Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection

Helen Meng; Jiawen Kang; Jinchao Li; Junan Li; Xixin Wu

REVIEW 3 major objections 6 minor 42 references

Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read ASR errors are not equal: stopwords contribute 60% of transcription errors but little discriminative signal, while task-related keywords—only 9% of errors—are pivotal for BERT-based Alzheimer's detection.

desk verdict Useful word-level breakdown of ASR errors in AD detection, but the causal importance ranking rests on a perturbation proxy and an analysis restricted to ASR-robust cases; still worth a referee. read the letter →

arxiv 2412.06332 v1 pith:KXI2KH4N submitted 2024-12-09 cs.CL cs.AIcs.LGq-bio.NC

Jiawen Kang , Junan Li , Jinchao Li , Xixin Wu , Helen Meng This is my paper · ORCID

classification cs.CLcs.AIcs.LGq-bio.NC

keywords Alzheimer'sdiseasespeechrecognitionpre-trainedlanguagemodelspokenprocessingASRerroranalysisstopwordstask-relatedkeywordsADReSS

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 explain why automatic transcriptions with a high word error rate can still support Alzheimer's disease detection as accurately as manual transcripts. Using the ADReSS Cookie Theft picture-description corpus, a BERT-based detector, and a tailor-made ASR system, it finds that a 33.9% word error rate leaves detection accuracy unchanged at 88%. The reason is that errors are not distributed evenly across word types: stopwords make up 60% of errors but carry little discriminative weight, while diagnosis-related keywords account for only 9% of errors and are the words that move BERT embeddings across the classification boundary. The implication is that for clinical screening, the quality of ASR should be measured by how well task-critical words survive, not by overall word error rate.

What carries the argument

The central object is the signed hyperplane offset $d = (w^T x + b)/\|w\|$ of the BERT [CLS] embedding (PCA-reduced to 108 dimensions) from the linear SVM decision boundary. The argument works by taking manual transcripts, randomly removing or substituting stopwords versus keywords, recomputing BERT embeddings, and tracking whether the average offset crosses zero, the decision boundary. This turns the qualitative question of which ASR errors matter into a quantitative measure of how far different word-level edits push the classifier's input.

What would settle it

A direct test would be to train two ASR variants, one whose errors fall mainly on stopwords and one whose errors fall mainly on task keywords at the same overall WER; if the keyword-error variant does not degrade AD detection accuracy more than the stopword-error variant, the central claim that keywords are pivotal would be falsified.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that ASR errors are strongly biased toward semantically light words, and the downstream BERT-based classifier is correspondingly insensitive to those errors. In the ASR-robust cases, stopwords are 24% of error types but 60% of error tokens, with a stopword WER of 32.4%, whereas task-related keywords such as 'cookie', 'jar', 'boy', 'girl', and 'water' have a WER of only 14.3% and account for just 9% of all errors. Controlled ablation experiments on manual transcripts show that randomly removing or substituting stopwords shifts BERT embeddings toward the SVM decision boundary without crossing it, while removing or substituting keywords crosses the boundary and flips the classification. The paper concludes that stopword errors are largely harmless to AD detection, keyword errors are pivotal, and the apparent paradox of high WER coexisting with unchanged accuracy is resolved by this asymmetry.

Load-bearing premise

The claim that stopwords are unimportant rests on random word-level edits to manual transcripts standing in for real ASR errors; if real errors are systematic, correlated with speaker traits, word position, or acoustic difficulty, their effect on the classifier could differ from what the ablation measures.

Editorial extensions

If this is right

If the central claim is right, ASR benchmarking for AD screening should report word error rates separately for task keywords and stopwords rather than a single overall WER.
A stopword-heavy ASR error profile can sustain 88% detection accuracy, so improving ASR on stopwords alone will not improve screening.
Recognition of the task-related keywords is the bottleneck; keyword WER at 14.3% is already low, so further gains should focus on the remaining keyword misses.
Because non-keyword and non-stopword errors also cross the boundary, the classifier is sensitive to semantic content broadly, meaning ASR evaluation should weight content words by their contribution to the downstream task.
Deletion or substitution of keywords, even a small fraction of errors, can flip a healthy-control transcript into the AD region, so error analyses should track keyword-level confusion rather than aggregate rates.

