Differentiating upper- and lower motor neuron diseases using automated acoustic analysis.

Differentiating upper- and lower motor neuron diseases using automated acoustic analysis - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Amyotroph Lateral Scler Frontotemporal Degener . Author manuscript; available in PMC: 2026 Jun 1. Published in final edited form as: Amyotroph Lateral Scler Frontotemporal Degener. 2025 Jun 1;26(7-8):691–701. doi: 10.1080/21678421.2025.2506444 Differentiating upper- and lower motor neuron diseases using automated acoustic analysis Justin Truong 1 , Leif Simmatis 1, 2, 3 , Timothy Pommee 1, 2 , Agessandro Abrahao 2, 4 , Kerry Adams 5 , Marvin Chum 6 , Colleen O’Connell 7 , Angela Genge 8 , Sanjay Kalra 9 , Benjamin Ritsma 10 , Kerri Schellenberg 11 , Christen Shoesmith 12 , Lorne Zinman 4, 13 , Gord Jewett 14 , Yana Yunusova 1, 2, 3 Author information Article notes Copyright and License information 1 Department of Speech-Language Pathology, Temerty Faculty of Medicine, University of Toronto, Toronto, ON, Canada 2 Sunnybrook Research Institute, Toronto, ON, Canada 3 Univeristy Health Network (UHN) – KITE 4 Division of Neurology, Department of Medicine, Sunnybrook Health Sciences and Harquail Centre for Neuromodulation, University of Toronto, Toronto, ON, Canada 5 Alberta Health Services, Kaye Edmonton Clinic ALS Programme, Edmonton, AB, Canada 6 Faculty of Health Sciences, McMaster University, Hamilton, ON, Canada 7 Department of Rehabilitation, Stan Cassidy Centre for Rehabilitation, Horizon Health Network Fredericton, NB, Canada 8 Department of Neurology and Neurosurgery, Faculty of Medicine and Health Sciences, McGill University, Montreal, QC, Canada 9 Division of Neurology, Department of Medicine, University of Alberta, Edmonton, AB, Canada 10 Department of Physical Medicine and Rehabilitation, Queen’s University, Providence Care, Kingston, ON, Canada 11 Division of Neurology, Department of Medicine, University of Saskatchewan, Saskatoon, SK, Canada 12 Department of Clinical Neurological Sciences, London Health Sciences Centre, Western University, London, ON, Canada 13 Rehabilitation Sciences Institute, Temerty Faculty of Medicine, University of Toronto, Toronto, ON, Canada 14 Department of Clinical Neurosciences, Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada ✉ Correspondence: yana.yunusova@utoronto.ca Issue date 2025 Nov. PMC Copyright notice PMCID: PMC12380423  NIHMSID: NIHMS2083680  PMID: 40450589 The publisher's version of this article is available at Amyotroph Lateral Scler Frontotemporal Degener Abstract Objective: Motor neuron diseases (MNDs) result in a spectrum of motor impairments, including considerable effects on speech function, which manifest as dysarthria—a motor speech disorder. Speech metrics are increasingly recognised as critical biomarkers with potential utility in disease diagnosis and phenotyping. This study aimed to (1) characterise acoustics of upper motor neuron (UMN) and lower motor neuron (LMN) dysarthria presentations in MNDs, and (2) identify relationships between bulbar disease severity scores and acoustic features, as these could collectively enable personalised approaches to management of these diseases. Methods: Data from 16 individuals with primary lateral sclerosis (PLS) representing UMN disease, 14 individuals with spinal and bulbar muscular atrophy (SBMA) representing LMN disease, and 25 neurologically healthy individuals were analyzed. Clinical measures were also collected from PLS and SBMA groups. All participants were remotely recorded performing passage reading, rapid syllable repetition, and vowel phonation. Fifty-two acoustic features were extracted representing articulation, phonation, prosody, resonance, and overall speech timing. Features were compared using Kruskal-Wallis tests for between-group comparisons and Spearman correlations between acoustic features and clinical scores. Results: Articulatory and prosodic features best differentiated PLS, SBMA and controls. Correlations were observed in the PLS group between the clinical score and various articulatory features, most notably those indexing tongue and jaw movements. Conclusions: Our study demonstrated that acoustic assessment could capture fingerprints of dysarthrias associated with PLS and SBMA. These findings also demonstrate the potential for remote speech assessment to characterize diverse dysarthria profiles and pave the way for creating ways for personalised disease management approaches in clinical care and trials. Keywords: Acoustic analysis, dysarthria, upper motor neuron, lower motor neuron Introduction Motor neuron diseases (MNDs) constitute a group of neurodegenerative disorders that impair the upper motor neuron (UMN) and/or lower motor neuron (LMN) systems, causing spasticity, muscle paralysis, fasciculations, and altered reflexes. Bulbar muscle weakness leads to dysarthria, a motor speech disorder. Primary lateral sclerosis (PLS) – a pure UMN condition – presents with spastic dysarthria (Statland et al., 2015), whereas Spinal and Bulbar Muscular Atrophy (SBMA or Kennedy's disease) affects the LMN system and presents with flaccid dysarthria (Kennedy et al., 1968). Amyotrophic lateral sclerosis (ALS), which affects both the UMN and LMN systems, manifests as a mixed dysarthria type, combining flaccid and spastic characteristics (Grad et al., 2017). Compared to ALS, which can progress rapidly, PLS and SBMA are associated with slower disease progression and overall better prognosis, with SBMA having a normal life expectancy (Atsuta et al., 2006). However, PLS and SBMA are often initially misdiagnosed as ALS (Breza & Koutsis, 2019; Statland et al., 2015), causing emotional distress for patients and families, and issues with healthcare resource allocation. Accurate diagnosis of these conditions would enable better estimation of disease progression and survival, as well as prescription of appropriate treatments - behavioural as well as pharmaceutical - and clinical trial enrichment (e.g., identification of individuals with expected progression). Early classification of dysarthria type can play an important role in the diagnostic process. Moreover, characterization of isolated spastic and flaccid dysarthrias can facilitate a better understanding of the mixed dysarthria associated with ALS, providing insight into UMN- or LMN-predominant ALS phenotypes. Currently, clinicians use auditory-perceptual assessment to distinguish between dysarthria types. Following the Mayo Clinic framework, slowed speaking rate, imprecise articulation, strained-strangled voice, and weakness of the oro-facial muscles are associated with spasticity, and characterise spastic dysarthria. In contrast, short phrases, breathiness, and articulatory imprecision corresponding to weakness and atrophy of tongue, palatal, laryngeal and facial musculature, are associated with flaccid dysarthria ( Darley et al., 1969a , 1969b ). However, it has been well established that auditory-perceptual assessments suffer from poor reliability, particularly at early disease stages, and require highly specialised training and expertise ( Fonville et al., 2008 ; Van Der Graaff et al., 2009 ). Although recent efforts