Syllable Features Research Articles

Comparing behaviours induced by natural memory retrieval and optogenetic reactivation of an engram ensemble in mice.

Memories are thought to be stored within sparse collections of neurons known as engram ensembles. Neurons active during a training episode are allocated to an engram ensemble ('engram neurons'). Memory retrieval is initiated by external sensory or internal cues present at the time of training reactivating engram neurons. Interestingly, optogenetic reactivation of engram ensemble neurons alone in the absence of external sensory cues is sufficient to induce behaviour consistent with memory retrieval in mice. However, there may exist differences between the behaviours induced by natural retrieval cues or artificial engram reactivation. Here, we compared two defensive behaviours (freezing and the syllable structure of ultrasonic vocalizations, USVs) induced by sensory cues present at training (natural memory retrieval) and optogenetic engram ensemble reactivation (artificial memory retrieval) in a threat conditioning paradigm in the same mice. During natural memory recall, we observed a strong positive correlation between freezing levels and distinct USV syllable features (characterized by an unsupervised algorithm, MUPET (Mouse Ultrasonic Profile ExTraction)). Moreover, we observed strikingly similar behavioural profiles in terms of freezing and USV characteristics between natural memory recall and artificial memory recall in the absence of sensory retrieval cues. Although our analysis focused on two behavioural measures of threat memory (freezing and USV characteristics), these results underscore the similarities between threat memory recall triggered naturally and through optogenetic reactivation of engram ensembles. This article is part of a discussion meeting issue 'Long-term potentiation: 50 years on'.

Zebra finch song parameters are affected by the breeding status of the male, but not temperature variability

Bird song is a crucial feature for mate choice and reproduction. Song can potentially communicate information related to the quality of the mate, through song complexity, structure or finer changes in syllable characteristics. It has been shown in zebra finches that those characteristics can be affected by various factors including motivation, hormone levels or extreme temperature. However, although the literature on zebra finch song is substantial, some factors have been neglected. In this paper, we recorded male zebra finches in two breeding contexts (before and after pairing) and in two ambient temperature conditions (stable and variable) to see how those factors could influence song production. We found strong differences between the two breeding contexts: compared to their song before pairing, males that were paired had lower song rate, syllable consistency, frequency and entropy, while surprisingly the amplitude of their syllables increased. Temperature variability had an impact on the extent of these differences, but did not directly affect the song parameters that we measured. Our results describe for the first time how breeding status and temperature variability can affect zebra finch song, and give some new insights into the subtleties of the acoustic communication of this model species.

An auditory perspective on phonological development in infancy.

The auditory system encodes the phonetic features of languages by processing spectro-temporal modulations in speech, which can be described at two time scales: relatively slow amplitude variations over time (AM, further distinguished into the slowest <8-16 Hz and faster components 16-500 Hz), and frequency modulations (FM, oscillating at higher rates about 600-10 kHz). While adults require only the slowest AM cues to identify and discriminate speech sounds, infants have been shown to also require faster AM cues (>8-16 Hz) for similar tasks. Using an observer-based psychophysical method, this study measured the ability of typical-hearing 6-month-olds, 10-month-olds, and adults to detect a change in the vowel or consonant features of consonant-vowel syllables when temporal modulations are selectively degraded. Two acoustically degraded conditions were designed, replacing FM cues with pure tones in 32 frequency bands, and then extracting AM cues in each frequency band with two different low-pass cut- off frequencies: (1) half the bandwidth (Fast AM condition), (2) <8 Hz (Slow AM condition). In the Fast AM condition, results show that with reduced FM cues, 85% of 6-month-olds, 72.5% of 10-month-olds, and 100% of adults successfully categorize phonemes. Among participants who passed the Fast AM condition, 67% of 6-month-olds, 75% of 10-month-olds, and 95% of adults passed the Slow AM condition. Furthermore, across the three age groups, the proportion of participants able to detect phonetic category change did not differ between the vowel and consonant conditions. However, age-related differences were observed for vowel categorization: while the 6- and 10-month-old groups did not differ from one another, they both independently differed from adults. Moreover, for consonant categorization, 10-month-olds were more impacted by acoustic temporal degradation compared to 6-month-olds, and showed a greater decline in detection success rates between the Fast AM and Slow AM conditions. The degradation of FM and faster AM cues (>8 Hz) appears to strongly affect consonant processing at 10 months of age. These findings suggest that between 6 and 10 months, infants show different developmental trajectories in the perceptual weight of speech temporal acoustic cues for vowel and consonant processing, possibly linked to phonological attunement.