Reading between the lines

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

A natural extension the paper does not draw: because ASR errors push embeddings toward the AD side of the boundary, stopword-heavy errors may systematically bias screening toward false positives.
The asymmetry suggests ASR confidence scores could be used to down-weight stopword errors when computing transcript-based features for clinical decision support.
The same word-class ablation could be applied to other clinical NLP tasks with a task-specific vocabulary; if the pattern generalizes, task-word preservation rather than overall WER should guide ASR development.
A testable prediction follows: an ASR language model biased toward Cookie Theft keywords should improve AD detection robustness more than an equal-WER reduction on general vocabulary.

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 paper gives a clear, quantitative answer to a question the field has been circling: when a 33.9% WER ASR transcript yields the same AD detection accuracy as a manual transcript, where do the errors actually land? The descriptive core is solid. Stopwords make up 60% of ASR errors, task keywords only 9%, and keyword WER is 14.3% versus 33.9% overall. That is a genuinely informative, reproducible observation that goes beyond prior claims that errors can be harmless or helpful. The alignment maps and the long-tail error distribution are good illustrations. Building on [31, 32] by explicitly separating stopwords from task keywords is a real extension, not just a rehash.

The soft spots are real but not fatal. First, the error-composition tables are computed only on the 40/48 ASR-robust cases, the ones where ASR did not flip the prediction. That is selection on the outcome: you are describing errors in cases where they had no impact, so concluding that those errors are harmless is somewhat circular. The 8 non-robust cases could be dominated by keyword errors, or even by stopword errors that happened to push the embedding across the boundary; we simply do not see them. The paper should at least report the error composition for those 8 cases separately.

Second, the ablation experiments randomly remove or substitute words in manual transcripts and measure the SVM hyperplane offset. That measures the classifier's sensitivity to hypothetical edits, not the actual effect of real ASR confusions. Real ASR errors are systematic and correlate with word acoustics and speaker characteristics. The paper is careful with its language, saying the results "suggest" rather than "prove," but the abstract's wording that stopwords "are shown to play a limited role" is a bit stronger than the evidence supports. A direct counterfactual on the actual ASR hypotheses—e.g., replacing the recognized keyword with the manual keyword and measuring the decision change—would nail the causal claim.

That said, the central argument holds up as an analysis rather than a definitive experiment. The finding that task keywords are both low-WER and high-importance is a useful practical hint for building ASR-robust AD screening: optimize ASR for content words, not for stopword accuracy. The paper is concise, honest about its assumptions, and the descriptive statistics are reproducible. I would send it to peer review. A solid revision would address the 8 excluded cases and add a direct ASR-error simulation. For a reading group on clinical speech analytics, it is worth a slot.

Referee Report

3 major / 6 minor

Summary. The paper investigates how automatic speech recognition (ASR) errors affect BERT-based Alzheimer's disease (AD) detection. Using a custom ASR system with 33.9% WER on the ADReSS-2020 test set, the authors reproduce the previous observation that ASR and manual transcriptions yield equivalent detection accuracy (88%). They then analyze error composition and report that, within the 40 ASR-robust cases where predictions did not change, stopwords comprise 60% of errors while task-related keywords comprise only 9%. Through word-ablation experiments on manual transcripts, they show that randomly removing or substituting stopwords shifts BERT embeddings toward the SVM decision boundary without crossing it, whereas removing or substituting keywords causes the embeddings to cross the boundary. The paper concludes that not all ASR errors are equally detrimental: stopword errors are largely harmless while keyword errors are pivotal for downstream AD detection.

Significance. The paper makes a useful descriptive contribution by documenting the non-uniform distribution of ASR errors across word types and by demonstrating that the BERT+SVM classifier's embedding is less sensitive to stopword edits than to keyword edits. It also reproduces the known non-linear WER-accuracy relationship and provides transparent error accounting with an externally sourced keyword list, which aids reproducibility. If the causal interpretation were fully supported, the findings would motivate ASR evaluation metrics that weight content-word errors more heavily in AD screening contexts. However, the causal claim currently rests on a proxy ablation rather than on a direct analysis of real ASR errors, so the significance is conditional on additional validation.

major comments (3)

