Direction of Cepstral Peak Prominence Change Moderates Divergent Kinematic Responses to Voice Stimulability Testing

Abstract

Objective:

Stimulability testing is frequently used in managing phonotraumatic vocal fold lesions (PVFLs) to identify strategies that may improve phonatory ease and quality, but little is known about what immediate physiological changes occur during stimulability testing. High-speed videoendoscopy (HSV) enables direct visualization of vocal fold kinematics, and cepstral peak prominence (CPP) is an objective acoustic correlate of voice quality and spectral periodicity. Though increases in CPP are traditionally interpreted as improvement, CPP may instead reflect movement toward the individual’s optimal physiological target. For some, therapeutic focus involves reducing pressed phonation and increasing airflow, which may decrease CPP despite perceptual improvement. For others, efficient resonance and balanced phonation could increase CPP. This study examined whether (1) average HSV parameters change after stimulability testing and (2) CPP direction reflects distinct kinematic patterns.

Methods:

Twenty-seven women with PVFLs were assessed before and after a standardized stimulability protocol using resonant voice facilitators. HSV was conducted on /i:/ in modal register using a transoral rigid scope. Exams were collected at 4000 frames per second. HSV parameters included closing quotient (ClQ), normalized maximum area declination rate (MADRn), speed index (SI), stiffness index (STI), phase asymmetry index (PSI), time periodicity (TP), and waveform symmetry index (WSI). CPP was extracted from simultaneous acoustic recordings. Paired t-tests examined pre- to post-changes in HSV variables and CPP. Exploratory analyses tested whether CPP direction of change (increase vs. decrease) moderated physiologic response (time × group).

Results:

As a single cohort, no significant pre/post changes were observed in HSV parameters or CPP. Time × CPP change direction interactions were significant for ClQ, SI, and STI and approached significance for MADRn. Reduced CPP was associated with slower, less forceful closure.

Conclusions:

Reduction of CPP after stimulability testing showed kinematic changes consistent with reduced hyperfunction evidenced by reduced closure abruptness and a shift toward less pressed phonation behavior. Increased CPP showed vibratory patterns associated with faster vocal fold closure, consistent with enhanced resonance. While both groups reported improvement, they achieved this through different physiological pathways. Lack of overall group change may conceal meaningful subgroups defined by therapeutic target. Findings provide preliminary support of CPP as an accessible indicator of individualized physiological shifts following stimulability testing.

INTRODUCTION

Phonotraumatic voice disorders arise from repetitive or excessive mechanical stress on the vocal folds, often resulting in benign lesions such as nodules and polyps.^1,2 These conditions are highly prevalent among individuals with high vocal demands and represent a substantial portion of clinical voice caseloads.^3,4 Behavioral voice therapy is a primary treatment approach for phonotrauma and has demonstrated effectiveness in improving both perceptual and functional voice outcomes.⁵ However, despite its clinical success, the underlying physiological mechanisms by which therapeutic techniques alter vocal function are not fully understood.

A key component of the voice therapy evaluation is voice stimulability testing, defined as the use of brief, structured voice facilitating tasks to elicit immediate changes in voice quality or ease of phonation.^6,7 Clinicians use stimulability responses for a number of reasons: to identify promising therapeutic approaches, to guide selection of treatment targets, to make prognostic predictions, and to facilitate buy-in and motivation for the therapy process.^6,8,9 Although changes in the sound or feel of the voice are often readily perceived during stimulability testing,¹⁰ a fundamental limitation is that the physiological mechanisms of these changes are not well characterized. In particular, it remains unclear whether perceived improvements reflect measurable alterations in vocal fold vibratory kinematics (e.g., periodicity or glottal closure patterns). This ambiguity poses an important challenge for clinical decision-making, as similar perceptual improvements may arise from different underlying physiological mechanisms. Consequently, clinicians may struggle to determine whether a stimulability response indicates progress toward an optimal treatment target or a compensatory pattern that could limit the overall therapeutic benefit. This highlights the need for more nuanced understanding of the relationship between treatment target, treatment task, and underlying physiological response, which can in turn guide individualized approaches to voice therapy.

From a research perspective, this lack of physiological specificity also constrains our ability to link therapeutic techniques to mechanisms of action,¹¹ limiting the development of more targeted and mechanism-driven interventions. Without a clear understanding of what is changing at the level of the sound source, it is difficult to interpret why certain voice facilitating tasks are effective for some individuals but not others. This may also impede the ability to predict which patients are most likely to benefit from a given therapeutic approach. Therefore, clarifying the physiological correlates of stimulability responses is important for advancing both the precision and efficacy of behavioral voice therapy.