Exploring the impact of tonal inventory on speech perception across languages: a study of MMN responses in tonal language speakers.

Previous research on the perception of segmental features of languages has established a correlation between the phoneme inventory of a language and its speakers' perceptual abilities, as indexed by discrimination tasks and Mismatch Negativity (MMN). Building on this background, the current study elucidated the relationship between perceptual ability and tonal inventory by utilizing two tonal languages. Two groups of participants were included in the present experiment: Mandarin speakers and Hakka-Mandarin speakers. Onset latency analysis revealed a significant difference in the Mandarin syllable condition, with Hakka-Mandarin speakers demonstrating earlier MMN latency than Mandarin speakers. This suggests a more efficient auditory processing mechanism in Hakka-Mandarin speakers. Both groups, however, showed similar MMN latency in the Hakka syllable condition. The interaction between language background and syllable type indicates that other factors, such as syllable sonority, also influence MMN responses. These findings highlight the importance of considering multiple phonemic inventories and syllable characteristics in studies of tonal perception.

Machine learning reveals that climate, geography, and cultural drift all predict bird song variation in coastal Zonotrichia leucophrys

Abstract Previous work has demonstrated that there is extensive variation in the songs of White-crowned Sparrow (Zonotrichia leucophrys) throughout the species range, including between neighboring (and genetically distinct) subspecies Z. l. nuttalli and Z. l. pugetensis. Using a machine learning approach to bioacoustic analysis, we demonstrate that variation in song is correlated with year of recording (representing cultural drift), geographic distance, and climatic differences, but the response is subspecies- and season-specific. Automated machine learning methods of bird song annotation can process large datasets more efficiently, allowing us to examine 1,913 recordings across ~60 years. We utilize a recently published artificial neural network to automatically annotate White-crowned Sparrow vocalizations. By analyzing differences in syllable usage and composition, we recapitulate the known pattern where Z. l. nuttalli and Z. l. pugetensis have significantly different songs. Our results are consistent with the interpretation that these differences are caused by the changes in characteristics of syllables in the White-crowned Sparrow repertoire. This supports the hypothesis that the evolution of vocalization behavior is affected by the environment, in addition to population structure.

Bio-inspired camouflage communication in air imitating birds songs

Radio communication is the most common means of communication nowadays, however, radio interference technology is becoming mature. Once radio communications are disrupted in the forest, people would be left in a passive position without any effective communication method. Sound is an also effective carrier of information. To realize effective communication in the forest environment, we proposed a bio-inspired camouflage communication strategy(BCCS) by imitating bird songs and took the Verditer Flycatcher as an example to demonstrate the feasibility of BCCS. We proposed a compound characteristics grouping coding method based on the tree structure to divide syllables into 32 feature groups. And five time–frequency(T-F) characteristics of syllables were extracted as grouping layers: syllable length, average frequency, frequency trend, the standard deviation of global frequency, and the difference of local frequency standard deviations. Then the correlation was used as the unique identification of a syllable within the group. Syllables in the feature group were screened to ensure high autocorrelation and low cross-correlation, which are the key basis in decoding. Each individual syllable can encode 8 bits of binary information through compound characteristics coding and correlation coding. Since the communication sequence is completely composed of the original bird songs and follows the regularities of distribution, BCCS has a good camouflage ability. Experiments were carried out to verify the feasibility and camouflage ability of BCCS in the forest environment.

Machine learning and statistical classification of birdsong link vocal acoustic features with phylogeny

Birdsong is a longstanding model system for studying evolution and biodiversity. Here, we collected and analyzed high quality song recordings from seven species in the family Estrildidae. We measured the acoustic features of syllables and then used dimensionality reduction and machine learning classifiers to identify features that accurately assigned syllables to species. Species differences were captured by the first 3 principal components, corresponding to basic frequency, power distribution, and spectrotemporal features. We then identified the measured features underlying classification accuracy. We found that fundamental frequency, mean frequency, spectral flatness, and syllable duration were the most informative features for species identification. Next, we tested whether specific acoustic features of species’ songs predicted phylogenetic distance. We found significant phylogenetic signal in syllable frequency features, but not in power distribution or spectrotemporal features. Results suggest that frequency features are more constrained by species’ genetics than are other features, and are the best signal features for identifying species from song recordings. The absence of phylogenetic signal in power distribution and spectrotemporal features suggests that these song features are labile, reflecting learning processes and individual recognition.

