Leveraging AM and FM Rhythm Spectrograms for Dementia Classification and Assessment

Jagabandhu Mishra; Md Sahidullah; Parismita Gogoi; Priyankoo Sarmah; Seema Khadirnaikar; Sishir Kalita; Soma Siddhartha; S. R. M. Prasanna; Vishwanath Pratap Singh

REVIEW 4 major objections 4 minor 37 references

Leveraging AM and FM Rhythm Spectrograms for Dementia Classification and Assessment

T0 review · 4 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read AM rhythm spectrograms outperform Mel spectrograms for dementia classification.

desk verdict Interesting domain application with a clear pipeline and released code, but the central claim is weakened by a decimal-point error in the reported improvements and by baselines that are cited rather than reproduced. read the letter →

arxiv 2506.00861 v2 pith:I7JVTRSU submitted 2025-06-01 eess.AS cs.SD

Parismita Gogoi , Vishwanath Pratap Singh , Seema Khadirnaikar , Soma Siddhartha , Sishir Kalita , Jagabandhu Mishra , Md Sahidullah , Priyankoo Sarmah

show 1 more author

S. R. M. Prasanna

This is my paper

classification eess.AScs.SD

keywords dementiadetectionrhythmformantanalysisspectrogramsamplitudemodulationfrequencyvisiontransformerBERTMMSEestimation

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 establish that rhythm spectrograms derived from Rhythm Formant Analysis (RFA) capture dementia-related changes in the low-frequency amplitude and frequency modulations of spontaneous speech, and that these annotation-free features beat standard acoustic baselines for classifying dementia and estimating MMSE. On the ADReSSo challenge data, handcrafted rhythm-spectrogram features reach 65.71% accuracy, a relative 14.2% improvement over eGeMAPS, and AM rhythm spectrograms in a vision-transformer/BERT fusion reach 74.29% accuracy, about 13.1% relative improvement over Mel spectrograms. If true, RFA offers a cheap, fully automatic speech biomarker for cognitive screening that does not require manual transcripts or forced alignment.

What carries the argument

The central object is the rhythm spectrogram, a time-frequency representation built from the low-frequency (<10 Hz) spectrum of the amplitude-modulation (AM) envelope and the F0-based frequency-modulation (FM) envelope of speech. Speech is split into a fixed grid of 100 overlapping 5-second segments, each segment's low-frequency spectrum is computed, and the spectra are stacked in time to form the spectrogram; rhythm formants are spectral peaks tracked across time, and their variance plus low-order 2D-DCT coefficients form the handcrafted features. In the data-driven path, the spectrograms with their delta and double-delta versions are fed as three channels to a pretrained vision transformer (ViT), whose output is concatenated with BERT embeddings of Wav2Vec2.0 ASR transcripts and passed through a single trainable fully connected layer. The machinery works by converting slow prosodic and articulatory variation, previously measured through manual syllable or word annotation, into an image-like representation that can be processed automatically.

What would settle it

Take the ADReSSo test set, measure total utterance duration, and retrain the handcrafted SVM with duration and pause proportion added as control features, or match the healthy and dementia groups on duration; if the 14.2% accuracy advantage over eGeMAPS disappears or falls below significance, the central claim is falsified.

Watch

Extended reading notes

Core claim

On its own terms, the discovery is that the long-term temporal modulations of the speech signal below 10 Hz, captured as AM and FM rhythm spectrograms, carry discriminating information about dementia. The authors show this information can be exploited two ways: as handcrafted variance and 2D-DCT features with an SVM, and as three-channel image inputs to a vision transformer whose embeddings are fused with BERT-based linguistic embeddings from ASR transcripts. In both schemes the rhythm-spectrogram features outperform the corresponding baselines (eGeMAPS for the handcrafted path, Mel spectrograms for the fusion path), while giving comparable MMSE regression. The authors claim this demonstrates RFA as a useful annotation-free alternative for dementia assessment.