to develop standardized clinical assessments for bulbar disease have been undertaken to promote the reliability of clinician-administered assessments ( Bouvier et al., 2023 ; Ball et al., 2022 ), these scales remain time-consuming to administer and thus have limited scalability. Acoustic analysis has proven valuable for objective characterization of dysarthrias in various neurological diseases ( Hecker et al., 2022 ). For example, acoustic measures have been well-established as means of detecting and tracking bulbar disease progression in ALS ( Green et al., 2013 ; Ramig et al., 1990 ; Rowe et al., 2021 ; Stegmann et al., 2020 ; Weismer et al., 1992 ; Yunusova et al., 2013 ). Early research on dysarthria phenotyping has emerged across MNDs in languages other than English. One study provided a direct comparison of acoustics in ALS, PLS, SBMA and controls, focusing on the production of vowel-glide syllables in French ( Lévêque et al., 2022 ). Their measures of Mel-frequency cepstral coefficients (total squared change in Mel-frequency cepstral coefficients/TSC-MFCCs), captured a reduced rate of vocal tract movement and distinguished flaccid from spastic and mixed dysarthria presentations. Fougeron et al., (2022) have additionally reported differences between MND variants for features extracted from the rapid syllable repetitions produced by French speakers. More recently, an Italian study ( Milella et al., 2023 ) evaluated acoustics in patients with ALS who were subdivided into UMN- or LMN-predominant phenotypes. The assessment of isolated vowels and rapid syllable repetitions indicated that measures of vowel space area and rate of rapid syllable repetitions distinguished these ALS phenotypes. Bulbar disease severity was not considered in the these studies. In this study, we examined acoustic differences in English-speaking people with PLS or SBMA relative to matched controls, across a large acoustic feature set selected to represent all physiological speech subsystems (i.e., overall timing, phonatory, articulatory, resonatory, and prosody) ( Green et al., 2013 ). We asked whether these features could distinguish between PLS and SBMA acoustic phenotypes and hypothesised that significant differences would be found across acoustic features– particularly in articulation and phonation – as suggested by the auditory-perceptual Mayo Clinic framework ( Darley et al., 1969a , 1969b ). We additionally explored the relationships between acoustic features and clinical scores obtained with ALS-Bulbar Dysfunction Index - Remote (ALSBDI-R) ( Bouvier et al., 2023 ) within each group, hypothesizing that features that captured a difference from healthy controls would also capture relevant dimensions of clinically established severity of bulbar impairment. The comparison of established acoustic features to clinical measures of bulbar function, such as the ALSBDI-R, aimed to provide a further opportunity to anchor these measures as objective and scalable indicators of bulbar function. Methods Participants and clinical assessment We recruited patient participants diagnosed with SBMA or PLS and neurologically healthy controls. All participants were fluent speakers of English, scored within the normal range on the Montreal Cognitive Assessment (MoCA; score ≥26) ( Nasreddine et al., 2005 ), and reported normal hearing. To qualify, individuals in the PLS group had to show an absence of clinical/electrodiagnostic LMN signs and be ≥4 years post symptom onset ( Turner et al., 2020 ). Individuals in the SBMA group had clinical features of SBMA and a confirmed pathogenic repeat expansion in the androgen receptor gene. All participants provided written, informed consent in accordance with the Declaration of Helsinki. For people with PLS, overall and bulbar disease severity was established using either the ALS Functional Rating Scale – Revised (ALSFRS-R, ( Cedarbaum et al., 1999 )) or the PLS Functional Rating Scale (PLSFRS) ( Mitsumoto et al., 2020 ). The latter scores were converted to ALSFRS-R scores using the process described previously ( Mitsumoto et al., 2020 ) to ensure comparability of scores across participants. Individuals in the SBMA group were clinically assessed using the SBMA Functional Rating Scale (SBMAFRS, ( Hashizume et al., 2015 )). The ALSFRS-R consists of 12 items scored from 0-4, with a maximum score of 48 points; the bulbar subscore is based on 3 items (i.e., speech, swallowing and salivation), with maximum score of 12 points. The SBMAFRS is related to the ALSFRS-R and consists of 14 items scored from 0-4, with a maximum score of 56 points; its bulbar subscore is based on 5 items, with the maximum score of 20 points. Additionally, all patients performed the Sentence Intelligibility Test (SIT) to obtain speaking rate (words per minute, WPM) and speech intelligibility (%) scores ( Yorkston et al., 1996 ), which have been established as means of stratifing patients by bulbar disease severity ( Allison et al., 2017 ). SIT data were collected as participants read 10 sentences of increasing word length. SIT ratings (WPM and % intelligibility) were then estimated by a trained research assistant. Owing to the greater variation in disease severity for the PLS group compared to the SBMA group, SIT scores were used to stratify the PLS group into high-intelligibility (>96%; “PLS-high”) and low-intelligibility (≤96%; “PLS-low”) subgroups, to mitigate potential effects of severity differences on whole-group comparisons. Further, bulbar disease profiles were evaluated using the clinician-administered ALS Bulbar Dysfunction Index – Remote (ALSBDI-R) ( Bouvier et al., 2023 ). The test was developed to effectively track bulbar disease progression in ALS. It comprises 35 items that are grouped into an auditory-perceptual examination, cranial nerve examination, and functional speech severity score. A total ALSBDI-R score is also computed. The test was conducted by a trained speech language pathologist (SLP) immediately prior to the online acoustic data collection. For ALSBDI-R higher scores indicated a more advanced bulbar disease. Participants were considered as mildly impaired if they had ALSBDI-R total scores of 7-24 points, and moderate if they had scores of 25-50 points, and severe if they scored >50. Speech tasks, recording and post-processing procedures Participants were assessed using 3 tasks: the Bamboo passage (see Appendix ), oral diadochokinesis (DDK), and maximum phonation task (MPT). Participants read the passage at their habitual speaking rate and volume, after a silent practice reading. The DDK task involved participants repeating the syllables /pa/ or /ta/ as fast and as clearly possible on one breath, also after a practice run. MPT evaluated the ability of participants to sustain phonation of the vowel /a/ for as long as possible. Data were recorded using a custom multimodal audio/video assessment app. The app was accessible via a web browser optimised for Google Chrome, and participants used their own personal computers (i.e., laptops or desktops using Windows ≥10 or MacOS ≥10.10) to access the app. The performance of the app feature extraction modules was previously validated in healthy younger participants and in individuals