In contemporary frameworks of voice therapy, such as the rehabilitation treatment specification system, treatment is conceptualized in terms of specific targets (i.e., what is being changed) and mechanisms of action (i.e., how change occurs).^12,13 In practice, however, the intended treatment target is not always directly observable, and patients may achieve perceptual improvements via different physiological strategies.¹¹ For example, a clinician may consider reduced vocal fold impact force to be a primary treatment target for a patient diagnosed with phonotraumatic hyperfunction. However, given that impact force itself is not directly observable, this clinician may assume that a clearer sounding voice signifies a reduction in vocal fold impact force, even though this perceptual change may be a consequence of a different or even maladaptive physiologic adjustment. Quantifying how stimulability-induced changes relate to underlying vocal fold kinematics could provide important insight into the mechanisms of therapeutic change and help clinicians refine treatment target selection in the initial stages of therapy.

High-speed videoendoscopy (HSV) provides direct visualization of vocal fold vibration at frame rates sufficient to capture cycle-to-cycle dynamics, enabling quantification of kinematic features such as vibratory periodicity, symmetry, and glottal closure characteristics.^14,15 At the same time, cepstral peak prominence (CPP), an acoustic measure that reflects the degree of harmonic energy in the acoustic voice signal¹⁶ and reflects phonation type on a continuum from breathy to pressed phonation,¹⁷ has been shown to correlate with vocal fold vibratory kinematics, including measures of glottal closure derived from HSV in healthy speakers.¹⁸ Because CPP has been shown to relate to aspects of these phonatory behaviors, it may serve as a noninvasive proxy for underlying physiological changes. In the context of stimulability testing, changes in CPP may reflect shifts in the implicit or unintentional treatment target being adopted by the patient. For example, increases in CPP may indicate greater vibratory periodicity or more complete closure, whereas decreases in CPP may reflect reduced hyperfunction or a shift toward less pressed phonation.¹⁸

The present study sought to investigate the relationship between stimulability-induced changes in underlying vocal fold kinematics in individuals with phonotraumatic voice disorders. The first aim was to determine whether group-level changes in HSV-derived measures occur following stimulability testing. We hypothesized that measures related to glottal closure and vibratory dynamics (e.g., closing speed and velocity) and vibratory symmetry and periodicity would change following stimulability testing when a perceptual improvement in voice quality and/or ease of phonation was reported. A second, exploratory aim was to examine whether changes in CPP moderate the physiological response to stimulability testing. Specifically, we explored whether individuals demonstrating increases versus decreases in CPP exhibit divergent patterns of change in HSV-derived kinematic measures, reflecting potentially different underlying treatment targets or mechanisms of action.

METHODS

Participants

This prospective study was approved by the UT Southwestern Institutional Review Board (STU-2022–0655). Women diagnosed with phonotraumatic vocal fold lesions at the UT Southwestern Voice Center by laryngologists based on videostroboscopy were recruited prior to initiating voice therapy or surgical management between January 2023 and January 2024. Twenty-seven participants provided informed consent and were included in the study. Mean (standard deviation) age was 35.6 (12.0) years. Four participants were diagnosed with a vocal fold polyp, 15 with nodules, and 8 with pseudocyst. There were no statistically significant differences in baseline CPP or HSV-derived parameters across lesion types. For descriptive purposes, a voice-specialized speech-language pathologist (SLP) rated audio samples using the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V) scales of Overall Severity, Roughness, Breathiness, and Strain.¹⁹

Data Collection Protocol

Participants underwent baseline recordings, simultaneously capturing acoustic output and laryngeal kinematics via HSV (described below), while producing a sustained /i:/. Following completion of baseline recordings, participants then underwent a brief stimulability testing protocol. Resonant voice facilitators that emphasize a “forward focus” of the voice (e.g., hums, trills, /z/, etc.) were introduced and modeled by a voice-specialized SLP. The participant was given an opportunity to produce the task first in isolation, and they were given cues consistent with resonant voice therapy, such as feeling anterior oral vibration, etc.^20–22 The clinician modeled a hum facilitator for all participants, which was the only facilitator needed for 23 (85%) of the participants to identify a change in the sound or feel of their voice. For four participants, an additional facilitator or cue was used: one responded to a lip trill cue, and the remaining three responded to respiratory/aerodynamic cues as a means to increase forward tone focus.

Once the participant reported that they either heard or felt an improvement in their voice with the facilitating task, they were instructed to practice with the tongue protruded between the lips and then transition into a sustained /i:/ on a comfortable pitch and volume with the tongue protruded to mimic the conditions of the transoral rigid endoscopy procedure described below. The participant was provided additional cues from the SLP as needed until they reported that their voice sounded or felt improved compared to their baseline voice. All participants included in this study reported an improvement in their voice immediately following stimulability testing. The SLP also reported that the voice quality and/or perceived ease of phonation had improved for all participants by the end of the stimulability testing session.

HSV Data Collection.