[Section 4.1 (Tables 4 and 5)] The error-composition statistics (60% stopwords, 9% keywords) are computed only on the 40/48 ASR-robust cases, i.e., cases where ASR errors did not change the prediction. This is close to circular for the central claim: by construction, these are cases where errors were not harmful. The 8 non-robust cases are excluded from the analysis and their error composition is not reported. If those 8 cases contain a high proportion of keyword or non-stopword errors, the conclusion would be strengthened; if they contain mostly stopword errors, the conclusion would be undermined. The paper should report the error composition for all 48 test cases, and specifically for the 8 non-robust cases, or justify why the restriction to robust cases is appropriate.
[Section 4.2 and Figure 4] The word-ablation experiments randomly remove or substitute words in manual transcripts, not actual ASR errors. Random edits do not mimic systematic ASR confusions (e.g., phonetic similarity, deletions and insertions correlated with word frequency and speaking style), and substituting a stopword with an unrelated content word changes text naturalness in ways that BERT embeddings may reflect. Moreover, the ablation uses edit ratios up to 100% of a word class, which far exceeds the per-speaker ASR error rate of roughly 34%. The resulting offset curves therefore measure the sensitivity of the fitted BERT+SVM decision boundary to arbitrary word-level perturbations, not the actual effect of ASR errors on downstream decisions. A direct counterfactual analysis on the real ASR transcriptions—for instance, correcting or perturbing actual recognized words by error type, or comparing manual vs. ASR embeddings for the same speaker—is needed to support the causal ranking of stopword versus keyword errors.
[Section 4.3 (Table 5) and Conclusions] The claim that keyword errors are 'pivotal' is based on the ablation crossing the SVM boundary when keywords are edited. However, the 40 ASR-robust cases contain 130 keyword errors (9% of all errors) yet yield unchanged predictions. If keyword errors were as pivotal as the ablation suggests, one would expect the robust cases to contain few or no keyword errors, or the errors to occur in non-predictive positions. The paper does not reconcile the frequency of keyword errors in robust cases with the strong sensitivity shown in Figure 4(b). Please quantify per-speaker keyword error counts and their relation to prediction stability, or provide an explanation for why the ablation's boundary crossing does not translate into real ASR-induced flips.

minor comments (6)

[Section 4.2 heading] The heading contains a typo: 'Anayisis' should be 'Analysis'.
[Abstract and Section 4.2] The abstract and introduction state that 'stopwords constitute 60% of errors' without noting that this figure comes from the ASR-robust subset; please qualify this claim in the abstract or clearly in the text.
[Table 3] Table 3 lists 36 keywords but the text says the list contains 39 task-related keywords; please verify the count or include the missing entries.
[Figure 4] The x-axis label 'Edit Ratio (x10%)' is ambiguous; consider labeling it as 'Edit ratio (×10)' or 'Edit ratio (1–10 = 10%–100%)'.
[Section 2.3] The PCA target dimensionality is set to 108 because it equals the training set size; this choice is mentioned but not justified as a design decision, and the sensitivity of the results to this hyperparameter is not explored.
[References] Reference [30] lists the title 'The far side of failure: Investigating the impact of speech recognition errors on subsequent dementia classification' but the cited work appears to have been published under a different title; please check the bibliographic details.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the stopword/keyword importance ranking comes from independent perturbation experiments, not from fitting or self-referential definitions.

full rationale

The paper's central claim is that stopword-dominant ASR errors are less harmful to BERT-based AD detection than rarer keyword errors. The derivation is not an input-output tautology: the detection model is fitted on manual transcriptions, the ASR errors are real system outputs, and the importance ranking is established by word-ablation experiments that measure signed distances to the fitted SVM hyperplane (Eq. 1). Those perturbation experiments are independent of the error-composition statistics; the conclusion that stopwords shift embeddings only toward but not across the boundary, while keyword edits cross it, is a computed empirical result rather than a restatement of a fitted parameter. The keyword list is sourced externally from [31], and the self-citations to [19] and [25] concern the construction of the ASR and detection systems, not the load-bearing evidence for the error-importance claim. The use of the 40/48 ASR-robust cases for error-composition statistics is a selection-bias limitation (those cases are by definition ones where errors did not change the prediction), but it is not a circular reduction because the perturbation analysis covers participants beyond that subset and provides independent support. No equation, parameter, or cited result is equivalent by construction to the paper's conclusions.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The headline numbers are descriptive WER statistics, which are directly measured. The causal importance ranking, however, depends on the fitted BERT+SVM model and on the word-perturbation proxy, so the free parameters and domain assumptions above carry the load.