Load-bearing premise

The load-bearing premise is that squeezing every utterance into the same fixed 100 five-second segments leaves the rhythm content comparable across speakers; if recording length or pause distribution differs systematically between people with and without dementia, the reported gains could come from duration rather than from rhythm.

Editorial extensions

If this is right

Handcrafted rhythm-spectrogram features could make dementia screening deployable on lightweight devices, since they require only an SVM rather than a large neural network.
The AM spectrogram's superiority over Mel suggests slow amplitude-modulation dynamics are a stronger acoustic marker for dementia than static spectral shape.
Using Wav2Vec2.0 ASR transcripts instead of manual transcriptions removes the need for expert transcription and makes the pipeline usable at scale.
Comparable MMSE regression with rhythm features offers a non-invasive route to estimating cognitive severity from speech.
Because RFA needs no syllable or word segmentation, the same pipeline can be applied to any spontaneous speech recording without annotation.

Reading between the lines

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

Editorial inference: because every utterance is squeezed into a fixed 100-segment grid, short and long recordings are rescaled differently; a duration-matched control would reveal whether part of the accuracy gain reflects recording length rather than genuine rhythm.
Editorial inference: the 13.1% AM-over-Mel margin is specific to this dataset and fusion configuration, so on other dementia corpora or with different pretrained transformers the ranking could shift.
Editorial inference: a natural next test is whether rhythm spectrograms generalize to mild cognitive impairment or frontotemporal dementia, where prosodic breakdown differs from Alzheimer's dementia.
Editorial inference: the two handcrafted feature families capture different aspects (variance yields better F1, DCT better accuracy), so a learned weighting of both could push performance further.

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

The new thing here is the transfer of RFA rhythm spectrograms to dementia speech. That domain shift is real, and the paper ships code, which is more than many in this area do. The handcrafted feature extraction is coherent, the ViT-BERT adaptation is reasonable, and the regression results are presented honestly enough. I read the stress-test note and largely agree with it.

The soft spots are the load-bearing ones. The abstract says the handcrafted features beat eGeMAPS by a relative 14.2%, but Table 3 says 65.71 vs 64.79, which is 1.42%. Same for the AM spectrogram versus Mel: 13.1% claimed, 1.31% actual. That is a factor-of-ten inflation, and it needs to be fixed outright. More importantly, the eGeMAPS number is taken from the ADReSSo challenge paper, and the Mel baseline is taken from the ViT-BERT paper; neither baseline was re-run under this paper's preprocessing, which includes concatenating only the subject's speech segments and then forcing a fixed 100-segment representation. That makes the comparison uncontrolled. The accuracy differences are also tiny: on a 71-utterance test set, one or two utterances separate the systems, and no significance test or confidence interval is reported. I would not call this a fatal flaw because the method could still be useful, but the central claim as stated is not established.

The fixed 100-segment resampling is a genuine worry too. Dementia speech has variable durations and long pauses. If the resampling rescues duration information or the pause pattern differently for short and long recordings, the classifier could be picking up recording length rather than rhythm formants. The paper does not address this.

Who is this for? Speech pathologists and dementia researchers who want a quick, annotation-free acoustic feature. It is not a breakthrough, and the modest gains might evaporate under a fair baseline, but the idea is worth testing. A serious referee should see this manuscript, but the authors must correct the arithmetic, re-run the baselines under identical preprocessing, and add at least some uncertainty estimate. Without those, the paper should not be accepted as is.

Referee Report

4 major / 4 minor

Summary. The paper introduces Rhythm Formant Analysis (RFA) based AM and FM rhythm spectrograms for dementia classification and MMSE regression on the ADReSSo dataset. Two families of features are proposed: handcrafted features (variance of rhythm formant trajectories and 2D-DCT coefficients) fed to SVM/SVR/DT, and data-driven features obtained by replacing Mel spectrograms with rhythm spectrograms in a ViT-BERT fusion system with ASR transcripts. The main claims are that the handcrafted features outperform eGeMAPS by a relative 14.2% in classification accuracy and that AM rhythm spectrograms outperform Mel spectrograms by about 13.1%, with comparable or slightly better regression performance.