HSV exams were acquired with a 70-degree rigid transoral endoscope (Storz HOPKINS II 8706CA) that was connected via a 28-mm KayPentax image coupler and a Storz adaptor mount to a high-speed camera (Phantom V311, Vision Research). The endoscope was powered by a xenon light source (Titan 400E). Images were acquired at 4000 frames per second at a resolution of 512 × 512 pixels as the participant held out a sustained /i:/ at comfortable speaking pitch and volume in their modal register.

HSV Data Analysis.

HSV videos were processed using the Glottis Analysis Tools software (version 2020) to conduct semi-automatic segmentation of images to identify the glottal area.^23,24 Analyses were completed on 200 consecutive vibratory cycles in steady state production of the vowel. Although the following HSV-derived measures were computed from direct vocal fold tissue movement rather than inverse-filtered glottal airflow, several are conceptually related to source parameters described in glottal flow and voice quality literature. In particular, measures reflecting the relative timing and velocity of opening and closing phases may parallel source-related constructs such as speed quotient, temporal skewing of the glottal flow waveform, and abruptness of glottal closure, which have previously been associated with a pressed or tense voice quality.^25–27 The parameters described below were selected to characterize physiologic dimensions potentially relevant to therapeutic shifts in phonation type during stimulability testing. A schematic of the glottal area waveform is provided in Figure 1, and Table 1 provides a summary of the HSV-derived measures.

Figure 1.

Representation of the glottal area waveform and its components.

Table 1.

High-speed video (HSV)-derived parameters estimated from the glottal area waveform and their physiologic interpretation.

HSV Parameter	Physiologic / Voice Source Interpretation
Mechanical Parameters
Maximum Area Declination Rate, normalized (MADRn)	Normalized measure of closing velocity and closure abruptness, related to glottal excitation characteristics
Amplitude-to-Length Ratio (ALR)	Magnitude of vocal fold vibratory excursion normalized to glottal length, reflecting the degree of glottal opening and vibratory amplitude
Closing Quotient (ClQ)	Relative duration of the closing phase of the glottal cycle, reflecting closure dynamics and abruptness of glottal closure
Speed Index (SI)	Relative timing balance between opening and closing phases of the glottal cycle, related to temporal skewing of the waveform and conceptually related to speed quotient
Stiffness Index (STI)	Rate of glottal area change normalized to vibratory amplitude, reflecting vibratory stiffness and rapidity of glottal excitation
Symmetry and Periodicity Parameters
Phase Asymmetry Index (PAI)	Left-right symmetry of vibratory phase
Time Periodicity (TP)	Cycle-to-cycle duration stability and vibratory periodicity
Waveform Symmetry Index (WSI)	Left-right symmetry of the shape and timing of the glottal area waveform

The following parameters related to the mechanical properties of the glottal cycle, as described in Schlegel et al.²⁸ and Patel et al.,²⁹ were extracted:

$$\mathit{A}\mathit{m}\mathit{p}\mathit{l}\mathit{i}\mathit{t}\mathit{u}\mathit{d}\mathit{e}‐\mathit{t}\mathit{o}‐\mathit{L}\mathit{e}\mathit{n}\mathit{g}\mathit{t}\mathit{h}\mathit{R}\mathit{a}\mathit{t}\mathit{i}\mathit{o}(\mathit{A}\mathit{L}\mathit{R})=\frac{(\text{max}glottalarea)-(\text{min}glottalarea)}{glottallength}$$

$$\mathit{C}\mathit{l}\mathit{o}\mathit{s}\mathit{i}\mathit{n}\mathit{g}\mathit{Q}\mathit{u}\mathit{o}\mathit{t}\mathit{i}\mathit{e}\mathit{n}\mathit{t}(\mathit{C}/\mathit{Q}))=\frac{durationoftheclosingphase}{durationofglottalcycle}$$

$$\mathit{S}\mathit{p}\mathit{e}\mathit{e}\mathit{d}\mathit{I}\mathit{n}\mathit{d}\mathit{e}\mathit{x}(\mathit{S}\mathit{I})=\frac{(durationoftheopeningphase)-(durationoftheclosingphase)}{durationoftheopenphase}$$

$$\mathit{S}\mathit{t}\mathit{i}\mathit{f}\mathit{f}\mathit{n}\mathit{e}\mathit{s}\mathit{s}\mathit{I}\mathit{n}\mathit{d}\mathit{e}\mathit{x}(\mathit{S}\mathit{T}\mathit{I})=\frac{\max \left(\frac{d}{dt}(glottalarea)\right)}{(\max glottalarea)-(\min glottalarea)}$$

