Mode-conditioned music learning and composition: a spiking neural network inspired by neuroscience and psychology

Menghaoran Tang; Qian Liang; Yi Zeng

REVIEW 3 major objections 6 minor 16 references

Mode-conditioned music learning and composition: a spiking neural network inspired by neuroscience and psychology

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that a brain-inspired spiking neural network, trained on four-part harmony and given Western mode/key theory as prior knowledge, develops pitch-connection statistics matching the Krumhansl-Schmuckler key profiles with…

desk verdict A mode-conditioned spiking network with a genuinely useful new dataset, but the KS-profile similarity claim needs a histogram baseline before it can be believed. read the letter →

arxiv 2411.14773 v2 pith:FG3CGJF2 submitted 2024-11-22 cs.SD cs.AIeess.ASq-bio.NC

Qian Liang , Yi Zeng , Menghaoran Tang This is my paper · ORCID

classification cs.SDcs.AIeess.ASq-bio.NC

keywords spikingneuralnetworkmusicgenerationmodeandkeyperceptionKrumhansl-SchmucklermodelSTDPsymbolicfour-partharmony

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to show that a brain-inspired spiking neural network can internalize Western musical modes and keys, and that the connections it forms between its mode/key neurons and its pitch neurons closely resemble the Krumhansl-Schmuckler (KS) key profiles from music psychology, with cosine similarities between 0.92 and 0.94. If true, the trained network is a mode-conditioned music generator whose internal architecture mirrors human tonal perception, offering a mechanistic account of how key hierarchies might emerge from exposure to music. The authors train the network on four-part harmony exercises and Bach chorales, then measure two connection statistics per pitch class (synaptic count and average synaptic weight) and compare them with the KS profiles. They also show the network can generate four-part pieces in a specified mode and key, with statistics closer to the training corpora than an earlier spiking baseline.

What carries the argument

The carrying mechanism is a two-subsystem spiking architecture: a music theory subsystem (MTS) with a mode cluster and a 24-group key cluster that encodes the prior knowledge of major/minor modes and all keys, and a sequential memory subsystem (SMS) with pitch and duration subnetworks that encode ordered notes. Learning proceeds by a synaptic creation rule (Eq. 6–7) that forms a connection whenever two neurons co-fire at least five times, followed by STDP weight updates (Eq. 8). The key comparison objects are two statistics computed from the trained MTS-to-pitch connections — Pitch Synaptic Count (PSC), the number of synapses per pitch class, and Pitch Average Synaptic Weight (PASW), the mean weight per pitch class — which are normalized and compared against the Krumhansl-Schmuckler key profiles as a proxy for the psychological tonal hierarchy.

What would settle it

Compute the cosine similarity between the raw pitch-class histogram of the SHTE and Bach corpora and the KS profiles. If those corpus-histogram similarities are close to the reported 0.92–0.94, then the PSC/PASW agreement is inherited from corpus statistics, and a control experiment without spiking dynamics — e.g., counting co-occurrences directly from note tokens — would reproduce the result.

Watch

Extended reading notes

Core claim

The central discovery claimed is that a structured spiking network, given Western mode/key theory as prior knowledge and trained on symbolic four-part music via synaptic creation and STDP, develops pitch-class connection statistics whose normalized profiles match the Krumhansl-Schmuckler key profiles. The match is quantified by cosine similarities of 0.93 and 0.92 for the harmony-exercise dataset (major PSC and minor PASW) and 0.94 and 0.94 for the chorale dataset, with about 91.5% of per-key similarities above 0.7. The same architecture, when seeded with a tonic chord and conditioned on a mode and key, generates four-part music whose diatonic pitch rate (0.86) and other statistical features resemble the training sets, and which outperforms the earlier spiking baseline in key adherence and pitch range. The authors interpret the connection-similarity result as evidence that the model's learned representation of tonal importance is consistent with the psychological model of key perception.

Load-bearing premise

The load-bearing premise is that the number and average strength of the synapses between mode/key neurons and pitch neurons reveal the model's notion of which pitches matter, rather than simply reflecting how often each pitch appears in the training music.