with ALS (Simmatis et al., 2023a; Simmatis, et al., 2023b). Participants were first guided through setup by an experienced SLP who ensured that variation in lighting, position, etc. was minimised, and then participants navigated through the recordings independently via on-screen instructions. Audio was recorded at 48 kHz in 16-bit depth. Sample quality verification and pre-processing was performed on each recording by an experienced research assistant. Based on the assessment of the signal-to-noise ratio (SNR) and clipping ( Deliyski et al., 2005 ), a number of recordings were excluded. See Table 1 for the final number included from each task in each group. Bamboo passage data were truncated to include only the voiced segments of the reading, to enable more accurate analysis of formants and voice/signal perturbation measures. DDK samples had leading and trailing silences manually removed before analysis. MPT recordings were trimmed to a 2-second interval within the middle of the recording ( Choi et al., 2012 ; Patel et al., 2018 ). Table 1. Number of recordings analysed per task and participant group. Task PLS SBMA HC Bamboo 16 14 25 DDK 13 14 13 MPT 12 12 15 Open in a new tab DDK = diadochokinesis. HC=healthy controls. MPT=maximum phonation task. PLS=primary lateral sclerosis. SBMA=spinal and bulbar muscular atrophy. Acoustic data analysis The acoustic pipeline included 52 features linked to physiological speech subsystems (overall timing, phonatory, resonatory, articulatory, and prosodic), which have been reported in prior literature ( Carignan, 2021 ; Lee et al., 2003 ; Liss et al., 2010 ; Rowe et al., 2021 ; Vogel et al., 2009 ). Respiratory subsystem features were not extracted because we wanted to maintain a sparse feature set and deemed a combined measure of speech and pause durations (speaking rate) to be an adequate measure for this purpose ( Barnett et al., 2020 ). Table 2 lists these features, organised by subsystem and category. Computation of features used a mixture of custom scripts and standard libraries (e.g., Parselmouth ( Jadoul et al., 2018 )), and code was implemented in Python 3.8. Table 2. The description of acoustic features by task and category. Speech Sub-system # of features Feature category Definition Articulatory Bamboo 8 F1 and F2 First formant (F1) and second formant (F2); measures indexing tongue range of motion (Mefferd and Green, 2010). Features extracted : F1 mean, F1 5 th prc (percentile), F1 95 th prc, F1 5 th -95 th prc, F2 mean, F2 5 th prc, F2 95 th prc, F2 5 th -95 th prc. 8 F1 and F2 speeds First derivative of F1 and F2 timeseries indexing tongue movement rates (e.g., Rowe et al., 2021 ). Features extracted : F1 speed median, F1 speed 5 th prc, F1 speed 95 th prc, F1 speed 5 th -95 th prc, F2 speed median, F2 speed 5 th prc, F2 speed 95 th prc, F2 speed 5 th -95 th prc 1 Articulatory precision A measure of articulatory precision, which captures the working articulatory space, while circumventing manual calculation of vowel space area ( Jiao et al., 2017 ). Features extracted : articulatory entropy 2 Vocal tract configuration Total squared change in Mel-frequency cepstral coefficients (TSC MFCCs); a measure indexing overall change in rate of change of vocal tract shape. Computed as the sum of the square of the first derivative of MFCCs across successive frames ( Slis et al., 2021 ). Features extracted : TSC MFCC mean and TSC MFCC coefficient of variation (CV) 1 Articulatory Rate Global measure of speech performance, evaluated in connected speech tasks, measured as the number of syllables produced per second, excluding pauses. Diadochokinesis (DDK) 2 Duration The total time between the start of the first and the end of the last DDK syllable. 4 Rate Rate of syllable production (syl/sec) ( Rong et al., 2015 ). Features extracted : DDK rate 4 Cycle-to-cycle variability (CTV) Measure indicating the consistency of production rates in syllable repetition tasks ( Tanchip et al., 2022 ). Features extracted : DDK cycle-to-cycle variability (CTV) for all syllables, DDK CTV for first 9 syllables (9syll). Prosodic Bamboo 8 Envelope Modulation Spectra (EMS) Based on the frequency of amplitude changes in the rectified low-pass filtered audio signal, used to quantify speech rhythmicity. We extracted features with an octave band centred around 1000 Hz, which ranges from 700-1300 Hz ( Liss et al., 2010 ). Features extracted : First Fast Fourier transform (FFT) peak amplitude (FFT ampli 1), First FFT peak frequency (FFT peak 1), second FFT peak amplitude (FFT ampli 2), second FFT peak frequency (FFT peak 2), energy below 4 Hz, energy above 4 Hz, ratio of energy below 4 Hz to above 4 Hz, energy between 3-6 Hz. Phonatory Bamboo + MPT 2 Cepstral measures An acoustic feature for evaluating voice quality. CPPS that has favourable qualities over other measures in connected speech over sustained phonation tasks ( Halberstam, 2004 ). Features extracted : CPPS 2 Voice Quality Multivariate measure of voice quality, validated for use in connected speech and sustained phonation, aggregated across 6 individual features ( Batthyany et al., 2022 ). Features extracted : Acoustic Voice Quality Index (AVQI) 2 Breathiness Multivariate measure of voice quality focusing on breathiness, validated for use in connected speech and sustained phonation, consisting of 9 features aggregated ( Barsties v. Latoszek et al., 2021 , 2023 ). Features extracted : Acoustic breathiness index (ABI) Bamboo 2 Fundamental Frequency (F0) Measure indexing voice stability. Mean and robust range (5th to 95th percentile values) of F0 extracted from connected speech tasks. Features extracted : F0 mean, F0 range. MPT 1 Jitter Measure of variations in speech frequency. PPQ5 jitter is extracted in the Multi-Dimensional Voice Program – a gold standard for the acoustic analysis of dysarthria. Feature extracted : PPQ5 jitter 1 Harmonic-to-noise ratio Measures the ratio of aperiodic to periodic sound (Dejonckere & Lebacq, 1987). Feature extracted : HNR-D Resonatory Bamboo 3 1/3 Octave Spectra Analysis Objective acoustic features validated in the context of hypernasality perception, focusing on bands centred around 630, 800, and 1000 Hz. These bands demonstrate higher intensity levels in hypernasal speech ( Lee et al., 2003 ; Vogel et al., 2009 ) and have been validated in the context of SBMA ( Tanaka et al., 2014 , 2019 ). Features extracted : OB 630 Hz, OB 800 Hz, OB 1000 Hz. Overall Timing Bamboo 1 Speech Rate Global measure of speech performance, evaluated in connected speech tasks, measured as the number of words produced per minute (WPM), includes pauses. Features extracted : speech rate Open in a new tab ABI=acoustic breathiness index, AVQI=acoustic voice quality index, CPPS=smoothed cepstral peak prominence, CTV=cycle-to-cycle variability, DDK=diadochokinesis, EMS=envelope modulation spectra, F0=fundamental frequency, F1=first formant, F2=second formant, HNR-D=harmonic-noise ratio (Dejonckere formula), OB=octave band, PPQ5=5-point pitch perturbation quotient, SBMA=spinal and bulbar muscular atrophy, TSC MFCC=total squared change in Mel-frequency cepstral coefficients, WPM=words per minute. Statistical analysis Group differences were