Effects of Intensive Speech Treatment for an Individual with Spastic Dysarthria Secondary to Stroke

Objective: This study investigated the impact of an intensive speech treatment on listener-rated communication success and functional outcome measures of communication for an individual with spastic dysarthria secondary to stroke. Method: A single-subject A-B-A-A experimental design was used to measure the effects of an intensive speech treatment that incorporated principles of motor learning to drive activity-dependent changes in neural plasticity. The primary dependent variables were listener-rated communication success (comprehensibility transcription in two conditions and listener perceptual ratings of speech and voice), and functional outcome measures as rated by the participant and his spouse. Secondary dependent variables included acoustic factors: vowel space area, phonatory stability, and vocal dB SPL during speech tasks. Results: Multiple comparisons with t-tests were used to determine statistically significant changes in primary and secondary dependent variables. Statistically significant changes (p<0.05) were present immediately post-treatment in listener perceptual ratings for speech naturalness in sentences (p=0.00), but demonstrated a preference for pre-treatment sustained vowel phonation (p=0.04). All functional outcome measures reflected the participant’s perception of increased communicative effectiveness, decreased psychosocial impacts of dysarthria, and increased social participation. There were statistically significant changes in secondary variables at post-treatment including phonatory stability in amplitude perturbation quotient (p=0.02), and vocal dB SPL during sustained vowel phonation (p=0.01), and sentence reading (p=0.03). Vowel space area increased by 13% at post-treatment. Three months following treatment, there were statistically significant changes in listener comprehensibility at the single word length (p=0.02) and sentence length (p=0.03), and listener perceptual ratings of speech naturalness (p=0.02). All functional outcome measures displayed maintained post-treatment effect. Vowel space area increased by 25% compared to pre-treatment. There were no statistically significant changes in phonatory stability or vocal dB SPL three months following treatment. Conclusions: Treatment outcomes were specific to the research participant’s individual characteristics. The improvements measured immediately postand three months following treatment cannot be generalized beyond this individual with dysarthria secondary to stroke. However, the positive treatment effects for STR03 indicated that individuals in the chronic stages of recovery with dysarthria can improve and maintain speech comprehensibility as well as increase communication effectiveness and reduce some of the negative emotional and social components of chronic dysarthria, even four years post-onset, warranting further investigation.

examined the impact of an intensive behavioral speech treatment that targeted clear speech with an adult who had spastic dysarthria secondary to a stroke. This first chapter of the thesis presents the background of the study, specifies the problem of the study, describes its significance and presents an overview of the methodology used.

Background
It is reported that approximately 795,000 individuals experience a new or recurrent stroke each year, but the fatality rate is in decline (Go et al., 2014).
Therefore, stroke is one of the leading causes of long-term disability in the United States. An estimate of the incidence of dysarthria post-stroke is around 40% (Flowers, 2013). Dysarthria is the collective term for a neurological speech disorder resulting from changes in strength, speed, range, steadiness, tone, or accuracy of speech movements. Dysarthria is further categorized and defined by the location of damage to the nervous system. Spastic dysarthria results from bilateral damage to the direct and indirect activation pathways in the central nervous system, which can result in changes to speech components including respiration, resonation, articulation, phonation, and prosodic variation (Duffy, 2012).
Very few studies have documented the efficacy of specific treatment approaches for individuals with dysarthria secondary to stroke (Sellars et al., 2005;Mackenzie, 2011). Even fewer studies describe specific treatment approaches for individuals over nine months post-onset . Many of the studies available emphasize the effects of treatment on acoustic factors of speech such as decibel sound pressure level (dB SPL), voice parameters, and vowel space area or the effects of treatment on listener intelligibility. A complete look at the effects of treatment should also include measurements of communication success and patient and/or family reported functional outcomes to determine the overall impact of treatment on activities of daily living.

Significance
Individuals with dysarthria secondary to stroke have reported feelings of marginalization and stigmatization, as well as emotional and social changes including changes in self-identity and relationships (Walshe et al., 2009). Social and emotional effects of dysarthria may be disproportionate to the severity of the communication disorder (Dickson et al., 2008) and can contribute to the negative impact of dysarthria on quality of life. Given the lack of research in this area and the significant social and emotional consequences associated with dysarthria after stroke, the purpose of this study was to determine the effect of a well-defined and intensive speech treatment for an individual with dysarthria secondary to stroke in the chronic stage of recovery with the goal of improving comprehensibility, and increasing participation in functional communication.