ClQ, SI, and ALR are dimensionless measures that reflect mechanical aspects of glottal dynamics. ClQ is defined as the ratio of the closing phase (i.e., spanning from peak opening to full closure) to the total duration of the glottal cycle. ClQ values range from 0 to 1, with smaller values indicating a relatively faster closing speed compared to larger values when overall cycle duration is comparable.³⁰ It should be distinguished from the closed quotient, which represents the portion of the glottal cycle during which the glottis is fully closed. In contrast, ClQ specifically describes the relative duration of the transition from maximum opening to closure. This parameter was included because it captures the rate of glottal closure and reflects the abruptness of vocal fold contact. Relatively rapid or abrupt closure has been associated with increased high-frequency harmonic excitation and pressed phonation patterns.¹⁸

SI describes the relative timing between the opening and closing phases of vocal fold vibration and is conceptually related to speed quotient and the glottal pulse shape commonly discussed in glottal flow analysis.²⁷ More positive SI values suggest greater temporal skewing with relatively faster closing time compared to opening. SI was incorporated to evaluate whether differences in phonation type are associated with changes in the temporal balance between opening and closing, which may reflect differences in vibratory efficiency.

ALR is the ratio of vocal fold displacement amplitude to glottal length, with both quantities measured in pixels. This metric was included as it represents the extent of vibratory motion normalized to glottal size, providing an index of how phonation type affects vibratory excursion.

STI, expressed in 1/s, represents the maximum rate of change in glottal area normalized by the amplitude of the glottal area waveform. Lower STI values indicate slower vibratory motion of the vocal folds. STI was included as a measure of waveform dynamics, reflecting the rate at which glottal area changes throughout the cycle.

Maximum area declination rate (MADR) was computed as a cycle-based parameter defined by the negative peak of the first derivative of the glottal area waveform, corresponding to the maximum velocity during the closing phase of the glottal cycle. It is conceptually linked to excitation strength of the acoustic source and spectral slope characteristics described in glottal flow literature.³¹ Because this metric reflects vibratory velocity, it provides indirect information about the mechanical forces associated with vocal fold closure. However, MADR is sensitive to noise and frame rate. To address this, a normalized form (MADRn) was calculated using the following equation to isolate waveform shape while minimizing the influence of amplitude and frequency variability:^32–34

$$\raisebox{1ex}{$\mathit{M}\mathit{A}\mathit{D}\mathit{R}\mathit{n}={}^1$}\!\left/ \!\raisebox{-1ex}{$\left(\frac{(\max glottalarea)-(\min glottalarea)}{MADR}\right)\times pi\times fundamentalfrequenc\text{y}$}\right.$$

As the value of the MADRn approaches 1, the waveform would be sinusoidal.

We also extracted kinematic parameters related to glottal symmetry and periodicity, as described in Patel et al.²⁹

$$\mathit{P}\mathit{h}\mathit{a}\mathit{s}\mathit{e}\mathit{A}\mathit{s}\mathit{y}\mathit{m}\mathit{m}\mathit{e}\mathit{t}\mathit{r}\mathit{y}\mathit{I}\mathit{n}\mathit{d}\mathit{e}\mathit{x}(\mathit{P}\mathit{A}\mathit{I})=\frac{\mid (timeof\text{max}displacementonleft)-(timeo\text{f max}displacementonright)\mid }{durationofglottalcycle}$$

$$\mathit{T}\mathit{i}\mathit{m}\mathit{e}\mathit{P}\mathit{e}\mathit{r}\mathit{i}\mathit{o}\mathit{d}\mathit{i}\mathit{c}\mathit{i}\mathit{t}\mathit{y}\left(\mathit{T}\mathit{P}\right)=\frac{durationmincycle}{durationofmaxcycle}$$

$$\begin{gathered} WaveformSymmetryIndex(WSI)=0.5\times [1+ \\ normalizedsimilaritybetweenleftandrightglottalareawaveform] \end{gathered}$$

Phase asymmetry index (PAI), time periodicity (TP), and waveform symmetry index (WSI) were included as measures reflecting symmetry and periodicity of vocal fold vibration. PAI quantifies temporal differences in the timing of peak displacement between the left and right vocal folds within a given cycle. Specifically, PAI is calculated as the absolute difference between the times at which the left and right vocal folds reach their maximum lateral displacement (or maximum hemiglottal area), normalized by the duration of the glottal cycle.^35,36 Values range from 0 to 1, with a value of 0 indicating perfect symmetry between the vocal folds and larger values indicating increasing phase asymmetry between the vocal folds. This measure was included to capture phase-based asymmetries in vocal fold motion.

TP reflects cycle-to-cycle consistency in glottal cycle duration. It is calculated as the ratio of the shorter to the longer duration between two consecutive cycles. TP values range from 0 to 1, with values closer to 1 indicating greater temporal regularity and values approaching 0 reflecting increased variability in successive cycle lengths. This parameter was included to assess the degree of periodicity in vocal fold vibration.