evaluated using the Kruskal-Wallis omnibus test ( Kruskal & Wallis, 1952 ), followed by Dunn’s post hoc tests for individual pairwise significant differences. For individual features that demonstrated at least an association (p<0.05) in pairwise comparisons, we analyzed the Spearman correlation (ρ) ( Spearman, 1904 ) between the feature values and ALSBDI-R total score. Across all statistical tests, multiple-comparisons correction was performed using the Benjamini-Hochberg method to control for the false discovery rate (FDR) ( Benjamini and Hochberg, 1995 ), and statistical significance was considered at p < 0.05 after correcting the p-values for FDR. Specifically, multiple-comparisons correction was applied to Kruskal-Wallis tests as well as Spearman correlation analyses. Given the small sample sizes and corresponding risk of over-interpreting the results, we did not perform statistical analyses on the PLS-high and PLS-low subgroups. Statistical tests (including the Kruskal-Wallis test and Spearman correlation) were performed in R version 3.6.1 using built-in functions and the rstatix package for Dunn’s test. Results Group comparisons Our sample comprised 25 controls as well as 14 participants diagnosed with SBMA, 16 participants diagnosed with PLS (split into high and low severity subgroups), recruited across several clinics in Canada. Table 3 describes clinical characteristics of the participants. Table 3. Summary of participant demographic and clinical information (medians & interquartile ranges). HC SBMA PLS PLS-High PLS-Low N (M/F) 25 (15/10) 14 (14/0) 16 (6/10) 7 (2/5) 9 (4/5) Age (years) 60 [11] 60.5 [17.5] 64 [9.25] 60 [7.5] 66 [12] SIT speech rate (WPM) – 166.81 [46.33] 112.21 [58.20] 117.44 [25.36] 85.94 [65.65] SIT speech intelligibility (%) – 97.27 [3.4] 93.64 [39.77] 97.27 [1.37] 70 [72.73] ALSFRS-R/ SBMA-FRS, total scores – 37.5/56 [8] 36.5/48 [12.5] 37/48 [3.5] 29/48 [17] ALSFRS-R/ SBMA-FRS, bulbar scores – 13.5/20 [4.5] 9/12 [2.25] 9/12 [1] 7/12 [4] ALSBDI-R total score - 8.0 [7.0] 23.0 [39.0] 11.0 [9.0] 47.0 [21.0] Open in a new tab ALSBDI-R=Amyotrophic Lateral Sclerosis Bulbar Dysfunction Index – Remote. ALSFRS-R=Amyotrophic Lateral Sclerosis Functional Rating Scale – Revised. HC=healthy controls. PLS=primary lateral sclerosis. SBMA=spinal and bulbar muscular atrophy. SBMAFRS=SBMA functional rating scale. SIT=sentence intelligibility test. WPM=words per minute. PLS-High refers to PLS patients with >96% intelligibility, PLS-Low refers to PLS patients with ≤96% intelligibility. Note that for ALSFRS-R scores, lower indicates a greater degree of impairment, whereas a higher score indicates a greater degree of impairment on the ALSBDI-R. The omnibus tests of each acoustic feature listed in the far-left column of Figure 1 revealed several between-groups differences. Significant group differences were noted in 18/30 articulatory subsystem features (8/20 from Bamboo, 5/10 from DDK), 2/10 phonatory subsystem features (2/5 from Bamboo, 0/5 from MPT), and 3/8 prosodic subsystem features (from Bamboo). No significant differences were observed for the overall timing or resonatory features. Figure 1. Open in a new tab A heatmap of p -values of the differences of each acoustic feature across groups, with Kruskal-Wallis omnibus and Dunn’s pairwise comparisons shown. Only statistically significant ( p <0.05) comparisons are displayed, with lower p -values indicated by darkening shades of red in each cell. Except where indicated (DDK and MPT), features were extracted from the Bamboo passage task. 9syll: first 9 syllables; abi: acoustic breathiness index; ampli: amplitude; amr: alternating motion rate; artic: articulatory; avqi: acoustic voice quality index; cpps: smoothed cepstral peak prominence; ctv: cycle-to-cycle variability; cv: coefficient of variation; ddk: diadochokinesis; f0: fundamental frequency; f1: first formant; f2: second formant; fft: fast Fourier transform; ent: entropy; hnr-d: harmonic-noise ratio Dejonckere; numsyll: number of syllables; prc: percentile; ob: octave band. Figure 1 (columns 2-4) also shows the results of the pairwise between-groups analyses. When compared to HC, the SBMA group primarily showed differences in prosodic features (3/8 features), whereas the PLS group primarily revealed differences in articulatory features (14/30 features). Phonatory subsystem differences were also identified for HC-PLS (1/10 features) but not for HC-SBMA (0/10 features). A comparison between SBMA and PLS revealed differences in N=11 features across all subsystems. Of these, differences were noted in 8/30 articulatory features, 2/10 phonatory features, and 2/8 prosodic features. Acoustic versus auditory-perceptual correlations and severity analyses After FDR correction, in the PLS group there were 5/15 (15 referring to the number of significant HC-PLS differences) significant correlations between acoustic features and the ALSBDI-R total score. The strongest individual correlations were with 95 th percentile of F1 speed, median F2 speed, and 5 th percentile of F2 speed (all ρ = −0.78). In the SBMA group, 0/3 (3 referring to the number of significant HC-SBMA differences) features had signifcant correlations. To further illustrate the effect of bulbar disease severity, we plotted the selected acoustic features for PLS-high and PLS-low subgroups against the same features for controls and the SBMA participants (see Figure 2 ). The plotted features were selected based on their importance in group difference analyses. Figure 2. Open in a new tab Boxplots of individual articulatory features for each group with individual data points shown with an overlying strip plot. PLS data stratified by intelligibility are also shown to the right of the vertical line of each plot. Axes differ in scale for each feature. Whiskers extend to the farthest point within ±1.5 interquartile range past the 25th and 75th percentiles of the box. Artic: articulatory; DDK: diadochokinesis; F2: second formant. Individual case analyses Since speech data in PLS/SBMA are rare, we opted to showcase individual analyses to demonstrate the potential of acoustic fingerprinting associated with these conditions. Figure 3 shows all acoustic features expressed relative to their normative values, as computed from the controls, for two individuals - 1 PLS and 1 SBMA - with mild bulbar disease as established on ALSBDI-R (Total score 10-11) and two individuals with the moderate bulbar disease (ALSBDI-R, Total score 25-28). Figure 3. Open in a new tab Dysarthria fingerprints for 4 different individuals from the present study. Panels A and B: Summaries of 2 individuals with ALSBDI scores of 10–11 (mild disease severity). Panels C and D: Summaries of 2 individuals with ALSBDI scores of 25–28 (moderate disease severity). Grey bands represent the region of ±1.96 standard deviations from the mean of control participants. Blue line markers indicate feature values within the normative region; whereas red markers indicate values outside the normative region. Values are truncated at ±10 standard deviations from the mean for improved visualization. The gap in Panel C is due to missing data. Note that resonatory subsystem values could not be computed for the moderate