congeneric songs. Our results indicate that, in areas where the studied Turtur doves live in sympatry, they do not exhibit consistent interspecific territoriality. However, we cannot exclude the possibility that the smaller tambourine dove avoids its larger congener during the process of territory establishment. The difference in tambourine doves' response toward the song of present (sympatric) or absent (allopatric) congeners suggests that the ability to discriminate between songs of similarly singing potential competitors is acquired through earlier interactions and learning. This plasticity in response supports the misdirected aggression hypothesis, which argues that interspecific territorialism emerges as a maladaptive by-product of signal similarity. However, on an evolutionary timescale, such an ability could be considered an adaptive cognitive tool useful for resolving competing interests with congeners.

Differentially private knowledge transfer for federated learning

Extracting useful knowledge from big data is important for machine learning. When data is privacy-sensitive and cannot be directly collected, federated learning is a promising option that extracts knowledge from decentralized data by learning and exchanging model parameters, rather than raw data. However, model parameters may encode not only non-private knowledge but also private information of local data, thereby transferring knowledge via model parameters is not privacy-secure. Here, we present a knowledge transfer method named PrivateKT, which uses actively selected small public data to transfer high-quality knowledge in federated learning with privacy guarantees. We verify PrivateKT on three different datasets, and results show that PrivateKT can maximally reduce 84% of the performance gap between centralized learning and existing federated learning methods under strict differential privacy restrictions. PrivateKT provides a potential direction to effective and privacy-preserving knowledge transfer in machine intelligent systems.

Challenges of mismatching timescales in longitudinal studies of collective behaviour

How individuals' prior experience and population evolutionary history shape emergent patterns in animal collectives remains a major gap in the study of collective behaviour. One reason for this is that the processes that can shape individual contributions to collective actions can happen over very different timescales from each other and from the collective actions themselves, resulting in mismatched timescales. For example, a preference to move towards a specific patch might arise from phenotype, memory or physiological state. Although providing critical context to collective actions, bridging different timescales remains conceptually and methodologically challenging. Here, we briefly outline some of these challenges, and discuss existing approaches that have already generated insights into the factors shaping individual contributions in animal collectives. We then explore a case study of mismatching timescales-defining relevant group membership-by combining fine-scaled GPS tracking data and daily field census data from a wild population of vulturine guineafowl (Acryllium vulturinum). We show that applying different temporal definitions can produce different assignments of individuals into groups. These assignments can then have consequences when determining individuals' social history, and thus the conclusions we might draw on the impacts of the social environment on collective actions. This article is part of a discussion meeting issue 'Collective behaviour through time'.

Mapping cryptic binding sites of drug targets to overcome drug resistance

The emergence of drug resistance is a primary obstacle for successful chemotherapy. Drugs that target cryptic binding sites (CBSs) represent a novel strategy for overcoming drug resistance. In this short communication, we explain and discuss how the discovery of CBSs and their inhibitors can overcome drug resistance.

Behavioural ecology at the spatial-social interface

Spatial and social behaviour are fundamental aspects of an animal's biology, and their social and spatial environments are indelibly linked through mutual causes and shared consequences. We define the 'spatial-social interface' as intersection of social and spatial aspects of individuals' phenotypes and environments. Behavioural variation at the spatial-social interface has implications for ecological and evolutionary processes including pathogen transmission, population dynamics, and the evolution of social systems. We link spatial and social processes through a foundation of shared theory, vocabulary, and methods. We provide examples and future directions for the integration of spatial and social behaviour and environments. We introduce key concepts and approaches that either implicitly or explicitly integrate social and spatial processes, for example, graph theory, density-dependent habitat selection, and niche specialization. Finally, we discuss how movement ecology helps link the spatial-social interface. Our review integrates social and spatial behavioural ecology and identifies testable hypotheses at the spatial-social interface.

Pairwise and high-order dependencies in the cryptocurrency trading network

In this paper we analyse the effects of information flows in cryptocurrency markets. We first define a cryptocurrency trading network, i.e. the network made using cryptocurrencies as nodes and the Granger causality among their weekly log returns as links, later we analyse its evolution over time. In particular, with reference to years 2020 and 2021, we study the logarithmic US dollar price returns of the cryptocurrency trading network using both pairwise and high-order statistical dependencies, quantified by Granger causality and O-information, respectively. With reference to the former, we find that it shows peaks in correspondence of important events, like e.g., Covid-19 pandemic turbulence or occasional sudden prices rise. The corresponding network structure is rather stable, across weekly time windows in the period considered and the coins are the most influential nodes in the network. In the pairwise description of the network, stable coins seem to play a marginal role whereas, turning high-order dependencies, they appear in the highest number of synergistic information circuits, thus proving that they play a major role for high order effects. With reference to redundancy and synergy with the time evolution of the total transactions in US dollars, we find that their large volume in the first semester of 2021 seems to have triggered a transition in the cryptocurrency network toward a more complex dynamical landscape. Our results show that pairwise and high-order descriptions of complex financial systems provide complementary information for cryptocurrency analysis.