WSI quantifies the similarity between the left and right hemiglottal area waveforms during a vibratory cycle. It is calculated from the cosine similarity between the left and right glottal area waveform cycles and scaled to range from 0 to 1. A value of 1 indicates identical waveform patterns and maximal symmetry. Lower values indicate increasing dissymmetry between the left and right waveform shapes. WSI was included as an index of spatial asymmetry in vocal fold vibration.

Acoustic Data Acquisition and Analysis.

Simultaneous acoustic voice recordings were obtained in a sound treated booth using a head-mounted unidirectional condenser microphone [AKG C520] positioned approximately 5 cm from the lips and an angle of 45 degrees away from the mouth. The microphone signal was input into a preamplifier [M-Audio M-Track Duo] and digitized at a 44.1 kHz sampling rate. Recordings were captured during steady state phonation both during baseline recordings and following stimulability testing. Audio signals were analyzed using Praat software [version 6.2.11]. CPP was calculated using the Praat CPPS function, as described by Murton et al. (2020) and Toles & Murton (2025).

For our exploratory research aim, we categorized participant data as having either a CPP increase or a CPP decrease following stimulability testing. In this case, CPP change direction was used as a proxy for therapy target. Our prior work has demonstrated that CPP tends to decrease with increased breathiness and increase with more pressed or hyperfunctional phonation patterns.^17,18 Therefore, an increase in CPP would suggest a target of increasing vocal fold closure or phonatory energy whereas a decrease in CPP would suggest a target of reducing vocal hyperfunction or increasing phonatory airflow. Fifteen participants demonstrated an increase in CPP following stimulability testing, and 12 demonstrated a CPP decrease. This was used as a categorical variable in the statistical analysis. Chi Square tests revealed that lesion type was not related to CPP change direction (p = .167).

Statistical Analysis

First, we conducted paired t test models to determine whether there were changes in HSV-derived kinematic measures following stimulability testing. For each model, the HSV parameter was the dependent variable, and Time (pre- vs post-stimulability testing) was the independent variable. Family-wise correction was applied to the alpha significance value based on the number of tests within the family of mechanical parameters (α = .010) and the family of symmetry/periodicity parameters (α = .0167). Cohen’s d effect sizes are presented, with the interpretation that 0.20 = small; 0.50 = medium; 0.80 = large.

Then, to examine whether changes in HSV kinematic measures following stimulability testing varied as a function of cepstral peak prominence (CPP) response, linear mixed-effects models were conducted using the GAMLj module in jamovi. Data were structured in long format with repeated observations (pre- and post-stimulability) nested within participants. Separate models were fit for each HSV outcome variable. Fixed effects included Time (pre vs post), CPP change direction group (CPP increase vs CPP decrease), and their interaction (Time × CPP group). Participant was included as a random intercept to account for within-participant dependence across repeated measurements. For this exploratory analysis, the primary effect of interest was the Time × CPP group interaction, which tested whether pre-to-post changes in HSV parameters differed between participants demonstrating increases versus decreases in CPP following stimulability testing.

RESULTS

For descriptive purposes, HSV parameters are visualized in box plots, stratified by lesion type, in Figure 2. Statistics for the paired t test comparisons are present. All comparisons were nonsignificant at the corrected significance values. Time Periodicity had a small-to-medium effect size (t(26) = 1.96, p = 0.061, d = 0.38), but it was not statistically significant.

Figure 2.

Boxplots illustrating high-speed videoendoscopic parameters at baseline (pre) and following stimulability testing (post) stratified by lesion type: polyps (in blue, left), nodules (in gray, middle), and pseudocysts (in orange, right). There were no statistically significant differences lesion types. MADRn = normalized maximum area declination rate.

Table 3 presents descriptive statistics for CAPE-V ratings and CPP values. Participants in the CPP increase group (i.e., the group for whom CPP increased following stimulability testing) demonstrated higher mean baseline breathiness ratings, whereas the CPP decrease group demonstrated higher baseline strain ratings. The CPP increase group had lower baseline CPP values compared to the CPP decrease group. A linear mixed-effects model revealed a significant Time × CPP group interaction for ClQ (F(1, 25) = 10.78, p = .003), Speed Index (F(1, 25) = 4.40, p = .046), and Stiffness Index (F(1, 25) = 4.96, p = .035) and a trending difference for MADRn (F(1, 25) = 2.74, p = .110). Estimated marginal means demonstrated that the CPP increase group exhibited a decrease in ClQ from pre- to post-stimulability testing (Δ = −0.04, 95% CI [−0.08, 0], p = 0.065, d = 0.77), whereas the CPP decrease group showed a statistically significant increase in ClQ (Δ = 0.06, 95% CI [0.01, 0.11], p = 0.013, d = 1.07). For SI, the CPP increase group exhibited an increase from pre- to post-stimulability testing (Δ = 0.04, 95% CI [−0.03, 0.10], p = 0.250, d = 0.47), and the CPP decrease group showed the opposite trend (Δ = −0.06, 95% CI [−0.14, 0.01], p = 0.091, d = 0.70). The CPP increase group also showed a post-stimulability testing increase in STI (Δ = 58.67, 95% CI [770.88, 188.22], p = 0.360, d = 0.37), whereas the CPP decrease group had lower STI values post-testing (Δ = −151.47, 95% CI [−296.31, −6.63], p = 0.041, d = 0.86). These interactions indicate that direction CPP change moderates the kinematic response to stimulability testing. Statistics for linear mixed models are presented in Table 4, and the interaction effects are visualized in Figure 3.