Methodology Overview
This Phase I study utilized a single-subject A-B-A-A experimental design (Robey, 2004). This design was selected because it was appropriate for making initial observations about the impact of an intensive speech treatment on an individual with spastic dysarthria secondary to stroke. The primary aim of the study was to determine the effect of treatment on listener-rated communication success and functional outcome measures. Changes in communication success from pre-to post-treatment and pre-to 3-months following treatment were assessed using listener comprehensibility ratings of the participant's speech in two conditions: 1) using the acoustic signal alone and 2) using the acoustic signal plus visual information as the participant spoke. Listeners also rated voice quality and speech to assess perceptual characteristics of voice and speech. The impact of treatment on functional outcome measures including participation in functional communication and communicative effectiveness were assessed using two patient and spouse-reported outcome measures, the Communicative Effectiveness Index-Modified (CETI-M; Yorkston et al., 1999), and the Dysarthria Impact Profile (DIP; Walshe et al., 2009). Additional qualitative input was obtained from the participant's and spouse's interviews pre-, post-, and 3months following treatment, and field notes taken during treatment. The following were the study hypotheses: Listener-rated Communication Success: 1) Listener comprehensibility ratings will increase following treatment using the acoustic signal alone.
2) Listener comprehensibility ratings will increase to a greater extent following treatment using the acoustic signal plus visual information.
3) Listeners will rate perceptual characteristics of voice and speech better following treatment when compared to pre-treatment.
Functional Outcome Measures 4) The participant and his spouse will rate communicative effectiveness higher following treatment.
5) The participant will rate psychosocial impacts of dysarthria lower following treatment.
6) The participant and his spouse will describe overall increases in social participation following treatment.
A secondary aim of the study was to evaluate the impact of treatment on acoustic variables of speech including the first two formants (F1 and F2) of the corner vowels /i/, /u/, and /a/, measures of phonatory stability, and vocal dB SPL during speaking tasks.

METHODOLOGY
The methods section of this thesis provides a study overview, information about the participant, protocol for the specific treatment approach, a rationale for the dependent variables of the study, explanation of assessment procedures, description of data analyses and statistical analyses, and a discussion about reliability in this study.

