• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

Visual reinforcement infant speech discrimination: Developing a method of performance analysis

Dissertation
Author: Tammy Fredrickson
Abstract:
While measures of speech perception are an important aspect of audiological assessment and validation of amplification fitting in older children and adults, no clinical method of speech perception assessment exists for infants and toddlers. Visual Reinforcement Infant Speech Discrimination (VRISD) is a tool that has been used to assess infant speech perception in studies for over 30 years and has been deemed to have potential for clinical use. Unfortunately, the foundational work to provide an appropriate protocol for VRISD's clinical use does not exist; the reliability and validity of VRISD have not been studied. In its current research form, VRISD consists of 30 trials and is designed not as an assessment of an individual child's ability to discriminate specific phoneme or consonant vowel comparisons, but as a technique to compare the mean performance of a group of children with another - by age, by native language, or to determine whether one comparison is more difficult than another. This dissertation aimed to begin investigation into the foundational work necessary to help VRISD become a clinical tool. VRISD was used to assess 15 normal-hearing infants' abilities to discriminate three speech sound contrasts (a/i, ba/pa, and da/ga). Results of infants. performance on these three contrasts indicated a hierarchy of difficulty, with the a/i contrast being the easiest and the da/ga contrast being the most difficult, as hypothesized. Results from each of the contrasts tested were analyzed in six different methods - five of which have been used in previously published VRISD studies - so that conclusions made from the different methods could be compared. The different methods of analysis sometimes led to differing conclusions as to the ability of an infant to discriminate a contrast. It was determined that the use of a criterion based on binomial probability was the best way to analyze performance in a manner that would yield reliable results as well as provide construct validity. The current study revealed that individual children have performance profiles that indicate that they do not perform consistently across 30 trials, particularly when they demonstrate mastery or correct performance consistently over the first 10-12 trials. This inconsistent performance has previously resulted in a significant number of children who have been unable to complete the task or whose performance is correct in the beginning of testing and then becomes incorrect, possibly because of habituation or boredom. There has previously been no universally accepted and well-defined criterion for establishing whether an individual child has mastered a specific discrimination task. This dissertation compares six different methods of determining VRISD performance. A criterion-based performance measure using binomial probability provides the best technique for construct validity.

TABLE OF CONTENTS

Table of Tables………………………………………………………………. xv Table of Figures……………………………………………………………… xvi Chapter 1: Introduction……………………………………………………… 1 Chapter 2: Review of the Literature……………………………………….. 4 Methods of Infant Speech Perception Assessment…………..…. 5 High Amplitude Sucking ................................................... 5 Visual Habituation ............................................................ 6 Observer-based Psychoacoustic Procedure ................... 8 Visual Reinforcement Infant Speech Discrimination ...... 11 Methods of Performance Analysis ............................................. 18 Percent Correct .............................................................. 18 Proportion Correct ......................................................... 22 Discrimination Index ...................................................... 25 Criterion ......................................................................... 30 Signal Detection Theory ................................................ 34 Comparison of Analysis Methods .................................. 38 Chapter 3: Methods……………………………………………………….... 47 Simulated Data……………………………………………………… 47 Subjects……………………………………………………………… 61 Stimuli…………………………………………………………………62 Calibration…………………………………………………………….64

viii

Instrumentation and Test Environment…………………………… 65 VRISD testing………………………………………………………... 65 Training……………………………………………………………….. 67 Testing………………………………………………………………… 68 Analysis………………………………………………………………. 69 Summary of Methods……………………………………………….. 69 Chapter 4: Results…………………………………………………………… 71 How do normal hearing infants perform on the three contrasts?.. 72 What is the best measure for a criterion-based stopping rule?.... 78 Individual Results and Comparisons ................................. 97 Chapter 5: Discussion………………………………………………………. 106 References………………………………………………………………...…. 122