Significance. If established, the proposal offers an annotation-free acoustic feature family for dementia assessment, which is practically valuable because RFA avoids manual syllable or word annotation. The paper is clearly structured, releases code, and evaluates on a standard challenge dataset (ADReSSo). The ViT-BERT variant with ASR transcripts also addresses a realistic deployment constraint. However, the headline superiority claims rest on a baseline comparison that is not demonstrated to be controlled, on a one-utterance accuracy gap, and on a factor-10 arithmetic error in the reported relative improvements. With a fair, statistically grounded comparison, the contribution could be a useful addition to speech-based dementia detection, but the current evidence is insufficient to support the abstract's claims.

major comments (4)

[Abstract and Section 4.3.1 (Table 3)] The relative improvements reported in the abstract and in Section 4.3.1 are inflated by a factor of 10. From Table 3, (65.71 - 64.79) / 64.79 = 1.42%, not 14.2%, for handcrafted features versus eGeMAPS, and (74.29 - 73.33) / 73.33 = 1.31%, not 13.1%, for AM spectrogram versus Mel spectrogram. The sentence in Section 4.3.1 stating 'a relative 13.09% improvement' is similarly inconsistent with the table. These are the paper's central quantitative claims and should be corrected, and any downstream conclusions should be adjusted accordingly.
[Section 4.1 and Table 3] The eGeMAPS comparison is not controlled. Table 3 lists 'eGeMAPS [22]' as the baseline, but the paper does not report re-computing eGeMAPS features on the same preprocessed utterances used for the proposed features. Section 4.1 describes isolating only the subject's speech segments via timestamps and concatenating them, and Section 2 imposes a fixed 100-segment representation; the ADReSSo challenge results in [22] may have used different preprocessing, features, and model selection. Without re-running eGeMAPS under identical preprocessing and the same SVM cross-validation scheme, the claimed superiority over eGeMAPS is not established. The same concern applies if the Mel-spectrogram ViT-BERT baseline values in Table 2 are quoted from [11] rather than produced by the authors' pipeline.
[Section 4.3.1 and Table 3] No significance testing or uncertainty quantification is reported for the handcrafted classification results. On the 71-utterance test set, 65.71% and 64.79% correspond to 47 and 46 correct utterances, respectively, a difference of a single subject. Moreover, the hyperparameter C=3 was selected on the development folds, and only the resulting test accuracy is reported. The authors should report confidence intervals, bootstrap or McNemar significance tests, and ideally the development-fold performance across C values, before claiming that rhythm-spectrogram features outperform eGeMAPS.
[Section 2] The fixed 100-segment extraction from a 5 s window is a load-bearing preprocessing choice that is not sufficiently justified or controlled. The text says the method 'extract[s] a fixed set of 100 segments from the AM envelope, ensuring consistent temporal segmentation regardless of variations in utterance duration,' which is ambiguous: if short and long utterances are rescaled to the same segment count, the features may encode duration or pause information rather than intrinsic rhythm; if instead a fixed step size is used, shorter utterances will not yield 100 segments. Since dementia speech typically has longer pauses and variable durations, the authors should clarify the resampling procedure, report duration statistics for HC and dementia groups, and show that the classification advantage is not explained by utterance duration or pause fraction.

minor comments (4)