Editorial extensions

If this is right

The trained connection statistics provide a quantitative bridge between neural network learning and the Krumhansl-Schmuckler key-finding algorithm, so the same network can be used to test how different training corpora shift tonal hierarchies.
Because generation is conditioned on mode and key through the MTS prior, the model supplies a concrete route for steering symbolic music generation toward a specified tonality without retraining.
The model's internal representation adapts to dataset-specific harmonic conventions (e.g., higher subdominant values on the harmony-exercise dataset), suggesting that deviations from KS profiles may be usable as fingerprints of a corpus's harmonic style.
The four-part generation results position the model as a brain-inspired baseline whose output statistics resemble the training corpora more closely than an earlier spiking system on key adherence.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A reader should treat the KS-similarity number as provisional: because PSC is defined as a binarized co-firing count, it is essentially a pitch-class frequency histogram of the training corpus, and a raw histogram may match the KS profiles just as closely without any spiking dynamics.
A direct control experiment comparing PSC/PASW with the corpus pitch-class histogram and with a non-spiking co-occurrence counter would separate what the spiking machinery contributes from what the music statistics already contain.
The generation quality is only compared against an earlier spiking baseline, not against statistical or deep-learning generators, so the claimed advantage in tonality characteristics and melodic adaptability would be strengthened by testing against a standard n-gram or Transformer baseline.
A behavioral extension would be to ask human listeners to rate the key clarity of generated versus corpus pieces; if listeners cannot distinguish the generated pieces' tonality from the training set, the conditioning claim would be validated perceptually.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a plausible, useful piece of work with one major unaddressed confound. The model is a spiking neural network that learns to generate four-part music conditioned on mode and key. What's actually new: the explicit music-theory subsystem (mode cluster and key cluster with hierarchical connections), the new SHTE dataset (193 textbook harmony exercises), and the fact that nobody else, as far as the cited literature goes, has done mode-conditioned SNN generation. The code is public. That is real.

The paper's central claim is that after training, the synapses between the mode/key clusters and the pitch subnetworks have cosine similarity 0.92–0.94 with the Krumhansl-Schmuckler profiles. The stress-test note is correct: PSC and PASW, as defined in Eq. 14–15, are essentially binarized and weighted counts of co-firing between the always-active mode/key neurons and the pitch neurons. Since those neurons fire whenever a pitch appears, PSC(k) reduces to a thresholded pitch-class histogram of the training corpus. The KS profiles are known to correlate with pitch-class usage. So the high cosine similarity may be nothing more than the model reflecting the corpus statistics. No histogram baseline or bag-of-pitches control is reported. This is not a fatal flaw, but it is a load-bearing gap: without such a control, the neuroscience/psychology interpretation is over-stated.

The generation evaluation has a similar issue. The aggregated KLD/OA comparisons to the training sets show the output resembles the data, which is nice, but there is no per-key compliance test, no statistical significance, and no listening test. The comparison to the SCSNN baseline is helpful, but again aggregate.

What the paper does well: it is transparent about missing keys (the zero values in Fig. 6), it gives the full architecture and parameters, and it is candid about dataset-specific deviations from KS. The SHTE dataset alone is worth having.

In short: the idea is good, the dataset is reusable, and the architecture is clearly described. The central claim needs a control baseline before publication. With that added, the paper would be a solid contribution to computational music cognition. I would send this to serious review rather than desk-reject it, with the clear expectation of a major revision.

Referee Report

3 major / 6 minor

Summary. The paper proposes a multi-region spiking neural network, implemented on BrainCog, for mode- and key-conditioned four-part music generation. A music theory subsystem encodes major/minor modes and 24 keys through hand-wired scale-degree connections, while a sequential memory subsystem encodes pitch and duration. Learning relies on activity-dependent synaptic creation and STDP. The authors define Pitch Synaptic Count (PSC) and Pitch Average Synaptic Weight (PASW) features of the trained connections, compare them against Krumhansl-Schmuckler (KS) key profiles, and evaluate generated pieces against SHTE and Bach datasets and an SCSNN baseline.