PLS participant and so they are omitted from all plots for consistency (refer to the online version for colour figure). The two individuals with mild disease presented with articulatory-predominant impairments. The individual with SBMA (panel A) had an elevated F2 range and F2 95 th percentile with respect to controls, whereas the individual with PLS (panel B) presented with elevated F2 features as well as a combination of reduced DDK rate and increased DDK CTV (for both /pa/ and /ta/). In the moderate disease pair, the individual with SBMA (panel C) had an impairment in DDK CTV, whereas the individual with PLS (panel D) had a combination of articulatory and prosodic impairments. Discussion In this study, we demonstrated that acoustic analysis was able to capture distinct characteristics of spastic and flaccid dysarthria associated with PLS and SBMA, respectively, using established acoustic features representing physiological speech subsystems that are typically affected in MNDs. We identified that spastic dysarthria in our sample of patients with PLS was most often associated with articulatory subsystem impairments, whereas flaccid dysarthria in SBMA was most often associated with prosodic subsystem impairments. Furthermore, we showed that salient acoustic features (i.e., that distinguished people with disease and controls) correlated with the clinician-administered ALSBDI-R, indicating their association with bulbar disease severity. Collectively, these result support the utility of acoustic analysis to characterise distinct dysarthria phenotypes and pave the way towards biomarkers of UMN and LMN impairments. We hypothesised that significant differences would be found across acoustic features – particularly in articulation and phonation, as suggested by the auditory-perceptual Mayo Clinic framework. We found strong performance of articulatory features in distinguishing the dysarthria types and their comparisons with controls. In contrast to our expectations based on the Mayo clinic framework, we did not observe notable deviations in phonatory features. Previous studies in individuals with UMN disease have also failed to identify some of the expected key findings of spastic dysarthria such as voice quality impairments, albeit in individuals with unilateral UMN injury typical of stroke ( Illner et al., 2023 ; Wang et al., 2009 ). The failure to identify phonatory changes aside from F0 mean and range may be due to the heterogeneity of our sample (particularly PLS) or due to the choice of features. For example, Kent et al. (2003) demonstrated that the phonatory impairments in UMN injury due to stroke arose in shimmer features (as well as smoothed pitch perturbation quotient [PPQ], which differed from the PPQ5 jitter that we employed). However, we elected to not extract shimmer features, nor a wide range of jitter features, given that shimmer in particular (and voice quality measures more broadly) are sensitive to microphone type (e.g. Fahed et al., 2022). SBMA showed differences in a prosodic feature representing speech rhythmicity instead ( Liss et al., 2010 ), which is important considering the very mild nature of our SBMA sample. It is important to note that existing literature on pure LMN and (non-stroke) UMN dysarthria is sparse; these results will need to be confirmed with larger and/or different groups of individuals. As in previous works, we found that patients with PLS presented with a notable articulatory impairment, detected using DDK-derived features ( Lévêque et al., 2022 ). The most significant articulatory features obtained from both DDK and passage tasks were assoccaited with formant speeds as compared to overall formant extents. This indicates that the dynamics of speech were the most affected, particularly in PLS. This also corresponds to previous observations of reduced articulator movement rates in MNDs ( Fougeron et al., 2022 ). But in contrast with previous work ( Lévêque et al., 2022 ), we did not find robust differences in TSC MFCCs between PLS and HC in DDKs, despite similar disease severities between our participant groups (ALSFRS-R bulbar score median ~9/12). Minimal bases for comparisons exist to our SBMA findings. Only two previous studies to our knowledge investigated voice changes in SBMA acoustically. Although Fougeron et al., 2022 identified moderate acoustic differences in jitter/shimmer and variance in F0 between individuals with SBMA and individuals with ALS, their individual analysis of deviance scores did not uniformly point to a voice quality impairment in their SBMA group. Tanaka et al., 2014 performed an analysis of SBMA patients with and without laryngospasm and note that harmonic-to-noise ratio (HNR) was the most discriminative feature between these two groups, however without reference to a control group it is challenging to generalise these findings to the present study. We additionally hypothesised that features that captured a difference from healthy controls would also capture relevant dimensions of clinically established severity of bulbar impairment. This was established in the present work both via correlations with clinical perceptual scores, as well as through PLS subgroup analysis and through individual investigation of cases stratified by ALSBDI-R severity. We observed stronger correlations in the PLS as compared to the SBMA group. This was not surprising considering the greater spread of bulbar disease severity in the former. The lower severity gradient in the SBMA group is notable, in that the most severe individual in our SBMA group had a moderate perceptual and mild acoustic impairment. Potentially, leveraging multimodal information (i.e., adding computer vision based kinematic analysis) may be able to further capture disease-related speech changes in this group, and identify further differentiators between dysarthria presentations. Our study had some important limitations. Primarily, the sample size in the present work was small, owing to the rarity of PLS and SBMA cases, even relative to ALS. Nevertheless, through careful literature-driven selection of acoustic features and cautious statistical analysis and interpretation, the present conclusions are supported by the exploratory nature of our analysis with the caveat that future, larger studies will be required for validation. The severity range of participants in the SBMA and PLS groups were also notably different, presenting a limitation. The SBMA group had relatively homogeneous disease severity, making inference of the progression of LMN-specific speech symptoms challenging. However, we were still able to identify acoustic signs that were specific to SBMA, beyond severity differences, supporting the acoustic phenotyping in these two groups. It is additionally important to acknowledge for future work that the choice of clinical scales is important as it was demonstrated in ALS that there may be a mismatch in perception of impairment early in bulbar disease by patients and assessment of early bulbar disease by experienced examiners ( Pawlukowska et al., 2019 ). Conclusion This study is one of the first efforts to demonstrate the potential of automatic acoustic analysis to establish signatures of UMN and