Study Overview
This study examined the administration of an intensive behavioral speech treatment that incorporates principles of motor learning to drive activity-dependent changes in neural plasticity that can contribute to our understanding of how motor learning theory applies to treatment of dysarthria and how we can administer effective treatment efficiently. The primary dependent variables of interest were speech comprehensibility in two conditions, listener perceptual ratings of voice and speech, and changes in communicative effectiveness, the psychosocial impact of dysarthria, and social participation based on questionnaire responses and interviews.
Speech comprehensibility was measured by listener transcriptions of phonetically balanced single word and sentence length materials (Kent et al. 1989;Nilsson 1994) using an audio recording of the participant alone, and using audio and video recordings of the participant. Perceptual voice quality and speech naturalness were measured with listener ratings of sustained vowel phonation and sentence reading samples comparing pre-and post-treatment and pre-and 3-month follow-up (FU).
Communicative effectiveness was measured using the CETI-M. The participant and his spouse rated communicative effectiveness in 10 different scenarios.  (Doyon and Benali, 2005). An individual stores past experiences and learns new behaviors through a process of neural plasticity. Neural plasticity has also been identified as the mechanism by which an individual rehabilitates and relearns processes following brain injury (Kleim & Jones, 2008). There are ten principles of experience-dependent neural plasticity defined by Kleim and Jones, 2008. These principles include, "use it or lose it", "use it and improve it", "specificity", "repetition matters", "intensity matters", "time matters", "salience matters", "age matters", "transference", and "interference". These principles were translated to serve as guidelines for behavioral treatment of motor systems, defined as principles of motor Implicit learning (if you are using a cue, why are you calling this "implicit learning" when you are explicitly highlighting clear speech?) was utilized through the use of a single cue for "clear speech" throughout the treatment, which minimized the cognitive load for the individual while allowing the clinician to change the way this is modeled based on the client's specific speech patterns. Augmented feedback was provided based on the needs of the client and decreased systematically throughout the treatment course to support generalization and increased independence (Duffy, 2012;Maas et al., 2008;Kleim & Jones, 2008).
Increasing intelligibility and naturalness are common goals of speech treatment for individuals with dysarthria. Providing cues for loudness, reducing rate of speech, and cueing for clear speech have been studied as ways to improve intelligibility for neurologically normal individuals as well as individuals with dysarthria secondary to multiple sclerosis and Parkinson's disease (Smijanic & Bradlow, 2009;Uchanski, 2005;Tjaden, 2014). Tjaden et al. (2014) established that speaker' intelligibility ratings increased with cues for either increased loudness or clear speech. A cue for clear speech may be a more effective cue for individuals with spastic dysarthria who may not benefit from a cue to "speak loud" or "slow down" speech due to the speech components and patterns that these individuals present with. Despite this evidence, there are very few studies reporting on the impact of a clear speech treatment protocol for individuals with dysarthria, or more specifically spastic dysarthria.
Cueing and modeling were important components of the treatment process.
Direct modeling can provide the participant with an understanding of what is meant by the cue for "clear speech". Cueing during non-speech tasks emphasized increasing or maintaining effort level. Appropriate cueing for non-speech tasks with the Iowa Oral Performance Instrument (IOPI) are "Push, push, push!" or "Go, go, go". Examples of appropriate cueing during speech tasks include "Remember to use your clear speech" and "Speak clearly". The participant received positive reinforcement following speech tasks such as "Great clear speech" and "That's the speech that people will understand". Cueing and modeling were decreased throughout the course of treatment to promote independence and increase carry-over outside of the clinic setting.
Data were collected during each session including kPa (pressure measurement) during lip and tongue IOPI exercises, duration of sustained vowel phonation, percentage of accurate articulation in minimal pair repetition, and the loudness of sustained vowel phonation, salient sentence reading, and the hierarchy reading task.
The consistent speech sound errors noted during pre-treatment evaluations included voicing errors, deletion errors, and vowel errors. STR03's speech sound errors were targeted through minimal pair tasks (i.e., pairs of words which differ by only one phoneme; e.g. bad and pad). Particular emphasis during the minimal pair task was placed on voicing errors due to their frequency in STR03's speech. The frequency of voiced/voiceless cognates in typical speech interfered with STR03's communication success in the pre-treatment evaluation. A list of minimal pair sets used during treatment is displayed in Appendix C. Homework consisting of treatment tasks and a carryover task (e.g. using clear speech to order movie tickets) were assigned each day to increase treatment intensity and promote generalization of clear speech to activities of daily living.

Dependent Variables
Primary Aim, Hypotheses 1-3: Listener-rated Communication Success Listener-rated communication success was measured using listener transcriptions of comprehensibility at the single word and sentence level in two conditions and using listener perceptual ratings of speech and voice. The goal of speech treatment is to increase communication success in functional conversation so outcome variables need to capture these functional changes.
Comprehensibility is differentiated from intelligibility because the listener is provided with the communication context of the utterance (Barefoot et al., 1993).
Measuring comprehensibility entails providing the listener with information other than the acoustic signal. This information may be in the form of semantic, syntactic, or physical context (Yorkston et al., 1996). Lindblom (1991) suggests that speech and listener perceptions of speech are adaptive to the needs of the situation. Therefore, speech perception is not always simply signal-dependent. Listener perception may require background knowledge or shared context when speech is disordered or distorted. Comprehensibility was selected as a primary variable because it provides the listener with some context for determining whether the participant was successful in conveying his message.
We compared how providing the listener with visual information through video and audio input impacts listener transcriptions of single word and sentence length material compared with audio input alone. Several other studies have used both audio and audio and video listener conditions for transcriptions (Keintz et al., 2007;Hunter et al., 1991;Garcia and Cannito, 1996).  (Threats, 2012).
Addressing the concerns of the individual receiving treatment is an essential component of the treatment process. Qualitative measurement of the participant's personal experience is critical for evaluating a treatment (Kovarsky, 2008). The participants' perceptions of treatment outcomes are particularly important due to the impact of acquired dysarthria on social participation and psychosocial factors (Dickson et al., 2008).
Communicative effectiveness is measured using the Communicative Effectiveness Index-Modified (CETI-M) in this study. Lomas et al. (1989) introduced the CETI as a measure of functional communication for adults with aphasia. The authors of the CETI demonstrated the measure's internal reliability (Split-half r=0.90), inter-rater reliability (r=0.73), test-retest reliability (r=0.94), and construct validity using an n of 22 (Lomas et al., 1989).