[Introduction] There is a typo in 'Mel-frequency cepstral co-coefficients' in the first paragraph of the Introduction; it should read 'coefficients.'
[Table 2] The F1-score standard deviation for the Mel baseline is reported as 0.006, while the accuracy standard deviation is 0.67; this disparity is surprising and should be checked for a typo or clarified in the text.
[Section 3.2 and Figure 3] The naming of the spectrogram channels is inconsistent: the text says 'AM and FM spectrograms, along with their ∆ and ∆∆ spectrograms, are each provided as three input channels to ViT,' but Table 2 lists C1, C2, C3 as the channel labels. Please unify the notation.
[Section 4.3.2] The regression comparison with eGeMAPS is reported only in prose ('eGeMAPS features achieve a lower RMSE of 6.09, as reported in [22]'); for completeness, this should be included in Table 4 with the same caveat about not being re-run on the authors' preprocessing.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: rhythm-spectrogram features are computed directly from the audio and compared with external baselines; the reported relative improvements are arithmetically inflated but not definitionally forced.

full rationale

The paper's derivation chain is self-contained with respect to the target labels. The AM/FM rhythm spectrograms are computed from the speech signal via Hilbert envelope and RAPT F0 extraction (Section 2), and the handcrafted features (variance of rhythm-formant trajectories and 2D-DCT coefficients) are functions of the audio alone; labels only enter at SVM/SVR training and hyperparameter selection on the development folds, which is ordinary supervised fitting rather than circularity. The data-driven ViT-BERT branch concatenates frozen pre-trained acoustic and linguistic embeddings with a two-neuron trainable layer, again without label-derived feature construction. The comparison baselines are external: eGeMAPS is cited from the ADReSSo challenge [22] and the Mel-spectrogram ViT-BERT result is either reproduced or cited from [11]; neither baseline is constructed from the proposed features. The self-citations [18,19] are background references for prior RFA applications and are not load-bearing for the central claim, and the core RFA method is attributed to Gibbon [15], an external source. The main quantitative concerns are that the abstract's 'relative improvement of 14.2%' and '13.1%' are inconsistent with Table 3, where the actual relative gains are about 1.42% and 1.31%, and that the eGeMAPS baseline is not re-run under the paper's subject-segment concatenation and fixed 100-segment resampling; however, these are arithmetic and experimental-control issues, not instances of a prediction reducing by construction to its inputs. No circular step is present.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The paper does not introduce new physical entities. The main load-bearing choices are inherited from prior RFA work (window length, segment count, low-frequency band), plus label-informed hyperparameter selection (N, C). The reliability of pitch tracking and pre-trained transformers on pathological speech is assumed.

free parameters (4)

Number of rhythm formants N = 6
Set to 6 in Section 3.1 without ablation or data-driven justification; the variance feature dimension depends on it.
DCT coefficient count C = 3 (best of 2, 3, 4)
Selected as the value that gives best classification accuracy on 5-fold development splits; this is a label-informed choice.
Analysis window length and segment count = 5 s window, 100 segments
Fixed in Section 2 to ensure consistent temporal segmentation regardless of utterance duration; not justified for pathological speech where pauses alter duration distributions.
Low-frequency band limit = 0-10 Hz
Taken from RFA [15] and applied to dementia speech; the paper does not test alternative band limits.

assumptions (4)

domain assumption Low-frequency AM and FM modulations below 10 Hz carry rhythm formant information that reflects articulatory and prosodic timing.
Invoked in Section 2 and inherited from Gibbon [15]; the paper assumes this transfers to pathological dementia speech.
domain assumption The RAPT pitch tracker produces reliable F0 contours on dementia speech with pauses and voiceless breaks.
Used in Section 2 to compute the FM envelope; no validation on pathological audio is provided.
domain assumption Pre-trained Wav2Vec2.0, ViT, and BERT models generalize to dementia speech and ASR transcripts without fine-tuning.
Section 3.2 uses these as non-trainable components; if ASR errors are correlated with dementia, the linguistic branch could introduce bias.
domain assumption ADReSSo dementia labels and MMSE scores are accurate ground truth.
Section 4.1 relies on challenge labels; no verification of label reliability.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Leveraging AM and FM Rhythm Spectrograms for Dementia Classification and Assessment." pith.science (2026). https://pith.science/paper/I7JVTRSU