LMN impairment in MNDs in English. Our results support the potential for novel acoustic approaches to provide critical clinical stratification, and the use of a novel remote data collection platform supports the utilization of such technologies in MND research. It will ultimately serve to provide tailored clinical speech management support to a wide diversity of patients and may become invaluable for therapeutic drug trials given further validation. Speech acoustics offer a non-invasive, affordable, and easily implemented means to assess and rapidly triage suspected ALS patients towards relevant clinical resources based on these aetiology-specific signs. Funding details. This work was supported by an ALS Society of Canada Discovery Grant, NIH R01DC013547, and The ALS Society of Alberta Jennifer Prest Bulbar ALS Fund. Biographies Justin Truong: JT completed his Honours Bachelor of Science degree in Neuroscience and Physiology and is currently a student in the Master of Science in Physical Therapy program at the University of Toronto. He was previously a research assistant in the Speech Production Lab at the University of Toronto. Leif Simmatis: LERS is a postdoctoral researcher in the Speech Production Lab at the University of Toronto, where his work focuses on the development of novel speech assessment technologies. He has a PhD in neuroscience from Queen’s University, where he specialised in systems neuroscience and human motor control. He has been funded by the Data Sciences Institute at the University of Toronto, Mitacs, and was recognised in 2024 as a promising New Investigator by the International Society of CNS Clinical Trials and Methodology. Timothy Pommée: TP holds a Master’s degree in Speech and Language Pathology with a specialization in Voice Therapy from the University of Liège (Belgium) and a PhD in Computer Sciences and Telecommunications from the Institut de Recherche en Informatique de Toulouse (France). After a postdoctoral fellowship at Sunnybrook Research Institute in Toronto (2022–2024), Timothy joined the University of Montreal as a research consultant. Agessandro Abrahao: AA’s research interests include clinical trials and epidemiological studies of ALS and other neurodegenerative diseases. He is also interested in the development of biomarkers for motor neuron function using transcranial magnetic stimulation and motor unit count techniques. Kerry Adams: KA holds a Master’s degree in Speech Language Pathology and has worked over 25 years with adult neurologic conditions. Since 2006, she has worked exclusively as the primarly SLP at the Kaye Edmonton Clinic ALS Programme. Marvin Chum: Dr. Marvin Chum holds a Master's of Science Degree in Clinical Research Design and Statistical Analysis. He is currently an associate professor at McMaster Unviersity in the Division of Neurology, where he has an active clinical practice providing care for patients with neuromuscular disorders. He is also the head of service for neurology at St. Joseph's Healthcare Hamilton, and the acting clinical lead for its neurodiagnostic service. Colleen O’Connell: CO is a professor of Physical Medicine & Rehabilitation and a member of the Canadian ALS Research Network. She is the medical director and department head of rehabilitation at the Stan Cassidy Centre for Rehabilitation, where she leads the New Brunswick ALS Clinic and research program. Angela Genge: AG completed her medical degree at the Memorial University of Newfoundland, joined the staff of the Montreal Neurological Hospital in 1994, and became Director of the ALS Clinic in 1998. Her interests in neurology focus on neuromuscular disease, and she works with the CRU and brings in clinical trials in neuromuscular disorders such as ALS, Myopathies, Neuropathies, Myasthenia Gravis and pain. Sanjay Kalra: SK is a professor and neurologist at the University of Alberta. His research focuses on multimodal biomarker development in ALS. He is the founding director of CALSNIC and CAPTURE ALS, both multicentre platforms for clinical and translational research in ALS/MND. Benjamin Ritsma: BR is an Assistant Professor at Queen’s University (Kingston, Ontario, Canada) in the Department of Physical Medicine and Rehabilitation, sub-specializing in neurological rehabilitation. He is the Clinical Director for Rehabiltation as well as the Co-Medical Director of the Adult Neuromuscular/ALS Clinic, Co-Founder/Co-Director of the Kingston Peripheral Nerve Clinic, and Medical Director of Stroke Rehabilitation. Kerri Schellenberg: KS is a Neuromuscular Neurologist and Associate Professor at the University of Saskatchewan. She is the Medical Director of the ALS/MND clinic and the Medical Director of the Royal University Hospital EMG lab She is the Co-Chair of the Canadian ALS Research Network. Christen Shoesmith: CS is Associate Professor, Neurology at Western University in London, Ontario. She is a neuromuscular neurologist and a Director of Neuromuscular Fellowships and Motor Neuron Diseases Clinic, London Health Sciences Centre. Lorne Zinman: LZ is a staff neurologist and director of the Sunnybrook ALS Clinic. His research aims to better characterise the clinical, electrophysiological and pathological phenotypes and variants of the disease to find a treatment. He also leads and participates in multicentre clinical trials evaluating the most promising ALS therapeutics. Gord Jewett: GJ is an assistant professor and neuromuscular neurologist at the Universtiy of Calgary. Dr. Jewett’s research is focused on digital and remotely assessable outcome measures as biomarkers in ALS and other neuromuscular diseases using advanced data science techniques including machine learning. Yana Yunusova: YY is a professor in the Department of Speech-Language Pathology and an Associate Director of the Rehabilitation Sciences Institute at the University of Toronto. She leads a team conducting research on the development of novel technologies for management of bulbar ALS. Appendix The Bamboo Passage Bamboo walls are getting to be very popular. They are strong, easy to use, and good-looking. They provide a good background and can create a look of a Japanese garden. Bamboo is a grass, and is one of the most rapidly growing grasses in the world. Many varieties of bamboo are grown in Asia, although it is also grown in America. Last year we bought a new home and have been working on the flower garden. In a few more days, we will be done with the bamboo wall in our garden. We have really enjoyed the project. Footnotes Disclosure statement. The authors report there are no competing interests to declare. Data availability statement. Data are available from the authors upon request. References Allison KM, Yunusova Y, Campbell TF, Wang J, Berry JD, & Green JR (2017). The diagnostic utility of patient-report and speech-language pathologists’ ratings for detecting the early onset of bulbar symptoms due to ALS. Amyotrophic Lateral Sclerosis and Frontotemporal Degeneration, 18(5–6), 358–366. 10.1080/21678421.2017.1303515 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Ball LJ, Geske JA, Burton E, Pattee GL. A clinical bulbar assessment scale (CBAS) for amyotrophic lateral sclerosis. Muscle Nerve. 