Zebra finches (Taeniopygia guttata) demonstrate cognitive flexibility in using phonology and sequence of syllables in auditory discrimination

Zebra finches rely mainly on syllable phonology rather than on syllable sequence when they discriminate between two songs. However, they can also learn to discriminate two strings containing the same set of syllables by their sequence. How learning about the phonological characteristics of syllables and their sequence relate to each other and to the composition of the stimuli is still an open question. We compared whether and how the zebra finches’ relative sensitivity for syllable phonology and syllable sequence depends on the differences between syllable strings. Two groups of zebra finches were trained in a Go-Left/Go-Right task to discriminate either between two strings in which each string contained a unique set of song syllables (‘Different-syllables group’) or two strings in which both strings contained the same set of syllables, but in a different sequential order (‘Same-syllables group’). We assessed to what extent the birds in the two experimental groups attend to the spectral characteristics and the sequence of the syllables by measuring the responses to test strings consisting of spectral modifications or sequence changes. Our results showed no difference in the number of trials needed to discriminate strings consisting of either different or identical sets of syllables. Both experimental groups attended to changes in spectral features in a similar way, but the group for which both training strings consisted of the same set of syllables responded more strongly to changes in sequence than the group for which the training strings consisted of different sets of syllables. This outcome suggests the presence of an additional learning process to learn about syllable sequence when learning about syllable phonology is not sufficient to discriminate two strings. Our study thus demonstrates that the relative importance of syllable phonology and sequence depends on how these features vary among stimuli. This indicates cognitive flexibility in the acoustic features that songbirds might use in their song recognition.

Syllable Structure in Setswana Personal Names

The study analyses the characteristics of syllables in personal names of the Setswana (Tswana) language (Sotho subgroup of the Bantu languages, Southern Africa). The authors focus on 1,001 most frequent names extracted from a dataset of 1,093,265 names using Wordsmith Tools. Of all the studied names, 55.5% are male, while 40.2% are female and 4.3% — unisex names. Most Setswana names have three syllables followed by those with four syllables (more than 70% of the studied personal names are either three or four syllable names), five syllable names are the third most common, while six syllable names are rare in Setswana. The study reveals that in female names, the first and penultimate (perceptually most prominent) syllables statistically more often end in front vowels, while in male names, in the same positions, back vowels are more frequent. At the same time, the last phoneme in most female names is a back vowel, while most of male names end in front vowels. Female names also more often terminate in syllabic constants compared to male names. All this suggests that in Setswana, there are distinct patterns for male and female names, which means that Setswana speakers are phonologically aware of the relations between the gender of a person and the phonological form of his or her name. These patterns found in Setswana are compared to what is known about the syllable structure of personal names in English and French. The results demonstrate that the phonological patterns in Setswana personal names may display sound symbolism.

Low-Power Feature-Attention Chinese Keyword Spotting Framework with Distillation Learning

In this paper, we propose a novel Low-Power Feature-Attention Chinese Keyword Spotting Framework based on a depthwise separable convolution neural network (DSCNN) with distillation learning to recognize speech signals of Chinese wake-up words. The framework consists of a low-power feature-attention acoustic model and its learning methods. Different from the existing model, the proposed acoustic model based on connectionist temporal classification (CTC) focuses on the reduction of power consumption by reducing model network parameters and multiply-accumulate (MAC) operations through our designed feature-attention network and DSCNN. In particular, the feature-attention network is specially designed to extract effective syllable features from a large number of MFCC features. This could refine MFCC features by selectively focusing on important speech signal features and removing invalid speech signal features to reduce the number of speech signal features, which helps to significantly reduce the parameters and MAC operations of the whole acoustic model. Moreover, DSCNN with fewer parameters and MAC operations compared with traditional convolution neural networks is adopted to extract effective high-dimensional features from syllable features. Furthermore, we apply a distillation learning algorithm to efficiently train the proposed low-power acoustic model by utilizing the knowledge of the trained large acoustic model. Experimental results thoroughly verify the effectiveness of our model and show that the proposed acoustic model still has better accuracy than other acoustic models with the lowest power consumption and smallest latency measured by NVIDIA JETSON TX2. It has only 14.524 KB parameters and consumes only 0.141 J energy per query and 17.9 ms latency on the platform, which is hardware-friendly.