@misc{pith2026250600861,
  author       = {Pith},
  title        = {Pith review of: Leveraging AM and FM Rhythm Spectrograms for Dementia Classification and Assessment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/I7JVTRSU}},
  note         = {Machine review of arXiv:2506.00861}
}

abstract

This study explores the potential of Rhythm Formant Analysis (RFA) to capture long-term temporal modulations in dementia speech. Specifically, we introduce RFA-derived rhythm spectrograms as novel features for dementia classification and regression tasks. We propose two methodologies: (1) handcrafted features derived from rhythm spectrograms, and (2) a data-driven fusion approach, integrating proposed RFA-derived rhythm spectrograms with vision transformer (ViT) for acoustic representations along with BERT-based linguistic embeddings. We compare these with existing features. Notably, our handcrafted features outperform eGeMAPs with a relative improvement of $14.2\%$ in classification accuracy and comparable performance in the regression task. The fusion approach also shows improvement, with RFA spectrograms surpassing Mel spectrograms in classification by around a relative improvement of $13.1\%$ and a comparable regression score with the baselines.

Figures

Figures reproduced from arXiv: 2506.00861 by the authors.

**Figure 2.** Illustration of AM and FM rhythm spectrograms of speech utterance from healthy control (HC) and dementia. acterization of AM and FM rhythm spectrograms. In the first approach, we compute the variance of rhythm formants over time from the rhythm spectrograms and explore the 2D discrete cosine transform-based joint spectro-temporal representation of rhythm spectrograms. These handcrafted features are then fed into a… view at source ↗

**Figure 3.** End-to-end pipeline for dementia detection using ViTBERT fusion system. each utterance is represented by a 2N + 2 × C 2 -dimensional feature vector, which is subsequently used for classification and regression with machine learning models. 3.2. Data-driven characterization of rhythm spectrogram using ViT-BERT for classification The ViT-BERT fusion model, trained using both acoustic and linguistic evidence, has rece… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

37 extracted references · 37 canonical work pages

[22]

The rhythms of rhythm,

D. Gibbon, “The rhythms of rhythm,” Journal of the International Phonetic Association, p. 1–33, 2021

work page 2021
[11]

Evaluation of speech-based pro- tocol for detection of early-stage dementia

A. Satt, A. Sorin, O. Toledo-Ronen, O. Barkan, I. Kompatsiaris, A. Kokonozi, and M. Tsolaki, “Evaluation of speech-based pro- tocol for detection of early-stage dementia.” inInterspeech, 2013, pp. 1692–1696

work page 2013
[1]

Introduction Dementia describes a cluster of neurodegenerative condi- tions characterized by progressive cognitive decline, with Alzheimer’s disease (AD) being the most prevalent cause [1]. While memory loss is often considered the primary clinical hall- mark, speech, and language impairments emerge early and can manifest as hesitations, disrupted rhythm,...

work page Pith review arXiv 2025
[2]

The overall block diagram illustrat- ing the computation process is shown in Figure 1 and described as follows

Rhythm spectrogram computation We use the RFA method reported in [15] to compute the AM and FM rhythm spectrograms. The overall block diagram illustrat- ing the computation process is shown in Figure 1 and described as follows. Computation of AM rhythm spectrogram. The speech signal is first normalized using its maximum absolute value. An absolute Hilbert...

work page
[3]

The embeddings extracted from ViT-BERT are further used for regression with machine learning models

Proposed approach for dementia detection and assessment We employ AM and FM rhythm spectrograms for dementia detection and assessment using two approaches: (1) extract- ing handcrafted features for classification and regression with machine learning models, and (2) leveraging a data-driven ap- proach with the ViT-BERT acoustic-linguistic end-to-end (E2E) ...

work page
[4]

Cookie Theft

Experimental setup and results 4.1. Dataset description We used Alzheimer’s Dementia Recognition through Sponta- neous Speech Only (ADReSSo) [22] in this study. This dataset was originally introduced as part of the ADReSSo Challenge 2021 [22], aiming to detect dementia and assess cognitive de- cline using speech alone. It consists of audio recordings of p...