free parameters (2)

PCA target dimensionality = 108
Set to the training-set size following [19]; determines the SVM input space and hence all offset measurements.
SVM regularization C = 1
Chosen without tuning following [19]; the decision boundary it produces is the reference used to define word importance.

assumptions (4)

domain assumption BERT [CLS] embedding of the whole transcript is a sufficient representation for AD detection.
The detection pipeline and all importance measurements are defined through this embedding (Section 2.3).
ad hoc to paper Random removal or substitution of words in manual transcripts is a valid proxy for how ASR errors alter BERT embeddings.
Used throughout Sections 4.2 and 4.3 to attribute causal roles to word types, but actual ASR errors are not random.
domain assumption The 39-keyword list from [31] accurately captures Cookie Theft picture elements relevant to AD diagnosis.
Keyword importance conclusions rest on this external list (Table 3).
domain assumption The SVM decision boundary trained on manual transcriptions is a valid measure of clinical discriminability.
Hyperplane offsets in Equation (1) and Figure 4 interpret model sensitivity as diagnostic relevance.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection." pith.science (2026). https://pith.science/paper/KXI2KH4N

@misc{pith2026241206332,
  author       = {Pith},
  title        = {Pith review of: Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KXI2KH4N}},
  note         = {Machine review of arXiv:2412.06332}
}

read the original abstract

Automatic Speech Recognition (ASR) plays an important role in speech-based automatic detection of Alzheimer's disease (AD). However, recognition errors could propagate downstream, potentially impacting the detection decisions. Recent studies have revealed a non-linear relationship between word error rates (WER) and AD detection performance, where ASR transcriptions with notable errors could still yield AD detection accuracy equivalent to that based on manual transcriptions. This work presents a series of analyses to explore the effect of ASR transcription errors in BERT-based AD detection systems. Our investigation reveals that not all ASR errors contribute equally to detection performance. Certain words, such as stopwords, despite constituting a large proportion of errors, are shown to play a limited role in distinguishing AD. In contrast, the keywords related to diagnosis tasks exhibit significantly greater importance relative to other words. These findings provide insights into the interplay between ASR errors and the downstream detection model.

Figures

Figures reproduced from arXiv: 2412.06332 by the authors.

**Figure 1.** The top 20 error word distribution of the ASR system. 4. Analysis of speech recognition errors Based on the detection system in Section 3, this section characterizes the composition of ASR errors and investigates why a high word error rate did not significantly degrade AD detection performance. We will start with error distribution and then examine certain types of errors using the test set transcriptions. 4.1. Ba… view at source ↗

**Figure 2.** Samples of alignment map for participants s170 (a) and s179 (b). Squares represent correctly transcribed words, while crosses (’x’) indicate ASR errors. Blue color for stopwords, red for keywords, and gray for other words [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** BERT embedding variations as stopwords are incrementally removed/substituted from manual transcriptions (participant s172). Light to dark blue: fewer to more stopwords removed/substituted. ’x’: ASR transcript. White/gray: health/AD decision regions. Furthermore, to reduce randomness and quantify this experiment across participants, we calculate the hyperplane offset for each participant and each removal/substitu… view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Average hyperplane offset for transcription embeddings as stopwords (a) and keywords (b) are incrementally removed or substituted. Edit ratio represents the percentage of words removed/substituted. how embeddings move toward or against the decision boundary. We aver…

**Figure 5.** Figure 5: BERT embedding variations as keywords are incrementally removed/substituted from manual transcriptions (participant s172). Light to dark blue: fewer to more keywords removed/substituted. ’x’: ASR transcript. White/gray: health/AD decision regions. 5. Conclusions We …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 37 canonical work pages

[1]

Introduction Alzheimer’s disease (AD) is a neurodegenerative disorder char- acterized by progressive cognitive impairment, including deteri- oration in memory, attention, and executive function. Due to the irreversible progression of AD pathology [1], early detection and diagnosis play a pivotal role in facilitating timely interven- tion and management, c...