Automated call detection for acoustic surveys with structured calls of varying length

AbstractWhen recorders are used to survey acoustically conspicuous species, identification calls of the target species in recordings is essential for estimating density and abundance. We investigate how well deep neural networks identify vocalisations consisting ofphrasesof varying lengths, each containing a variable number ofsyllables. We use recordings of Hainan gibbonNomascus hainanusvocalisations to develop and test the methods.We propose two methods for exploiting the two‐level structure of such data. The first combines convolutional neural network (CNN) models with a hidden Markov model (HMM) and the second uses a convolutional recurrent neural network (CRNN). Both models learn acoustic features of syllables via a CNN and temporal correlations of syllables into phrases either via an HMM or recurrent network. We compare their performance to commonly used CNNs LeNet and VGGNet, and support vector machine (SVM). We also propose a dynamic programming method to evaluate how well phrases are predicted. This is useful for evaluating performance when vocalisations are labelled by phrases, not syllables.Our methods perform substantially better than the commonly used methods when applied to the gibbon acoustic recordings. The CRNN has anF‐score of 90% on phrase prediction, which is 18% higher than the best of the SVM or LeNet and VGGNet methods. HMM post‐processing raised theF‐score of these last three methods to as much as 87%. The number of phrases is overestimated by CNNs and SVM, leading to error rates between 49% and 54%. With HMM, these error rates can be reduced to 0.4% at the lowest. Similarly, the error rate of CRNN's prediction is no more than 0.5%.CRNNs are better at identifying phrases of varying lengths composed of a varying number of syllables than simpler CNN or SVM models. We find a CRNN model to be best at this task, with a CNN combined with an HMM performing almost as well. We recommend that these kinds of models are used for species whose vocalisations are structured into phrases of varying lengths.

Diversity Monitoring of Coexisting Birds in Urban Forests by Integrating Spectrograms and Object-Based Image Analysis

In the context of rapid urbanization, urban foresters are actively seeking management monitoring programs that address the challenges of urban biodiversity loss. Passive acoustic monitoring (PAM) has attracted attention because it allows for the collection of data passively, objectively, and continuously across large areas and for extended periods. However, it continues to be a difficult subject due to the massive amount of information that audio recordings contain. Most existing automated analysis methods have limitations in their application in urban areas, with unclear ecological relevance and efficacy. To better support urban forest biodiversity monitoring, we present a novel methodology for automatically extracting bird vocalizations from spectrograms of field audio recordings, integrating object-based classification. We applied this approach to acoustic data from an urban forest in Beijing and achieved an accuracy of 93.55% (±4.78%) in vocalization recognition while requiring less than ⅛ of the time needed for traditional inspection. The difference in efficiency would become more significant as the data size increases because object-based classification allows for batch processing of spectrograms. Using the extracted vocalizations, a series of acoustic and morphological features of bird-vocalization syllables (syllable feature metrics, SFMs) could be calculated to better quantify acoustic events and describe the soundscape. A significant correlation between the SFMs and biodiversity indices was found, with 57% of the variance in species richness, 41% in Shannon’s diversity index and 38% in Simpson’s diversity index being explained by SFMs. Therefore, our proposed method provides an effective complementary tool to existing automated methods for long-term urban forest biodiversity monitoring and conservation.

Dynamic Time Warping Features Extraction Design for Quranic Syllable-based Harakaat Assessment

The use of technological speech recognition systems with a variety of approaches and techniques has grown exponentially in varieties of human-machine interaction applications. The assessment for Qur'anic recitation errors based on syllables utterance is used to meet the Tajweed rules which generally consist of Harakaat (prolonging). The digital transformation of Quranic voice signals with identification of Tajweed-based recitation errors of Harakaat is the main research work in this paper. The study focused on speech processing implemented using the representation of Quranic Recitation Speech Signals (QRSS) in the best digital format based on Al-Quran syllables and feature extraction design to reveal similarities or differences in recitation (based on Al-Quran syllables) between experts and student. The method of Dynamic Time Warping (DTW) is used as Short Time Frequency Transform (STFT) of QRSS syllable feature for Harakaat measurement. Findings from this paper include an approach based on human-guidance threshold classification that is used specifically to evaluate Harakaat based on the syllables of the Qur'an. The threshold classification performance obtained for Harakaat is above 80% in the training and testing stages. The results of the analysis at the end of the experiment have concluded that the threshold classification method for Minimum Path Cost (MPC) feature parameters can be used as an important feature to evaluate the rules of Tajwid Harakaat embedded in syllables.