Significance. If the KS-similarity claim were established, the model would be an interpretable, brain-inspired generative architecture whose internal connectivity mirrors psychological tonal hierarchies. The paper contributes a new SHTE dataset, releases code on GitHub, and benchmarks against an existing SNN baseline with quantitative KLD/OA metrics. However, the central similarity result is currently not distinguished from trivial pitch-class corpus statistics, and the generation evaluation aggregates over keys without per-key compliance tests or significance statements. The conceptual claims therefore outrun the evidence provided.

major comments (3)

[Sec. 5.1, Eqs. (14)-(15) together with Eqs. (6)-(7)] PSC and PASW are defined as counts and average weights of synapses created when mode/key-cluster neurons and pitch neurons co-fire at least five times. Because the mode/key clusters are driven by the input note pitches themselves (Eqs. (1)-(2)), a synapse between mode neuron k and pitch-class-k neurons is created essentially whenever pitch class k occurs in a piece labeled with that mode/key. PSC(k) is therefore a thresholded pitch-class co-occurrence count of the training corpus, and PASW(k) is an STDP-weighted version of the same statistic. The reported cosine similarities of 0.92-0.94 to KS profiles may simply reflect the pitch-class distributions of Western tonal music in SHTE and Bach, rather than any emergent property of the spiking dynamics, synaptic-creation rule, or conditioning architecture. The paper does not report a baseline control, for example the cosine similarity between raw (or thresholded) pitch-class histograms of the training corpora per mode and the corresponding KS profiles. Such a control is necessary to support the paper's central claim that the trained connection framework is closely similar to the Krumhansl-Schmuckler model.
[Sec. 5.2, Table 3 and Sec. 5.2.2] The generation evaluation aggregates 50 generated samples across 'different keys' and reports mean KLD/OA values against SHTE and Bach. These aggregate statistics do not test the conditioning claim: they do not show that a piece generated in a specified key has pitch-class statistics closer to that key than to other keys, and no per-key diatonic pitch rate or per-key KLD is reported. Without per-key compliance tests and associated significance or confidence statements, the claim that the model generates music with the characteristics of the given modes and keys is not established.
[Sec. 5.1, Fig. 6 and Sec. 5.2] Figure 6 reports zero cosine similarities for C# major, F# major, C# minor, Ab minor, A# minor, and D# minor because no training pieces exist in those keys, as confirmed by Table 2. The paper nonetheless claims that the model can generate music in various modes and keys. Section 5.2 does not state which keys were used for the 50 generated samples. If the generated set excludes the unseen keys, the generalization claim is unsupported; if it includes them, per-key results must be reported. Please clarify and, if appropriate, evaluate unseen keys separately.

minor comments (6)

[Eq. (14) caption] The caption contains a typo: 'pitch lass' should be 'pitch class'.
[Sec. 3.4] 'Winner-Takes-All priciple' should be 'Winner-Takes-All principle'.
[Sec. 5.1] The word 'Phenorminans' appears in the bullet list; it should likely be 'phenomenon' or similar.
[Sec. 5.1] The acronym is inconsistently written as both 'PSWA' and 'PASW' in the text; please unify.
[Sec. 3.3.1] The phrase 'oscillatory times' in the description of Eq. (6) is confusing; 'coincidence counts' or 'co-firing counts' would be clearer.
[Sec. 5.1, Fig. 6] The statement 'Approximately 91.5% of the values are above 0.7' should be recomputed and justified, since the figure includes zero values for absent keys.

Circularity Check

1 steps flagged · score 6.0 of 10

The KS-similarity evidence (Sec. 5.1) reduces by construction to a binarized pitch-class histogram of the training corpus, so the claimed cognitive/architectural match is a renaming of known corpus statistics rather than a test of the spiking model.

renaming known result [Sec. 3.3.1 (Eqs. 6-7), Sec. 5.1 (Eqs. 14-15 and cosine-similarity results)]
"Pitch Synaptic Count(PSC) refers to the total number for each neuron representing different pitch class in the mode clusters and the corresponding neurons in the pitch subnetwork... 𝑃𝑆𝐶𝑘=∑4𝑗=1∑128𝑟%12=𝑘∑𝑁𝑠𝑐𝑘,𝑗𝑟𝑠 (14), where 𝑐𝑖𝑗 = {1,𝑜 ≥5 0,𝑒𝑙𝑠𝑒 (7). After training, we normalize to PSC, PASW and KS pitch profile scores."