work page
[2]

Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection

Approach 2.1. Data The data used in this work comes from the Alzheimer’s Demen- tia Recognition Through Spontaneous Speech (ADReSS) Chal- lenge 2020 corpus [33]. This challenge selects a sub-task of Pitt Corpus in the DementiaBank database [34], which requires all arXiv:2412.06332v1 [cs.CL] 9 Dec 2024 the participants to describe the Cookie Theft picture ...

work page Pith review arXiv 2020
[3]

TP , TN, FP , FN

AD Detection with ASR Transcriptions The non-linear dependency between ASR WER and AD de- tection accuracy has been observed in multiple studies [19, 29, 30, 31]. To look into this observation, we first set up a subject system using the above-mentioned settings. Table 1 shows the classification results of the subject detection system. Note that the detect...

work page
[4]

tokens” represents the count of words and “types

Analysis of speech recognition errors Based on the detection system in Section 3, this section char- acterizes the composition of ASR errors and investigates why a high word error rate did not significantly degrade AD detec- tion performance. We will start with error distribution and then examine certain types of errors using the test set transcriptions. ...

work page
[5]

Specifically, we explored why transcriptions with notable word error rates could yield detection accuracy equivalent to that of manual transcriptions

Conclusions We performed a series of analyses to investigate the effect of ASR transcription errors in BERT-based Alzheimer’s Disease (AD) detection. Specifically, we explored why transcriptions with notable word error rates could yield detection accuracy equivalent to that of manual transcriptions. We have shown that not all ASR errors are equally detrim...

work page
[6]

T45- 407/19N) and the CUHK Stanley Ho Big Data Decision Research Centre

Acknowledgements This work is supported by the HKSARG Research Grants Council’s Theme-based Research Grant Scheme (Project No. T45- 407/19N) and the CUHK Stanley Ho Big Data Decision Research Centre

work page
[7]

World alzheimer report 2019: Attitudes to dementia, a global survey: Public health: Engaging people in adrd research,

C. Lynch, “World alzheimer report 2019: Attitudes to dementia, a global survey: Public health: Engaging people in adrd research,” Alzheimer’s & Dementia, 2020

work page 2019
[8]

The montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment,

Z. S. Nasreddine, N. A. Phillips, V . B ´edirian, S. Charbonneau, V . Whitehead, I. Collin, J. L. Cummings, and H. Chertkow, “The montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment,” Journal of the American Geriatrics Society, vol. 53, no. 4, pp. 695–699, 2005

work page 2005

Show all 42 references

[9]

“frog, where are you?

J. Reilly, M. Losh, U. Bellugi, and B. Wulfeck, ““frog, where are you?” narratives in children with specific language impairment, early focal brain injury, and williams syndrome,” Brain and lan- guage, vol. 88, no. 2, pp. 229–247, 2004

2004
[10]

Ten years of research on automatic voice and speech analysis of people with alzheimer’s disease and mild cog- nitive impairment: a systematic review article,

I. Mart ´ınez-Nicol´as, T. E. Llorente, F. Mart ´ınez-S´anchez, and J. J. G. Meil ´an, “Ten years of research on automatic voice and speech analysis of people with alzheimer’s disease and mild cog- nitive impairment: a systematic review article,” Frontiers in Psy- chology, vol...

2021
[11]

Integrated and enhanced pipeline system to support spoken language analytics for screen- ing neurocognitive disorders,

H. Meng, B. Mak, M.-W. Mak, H. Fung, X. Gong, T. Kwok, X. Liu, V . Mok, P. Wong, J. Wooet al., “Integrated and enhanced pipeline system to support spoken language analytics for screen- ing neurocognitive disorders,” Interspeech, 2023

2023
[12]

Spoken language biomarkers for detecting cognitive impairment,

T. Alhanai, R. Au, and J. Glass, “Spoken language biomarkers for detecting cognitive impairment,” 2017

2017
[13]

Linguistic fea- tures identify alzheimer’s disease in narrative speech,

K. C. Fraser, J. A. Meltzer, and F. Rudzicz, “Linguistic fea- tures identify alzheimer’s disease in narrative speech,” Journal of Alzheimer’s Disease, 2016