Secondary Aim: Acoustic Factors
Acoustic measurements in this study included vowel space area, phonatory stability, and vocal dB SPL. The selected acoustic measurements were analyzed for the purpose of understanding potential factors contributing to changes in listener comprehensibility ratings. There is no direct correlation between perceptual features and acoustic variables but acoustic analysis can be informative and supportive of perceptual findings (Kent et al., 1999).
Vowel formants are important measurements in the analysis of speech production as they have been linked to articulatory precision. Vowel space area was determined by measurement of the first and second formants (F1 and F2) of three corner vowels: /a/, /i/, and /u/ in the sentence "The boot on top is packed to keep".
These three corner vowels are selected because of their representation of extreme articulatory movements of the tongue.  demonstrated that lower intelligibility ratings were associated with greater overlap among vowel formants, relating to "reduced articulatory working space" (192). Vowel space area analysis will help to determine the impacts of treatment on articulatory precision in speech production. Kent et al. (2003) validated the use of the Multidimensional Voice Profile (MDVP Advanced; CSL 4500) to assess voice data collected from individuals with dysarthria secondary to hemispheric and brainstem stroke. This study identified several potentially deviating acoustic measurements associated with this population such as variation in fundamental frequency (vf0,) smoothed pitch perturbation quotient (sPPQ), absolute shimmer (ShdB), relative shimmer (Shim), smoothed amplitude perturbation quotient (sAPQ), peak amplitude variation (vAm), and amplitude perturbation quotient (APQ). All of these acoustic measurements fall into categories of either frequency perturbation or amplitude parameters and are considered measures of phonatory stability.
Vocal loudness is determined by the intensity of the sound signal, which is measured in dB SPL. The speaker's vocal loudness impacts the listener's understanding of the message.

Assessment Procedures
Dependent variables were assessed three times during the study. Each of the three evaluations included four consecutive days of testing. Initial data collection took place immediately prior to treatment (Pre), the second occurred during the week immediately following completion of treatment (Post), and the third was a follow-up evaluation, which took place three months following treatment (FU).

Data Analyses
Primary Aim, Hypotheses 1-3: Listener-rated Communication Success A total of sixty listeners with normal hearing and no history of neurological disorder or head injury assessed comprehensibility by transcribing single word and sentence length materials. One group of thirty listeners transcribed words and sentences from audio input only, and one group of thirty listeners transcribed using both audio and visual input to measure and compare comprehensibility conditions.
Ten listeners from each group transcribed pre-treatment samples, ten listeners from each group transcribed post-treatment samples, and ten listeners from each group transcribed FU-treatment samples. and five samples out of the group were randomly selected as repeated measures for determination of intra-rater reliability.
Primary Aim, Hypotheses 4-6: Functional Outcome Measures The impact of treatment on the functional outcomes was measured in three ways. The CETI-M was used to provide a quantitative measure of change in the level of communicative effectiveness in daily living situations over the treatment course (Lomas et al. 1999;Yorkston et al., 1999). Voice dysfunction and targeted acoustic parameters of voice were assessed using acoustic software, MDVP. MDVP was used to analyze phonatory stability measures during sustained vowel phonations. Vocal sound pressure level (dB SPL) during speech tasks was collected throughout the evaluation sessions. Vocal sound pressure level was also measured during each treatment session using an SLM.

Statistical Analyses
Multiple comparisons with t-tests determined the significance of any changes to the dependent variables following treatment at Post or FU evaluations. Effect size using Cohen's d determined the magnitude of treatment effect. Average percentage and standard deviation of listener ratings for sustained vowel phonation and sentence reading were calculated to determine overall listener preference and the magnitude of preference for samples. The means of F1, F2, and vowel duration from 20 corner vowels repeated in "The boot on top is packed to keep" were used to create pre-, post-, and 3-month follow-up mean vowel space area, calculate vowel space area change, and determine changes in vowel duration.