Table 3.

Means (standard deviations) of CAPE-V parameters and CPP values at baseline (pre) and after stimulability testing (post) in the total cohort and in groups separated by whether CPP increased or decreased post-stimulability testing.

Parameter	Total Cohort (n = 27)		CPP Increase Group (n = 15)		CPP Decrease Group (n = 12)
	Pre	Post	Pre	Post	Pre	Post
CAPE-V Parameter
Overall Severity	34.41 (14.26)	25.82 (16.83)	34.87 (16.57)	22.13 (11.25)	33.83 (11.43)	16.58 (11.17)
Roughness	25.82 (16.83)	13.59 (13.03)	27.00 (19.84)	16.40 (14.46)	24.33 (12.81)	10.08 (10.54)
Breathiness	20.78 (16.86)	11.33 (10.17)	30.00 (15.78)	14.13 (9.65)	9.25 (9.71)	7.83 (10.08)
Strain	26.44 (11.90)	12.52 (6.04)	21.27 (11.13)	13.60 (6.32)	32.92 (9.75)	11.17 (5.64)
CPP (dB)	12.70 (2.31)	13.07 (1.99)	11.70 (2.23)	13.75 (1.92)	13.95 (1.78)	12.22 (1.81)

CAPE-V = Consensus Auditory-Perceptual Evaluation of Voice; CPP = Cepstral Peak Prominence

Table 4.

Linear mixed model statistics for fixed effects of Time, Group, and their interaction on each high-speed videoendoscopy-derived kinematic parameter.

Kinematic Parameter	Effect	Estimate	SE	95% CI	t	p
MADRn	Intercept	1.59	0.06	1.48, 1.70	28.23	< .001
	Time	0.08	0.10	−0.11, 0.27	0.82	0.423
	CPP Group	−0.22	0.11	−0.44, 0.02	−1.88	0.072
	Time × CPP Group	0.32	0.19	−0.07, 0.70	1.66	0.110
Amplitude-to-Length Ratio	Intercept	13.67	1.08	11.50, 15.85	12.62	< .001
	Time	0.10	1.03	−1.08, 3.08	0.97	0.344
	CPP Group	1.20	2.17	−3.16, 5.55	0.55	0.585
	Time × CPP Group	0.56	2.07	−3.60, 4.72	0.27	0.789
Closing Quotient	Intercept	0.38	0.014	0.35, 0.40	26.73	< .001
	Time	0.01	0.06	−0.02, 0.04	0.71	0.486
	CPP Group	0.03	0.03	−0.03, 0.09	1.10	0.282
	Time × CPP Group	0.10	0.03	0.04, 0.17	3.28	0.003
Speed Index	Intercept	0.17	0.02	0.12, 0.22	7.38	< .001
	Time	−0.01	0.02	−0.06, 0.04	−0.53	0.603
	CPP Group	−0.09	0.05	−0.18, 0.00	−1.95	0.062
	Time × CPP Group	−0.10	0.05	−0.20, −0.00	−2.10	0.046
Stiffness Index	Intercept	1163.90	34.64	1094.26, 1233.54	33.60	< .001
	Time	−46.40	47.18	−141.26, 48.45	−0.98	0.335
	CPP Group	−85.19	69.27	−224.48, 54.09	−1.23	0.230
	Time × CPP Group	−210.14	94.35	−399.85, −20.42	−2.23	0.035
Phase Asymmetry Index	Intercept	0.07	0.01	0.05, 0.08	8.12	< .001
	Time	−0.00	0.01	−0.02, 0.02	−0.28	0.783
	CPP Group	0.02	0.02	−0.01, 0.05	1.18	0.250
	Time × CPP Group	−0.03	0.02	−0.07, 0.01	−1.40	0.173
Time Periodicity	Intercept	0.96	0.00	0.95, 0.96	471.78	< .001
	Time	0.01	0.00	−0.00, 0.01	1.83	0.079
	CPP Group	0.00	0.00	−0.01, 0.01	0.25	0.807
	Time × CPP Group	−0.01	0.01	−0.02, 0.01	−1.04	0.309
Waveform Symmetry Index	Intercept	0.94	0.01	0.92, 0.96	99.17	< .001
	Time	0.01	0.02	−0.02, 0.04	0.68	0.500
	CPP Group	−0.02	0.02	−0.06, 0.02	−1.11	0.279
	Time × CPP Group	0.04	0.03	−0.02, 0.10	1.26	0.218

Note. MADRn = Normalized Maximum Area Declination Rate; CPP = Cepstral Peak Prominence

p values for interaction effects that were statistically significant are bolded.