2016
[14]

Speech reveals future risk of developing dementia: Predictive dementia screening from biographic interviews,

J. Weiner, C. Frankenberg, J. Schr ¨oder, and T. Schultz, “Speech reveals future risk of developing dementia: Predictive dementia screening from biographic interviews,” in ASRU. IEEE, 2019

2019
[15]

Verbal fluency in normal ag- ing and cognitive decline: Results of a longitudinal study,

C. Frankenberg, J. Weiner, M. Knebel, A. Abulimiti, P. Toro, C. J. Herold, T. Schultz, and J. Schr¨oder, “Verbal fluency in normal ag- ing and cognitive decline: Results of a longitudinal study,” Com- puter Speech & Language, 2021

2021
[16]

Exploiting multi- modal features from pre-trained networks for alzheimer’s demen- tia recognition,

J. Koo, J. H. Lee, J. Pyo, Y . Jo, and K. Lee, “Exploiting multi- modal features from pre-trained networks for alzheimer’s demen- tia recognition,” in INTERSPEECH, 2020

2020
[17]

Comparing pre-trained and feature-based models for prediction of alzheimer’s disease based on speech,

A. Balagopalan, B. Eyre, J. Robin, F. Rudzicz, and J. Novikova, “Comparing pre-trained and feature-based models for prediction of alzheimer’s disease based on speech,” Frontiers in aging neu- roscience, 2021

2021
[18]

Automated recognition of alzheimer’s dementia using bag-of-deep-features and model ensembling,

Z. S. Syed, M. S. S. Syed, M. Lech, and E. Pirogova, “Automated recognition of alzheimer’s dementia using bag-of-deep-features and model ensembling,” IEEE Access, 2021

2021
[19]

To bert or not to bert: comparing speech and language-based approaches for alzheimer’s disease detection,

A. Balagopalan, B. Eyre, F. Rudzicz, and J. Novikova, “To bert or not to bert: comparing speech and language-based approaches for alzheimer’s disease detection,” arXiv preprint arXiv:2008.01551, 2020

2008 arXiv
[20]

Disflu- encies and fine-tuning pre-trained language models for detection of alzheimer’s disease

J. Yuan, Y . Bian, X. Cai, J. Huang, Z. Ye, and K. Church, “Disflu- encies and fine-tuning pre-trained language models for detection of alzheimer’s disease.” in INTERSPEECH, 2020

2020
[21]

Temporal integra- tion of text transcripts and acoustic features for alzheimer’s diag- nosis based on spontaneous speech,

M. Martinc, F. Haider, S. Pollak, and S. Luz, “Temporal integra- tion of text transcripts and acoustic features for alzheimer’s diag- nosis based on spontaneous speech,” Frontiers in Aging Neuro- science, 2021

2021
[22]

BERT: Pre- training of deep bidirectional transformers for language under- standing,

J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “BERT: Pre- training of deep bidirectional transformers for language under- standing,” 2019

2019
[23]

Roberta: A robustly optimized bert pretraining approach,

Y . Liu, M. Ott, N. Goyal, J. Du, M. Joshi, D. Chen, O. Levy, M. Lewis, L. Zettlemoyer, and V . Stoyanov, “Roberta: A robustly optimized bert pretraining approach,” 2019

2019
[24]

Automated screening for alzheimer’s dementia through spontaneous speech

M. S. S. Syed, Z. S. Syed, M. Lech, and E. Pirogova, “Automated screening for alzheimer’s dementia through spontaneous speech.” in Interspeech, vol. 2020, 2020, pp. 2222–6

2020
[25]

A comparative study of acoustic and linguistic fea- tures classification for alzheimer’s disease detection,

J. Li, J. Yu, Z. Ye, S. Wong, M. Mak, B. Mak, X. Liu, and H. Meng, “A comparative study of acoustic and linguistic fea- tures classification for alzheimer’s disease detection,” in ICASSP. IEEE, 2021, pp. 6423–6427

2021
[26]

Leveraging pretrained representations with task-related key- words for alzheimer’s disease detection,

J. Li, K. Song, J. Li, B. Zheng, D. Li, X. Wu, X. Liu, and H. Meng, “Leveraging pretrained representations with task-related key- words for alzheimer’s disease detection,” 2023