Measurement Reliability
The clinician who administered the treatment (CP) did not participate in evaluations to limit potential bias. Intra-rater reliability was calculated using percent agreement for vowel space area analysis on 25% of the data at 2-4 months following the initial analysis. There is typical agreement in the literature that percent agreement above 70% is acceptable (Stemler, 2004). Intra-rater reliability for vowel space area using PRAAT formant analysis was 87.5%, calculated based on differences in formant data over 50 Hz during the second analysis. Intra-rater reliability for vowel duration using PRAAT was 75%, calculated based on differences in duration data over 50 ms during the second analysis.
Listener studies were conducted in the IAC treated sound booth. Participants listened to samples at a consistent volume. A random number generator was used to randomize HINT sentences repeated during evaluation tasks and presented to listeners during the transcription task. Individual rater variability for each component of the listener transcription task is displayed in Appendix E. Any individual listener percentage that was two standard deviations below or above the mean was extracted from the data set to reduce the effects of inter-rater variability.
Listeners participating in the perceptual rating task evaluated a randomized selection of 20 pairs of sustained vowel phonations and 20 pairs of sentence repetitions ("The boot on top is packed to keep") collected during evaluations. Twenty percent of sentence pair and sustained vowel phonation combinations were randomly selected and repeated to determine intra-rater reliability with this task. Intra-rater and inter-rater reliability for the listener preference study was calculated using ReCal 0.1 Alpha, a statistics application on the Internet (Freelon 2010;Freelon 2013), which performed a calculation of average pairwise percent agreement and Cohen's Kappa (Dewey 1983). Cohen's Kappa was designed as a reliability measurement to eliminate the amount that raters may agree by chance alone. Landis and Koch (1977) suggested that Cohen's Kappa coefficients between 0.41-0.60 represent moderate agreement, and coefficients above 0.60 represent substantial agreement. However, other studies suggest greater stringency when interpreting inter-rater and intra-rater reliability coefficients.
Listener intra-rater reliability for the sustained vowel phonation listener preference task was 74%, with an average pairwise Cohen's Kappa of r=0.61. Listener intra-rater reliability for the sentence reading listener preference task was 74%, with an average pairwise Cohen's Kappa of r=0.53. Listener inter-rater reliability for sustained vowel phonation preference at pre-post and pre-FU was 60.5% and Cohen's Kappa was r=0.40. Listener inter-rater reliability for sentence reading preference at pre-post and pre-FU was 74.8% and Cohen's Kappa was r=0.54.
STR03 did not receive any co-occurring speech treatment during the treatment phase of this study. However, he received speech, physical therapy, and occupational therapy 2 days/week for two months following treatment (between post-treatment and 3-month follow-up evaluations).

RESULTS
The results of this study are presented in five categories: treatment data, biweekly probe data, listener-rated communication success (comprehensibility in audio-only and audio+visual conditions and listener perceptual ratings), functional outcome measures (CETI-M, DIP, and interview), and acoustic variables of speech and voice (vowel space area, phonatory stability, and vocal dB SPL).

Treatment Data
The data collected during each treatment session for vocal dB SPL and lip and tongue pressure in kPa were compiled for an average per week to determine trend changes from week 1 to week 6. Lip pressure increased by an average of 0.6 kPa.

Biweekly Probe Data
Vocal dB SPL data were taken during the sentence reading task and the picture description task and compared to the pre-treatment evaluation data for these tasks.
These data demonstrate an overall decrease in vocal dB SPL during the biweekly probes as compared to the pre-treatment evaluation. The summary vocal dB SPL data for biweekly probes during sentence reading and picture description is displayed in Table 2. Sustained vowel phonation from the biweekly probes was analyzed through MDVP and compared to the pre-treatment evaluation data. The data collected at the biweekly probes displayed considerable variability. Few patterns emerged from this data, aside from a considerable decrease in vAM displayed at all three probes.
Summary MDVP data from biweekly probes is displayed in Table 3.