2022. Dec;66(6):694–701. doi: 10.1002/mus.27738. Epub 2022 Oct 26. [ DOI ] [ PubMed ] [ Google Scholar ] Barnett C, Green JR, Marzouqah R, Stipancic KL, Berry JD, Korngut L, Genge A, Shoesmith C, Briemberg H, Abrahao A, Kalra S, Zinman L, Yunusova Y. Reliability and validity of speech & pause measures during passage reading in ALS. Amyotroph Lateral Scler Frontotemporal Degener. 2020. Feb;21(1-2):42–50. doi: 10.1080/21678421.2019.1697888. Epub 2019 Dec 6. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Barsties v. Latoszek B, Englert M, Lucero JC, & Behlau M (2023). The Performance of the Acoustic Voice Quality Index and Acoustic Breathiness Index in Synthesized Voices. Journal of Voice, 37(5), 804.e21-804.e28. 10.1016/j.jvoice.2021.05.005 [ DOI ] [ PubMed ] [ Google Scholar ] Barsties v. Latoszek B, Kim GH, Delgado Hernández J, Hosokawa K, Englert M, Neumann K, & Hetjens S (2021). The validity of the Acoustic Breathiness Index in the evaluation of breathy voice quality: A Meta-Analysis. Clinical Otolaryngology, 46(1), 31–40. 10.1111/coa.13629 [ DOI ] [ PubMed ] [ Google Scholar ] Batthyany C, Latoszek BBV, & Maryn Y (2022). Meta-Analysis on the Validity of the Acoustic Voice Quality Index. Journal of Voice, 38(6), 1527.e1–1527.e19. 10.1016/j.jvoice.2022.04.022 [ DOI ] [ PubMed ] [ Google Scholar ] Benjamini Y and Hochberg Y (1995), Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B (Methodological), 57: 289–300. 10.1111/j.2517-6161.1995.tb02031.x [ DOI ] [ Google Scholar ] Bouvier L, Green JR, Tapia CB, Tilton-Bolowsky V, Maffei MF, Fless Z, Seaver K, Huynh A, Gutz SE, Martino R, Abrahao A, Berry J, Zinman L, Yunusova Y. Amyotrophic Lateral Sclerosis-Bulbar Dysfunction Index-Remote: Test-Retest and Interrater Reliability of Candidate Items. Am J Speech Lang Pathol. 2023. Aug 17;32(4S):1884–1900. doi: 10.1044/2023_AJSLP-22-00177. Epub 2023 Jul 26. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Carignan C. (2021). A practical method of estimating the time-varying degree of vowel nasalization from acoustic features. The Journal of the Acoustical Society of America, 149(2), 911–922. 10.1121/10.0002925 [ DOI ] [ PubMed ] [ Google Scholar ] Cedarbaum JM, Stambler N, Malta E, Fuller C, Hilt D, Thurmond B, & Nakanishi A (1999). The ALSFRS-R: A revised ALS functional rating scale that incorporates assessments of respiratory function. Journal of the Neurological Sciences, 169(1–2), 13–21. 10.1016/S0022-510X(99)00210-5 [ DOI ] [ PubMed ] [ Google Scholar ] Choi SH, Lee J, Sprecher AJ, & Jiang JJ (2012). The effect of segment selection on acoustic analysis. Journal of Voice, 26(1), 1–7. 10.1016/j.jvoice.2010.10.009 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Darley FL, Aronson AE, & Brown JR (1969a). Clusters of Deviant Speech Dimensions in The Dysarthrias. Journal of Speech and Hearing Research, 12(3), 462–496. [ DOI ] [ PubMed ] [ Google Scholar ] Darley FL, Aronson AE, & Brown JR (1969b). Differential diagnostic patterns of dysarthria. Journal of Speech and Hearing Research, 12(2), 246–269. 10.1044/jshr.1202.246 [ DOI ] [ PubMed ] [ Google Scholar ] Deliyski DD, Shaw HS, & Evans MK (2005). Adverse effects of environmental noise on acoustic voice quality measurements. Journal of Voice, 19(1), 15–28. 10.1016/j.jvoice.2004.07.003 [ DOI ] [ PubMed ] [ Google Scholar ] Fonville S, Van Der Worp HB, Maat P, Aldenhoven M, Algra A, & Van Gijn J (2008). Accuracy and inter-observer variation in the classification of dysarthria from speech recordings. Journal of Neurology, 255(10), 1545–1548. 10.1007/s00415-008-0978-4 [ DOI ] [ PubMed ] [ Google Scholar ] Fougeron C, Kodrasi I, & Laganaro M (2022). Differentiation of Motor Speech Disorders through the Seven Deviance Scores from MonPaGe-2.0.s. Brain Sciences, 12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Green JR, Yunusova Y, Kuruvilla MS, Wang J, Pattee GL, Synhorst L, Zinman L, & Berry JD (2013). Bulbar and speech motor assessment in ALS: Challenges and future directions. In Amyotrophic Lateral Sclerosis and Frontotemporal Degeneration (Vol. 14, Issues 7–8, pp. 494–500). 10.3109/21678421.2013.817585 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Halberstam B. (2004). Acoustic and Perceptual Parameters Relating to Connected Speech Are More Reliable Measures of Hoarseness than Parameters Relating to Sustained Vowels. 07055, 70–73. 10.1159/000077798 [ DOI ] [ PubMed ] [ Google Scholar ] Hashizume A, Katsuno M, Suzuki K, Banno H, Suga N, Mano T, Araki A, Hijikata Y, Grunseich C, Kokkinis A, Hirakawa A, Watanabe H, Yamamoto M, Fischbeck KH, & Sobue G (2015). A functional scale for spinal and bulbar muscular atrophy: Cross-sectional and longitudinal study. Neuromuscular Disorders, 25(7), 554–562. 10.1016/j.nmd.2015.03.008 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Hecker P, Steckhan N, Eyben F, Schuller BW, Arnrich B, & Hecker P (2022). Voice Analysis for Neurological Disorder Recognition – A Systematic Review and Perspective on Emerging Trends. 4(July), 1–16. 10.3389/fdgth.2022.842301 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Illner V, Tykalova T, Skrabal D, Klempir J, & Rusz J (2023). Automated Vowel Articulation Analysis in Connected Speech Among Progressive Neurological Diseases , Dysarthria Types , and Dysarthria Severities. 66(August), 2600–2621. [ DOI ] [ PubMed ] [ Google Scholar ] Jadoul Y, Thompson B, & de Boer B (2018). Introducing Parselmouth: A Python interface to Praat. Journal of Phonetics, 71(2018), 1–15. 10.1016/j.wocn.2018.07.001 [ DOI ] [ Google Scholar ] Jiao Y, Berisha V, Liss J, Hsu SC, Levy E, & McAuliffe M (2017). Articulation entropy: An unsupervised measure of articulatory precision. IEEE Signal Processing Letters, 24(4), 485–489. 10.1109/LSP.2016.2633871 [ DOI ] [ Google Scholar ] Kruskal WH, & Wallis WA (1952). Use of ranks in one-criterion variance analysis. Journal of the American Statistical Association, 47(260), 583–621. [ Google Scholar ] Lee ASY, Ciocca V, & Whitehill TL (2003). Acoustic correlates of hypernasality. Clinical Linguistics and Phonetics, 17(4–5), 259–264. 10.1080/0269920031000080091 [ DOI ] [ PubMed ] [ Google Scholar ] Lévêque N, Slis A, Lancia L, Bruneteau G, & Fougeron C (2022). Acoustic Change Over Time in Spastic and/or Flaccid Dysarthria in Motor Neuron Diseases. Journal of Speech, Language, and Hearing Research, 65(5), 1767–1783. 10.1044/2022_JSLHR-21-00434 [ DOI ] [ PubMed ] [ Google Scholar ] Liss JM, LeGendre S, & Lotto AJ (2010). Discriminating Dysarthria Type From Envelope Modulation Spectra. Journal of Speech, Language, and Hearing Research, 53(5), 1246–1255. 10.1044/1092-4388(2010/09-0121).Discriminating [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Milella G, Sciancalepore D, Cavallaro G, Piccirilli G, Nanni AG, Fraddosio A, Errico ED, Paolicelli D, Fiorella ML, & Simone IL (2023). Acoustic Voice Analysis as a Useful Tool to Discriminate Different ALS Phenotypes. Biomedicines, 11. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Mitsumoto H, Chiuzan C, Gilmore M, Zhang Y, Simmons Z, Paganoni S, Kisanuki YY, Zinman L, Jawdat O, Sorenson E, Floeter MK, Pioro EP, Fernandes Filho JAM, Heitzman D, Fournier CN, Oskarsson B, Heiman-Patterson T, Maragakis N, Joyce N, … McHale B (2020). Primary lateral sclerosis (PLS) functional rating scale: PLS-specific clinimetric scale. Muscle and Nerve, 61(2), 163–172. 