ix

Appendix A: Subject 1……………………………………………………….130 Table 1. Subject 1‟s performance on each contrast .....................130 Figure 1. Performance on a/i – SDT ........................................... 131 Figure 2. Performance on ba/pa – SDT ...................................... 132 Figure 3. Performance on da/ga – SDT ...................................... 133 Figure 4. Cumulative distribution function – a/i ............................ 134 Figure 5. Cumulative distribution function – ba/pa ....................... 134 Figure 6. Cumulative distribution function – da/ga ....................... 135 Appendix B: Subject 2……………………………………………………….. 136 Table 1. Subject 2‟s performance on each contrast ..................... 136 Figure 1. Performance on a/i – SDT ........................................... 137 Figure 2. Performance on ba/pa – SDT ...................................... 138 Figure 3. Performance on da/ga – SDT ...................................... 139 Figure 4. Cumulative distribution function – a/i ............................ 140 Figure 5. Cumulative distribution function – ba/pa ....................... 140 Figure 6. Cumulative distribution function – da/ga ....................... 141 Appendix C: Subject 3………………………………………………………. 142 Table 1. Subject 3‟s performance on each contrast ..................... 142 Figure 1. Performance on a/i – SDT ........................................... 143 Figure 2. Performance on ba/pa – SDT ...................................... 144 Figure 3. Performance on da/ga – SDT ...................................... 145 Figure 4. Cumulative distribution function – a/i ............................ 146 Figure 5. Cumulative distribution function – ba/pa ....................... 146

x

Figure 6. Cumulative distribution function – da/ga ....................... 147 Appendix D: Subject 4………………………………………………………. 148 Table 1. Subject 4‟s performance on each contrast ..................... 148 Figure 1. Performance on a/i – SDT ............................................. 149 Figure 2. Performance on ba/pa – SDT ........................................ 150 Figure 3. Performance on da/ga – SDT ........................................ 151 Figure 4. Cumulative distribution function – a/i ............................. 152 Figure 5. Cumulative distribution function – ba/pa ........................ 152 Figure 6. Cumulative distribution function – da/ga ........................ 153 Appendix E: Subject 5………………………………………………………....154 Table 1. Subject 5‟s performance on each contrast .......................154 Figure 1. Performance on a/i – SDT ............................................. 155 Figure 2. Performance on ba/pa – SDT ........................................ 156 Figure 3. Performance on da/ga – SDT ........................................ 157 Figure 4. Cumulative distribution function – a/i ............................. 158 Figure 5. Cumulative distribution function – ba/pa ........................ 158 Figure 6. Cumulative distribution function – da/ga ........................ 159 Appendix F: Subject 6………………………………………………………... 160 Table 1. Subject 6‟s performance on each contrast ........................160 Figure 1. Performance on a/i – SDT .............................................. 161 Figure 2. Performance on ba/pa – SDT ........................................ 162 Figure 3. Performance on da/ga – SDT ........................................ 163 Figure 4. Cumulative distribution function – a/i ............................. 164

xi

Figure 5. Cumulative distribution function – ba/pa ....................... 164 Figure 6. Cumulative distribution function – da/ga ....................... 165 Appendix G: Subject 7………………………………………………………. 166 Table 1. Subject 7‟s performance on each contrast .......................166 Figure 1. Performance on a/i – SDT ............................................. 167 Figure 2. Performance on ba/pa – SDT ........................................ 168 Figure 3. Performance on da/ga – SDT ........................................ 169 Figure 4. Cumulative distribution function – a/i ............................. 170 Figure 5. Cumulative distribution function – ba/pa ........................ 170 Figure 6. Cumulative distribution function – da/ga ........................ 171 Appendix H: Subject 8……………………………………………………….. 172 Table 1. Subject 8‟s performance on each contrast ...................... 172 Figure 1. Performance on a/i – SDT ............................................. 173 Figure 2. Performance on ba/pa – SDT ........................................ 174 Figure 3. Performance on da/ga – SDT ........................................ 175 Figure 4. Cumulative distribution function – a/i ............................. 176 Figure 5. Cumulative distribution function – ba/pa ........................ 176 Figure 6. Cumulative distribution function – da/ga ........................ 177 Appendix I: Subject 9……………………………………………………….... 178 Table 1. Subject 9‟s performance on each contrast ...................... 178 Figure 1. Performance on a/i – SDT ............................................. 179 Figure 2. Performance on ba/pa – SDT ........................................ 180 Figure 3. Performance on da/ga – SDT ........................................ 181