Hypothesis 1: Audio-Only Condition-Single Word Comprehensibility
There was not a statistically significant difference between single word comprehensibility measured during the audio-only condition from pre-to posttreatment (p=0.22). Single word comprehensibility, however, increased significantly from pre-to FU-treatment (p=0.02) with a medium effect size (r=0.58). Quantitative changes of single word percent comprehensibility in the audio-only condition from pre-, post-, and follow-up-treatment evaluations are displayed in Table 4.   Table 5. There was not a statistically significant difference between single word comprehensibility in the video condition from pre-to post-treatment (p=0.32). The difference between pre-and FU-treatment evaluations was also not statistically significant during this condition (p=0.38). Quantitative changes of single word percent comprehensibility in the video condition from pre-, post-, and follow-up treatment evaluations are displayed in Table 6. There was no statistically significant difference between sentence comprehensibility in the video condition from pre-to post-treatment (p=0.25). The difference between pre-and FU-treatment evaluations was also not statistically significant during this condition (p=0.25). Quantitative changes in sentence percent comprehensibility in the video condition from pre-, post-, and FU-treatment evaluations are displayed in Table 7. There was a statistically significant preference for pre-treatment sustained vowel phonation compared with post-treatment sustained vowel phonation (p=0.04) with a large effect size (r=0.80). Table 8 illustrates the individual listener preference ratings including the frequency and magnitude of preference for the pre-treatment sustained vowel phonations. There was a statistically significant preference for post-treatment sentence reading compared with pre-treatment sentence repetitions (p=0.00) with a large effect size of (r=0.98). Table 8 illustrates the individual listener preference ratings for posttreatment sentences.  There was a statistically significant preference for sentence reading at FUtreatment (p=0.02) with a large effect size (r=0.86). Table 11 shows the individual listener preference ratings for FU-treatment samples of sentence repetitions.

Primary Aim: Functional Outcome Measures
Hypothesis 4  "Dysarthria relative to other worries and concerns", in which he is asked to rank his dysarthria within four other personal and health related concerns. STR03 reported that speech was a primary concern during pre-and post-treatment evaluations, but that the  Throughout the treatment, STR03 reported that he required a high level of effort to speak. Discussion about this throughout treatment indicated that he felt as though he would speak more frequently if it didn't require so much effort. His wife also commented during the FU evaluation that the effort level for speaking was STR03's biggest complaint for the three years following his stroke. STR03 temporally located a substantial change in the amount of effort required during his FU evaluation: "the effort has gone away…since we finished."

Vowel Space Analysis
Pre-, post-, and follow-up vowel triangles were obtained by analyzing F1 and F2 values of vowels /u, a, i/ to calculate vowel space area. Vowel space area for pretreatment was 111,645 Hz 2 and 120,150 Hz 2 at post-treatment, indicating an increase of 8,505 Hz 2 (13%). Vowel space area continued to increase at the 3-month follow-up evaluation, to 139,933 Hz 2 , indicating an increase of 28,288 Hz 2 (25%) from pretreatment. Figure 1 is a visual depiction of pre, post-, and follow-up evaluation visual space areas.

Figure 3. Vowel Space Area at Pre-, Post-, and Follow-Up Evaluations
There were statistically significant changes in F1 and F2 values during the post-and FU evaluations. All values for F1 and F2 /u, a, i/ changed significantly at FU. Quantitative changes in F1 and F2 for /u/, /a/, and /i/ are illustrated in Table 12.