Under Eqs. 1-3, the mode/key-cluster neuron for pitch class k fires whenever a note of that pitch class occurs in a piece labeled with that mode/key, and the corresponding pitch-subnetwork neuron fires when that pitch occurs. The co-firing count o in Eq. 6 is therefore the number of times pitch class k appears in the training corpus (thresholded at 5), and Eq. 7 converts this count into a binary synapse. Eq. 14 sums those synapses into PSC(k); PASW is the same co-occurrence statistic weighted by STDP (Eqs. 8 and 15). Consequently, the cosine similarities of 0.92-0.94 reported in Sec. 5.1 compare the corpus's pitch-class histogram to the KS profiles, not an emergent property of the spiking dynamics or the mode-conditioning architecture.

full rationale

The paper's central evaluation claim is that the trained network's connection architecture resembles the Krumhansl-Schmuckler key-perception model. That claim is supported by PSC and PASW, which are defined from Eq. 7's co-firing rule. Because mode/key neurons and pitch neurons fire exactly in response to the pitch content of the training pieces, PSC/PASW are deterministic functions of the corpus's pitch-class co-occurrence statistics. Comparing those statistics to KS profiles is an external comparison, so it is not circular in the strict sense of fitting the benchmark, but it is circular in the weaker sense that the measured quantity is constructed from the same corpus statistics that are already known to correlate with KS profiles. The absence of a simple histogram baseline means the spiking network's specific mechanisms are not responsible for the reported similarity. The remaining contributions, such as the mode-conditioned generation experiments and the comparison to the SCSNN baseline, are independent of this issue. One partial circularity is therefore present, warranting a score of 6 rather than a higher score.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

No new physical or natural entities are introduced. The mode cluster, key cluster, and memory subnetworks are computational modules, not unobserved entities. The central claim depends on free parameters such as the input scale factors, Izhikevich constants, STDP constants, and the connection-creation threshold, most of which are chosen by hand and some of which are unreported.

free parameters (6)

Input current scale factors alpha_Mr, alpha_Ks, alpha_P, alpha_D = 50, 50, 30, 30
Eq.1-4 set these values by hand. They control spike rates and therefore which co-firing pairs cross the connection-creation threshold, influencing both PSC and generated note choices.
Izhikevich neuron parameters a, b, c, d, V_th = 0.1, 0.2, -65, 30, 30
Eq.5; chosen by hand to produce a regular spiking pattern. These values affect spike timing and hence STDP updates and connection creation.
Synaptic creation threshold o >= 5 = 5
Eq.7; chosen by hand. This threshold directly determines which connections are counted in PSC and therefore the main comparison to the KS model. No sensitivity analysis is provided.
STDP amplitudes and time constants A+, A-, tau+, tau-
Eq.8 uses these STDP parameters, but the paper never reports their values. They are required to reproduce the learned weights.
Initial synaptic weight random distribution
Section 3.3.1 states that synaptic weights are initially random, but the distribution and random seed are not specified, so exact reproduction across runs is not possible.
Winner-takes-all integration window T
Eq.13 sums spikes over T time steps to choose the generated note, but T is not specified.

assumptions (5)

domain assumption The Izhikevich neuron model and STDP learning capture the relevant music-cognition mechanisms.
The paper motivates the model through brain regions and plasticity but does not validate that these simplified neural dynamics are sufficient or necessary for mode and key perception.
domain assumption The Krumhansl-Schmuckler key profiles are the correct external benchmark for human key perception.
The central comparison in Section 5.1 treats KS profiles as ground truth. This is a reasonable, standard music-psychology benchmark, but it is only one model of key perception.
ad hoc to paper Co-firing of a mode/key neuron and a pitch neuron, counted in Eq.6-7, creates a synapse whose number and weight reflect tonal importance.
This assumption is load-bearing for PSC/PASW and the KS comparison, but it is not derived from neuroscience or music psychology. It equates synaptic count with perceptual salience.
domain assumption Western 12-TET major and minor mode theory is the appropriate representational framework.
The music theory subsystem hard-codes the 12-pitch-class Western major and minor modes, which limits scope to this tradition but is appropriate for the chosen datasets.
domain assumption The o >= 5 coincidence rule approximates growth-cone-guided synaptic formation.
The paper explicitly simplifies the neuroscience of synaptic formation to a spike-coincidence rule. This is a modeling choice, not an established equivalence.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Mode-conditioned music learning and composition: a spiking neural network inspired by neuroscience and psychology." pith.science (2026). https://pith.science/paper/FG3CGJF2