2023
[27]

Exploring linguistic feature and model combination for speech recognition based automatic ad detection,

Y . Wang, T. Wang, Z. Ye, L. Meng, S. Hu, X. Wu, X. Liu, and H. Meng, “Exploring linguistic feature and model combination for speech recognition based automatic ad detection,” INTER- SPEECH, 2022

2022
[28]

Alzheimer’s dis- ease detection from spontaneous speech through combining lin- guistic complexity and (dis) fluency features with pretrained lan- guage models,

Y . Qiao, X. Yin, D. Wiechmann, and E. Kerz, “Alzheimer’s dis- ease detection from spontaneous speech through combining lin- guistic complexity and (dis) fluency features with pretrained lan- guage models,” in INTERSPEECH, 2021

2021
[29]

Speech recognition in alzheimer’s disease and in its assessment

L. Zhou, K. C. Fraser, and F. Rudzicz, “Speech recognition in alzheimer’s disease and in its assessment.” in Interspeech, vol. 2016, 2016, pp. 1948–1952

2016
[30]

Dementia detection using automatic analysis of conver- sations,

B. Mirheidari, D. Blackburn, T. Walker, M. Reuber, and H. Chris- tensen, “Dementia detection using automatic analysis of conver- sations,” Computer Speech & Language, vol. 53, pp. 65–79, 2019

2019
[31]

Development of the cuhk elderly speech recognition system for neurocognitive disorder detection using the dementia- bank corpus,

Z. Ye, S. Hu, J. Li, X. Xie, M. Geng, J. Yu, J. Xu, B. Xue, S. Liu, X. Liu et al., “Development of the cuhk elderly speech recognition system for neurocognitive disorder detection using the dementia- bank corpus,” in ICASSP. IEEE, 2021, pp. 6433–6437

2021
[32]

In- vestigation of data augmentation techniques for disordered speech recognition,

M. Geng, X. Xie, S. Liu, J. Yu, S. Hu, X. Liu, and H. Meng, “In- vestigation of data augmentation techniques for disordered speech recognition,” arXiv preprint arXiv:2201.05562, 2022

2022 arXiv
[33]

Per- sonalized adversarial data augmentation for dysarthric and elderly speech recognition,

Z. Jin, M. Geng, J. Deng, T. Wang, S. Hu, G. Li, and X. Liu, “Per- sonalized adversarial data augmentation for dysarthric and elderly speech recognition,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2023

2023
[34]

Bayesian parametric and architec- tural domain adaptation of lf-mmi trained tdnns for elderly and dysarthric speech recognition

J. Deng, F. R. Gutierrez, S. Hu, M. Geng, X. Xie, Z. Ye, S. Liu, J. Yu, X. Liu, and H. Meng, “Bayesian parametric and architec- tural domain adaptation of lf-mmi trained tdnns for elderly and dysarthric speech recognition.” in Interspeech, 2021, pp. 4818– 4822

2021
[35]

Conformer based elderly speech recog- nition system for alzheimer’s disease detection,

T. Wang, J. Deng, M. Geng, Z. Ye, S. Hu, Y . Wang, M. Cui, Z. Jin, X. Liu, and H. Meng, “Conformer based elderly speech recog- nition system for alzheimer’s disease detection,” arXiv preprint arXiv:2206.13232, 2022

2022 arXiv
[36]

The far side of failure: In- vestigating the impact of speech recognition errors on subsequent dementia classification,

C. Li, T. Cohen, and S. Pakhomov, “The far side of failure: In- vestigating the impact of speech recognition errors on subsequent dementia classification,” arXiv preprint arXiv:2211.07430, 2022

2022 arXiv
[37]

Useful blunders: Can automated speech recognition errors improve downstream demen- tia classification?