The Time effect was pre-post stimulability testing with the pre-testing being the reference value.

The CPP Group effect was the direction of change in CPP from pre- to post- stimulability testing with CPP decrease being the reference value.

Figure 3.

Plots visualizing means and 95% confidence intervals of high-speed videoendoscopy-derived kinematic parameters pre- and post-stimulability testing by CPP change direction group for interactions that were statistically significant or near-significant. The group who demonstrated a decrease in CPP following stimulability testing are presented in light orange, and the group whose CPP increased are shown in blue.

DISCUSSION

The purpose of this study was to examine whether stimulability testing results in measurable changes in vocal fold kinematics in individuals with phonotraumatic voice disorders. Our secondary objective was to explore whether changes in CPP moderate these physiological responses. Although the primary hypothesis of group-level change in vocal fold kinematics was not supported, findings from the exploratory analysis suggest that CPP change direction may provide meaningful insight into divergent kinematic responses to stimulability testing.

With respect to the first aim, none of the HSV parameters demonstrated statistically significant group-level changes following stimulability testing. Superficially, these findings suggest that stimulability does not produce uniform physiological changes across individuals. However, this interpretation is at odds with the clinical observation that all participants in the study reported a perceptible change in the sound and/or feel of their voice during and after stimulability testing. One potential interpretation for this finding is that group-level analyses obscure meaningful individual variability, effectively washing out divergent patterns of physiological kinematic change. In other words, participants may achieve perceptual improvement in their voice through different underlying mechanisms, resulting in minimal observable change in kinematic measures when averaged across the sample. This discrepancy motivated further exploration of whether a moderating variable, such as the direction of change in CPP, could help explain this heterogeneity.

Findings from the second, exploratory aim provided useful clinical context. CPP change direction appeared to moderate changes in several of the HSV parameters that are related to vocal fold closure and vibratory dynamics, including ClQ, SI, STI, and MADRn, each of which demonstrated significant or near-significant interaction effects with visual divergence in trends between groups. Specifically, participants whose CPP increased demonstrated faster glottal closure patterns after stimulability testing, whereas participants whose CPP decreased demonstrated less abrupt glottal closure post-testing. In contrast, measures reflecting symmetry and periodicity were not moderated by CPP change. These findings suggest that CPP change direction may be more closely related to vibratory timing properties than to symmetry- or periodicity-based features. Therefore, in the context of stimulability testing, CPP may serve as a clinically accessible indicator of underlying changes in vocal fold contact and vibratory efficiency. These results are consistent with prior work demonstrating associations between CPP and aspects of glottal closure and vibratory kinematics.^29,34,37,38 The present findings extend beyond the previous work by suggesting that the direction of CPP change may reflect qualitatively different physiological response patterns, rather than simply indicating improvement or worsening along a single continuum.From a clinical perspective, the distinction of CPP direction of change is highly relevant. CPP is relatively simple to obtain in clinical settings and is recommended as part of a standard acoustic assessment battery.^39,40 If CPP change direction reflects underlying differences in vocal fold closure behavior, clinicians may be able to use this information to infer the implicit treatment target being adopted during stimulability testing. For example, a decrease in CPP during the perceived “improved” voice condition may reflect a reduction in hyperfunctional or pressed phonation patterns, potentially corresponding to decreased vocal fold collision forces and improved tissue loading conditions in individuals with phonotrauma.³⁴ Conversely, an increase in CPP during the “improved” voice condition may reflect increased harmonic energy and more efficient vocal fold vibration, possibly indicating a shift away from compensatory patterns and characterized by excessive airflow or aperiodicity. These interpretations are not mutually exclusive and likely reflect different pathways to perceptual improvement, depending on the individual’s baseline voice production strategy.

Descriptive auditory-perceptual findings further support this interpretation (see Table 3). Participants demonstrating decreased CPP following stimulability testing tended to present with greater baseline strain, whereas those demonstrating an increase in CPP tended to exhibit greater baseline breathiness. This pattern bolsters the interpretation that individuals with more hyperfunctional or pressed baseline phonation may achieve perceptual improvement through less abrupt glottal closure and increased airflow release. In contrast, those with breathier or less efficient baseline phonation may improve by enhancing glottal closure efficiency and harmonic energy. Although auditory-perceptual ratings were only supplied for descriptive purposes in the current study, these patterns should be explored in future work to further investigate whether baseline phonatory characteristics influence the physiologic pathway through which immediate improvement can be achieved.