xii

Figure 4. Cumulative distribution function – a/i ............................. 182 Figure 5. Cumulative distribution function – ba/pa ........................ 182 Figure 6. Cumulative distribution function – da/ga ........................ 183 Appendix J: Subject 10…………………………………………………….... 184 Table 1. Subject 10‟s performance on each contrast .................... 184 Figure 1. Performance on a/i – SDT ............................................. 185 Figure 2. Performance on ba/pa – SDT ........................................ 186 Figure 3. Performance on da/ga – SDT ........................................ 187 Figure 4. Cumulative distribution function – a/i ............................. 188 Figure 5. Cumulative distribution function – ba/pa ........................ 188 Figure 6. Cumulative distribution function – da/ga ........................ 189 Appendix K: Subject 11…………………………………………………….... 190 Table 1. Subject 11‟s performance on each contrast .................... 190 Figure 1. Performance on a/i – SDT ............................................. 191 Figure 2. Performance on ba/pa – SDT ........................................ 192 Figure 3. Performance on da/ga – SDT ........................................ 193 Figure 4. Cumulative distribution function – a/i ............................. 194 Figure 5. Cumulative distribution function – ba/pa ........................ 194 Figure 6. Cumulative distribution function – da/ga ........................ 195 Appendix L: Subject 12……………………………………………………..... 196 Table 1. Subject 12‟s performance on each contrast .................... 196 Figure 1. Performance on a/i – SDT ............................................. 197 Figure 2. Performance on ba/pa – SDT ........................................ 198

xiii

Figure 3. Performance on da/ga – SDT ........................................ 199 Figure 4. Cumulative distribution function – a/i ............................. 200 Figure 5. Cumulative distribution function – ba/pa ........................ 200 Figure 6. Cumulative distribution function – da/ga ........................ 201 Appendix M: Subject 13……………………………………………………... 202 Table 1. Subject 13‟s performance on each contrast ................... 202 Figure 1. Performance on a/i – SDT ............................................. 203 Figure 2. Performance on ba/pa – SDT ........................................ 204 Figure 3. Performance on da/ga – SDT ........................................ 205 Figure 4. Cumulative distribution function – a/i ............................. 206 Figure 5. Cumulative distribution function – ba/pa ........................ 206 Figure 6. Cumulative distribution function – da/ga ........................ 207 Appendix N: Subject 14……………………………………………………... 208 Table 1. Subject 14‟s performance on each contrast ................... 208 Figure 1. Performance on a/i – SDT ............................................. 209 Figure 2. Performance on ba/pa – SDT ........................................ 210 Figure 3. Performance on da/ga – SDT ........................................ 211 Figure 4. Cumulative distribution function – a/i ............................. 212 Figure 5. Cumulative distribution function – ba/pa ........................ 212 Figure 6. Cumulative distribution function – da/ga ........................ 213 Appendix O: Subject 15……………………………………………………... 214 Table 1. Subject 15‟s performance on each contrast .................... 214 Figure 1. Performance on a/i – SDT ............................................. 215

xiv

Figure 2. Performance on ba/pa – SDT ........................................ 216 Figure 3. Performance on da/ga – SDT ........................................ 217 Figure 4. Cumulative distribution function – a/i ............................. 218 Figure 5. Cumulative distribution function – ba/pa ........................ 218 Figure 6. Cumulative distribution function – da/ga ........................ 219 Appendix P: Pseudo Subject Responses Table 1. Simulated trial-by-trial response data for 15 pseudo subjects............................................................................... 220