@misc{pith2026241114773,
  author       = {Pith},
  title        = {Pith review of: Mode-conditioned music learning and composition: a spiking neural network inspired by neuroscience and psychology},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FG3CGJF2}},
  note         = {Machine review of arXiv:2411.14773}
}

read the original abstract

Musical mode is one of the most critical element that establishes the framework of pitch organization and determines the harmonic relationships. Previous works often use the simplistic and rigid alignment method, and overlook the diversity of modes. However, in contrast to AI models, humans possess cognitive mechanisms for perceiving the various modes and keys. In this paper, we propose a spiking neural network inspired by brain mechanisms and psychological theories to represent musical modes and keys, ultimately generating musical pieces that incorporate tonality features. Specifically, the contributions are detailed as follows: 1) The model is designed with multiple collaborated subsystems inspired by the structures and functions of corresponding brain regions; 2)We incorporate mechanisms for neural circuit evolutionary learning that enable the network to learn and generate mode-related features in music, reflecting the cognitive processes involved in human music perception. 3)The results demonstrate that the proposed model shows a connection framework closely similar to the Krumhansl-Schmuckler model, which is one of the most significant key perception models in the music psychology domain. 4) Experiments show that the model can generate music pieces with characteristics of the given modes and keys. Additionally, the quantitative assessments of generated pieces reveals that the generating music pieces have both tonality characteristics and the melodic adaptability needed to generate diverse and musical content. By combining insights from neuroscience, psychology, and music theory with advanced neural network architectures, our research aims to create a system that not only learns and generates music but also bridges the gap between human cognition and artificial intelligence.

Figures

Figures reproduced from arXiv: 2411.14773 by the authors.

**Figure 1.** The architecture of the brain-inspired mode learning spiking neural network. The model contains the music theory subsystem(MTS) and the sequential memory subsystem(SMS). The MTS contains the mode cluster and key clusters, which is responsible for encoding the related music knowledge. The SMS receives the symbolic representation of the pitches and the durations, encoding and memorizing the relationships of the ordere… view at source ↗

**Figure 2.** The musical theory subsystem, (A) describes how a neuron group in first layer in the mode cluster represents the major mode. Seven neurons drawn by orange parallelograms encodes the diatonic tones and the gray ones encodes the chromatic tones; (B) employs the key of G major to illustrate the principle, neurons drawn by green parallelograms encodes diatonic tones, G, A, B, C, D, E, #F, the gray parallelograms also en… view at source ↗

**Figure 3.** The learning process of music guided by the mode theory. where, the 𝑡 𝑓 𝑖 and 𝑡 𝑛 𝑗 represent the spike time 𝑓 and 𝑛 of postsynaptic neuron 𝑖 and presynaptic neuron 𝑗, respectively. When the oscillatory times 𝑜 ≥ 5 of these neuron 𝑖 and 𝑗, a new synaptic connection is formed, and is represented as 𝑐𝑖𝑗. This simple rule is defined to describe that neurons oscillate together are more likely to form a synaptic connecti… view at source ↗

Figures from the paper (6 more)

**Figure 4.** Figure 4: The generation process of a musical piece in G minor begins with a tonic chord as the seed. The model receives this seed input and activates the neurons representing pitches that distribute in four parts, guided by the active neurons that represent G minor, step by ste…

**Figure 5.** Figure 5: The comparative analysis between the connection architecture of our model and the Krumhansl-Schmuckler model is presented as follows: Panel (A) shows the average synaptic weight and synaptic count of neurons within the major cluster and pitch subnetwork trained on the …

**Figure 6.** Figure 6: The results of cosine similarities for 24 keys between the two features(PSC and PASW) and the KS model under the SHTE and Bach Datasets. or descending semitone steps, creating smoother melodic transitions. The use of semitones introduces chromatic tones, which break up…

**Figure 7.** Figure 7: A generated sample in A Major, with red circles indicating semitone step ascending or descending for Smoother Melodic Transitions. 5.2.1. Metrics To further assess and analyze the quality of the generated music pieces, we employ a comprehensive objective method which i…

**Figure 8.** Figure 8: The comparison of the mean values between our model and the SCSNN baseline for the PC, DPR, PR, and PI metrics. The figure shows that, our model demonstrates a significantly higher mean PC, indicating that it generates music with a broader variety of pitches compared …

**Figure 9.** Figure 9: The performance comparison of KLDs and OAs for the features between our model and the baseline. capability is essential for our future work on harmonic learning, emotion perception, and stylistic composition. 6. Conclusion In this paper, we present a brain-inspired spi…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

16 extracted references · 9 canonical work pages

[1]

Synapticmodificationsinculturedhippocampal neurons:dependenceonspiketiming,synapticstrength,andpostsynap- tic cell type