In cases where multiple voice facilitators result in perceptually “improved” voice, subtle differences in CPP may help clarify which facilitators are the most promising therapeutic agents. In other words, improvements achieved with one facilitator versus another may be perceptually indistinguishable⁴¹ yet differentiable using CPP. These differences may guide clinicians in selecting one approach over another for a given patient. Notably, the direction of CPP change may also help identify instances in which a patient’s perception of “improvement” reflects a maladaptive strategy. For example, if an individual with phonotraumatic pathology and pressed phonation at baseline further increases CPP during a facilitating task, they might perceive this as an “improved voice” despite this pattern potentially reflecting increased hyperfunction. In such cases, an increase in CPP beyond an expected or desirable range may signal increased vocal fold contact or collision forces, suggesting that the facilitator may reinforce, rather than alleviate, the underlying pathology. While these scenarios are outside of the scope of this study, the current findings highlight a potential role for CPP not only in treatment selection but also in patient education, helping to align patient perceptions of improvement with physiologically beneficial voice production strategies.

More broadly, CPP may be particularly valuable when evaluating subtle changes in vocal strain, a perceptual dimension known to have relatively low intra- and inter-rater reliability.^41–43 In this context, CPP may provide an objective complement to perceptual judgment, enhancing clinical decision-making when perceptual cues are ambiguous. A potential future direction is the establishment of minimal clinically important difference values for CPP, which would allow clinicians to determine whether observed changes reflect meaningful improvements in voice production. Such thresholds could further refine interpretation of CPP changes during stimulability testing, helping to determine whether changes are real and clinically meaningful.

The findings presented here have important implications for treatment planning and clinical decision-making. If perceptual improvements can arise from different underlying physiological mechanisms, identifying these mechanisms could help clinicians individualize therapy targets and select more appropriate intervention strategies aligned with optimal vocal function. As such, CPP may serve not only as an outcome measure but also as a clinically accessible indicator of treatment response at the level of the sound source.

Limitations

This study has several limitations that should be considered when interpreting the findings. First, the sample size was modest, particularly limiting statistical power in the second aim for detecting interaction effects. Therefore, findings should be interpreted as exploratory. Replication with larger sample sizes are needed. Second, the dichotomization of CPP change simplifies what is inherently a continuous variable and may obscure more nuanced relationships. Third, although HSV provides detailed information about vocal fold kinematics, it does not capture all aspects of laryngeal and supraglottic behavior that may contribute to perceptual voice changes. Future work incorporating additional physiological measures and larger samples will be important for confirming and extending these findings.

CONCLUSION

Group-level kinematic changes following stimulability testing were not observed. However, reduction of CPP after stimulability testing showed kinematic changes characterized by slower and less abrupt glottal closure behavior, patterns potentially consistent with reduced vocal hyperfunction and greater airflow release. In contrast, increased CPP showed vibratory patterns associated with relatively faster and perhaps more efficient vocal fold closure, suggestive of enhanced harmonic excitation and resonant voice production. Although both groups reported improvement, they achieved this through different physiological pathways. Lack of overall group change may conceal meaningful subgroups defined by therapeutic target. Findings provide preliminary support of CPP as an accessible indicator of individualized physiological shifts following stimulability testing.

Table 2.

Statistics for paired-samples t tests comparing high speed videoendoscopy-derived kinematic parameters before and after stimulability testing.

Kinematic Parameter	Baseline Mean (SD)	Post-Stimulability Testing Mean (SD)	t	p	d
Mechanical Parameters
Maximum Area Declination Rate, normalized	1.63 (0.06)	1.57 (0.09)	−0.62	0.544	0.12
Amplitude-to-Length Ratio	13.12 (1.07)	14.09 (1.28)	0.96	0.347	0.18
Closing Quotient	0.37 (0.02)	0.38 (0.02)	0.29	0.771	0.06
Speed Index	0.18 (0.02)	0.17 (0.03)	−0.28	0.783	0.05
Stiffness Index	1185.99 (44.25)	1151.27 (41.53)	−0.69	0.496	0.13
Symmetry and Periodicity Parameters
Phase Asymmetry Index	0.07 (0.01)	0.07 (0.01)	−0.12	0.905	0.02
Time Periodicity	0.95 (0.003)	0.96 (0.002)	1.96	0.061	0.38
Waveform Symmetry Index	0.94 (0.02)	0.05 (0.01)	0.54	0.593	0.10

Note. There were no statistically significant pre-post differences across all parameters.

DATA AVAILABILITY STATEMENT

The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.