Psychometric Properties of Rapid Word-Based Rate Measures in the Assessment of Bulbar Amyotrophic Lateral Sclerosis: Comparisons With Syllable-Based Rate Tasks

Abstract

Purpose Rapid maximum performance repetition tasks have increasingly demonstrated their utility as clinimetric markers supporting diagnosis and monitoring of bulbar disease in amyotrophic lateral sclerosis (ALS). A recently developed protocol uses novel real-word repetitions instead of traditional nonword/syllable sequences in hopes of improving sensitivity to motor speech impairments by adding a phonological target constraint that would activate a greater expanse of the motor speech neuroanatomy. This study established the psychometric properties of this novel clinimetric protocol in its assessment of bulbar ALS and compared performance to traditional syllable sequence dysdiadochokinetic (DDK) tasks. Specific objectives were to (a) compare rates between controls and speakers with symptomatic versus presymptomatic bulbar disease, (b) characterize their discriminatory ability in detecting presymptomatic bulbar disease compared to healthy speech, (c) determine their articulatory movement underpinnings, and (d) establish within-individual longitudinal changes. Method DDK and novel tongue ("ticker"-TAR) and labial ("pepper"-LAR) articulatory rates were compared between n = 18 speakers with presymptomatic bulbar disease, n = 10 speakers with symptomatic bulbar disease, and n = 13 healthy controls. Bulbar disease groups were determined by a previously validated speaking rate cutoff. Discriminatory ability was determined using receiver operating characteristic analysis. Within-individual change over time was characterized in a subset of 16 participants with available longitudinal data using linear mixed-effects models. Real-time articulatory movements of the tongue front, tongue dorsum, jaw, and lips were captured using 3-D electromagnetic articulography; effects of movement displacement and speed on clinimetric rates were determined using stepwise linear regressions. Results All clinimetric rates (traditional DDK tasks and novel tasks) were reduced in speakers with symptomatic bulbar disease; only TAR was reduced in speakers with presymptomatic bulbar disease and was able to detect this group with an excellent discrimination ability (area under the curve = 0.83). Kinematic analyses revealed associations with expected articulators, greater motor complexity, and differential articulatory patterns for the novel real-word repetitions than their DDK counterparts. Only LAR significantly declined longitudinally over the disease course. Conclusion Novel real-word clinimetric rate tasks evaluating tongue and labial articulatory dysfunction are valid and effective markers for early detection and tracking of bulbar disease in ALS.

Figure 1.

Kinematic signals and parsing landmarks for the early 10 repetitions of labial articulation rate (LAR). The kinematic signal used for parsing LAR, DDK-pataka (dysdiadochokinetic task of /pataka/ syllable repetition), and DDK-ba (dysdiadochokinetic task of /ba/ syllable repetition) was the upper lip and lower lip aperture (UL–LL, shown in bold). The kinematic signal used for parsing DDK-ta (dysdiadochokinetic task of /ta/ syllable repetition) and tongue articulation rate was the tongue tip distance trace (T1d). The first cycle of each sequence was omitted from analyses.

Figure 2.

Clinimetric rate variables for the different groups presented as boxplots, indicating the median and quartiles with whiskers reaching up to 1.5 times the interquartile range. The violin plot outlines illustrate kernel probability density, that is, the width of the shaded area represents the proportion of the data located there. Significant pairwise comparisons with Tukey's adjusted p values are indicated with square brackets TAR = tongue articulation rate; LAR = labial articulation rate; DDK-pataka = dysdiadochokinetic task of /pataka/ syllable repetition; DDK-ta = dysdiadochokinetic task of /ta/ syllable repetition; DDK-ba = dysdiadochokinetic task of /ba/ syllable repetition; SPS = syllables per second; ALS = amyotrophic lateral sclerosis.

Figure 3.

Scatter plots and regression lines showing associations between statistically significant articulatory kinematic predictors and tongue articulation rate (TAR)—that is, “ticker” (fitted values) in speakers with amyotrophic lateral sclerosis. TAR values are on the x-axis and reversed for easier interpretation of articulatory behaviors with pathologic rate decline. Best-fit kinematic predictors of rate were determined by Akaike information criterion–based stepwise linear regression modeling. SPS = syllables per second.

Figure 4.

Scatter plots and regression lines showing associations between statistically significant articulatory kinematic predictors and labial articulation rate (LAR)—that is, “pepper” (fitted values) in speakers with amyotrophic lateral sclerosis. LAR values are on the x-axis and reversed for easier interpretation of articulatory behaviors with pathologic rate decline. Best-fit kinematic predictors of rate were determined by Akaike information criterion–based stepwise linear regression modeling. SPS = syllables per second.

Figure 5.

Receiver operating characteristic (ROC) curves and area under the curve (AUC) values (95% CI) displaying performance of rate measures in differentiating speakers with presymptomatic bulbar amyotrophic lateral sclerosis (ALS) from healthy controls. TAR = tongue articulation rate; DDK-pataka = dysdiadochokinetic task of /pataka/ syllable repetition; LAR = labial articulation rate; DDK-ba = dysdiadochokinetic task of /ba/ syllable repetition; DDK-ta = dysdiadochokinetic task of /ta/ syllable repetition.

Figure 6.

Line plots show within-individual longitudinal change in labial articulation rate (LAR; in syllables per second [SPS]) per subject over time (raw values, n = 16). Statistically significant regression line for time in months is shown in bold. The other rate measures did not significantly change over time.