work page 2021
[5]

Conclusions In this study, we explored the use of rhythm spectrograms through both handcrafted and data-driven representations for dementia detection. Experimental results demonstrate that the proposed characterization of rhythm spectrograms achieves su- perior performance in dementia classification and comparable results in MMSE score prediction. In the ...

work page
[6]

Acknowledgements The authors wish to acknowledge CSC – IT Center for Science, Finland, for computational resources

work page

Show all 37 references

[7]

Using the outputs of different automatic speech recognition paradigms for acoustic-and BERT-based Alzheimer’s dementia detection through spontaneous speech

Y . Pan, B. Mirheidari, J. M. Harris, J. C. Thompson, M. Jones, J. S. Snowden, D. Blackburn, and H. Christensen, “Using the outputs of different automatic speech recognition paradigms for acoustic-and BERT-based Alzheimer’s dementia detection through spontaneous speech.” in In...

2021
[8]

Dementia,

W. H. Organization, “Dementia,” WHO Fact Sheets , 2023. [Online]. Available: https://www.who.int/news-room/fact-sheets/ detail/dementia

2023
[9]

Speech and language impairments in dementia,

B. Klimova and K. Kuca, “Speech and language impairments in dementia,” Journal of Applied Biomedicine , vol. 14, no. 2, pp. 97–103, 2016

2016
[10]

Alzheimer’s dementia recognition through spontaneous speech,

S. Luz, F. Haider, S. de la Fuente Garcia, D. Fromm, and B. MacWhinney, “Alzheimer’s dementia recognition through spontaneous speech,” Frontiers in Computer Science , vol. 3, p. 780169, 2021

2021
[12]

Lexical and acoustic speech features relating to alzheimer disease pathology,

S. Cho, K. A. Q. Cousins, S. Shellikeri, S. Ash, D. J. Irwin, M. Y . Liberman, M. Grossman, and N. Nevler, “Lexical and acoustic speech features relating to alzheimer disease pathology,” Neurol- ogy, vol. 99, no. 4, pp. e313–e322, 2022

2022
[13]

Speech rate and syntactic complexity effects on the auditory comprehension of alzheimer patients,

C. K. Tomoeda, K. A. Bayles, D. R. Boone, A. W. Kaszniak, and T. J. Slauson, “Speech rate and syntactic complexity effects on the auditory comprehension of alzheimer patients,” Journal of Com- munication Disorders, vol. 23, no. 2, pp. 151–161, 1990

1990
[14]

Automatic measurement of prosody in behavioral variant ftd,

N. Nevler, S. Ash, C. Jester, D. J. Irwin, M. Liberman, and M. Grossman, “Automatic measurement of prosody in behavioral variant ftd,” Neurology, vol. 89, no. 7, pp. 650–656, 2017

2017
[15]

A two- step attention-based feature combination cross-attention system for speech-based dementia detection,

Y . Pan, B. Mirheidari, D. Blackburn, and H. Christensen, “A two- step attention-based feature combination cross-attention system for speech-based dementia detection,” IEEE Transactions on Au- dio, Speech and Language Processing, 2025

2025
[16]

Acoustic speech analysis in alzheimer’s disease: A systematic review and meta-analysis,

S. Saeedi, S. Hetjens, M. Grimm, and B. B. v. Latoszek, “Acoustic speech analysis in alzheimer’s disease: A systematic review and meta-analysis,” The Journal of Prevention of Alzheimer’s Disease, vol. 11, no. 6, pp. 1789–1797, 2024

2024
[17]

Automatic detection of Alzheimer’s disease using spontaneous speech only,

J. Chen, J. Ye, F. Tang, and J. Zhou, “Automatic detection of Alzheimer’s disease using spontaneous speech only,” in Proc. In- terspeech, 2021, p. 3830

2021
[18]