C. Li, W. Xu, T. Cohen, and S. Pakhomov, “Useful blunders: Can automated speech recognition errors improve downstream demen- tia classification?” Journal of Biomedical Informatics, vol. 150, p. 104598, 2024

2024
[38]

Impact of asr on alzheimer’s disease detection: All errors are equal, but deletions are more equal than others,

A. Balagopalan, K. Shkaruta, and J. Novikova, “Impact of asr on alzheimer’s disease detection: All errors are equal, but deletions are more equal than others,” in W-NUT, 2020, pp. 159–164

2020
[39]

Alzheimer’s Dementia Recognition Through Spontaneous Speech: The ADReSS Challenge,

S. Luz, F. Haider, S. de la Fuente, D. Fromm, and B. MacWhin- ney, “Alzheimer’s Dementia Recognition Through Spontaneous Speech: The ADReSS Challenge,” INTERSPEECH, 2020

2020
[40]

The natural history of alzheimer’s disease: description of study cohort and accuracy of diagnosis,

J. T. Becker, F. Boiler, O. L. Lopez, J. Saxton, and K. L. McGo- nigle, “The natural history of alzheimer’s disease: description of study cohort and accuracy of diagnosis,” Archives of neurology , 1994

1994
[41]

Kaplan, H

E. Kaplan, H. Goodglass, and S. Weintraub, Boston Naming Test. Philadelphia, PA: Lea & Febiger, 1983

1983
[42]

NLTK: The natural language toolkit,

S. Bird and E. Loper, “NLTK: The natural language toolkit,” in Proceedings of the ACL Interactive Poster and Demonstration Sessions. Barcelona, Spain: Association for Computational Linguistics, Jul. 2004, pp. 214–217. [Online]. Available: https://aclanthology.org/P04-3031

2004

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

Introduction Alzheimer’s disease (AD) is a neurodegenerative disorder char- acterized by progressive cognitive impairment, including deteri- oration in memory, attention, and executive function. Due to the irreversible progression of AD pathology [1], early detection and diagnosis play a pivotal role in facilitating timely interven- tion and management, c...

work page

[2] [2]

Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection

Approach 2.1. Data The data used in this work comes from the Alzheimer’s Demen- tia Recognition Through Spontaneous Speech (ADReSS) Chal- lenge 2020 corpus [33]. This challenge selects a sub-task of Pitt Corpus in the DementiaBank database [34], which requires all arXiv:2412.06332v1 [cs.CL] 9 Dec 2024 the participants to describe the Cookie Theft picture ...

work page Pith review arXiv 2020

[3] [3]

TP , TN, FP , FN

AD Detection with ASR Transcriptions The non-linear dependency between ASR WER and AD de- tection accuracy has been observed in multiple studies [19, 29, 30, 31]. To look into this observation, we first set up a subject system using the above-mentioned settings. Table 1 shows the classification results of the subject detection system. Note that the detect...

work page

[4] [4]

tokens” represents the count of words and “types

Analysis of speech recognition errors Based on the detection system in Section 3, this section char- acterizes the composition of ASR errors and investigates why a high word error rate did not significantly degrade AD detec- tion performance. We will start with error distribution and then examine certain types of errors using the test set transcriptions. ...

work page

[5] [5]

Specifically, we explored why transcriptions with notable word error rates could yield detection accuracy equivalent to that of manual transcriptions

Conclusions We performed a series of analyses to investigate the effect of ASR transcription errors in BERT-based Alzheimer’s Disease (AD) detection. Specifically, we explored why transcriptions with notable word error rates could yield detection accuracy equivalent to that of manual transcriptions. We have shown that not all ASR errors are equally detrim...

work page

[6] [6]

T45- 407/19N) and the CUHK Stanley Ho Big Data Decision Research Centre

Acknowledgements This work is supported by the HKSARG Research Grants Council’s Theme-based Research Grant Scheme (Project No. T45- 407/19N) and the CUHK Stanley Ho Big Data Decision Research Centre

work page

[7] [7]

World alzheimer report 2019: Attitudes to dementia, a global survey: Public health: Engaging people in adrd research,

C. Lynch, “World alzheimer report 2019: Attitudes to dementia, a global survey: Public health: Engaging people in adrd research,” Alzheimer’s & Dementia, 2020

work page 2019

[8] [8]

The montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment,

Z. S. Nasreddine, N. A. Phillips, V . B ´edirian, S. Charbonneau, V . Whitehead, I. Collin, J. L. Cummings, and H. Chertkow, “The montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment,” Journal of the American Geriatrics Society, vol. 53, no. 4, pp. 695–699, 2005

work page 2005