Bi,G.q.,Poo,M.m.,1998. Synapticmodificationsinculturedhippocampal neurons:dependenceonspiketiming,synapticstrength,andpostsynap- tic cell type. Journal of neuroscience 18, 10464–10472. doi:10.1523/ JNEUROSCI.18-24-10464.1998. Bi, Y.,

arXiv 1998
[10]

FrontiersinComputationalNeuroscience14,51

Temporal-sequential learning with a brain-inspired spiking neural network and its application to musical memory. FrontiersinComputationalNeuroscience14,51. doi: 10.3389/ fncom.2020.00051. MacDonald, C.J., Lepage, K.Q., Eden, U.T., Eichenbaum, H.,

arXiv 2020
[14]

Patterns 4, 100789

Braincog: A spiking neural network based, brain-inspired cognitive intelligence engine for brain-inspired ai and brain simulation. Patterns 4, 100789. doi:10.1016/j.patter.2023. 100789. Zhao, Z.,

doi:10.1016/j.patter.2023 2023
[15]

ArXiv abs/2407.08306

Adversarial-midibert: Symbolic music understanding model based on unbias pre-training and mask fine-tuning. ArXiv abs/2407.08306. URL: https://api.semanticscholar.org/CorpusID: 271097781. Zheng, J.Q., ming Poo, M.,

arXiv
[1569]

Jenkins, L.J., Ranganath, C.,

doi:10.1109/TNN.2003.820440. Jenkins, L.J., Ranganath, C.,

arXiv 2003
[2007]

Annual Review of Cell and Developmental Biology 23, 375–404

Calcium signaling in neuronal motil- ity. Annual Review of Cell and Developmental Biology 23, 375–404. doi:10.1146/annurev.cellbio.23.090506.123221. Q.Liang, Y.Zeng and MHR.Tang:Preprint submitted to Elsevier Page 15 of 15

work page arXiv
[2008]

Current opinion in neurobiology 18, 452–463

Music perception, pitch, and the auditory system. Current opinion in neurobiology 18, 452–463. Meier, B., Weiermann, B., Gutbrod, K., Stephan, M.A., Cock, J., Müri, R.M.,Kaelin-Lang,A.,2013. Implicittasksequencelearninginpatients with parkinson’s disease, frontal lesions and amnesia: The critical role offronto–striatalloops. Neuropsychologia51,3014–3024. ...

work page 2013
[2010]

Journal of Neuroscience 30, 15558–15565

Prefrontal and medial temporal lobe activity at encoding predicts temporal context memory. Journal of Neuroscience 30, 15558–15565. doi:10.1523/JNEUROSCI.1337-10.2010. Ji,S.,Yang,X.,Luo,J.,2023.Asurveyondeeplearningforsymbolicmusic generation: Representations, algorithms, evaluations, and challenges. ACM Computing Surveys 56, 1–39. Koelsch, S.,

work page doi:10.1523/jneurosci.1337-10.2010 2010

Show all 16 references

[2012]

arXiv preprint arXiv:1206.6392

Model- ing temporal dependencies in high-dimensional sequences: Application to polyphonic music generation and transcription. arXiv preprint arXiv:1206.6392 . Briot, J.P., Hadjeres, G., Pachet, F.D.,

arXiv
[2017]

Deepbach: a steerable model for bach chorales generation, in: International conference on machine learning, PMLR. pp. 1362–1371. Hernandez-Olivan,C.,Hernandez-Olivan,J.,Beltrán,J.R.,2022. Asurvey on artificial intelligence for music generation: Agents, domains and perspectives...

2022 doi arXiv
[2018]

Eck,D.,Schmidhuber,J.,2002.Afirstlookatmusiccompositionusinglstm recurrentneuralnetworks

Musegan: Multi- tracksequentialgenerativeadversarialnetworksforsymbolicmusicgen- erationandaccompaniment,in:ProceedingsoftheAAAIConferenceon Artificial Intelligence. Eck,D.,Schmidhuber,J.,2002.Afirstlookatmusiccompositionusinglstm recurrentneuralnetworks. IstitutoDalleMolleDiS...