10.1002/mus.26765 [ DOI ] [ PubMed ] [ Google Scholar ] Nasreddine Z, Phillips N, Bédirian V, Charbonneau S, Whitehead V, Colllin I, Cummings J, & Chertkow H (2005). The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. Journal of the American Geriatrics Society, 53, 695–699. 10.1111/j.1532-5415.2005.53221.x [ DOI ] [ PubMed ] [ Google Scholar ] Patel RR, Awan SN, Barkmeier-Kraemer J, Courey M, Deliyski D, Eadie T, Paul D, Švec JG, & Hillman R (2018). Recommended protocols for instrumental assessment of voice: American speech-language-hearing association expert panel to develop a protocol for instrumental assessment of vocal function. American Journal of Speech-Language Pathology, 27(3), 887–905. 10.1044/2018_AJSLP-17-0009 [ DOI ] [ PubMed ] [ Google Scholar ] Pawlukowska W, Baumert B, Gołab-Janowska M, Meller A, MacHowska-Sempruch K, Wełnicka A, Paczkowska E, Rotter I, MacHaliński B, & Nowacki P (2019). Comparative assessment and monitoring of deterioration of articulatory organs using subjective and objective tools among patients with amyotrophic lateral sclerosis. BMC Neurology, 19(1), 1–8. 10.1186/s12883-019-1484-2 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Ramig LO, Scherer RC, Klasner ER, Titze IR, & Horri Y (1990). Acoustic analysis of voice in amyotrophic lateral sclerosis: A longitudinal case study. Journal of Speech and Hearing Disorders, 55(1), 2–14. 10.1044/jshd.5501.02 [ DOI ] [ PubMed ] [ Google Scholar ] Rong P, Yunusova Y, Wang J, & Green JR (2015). Predicting early bulbar decline in amyotrophic lateral sclerosis: A speech subsystem approach. Behavioural Neurology, 2015. 10.1155/2015/183027 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Rowe HP, Stipancic KL, Lammert AC, & Green JR (2021). Validation of an acoustic-based framework of speech motor control: Assessing criterion and construct validity using kinematic and perceptual measures. Journal of Speech, Language, and Hearing Research, 64(12), 4736–4753. 10.1044/2021_JSLHR-21-00201 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Simmatis L, Naeini A, & Yue K (2023). Analytical Validation of a Webcam-Based Assessment of Speech Kinematics : Digital Biomarker Evaluation following the V3 Framework. 7–17. 10.1159/000529685 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Simmatis L, Pommeé T, & Yunusova Y (2023). Multimodal Assessment of Bulbar Amyotrophic Lateral Sclerosis (ALS) Using a Novel Remote Speech Assessment App. August, 2378–2382. 10.21437/interspeech.2023-2093 [ DOI ] [ Google Scholar ] Slis A, Lévêque N, Fougeron C, Pernon M, Assal F, & Lancia L (2021). Analysing spectral changes over time to identify articulatory impairments in dysarthria. The Journal of the Acoustical Society of America, 149(2), 758–769. 10.1121/10.0003332 [ DOI ] [ PubMed ] [ Google Scholar ] Spearman C. (1904). The Proof and Measurement of Association between Two Things. The American Journal of Psychology, 15(1), 72–101. [ PubMed ] [ Google Scholar ] Stegmann GM, Hahn S, Liss J, Shefner J, Rutkove S, Shelton K, Duncan CJ, & Berisha V (2020). Early detection and tracking of bulbar changes in ALS via frequent and remote speech analysis. Npj Digital Medicine, 3(1), 1–5. 10.1038/s41746-020-00335-x [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Tanaka S, Banno H, Katsuno M, Suzuki K, Suga N, Hashizume A, Mano T, Araki A, Watanabe H, Adachi H, Tatsumi H, Yamamoto M, & Sobue G (2014). Distinct acoustic features in spinal and bulbar muscular atrophy patients with laryngospasm. Journal of the Neurological Sciences, 337(1–2), 193–200. 10.1016/j.jns.2013.12.010 [ DOI ] [ PubMed ] [ Google Scholar ] Tanaka S, Hashizume A, Hijikata Y, Yamada S, Ito D, Nakayama A, Kurita K, Yogo H, Banno H, Suzuki K, Yamamoto M, Sobue G, & Katsuno M (2019). Nasometric Scores in spinal and bulbar muscular atrophy - Effects of palatal lift prosthesis on dysarthria and dysphagia. Journal of the Neurological Sciences, 407(September), 116503. 10.1016/j.jns.2019.116503 [ DOI ] [ PubMed ] [ Google Scholar ] Tanchip C, Guarin DL, McKinlay S, Barnett C, Kalra S, Genge A, Korngut L, Green JR, Berry J, Zinman L, Yadollahi A, Abrahao A, & Yunusova Y (2022). Validating Automatic Diadochokinesis Analysis Methods Across Dysarthria Severity and Syllable Task in Amyotrophic Lateral Sclerosis. Journal of Speech, Language, and Hearing Research, 65(3), 940–953. 10.1044/2021_JSLHR-21-00503 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Turner MR, Barohn RJ, Corcia P, Fink JK, Harms MB, Kiernan MC, Ravits J, Silani V, Simmons Z, Statland J, Van Den Berg LH, & Mitsumoto H (2020). Primary lateral sclerosis: Consensus diagnostic criteria. Journal of Neurology, Neurosurgery and Psychiatry, 91(4), 373–377. 10.1136/jnnp-2019-322541 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Van Der Graaff M, Kuiper T, Zwinderman A, Van De Warrenburg B, Poels P, Offeringa A, Van Der Kooi A, Speelman H, & De Visser M (2009). Clinical identification of dysarthria types among neurologists, residents in neurology and speech therapists. European Neurology, 61(5), 295–300. 10.1159/000206855 [ DOI ] [ PubMed ] [ Google Scholar ] Vogel AP, Ibrahim HM, Reilly S, & Kilpatrick N (2009). A comparative study of two acoustic measures of hypernasality. Journal of Speech, Language, and Hearing Research, 52(6), 1640–1651. 10.1044/1092-4388(2009/08-0161) [ DOI ] [ PubMed ] [ Google Scholar ] Wang YT, Kent RD, Kent JF, Duffy JR, & Thomas JE (2009). Acoustic analysis of voice in dysarthria following stroke. Clinical Linguistics and Phonetics, 23(5), 335–347. 10.1080/02699200802688604 [ DOI ] [ PubMed ] [ Google Scholar ] Weismer G, Martin R, Kent RD, & Kent JF (1992). Formant trajectory characteristics of males with amyotrophic lateral sclerosis. Journal of the Acoustical Society of America, 91(2), 1085–1098. 10.1121/1.402635 [ DOI ] [ PubMed ] [ Google Scholar ] Yorkston KM, Beukelman D, Hakel M, & Dorsey M (1996). Speech intelligibility test. In Lincoln NE: Institute for Rehabilitation Science and Engineering at Madonna Rehabilitation Hospital. [ Google Scholar ] Yunusova Y, Green JR, Lindstrom MJ, Pattee GL, & Zinman L (2013). Speech in ALS: Longitudinal Changes in Lips and Jaw Movements and Vowel Acoustics HHS Public Access. In J Med Speech Lang Pathol (Vol. 21, Issue 1). [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Data Availability Statement Data are available from the authors upon request. ACTIONS View on publisher site PDF (1.3 MB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Using Neurogram Similarity Index Measure (NSIM) to Model Hearing Loss and Cochlear Neural Degeneration#1142542
• Real-time shear wave elastography and high-frequency ultrasound in the assessment of treatment outcomes for triceps surae muscle injury.#162033
• Comparison of BMD Measurement by DEXA to BeamMed Speed-of-Sound Measurement at Forearm in Patients With Gaucher Disease#1031384
• Development of an artificial intelligence powered software for automated analysis of skeletal muscle ultrasonography#1098699
• Comparison of 3 Methods to Diagnose Endolymphatic Hydrops in Meniere's Disease#827380