Syllable isochrony and the prosodic features of stop syllables in Cantonese

AbstractCantonese is a syllable-timed language: that is, the syllable is the isochronous unit of speech. However, in Cantonese, there is a type of closed syllable with the stop codas [-p], [-t], or [-k] (i.e. syllables with the so called “entering-tones”) which sound much shorter than other syllables. On the surface, the shorter duration of stop syllables and the general prosodic feature of syllable-isochrony seem to conflict. This study conducted acoustic investigations of stop syllables in Cantonese in different contexts (i.e. in isolated form, in disyllabic words, and in disyllabic words located at the beginning, middle, and final positions of sentences). The results showed that stop syllables alone are shorter than non-stop syllables in various contexts. However, in disyllabic words or in sentences, there is a supplementary lengthening effect immediately after the stop syllables: there is more acoustic blank, and in some circumstances the initial of the following syllable is lengthened. Therefore, we propose that the phonetic realization of syllable isochrony is beyond the syllable itself in Cantonese. The results and discussions of this study may also shed light on the problem of the disappearance of “entering tones” from various Chinese dialects.

Isolation-Induced Ultrasonic Vocalization in Environmental and Genetic Mice Models of Autism.

Studies in rodent models suggest that calls emitted by isolated pups serve as an early behavioral manifestation of communication deficits and autistic like behavior. Previous studies in our labs showed that gestational exposure to the pesticide chlorpyrifos (CPF) and the Mthfr-knock-out mice are associated with impaired social preference and restricted or repetitive behavior. To extend these studies, we examine how pup communication via ultrasonic vocalizations is altered in these ASD models. We implemented an unsupervised hierarchical clustering method based on the spectral properties of the syllables in order to exploit syllable classification to homogeneous categories while avoiding over-categorization. Comparative exploration of the spectral and temporal aspects of syllables emitted by pups in two ASD models point to the following: (1) Most clusters showed a significant effect of the ASD factor on the start and end frequencies and bandwidth and (2) The highest percent change due to the ASD factor was on the bandwidth and duration. In addition, we found sex differences in the spectral and temporal properties of the calls in both control groups as well as an interaction between sex and the gene/environment factor. Considering the basal differences in the characteristics of syllables emitted by pups of the C57Bl/6 and Balb/c strains used as a background in the two models, we suggest that the above spectral-temporal parameters start frequency, bandwidth, and duration are the most sensitive USV features that may represent developmental changes in ASD models.

Comparison of speech perception lists used to evaluate Japanese single syllable hearing

Objective: In Japan, the 57S and 67S single syllable lists have been used to test speech perception in cochlear implant (CI) users. However, due to advances in implants and processors, these lists have become too simple for CI users. In 2019, the Japan Otological Society created a new list, referred to as the iCI2004. The objective of this study was to compare the performance of these lists.Methods: The Japanese single syllable speech perception tests, iCI2004, 57S, and 67S, were administered to 70 patients with CI users. The effects of single syllable characteristics on the test scores were examined. The type and number of single syllables used in each list were different. Therefore, we compared the scores for normal lists, shared single syllables, and non-shared single syllables.Results: The average test results were 52% for iCI2004, 64% for 57S, and 77% for 67S; 67S performed the best, followed by 57S and iCI2004. The test results were significantly different. In a comparison of shared single syllables, the average scores were 58% for iCI2004 and 63% for 57S (45 pieces). A comparison of iCI2004 and 67S (17 pieces) showed that the average scores were 63% for iCI2004 and 75% for 67S. A comparison of 57S and 67S (20 pieces) showed that the average score for 57S was 70% and the average scored for 67S was 77%. No significant differences were detected under all conditions. Based on these results, the type of single syllable adopted affected the result.Conclusion: The data gives no clear indication that the selection of the word table or presentation of sound pressure affects listening to sound for CI users in Japan. Based on the type and number of single syllables used, iCI2004 seems appropriate for evaluation of hearing in patients using CI.