2002
[2020]

1180–1188

Pop music transformer: Beat-based modeling and generation of expressive pop piano compositions, in: Proceedings of the 28th ACM international conference on multimedia, pp. 1180–1188. Izhikevich,E.,2003. Simplemodelofspikingneurons. IEEETransactions on Neural Networks 14,

2003
[2021]

doi:10.3389/fnsys.2021.639484

Stylistic composition of melodies based on a brain-inspiredspikingneuralnetwork.Frontiersinsystemsneuroscience 15, 639484. doi:10.3389/fnsys.2021.639484. Liang, Q., Zeng, Y., Xu, B.,

2021
[2022]

Nature Reviews Neuroscience 23, 287–305

Music in the brain. Nature Reviews Neuroscience 23, 287–305. Q.Liang, Y.Zeng and MHR.Tang:Preprint submitted to Elsevier Page 14 of 15 SNN-based mode-conditioned music learning and composition Waite,E.,2016. Generatinglong-termstructureinsongsandstories. https: //magenta.tenso...

2016
[2023]

arXiv.org doi:10.48550/arxiv.2307.11078

Brain2music: Reconstructing music from human brain activity. arXiv.org doi:10.48550/arxiv.2307.11078. Dong, H.W., Hsiao, W.Y., Yang, L.C., Yang, Y.H.,

doi
[2024]

arXiv preprint arXiv:2410.02084

Generatingsymbolicmusicfromnaturallanguagepromptsusinganllm- enhanced dataset. arXiv preprint arXiv:2410.02084 . Yang, L.C., Chou, S.Y., Yang, Y.H.,

arXiv

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Synapticmodificationsinculturedhippocampal neurons:dependenceonspiketiming,synapticstrength,andpostsynap- tic cell type

Bi,G.q.,Poo,M.m.,1998. Synapticmodificationsinculturedhippocampal neurons:dependenceonspiketiming,synapticstrength,andpostsynap- tic cell type. Journal of neuroscience 18, 10464–10472. doi:10.1523/ JNEUROSCI.18-24-10464.1998. Bi, Y.,

arXiv 1998

[2] [10]

FrontiersinComputationalNeuroscience14,51

Temporal-sequential learning with a brain-inspired spiking neural network and its application to musical memory. FrontiersinComputationalNeuroscience14,51. doi: 10.3389/ fncom.2020.00051. MacDonald, C.J., Lepage, K.Q., Eden, U.T., Eichenbaum, H.,

arXiv 2020

[3] [14]

Patterns 4, 100789

Braincog: A spiking neural network based, brain-inspired cognitive intelligence engine for brain-inspired ai and brain simulation. Patterns 4, 100789. doi:10.1016/j.patter.2023. 100789. Zhao, Z.,

doi:10.1016/j.patter.2023 2023

[4] [15]

ArXiv abs/2407.08306

Adversarial-midibert: Symbolic music understanding model based on unbias pre-training and mask fine-tuning. ArXiv abs/2407.08306. URL: https://api.semanticscholar.org/CorpusID: 271097781. Zheng, J.Q., ming Poo, M.,

arXiv

[5] [1569]

Jenkins, L.J., Ranganath, C.,

doi:10.1109/TNN.2003.820440. Jenkins, L.J., Ranganath, C.,

arXiv 2003

[6] [2007]

Annual Review of Cell and Developmental Biology 23, 375–404

Calcium signaling in neuronal motil- ity. Annual Review of Cell and Developmental Biology 23, 375–404. doi:10.1146/annurev.cellbio.23.090506.123221. Q.Liang, Y.Zeng and MHR.Tang:Preprint submitted to Elsevier Page 15 of 15

work page arXiv

[7] [2008]

Current opinion in neurobiology 18, 452–463

Music perception, pitch, and the auditory system. Current opinion in neurobiology 18, 452–463. Meier, B., Weiermann, B., Gutbrod, K., Stephan, M.A., Cock, J., Müri, R.M.,Kaelin-Lang,A.,2013. Implicittasksequencelearninginpatients with parkinson’s disease, frontal lesions and amnesia: The critical role offronto–striatalloops. Neuropsychologia51,3014–3024. ...

work page 2013

[8] [2010]

Journal of Neuroscience 30, 15558–15565

Prefrontal and medial temporal lobe activity at encoding predicts temporal context memory. Journal of Neuroscience 30, 15558–15565. doi:10.1523/JNEUROSCI.1337-10.2010. Ji,S.,Yang,X.,Luo,J.,2023.Asurveyondeeplearningforsymbolicmusic generation: Representations, algorithms, evaluations, and challenges. ACM Computing Surveys 56, 1–39. Koelsch, S.,

work page doi:10.1523/jneurosci.1337-10.2010 2010