Detecting dementia from speech and transcripts using transformers,

L. Ilias, D. Askounis, and J. Psarras, “Detecting dementia from speech and transcripts using transformers,” Computer Speech & Language, vol. 79, p. 101485, 2023

2023
[19]

A systematic review of ex- pressive and receptive prosody in people with dementia,

C. Oh, R. J. Morris, and X. Wang, “A systematic review of ex- pressive and receptive prosody in people with dementia,” Journal of Speech, Language, and Hearing Research, vol. 64, no. 10, pp. 3803–3825, 2021

2021
[20]

Prosody analysis as a tool for differential diagnosis of cognitive impairment,

——, “Prosody analysis as a tool for differential diagnosis of cognitive impairment,” in Proceedings of Meetings on Acoustics, vol. 50, no. 1. AIP Publishing, 2022

2022
[21]

The F0 contour is smoothed to pro- duce the FM envelope

Python package). The F0 contour is smoothed to pro- duce the FM envelope. V oiceless segments and pauses appear as breaks in the contour but are preserved as valid components for subsequent spectral analysis. As with the AM envelope, a 5 s window with overlapping steps is used...
[23]

The Future of Prosody: It’s about Time,

——, “The Future of Prosody: It’s about Time,” in Proc. Speech Prosody 2018, 2018, pp. 1–9

2018
[24]

Quantifying and correlating rhythm for- mants in speech,

D. Gibbon and P. Li, “Quantifying and correlating rhythm for- mants in speech,” in Proc. Linguistic Patterns in Spontaneous Speech (LPSS). Taipei, Academia Sinica, 2019

2019
[25]

Cross-linguistic rhythm analysis of Mising and Assamese,

P. Gogoi, P. Sarmah, and S. R. M. Prasanna, “Cross-linguistic rhythm analysis of Mising and Assamese,” ACM Trans. Asian Low-Resour. Lang. Inf. Process., vol. 23, no. 10, pp. 1–18, Oct
[26]

Intelligibility assessment of cleft lip and palate speech using Gaussian posteri- ograms based on joint spectro-temporal features,

S. Kalita, S. Mahadeva Prasanna, and S. Dandapat, “Intelligibility assessment of cleft lip and palate speech using Gaussian posteri- ograms based on joint spectro-temporal features,” The Journal of the Acoustical Society of America, vol. 144, no. 4, pp. 2413–2423, 2018

2018
[27]

Analyzing long-term rhythm variations in mising and assamese using frequency domain correlates,

——, “Analyzing long-term rhythm variations in mising and assamese using frequency domain correlates,” International Journal of Asian Language Processing , no. just accepted, 2025. [Online]. Available: https://doi.org/10.1142/S2717554525500018

2025 doi
[28]

A robust algorithm for pitch tracking (rapt),

D. Talkin and W. B. Kleijn, “A robust algorithm for pitch tracking (rapt),” Speech coding and synthesis, vol. 495, p. 518, 1995

1995
[29]

r9y9/pysptk: 0.1. 14,

R. Yamamoto, J. Felipe, and M. Blaauw, “r9y9/pysptk: 0.1. 14,” URL: https://github. com/r9y9/pysptk, 2019

2019
[30]

Detecting cognitive decline using speech only: The ADReSSo Challenge,

S. Luz, F. Haider, S. de la Fuente, D. Fromm, and B. MacWhin- ney, “Detecting cognitive decline using speech only: The ADReSSo Challenge,” in Proc. Interspeech, 2021

2021
[31]

SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python,

P. Virtanen, R. Gommers, E. Oliphant, Travis et al., “SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python,” Nature Methods, vol. 17, pp. 261–272, 2020

2020
[32]

K. R. Rao and P. Yip, Discrete Cosine Transform: Algorithms, Advantages, Applications. Academic Press, 1990

1990
[33]

Acoustic and percep- tual evaluation of category goodness of /t/ and /k/ in typical and misarticulated children’s speech,