Linguodidactic framework of teaching russian intonation to a multinational audience in light of experimental data

The article deals with the linguodidactic framework of teaching Russian intonation to foreign students, in particular, the intonation of the general question which presents the most difficulty for students of different nationalities. The authors analyze an experiment aimed at identifying the difficulties in intoning the general question for speakers of tone and non-tone languages. During the experiment, the authors identify universal (core) intonational characteristics that manifest regardless of nation-specific (idiomatic) language features. Such features include not only the set of secondary melodic traits but also indicators of intensity and duration as well as formant features of nuclear syllables.

Learning Syllables Using Conv-LSTM Model for Swahili Word Representation and Part-of-speech Tagging

The need to capture intra-word information in natural language processing (NLP) tasks has inspired research in learning various word representations at word, character, or morpheme levels, but little attention has been given to syllables from a syllabic alphabet. Motivated by the success of compositional models in morphological languages, we present a Convolutional-long short term memory (Conv-LSTM) model for constructing Swahili word representation vectors from syllables. The unified architecture addresses the word agglutination and polysemous nature of Swahili by extracting high-level syllable features using a convolutional neural network (CNN) and then composes quality word embeddings with a long short term memory (LSTM). The word embeddings are then validated using a syllable-aware language model ( 31.267 ) and a part-of-speech (POS) tagging task ( 98.78 ), both yielding very competitive results to the state-of-art models in their respective domains. We further validate the language model using Xhosa and Shona, which are syllabic-based languages. The novelty of the study is in its capability to construct quality word embeddings from syllables using a hybrid model that does not use max-over-pool common in CNN and then the exploitation of these embeddings in POS tagging. Therefore, the study plays a crucial role in the processing of agglutinative and syllabic-based languages by contributing quality word embeddings from syllable embeddings, a robust Conv–LSTM model that learns syllables for not only language modeling and POS tagging, but also for other downstream NLP tasks.

Auditory detection is modulated by theta phase of silent lip movements.

Audiovisual speech perception relies, among other things, on our expertise to map a speaker's lip movements with speech sounds. This multimodal matching is facilitated by salient syllable features that align lip movements and acoustic envelope signals in the 4–8 ​Hz theta band. Although non-exclusive, the predominance of theta rhythms in speech processing has been firmly established by studies showing that neural oscillations track the acoustic envelope in the primary auditory cortex. Equivalently, theta oscillations in the visual cortex entrain to lip movements, and the auditory cortex is recruited during silent speech perception. These findings suggest that neuronal theta oscillations may play a functional role in organising information flow across visual and auditory sensory areas. We presented silent speech movies while participants performed a pure tone detection task to test whether entrainment to lip movements directs the auditory system and drives behavioural outcomes. We showed that auditory detection varied depending on the ongoing theta phase conveyed by lip movements in the movies. In a complementary experiment presenting the same movies while recording participants' electro-encephalogram (EEG), we found that silent lip movements entrained neural oscillations in the visual and auditory cortices with the visual phase leading the auditory phase. These results support the idea that the visual cortex entrained by lip movements filtered the sensitivity of the auditory cortex via theta phase synchronization.

The Adjustment of Anterior Forebrain Pathway (AFP) to Birdsong Is Phased during Song Learning and Maintenance.

Anterior forebrain pathway (AFP), a basal ganglia-dorsal forebrain circuit, significantly impacts birdsong, specifically in juvenile or deaf birds. Despite many physiological experiments supporting AFP's role in song production, the mechanism underlying it remains poorly understood. Using a computational model of the anterior forebrain pathway and song premotor pathway, we examined the dynamic process and exact role of AFP during song learning and distorted auditory feedback (DAF). Our simulation suggests that AFP can adjust the premotor pathway structure and syllables based on its delayed input to the robust nucleus of the archistriatum (RA). It is also indicated that the adjustment to the synaptic conductance in the song premotor pathway has two phases: normal phases where the adjustment decreases with an increasing number of trials and abnormal phases where the adjustment remains stable or even increases. These two phases alternate and impel a specific effect on birdsong based on AFP's specific structures, which may be associated with auditory feedback. Furthermore, our model captured some characteristics shown in birdsong experiments, such as similarities in pitch, intensity, and duration to real birds and the highly abnormal features of syllables during DAF.