Primary Aim, Hypotheses 1-3: Listener-rated communication success
The six-week intensive treatment appeared to be a feasible intervention for increasing listener-rated communication success for the individual in this study. There were increases in the audio-only comprehensibility condition with sentences and single words at post-treatment, but the changes were not statistically significant. The increase in comprehensibility was supported by the listener perceptual study at posttreatment, which revealed a statistically significant preference for post-treatment sentences. However, both sentence length and single word comprehensibility significantly increased from the pre-treatment level to the 3-month follow-up. This demonstrated that the participant continued to make progress following treatment. percentage in this condition may also have made it more difficult to measure a statistically significant improvement. Therefore, the measurement of comprehensibility using visual information plus audio information was a less sensitive measurement of treatment effectiveness than the audio-only condition for this study at the sentence level.
The single word comprehensibility measurement taken at follow-up from the audio+visual condition was 2% greater than that taken during the audio-only condition. Comprehensibility measured during sentence length materials in the audio+visual condition was approximately 0.6% greater at FU than the audio-only condition. Therefore, the treatment had a clinically meaningful effect of increasing understandability using acoustic information alone to a level consistent with communication supported by visual information. This could have a meaningful effect on functional communication and conversation with others, in which visual information is not consistently available such as conversation while driving in a car, conversation on the phone, or conversation while walking/pushing a wheelchair.
Listener perceptual ratings identified significant preference for post-and FUtreatment speech samples when compared to pre-treatment. This indicated that the participant's speech was perceived as more natural following treatment, which was reflected in the increase in listener comprehensibility ratings. However, listener preference for voice quality during sustained phonation was greater at pre-treatment when compared to post-treatment. Due to STR03's baseline increased vocal loudness, it is possible that listener preference ratings for voice quality were related to the statistically significant increases in loudness recorded during post-treatment evaluations. The subsequent decrease in vocal loudness from post-to FU-treatment coincided with the preference for vocal quality in FU sustained vowel phonation, and an increase in comprehensibility ratings at FU-treatment. Intra-rater and inter-rater reliability for listener perceptual rating tasks were challenges in this study. Listeners demonstrated moderate-substantial intra-rater reliability for perception of voice in sustained vowel phonation and speech naturalness in sentences. Listeners demonstrated weak-moderate inter-rater reliability for perceptual ratings of voice and speech, respectively. The listener perceptual rating task was subjective, and listeners demonstrated poorer reliability with rating voice quality in sustained vowel phonations than rating speech naturalness in sentence reading. These challenges with reliability highlight the difficulty with using perceptual measures as treatment effectiveness variables.

Primary Aim, Hypotheses 4-6: Functional Outcome Measures
The participant and his spouse reported increases in communicative effectiveness and decreases in the psychosocial impacts associated with dysarthria following treatment. STR03 and his wife reported increases in the quantity of information he provided in conversation, and the frequency with which he contributed.
The six-week intensive treatment appeared to provide social stimulation, practice with specific speech tasks, and a decreased level of effort necessary for speech, which

4.3c Vocal dB SPL
The treatment had a statistically significant effect of increasing vocal dB SPL for speech tasks including sentence reading and sustained vowel phonation. These increases displayed large effect sizes. This finding was unexpected because increased loudness was not directly trained during treatment. STR03 presented with loudness levels greater than normal limits at pre-treatment, which was consistent with his diagnosis of spastic dysarthria. Decreased loudness and easy-onset of phonation during treatment tasks was modeled, but not directly stated throughout treatment to preserve the singular cue for "clear speech".
STR03 presented with laryngeal tension, and severe strain-strangled voice quality, which likely contributed to greater loudness during speech tasks. He received cues to bring the effort to his lips and tongue, and away from his throat during treatment tasks.
He frequently produced several utterances following cueing with reduced vocal loudness, but did not achieve independence from this cue during the treatment course.
The increased vocal dB SPL level during the post-treatment is consistent with continued dependence on cues for decreased loudness in the presence of the high effort training necessary for clear speech.
The decreased loudness at the 3-month FU back to baseline level suggests that

Limitations
There are inherent limitations related to single-subject research designs.
Findings are specific to the individual, and therefore, cannot be further generalized to other individuals with the same disorder. However, positive treatment results can provide a rationale for future investigatory research. Inherent small sample sizes in data collection for a single-subject design relate to challenges with internal validity.

Vowel Space Analysis
Paragraph Reading Participant will read through the Farm Passage, (Crystal & House, 1982) Loudness data and specific sound errors Picture Descriptions Participant will describe the picnic scene from the Western Aphasia Battery (WAB), (Kertesz, 1982) Task Description Patient will describe how to do a stated task (e.g. "Describe how to make a peanut butter and jelly sandwich") Hearing in noise test Participant will repeat a series of sentences (Nilsson, 1994) Sentence-level speaker intelligibility and comprehensibility Single word reading Participant will read through a series of 70 single words (Kent et al., 1989)

STR03 IOPI Lip and Tongue Pressure
Lip Tongue