S. Str ¨ombergsson, G. Salvi, and D. House, “Acoustic and percep- tual evaluation of category goodness of /t/ and /k/ in typical and misarticulated children’s speech,” The Journal of the Acoustical Society of America, vol. 137, no. 6, pp. 3422–3435, 2015

2015
[35]

Evaluating web-based automatic transcription for Alzheimer speech data: Transcript comparison and machine learning analysis,

T. Soroski et al., “Evaluating web-based automatic transcription for Alzheimer speech data: Transcript comparison and machine learning analysis,” JMIR Aging, vol. 5, 2022

2022
[36]

wav2vec 2.0: A framework for self-supervised learning of speech representa- tions,

A. Baevski, Y . Zhou, A. Mohamed, and M. Auli, “wav2vec 2.0: A framework for self-supervised learning of speech representa- tions,” in Advances in neural information processing systems , 2020

2020
[37]

The Geneva minimalistic acoustic parameter set (GeMAPS) for voice research and affective computing,

F. Eyben et al., “The Geneva minimalistic acoustic parameter set (GeMAPS) for voice research and affective computing,” IEEE Transactions on Affective Computing, vol. 7, no. 2, pp. 190–202, 2015

2015
[2024]

Available: https://doi.org/10.1145/3694785

[Online]. Available: https://doi.org/10.1145/3694785

doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [22]

The rhythms of rhythm,

D. Gibbon, “The rhythms of rhythm,” Journal of the International Phonetic Association, p. 1–33, 2021

work page 2021

[2] [11]

Evaluation of speech-based pro- tocol for detection of early-stage dementia

A. Satt, A. Sorin, O. Toledo-Ronen, O. Barkan, I. Kompatsiaris, A. Kokonozi, and M. Tsolaki, “Evaluation of speech-based pro- tocol for detection of early-stage dementia.” inInterspeech, 2013, pp. 1692–1696

work page 2013

[3] [1]

Introduction Dementia describes a cluster of neurodegenerative condi- tions characterized by progressive cognitive decline, with Alzheimer’s disease (AD) being the most prevalent cause [1]. While memory loss is often considered the primary clinical hall- mark, speech, and language impairments emerge early and can manifest as hesitations, disrupted rhythm,...

work page Pith review arXiv 2025

[4] [2]

The overall block diagram illustrat- ing the computation process is shown in Figure 1 and described as follows

Rhythm spectrogram computation We use the RFA method reported in [15] to compute the AM and FM rhythm spectrograms. The overall block diagram illustrat- ing the computation process is shown in Figure 1 and described as follows. Computation of AM rhythm spectrogram. The speech signal is first normalized using its maximum absolute value. An absolute Hilbert...

work page

[5] [3]

The embeddings extracted from ViT-BERT are further used for regression with machine learning models

Proposed approach for dementia detection and assessment We employ AM and FM rhythm spectrograms for dementia detection and assessment using two approaches: (1) extract- ing handcrafted features for classification and regression with machine learning models, and (2) leveraging a data-driven ap- proach with the ViT-BERT acoustic-linguistic end-to-end (E2E) ...

work page

[6] [4]

Cookie Theft

Experimental setup and results 4.1. Dataset description We used Alzheimer’s Dementia Recognition through Sponta- neous Speech Only (ADReSSo) [22] in this study. This dataset was originally introduced as part of the ADReSSo Challenge 2021 [22], aiming to detect dementia and assess cognitive de- cline using speech alone. It consists of audio recordings of p...

work page 2021

[7] [5]

Conclusions In this study, we explored the use of rhythm spectrograms through both handcrafted and data-driven representations for dementia detection. Experimental results demonstrate that the proposed characterization of rhythm spectrograms achieves su- perior performance in dementia classification and comparable results in MMSE score prediction. In the ...

work page

[8] [6]

Acknowledgements The authors wish to acknowledge CSC – IT Center for Science, Finland, for computational resources

work page