xv

LIST OF TABLES

Table 1. Performance analysis results for three subjects….…………..… 38 Table 2. 2 x 2 contingency table for VRISD ……………………………….. 50 Table 3. Pseudo-subject 7 performance .................................................. 53 Table 4. Pseudo-subject 4 performance …………………………………… 55 Table 5. Pseudo-subject 15 performance ………………………….……… 56 Table 6. Pseudo-subject 5 performance …………………………………… 57 Table 7. Pseudo-subject 12 performance ………………………………….. 58 Table 8. Pseudo-subject 6 performance ……………………….……….……59 Table 9. Pseudo-subject 9 performance ………………………………….… 60 Table 10. Mean group performance for three contrasts …………….………73 Table 11. Correlations between analysis measures – a/i …………….….…80 Table 12. Correlations between analysis measures – ba/pa ……….….…. 82 Table 13. Correlations between analysis measures – da/ga ……….….…. 85 Table 14. Mean group performance on each of the three contrasts .…..… 87 Table 15. Number of subjects who reached binomial probability criterion . 94 Table 16. Individual performance for a/i contrast ………………………..… 99 Table 17. Individual performance for ba/pa contrast ……………………… 100 Table 18. Individual performance for da/ga contrast ……………………… 101 Table 19. Minimum and maximum number of trials necessary to demonstrate successful discrimination ……………………………………………. 104

xvi

LIST OF FIGURES Figure 1. Signal Detection Theory sensitivity index………………………… 35 Figure 2. Pseudo-subject 7 performance ................................................... 53 Figure 3. Pseudo-subject 4 performance ……………….……………..……. 54 Figure 4. Pseudo-subject 15 performance ………………………………….. 56 Figure 5. Pseudo-subject 5 performance ………………….………………… 57 Figure 6. Pseudo-subject 12 performance …………………….…………….. 58 Figure 7. Pseudo-subject 6 performance …………………….……………… 59 Figure 8. Pseudo-subject 9 performance ……………………………….…… 60 Figure 9. Sound booth configuration ………………………….……………… 66 Figure 10. Individual subject performance (percent correct) on a/i …..….. . 74 Figure 11. Individual subject performance (percent correct ) on ba/pa …….75 Figure 12. Individual subject performance (percent correct) on da/ga ….….76 Figure 13. Performance – percent correct vs binomial probability criterion (a/i)………………………………………………………………………….89 Figure 14. Performance – percent correct vs binomial probability criterion (ba/pa) ………………………………………………………………….… 91 Figure 15. Performance – percent correct vs binomial probability criterion (da/ga) ………………………………………………………………….… 93 Figure 16. Subject performance (a/i - binomial probability criterion) ..…….. 95 Figure 17. Subject performance (ba/pa – binomial probability criterion)….. 96 Figure 18. Subject performance (da/ga – binomial probability criterion) ….. 97 Figure 19. Subject 4 performance on da/ga ………………………………… 102

xvii

Figure 20. Subject 6 performance on da/ga ………………………………… 103 Figure 21. Infant with auditory dys-synchrony – cumulative distribution function (a/i)…………………………………… ………………………. 116 Figure 22. Infant with hearing loss (aided) – binomial probability and percent correct (a/i) …………………………………………………………….. 117

1

CHAPTER 1: INTRODUCTION Universal newborn hearing screening has become commonplace in the United States and other areas of the world over the last 10 years. According to the National Center for Hearing Assessment and Management, 42 states (as well as the District of Columbia and Puerto Rico) currently have laws or statutes that make newborn hearing screening mandatory ("EHDI Legislation," 2009). The Centers for Disease Control and Prevention reports that 94% of newborns in the United States and its territories received hearing screenings in 2007 ("Early Hearing Detection and Intervention (EHDI) Program," 2010). Based on data reported, approximately 6% of those infants screened in 2007 were diagnosed with permanent hearing loss (1.2 per 1000 babies screened). Those infants identified with permanent hearing loss are fit with amplification earlier than was possible prior to newborn hearing screening, with an average age of identification of 3 months or lower in many states ("Summary of Infants Diagnosed Before 3 Months of Age ( Year 2007)," 2009). While validation of the fitting of amplification in older children and adults can be made using both Real Ear measurements and speech perception measures, infants and young children are fit with amplification through prescriptive formulas and validation of fitting is evaluated through Real Ear measurements (typically Real Ear to Coupler (RECD) measurements) and observations of their behavior. There is no standard method of speech perception assessment for this population that has been

2

incorporated into the clinical battery of audiometric testing. Typically, infants‟ and young children‟s amplification benefit is assessed via methods such as clinical observation, behavioral observation audiometry, visual reinforcement audiometry and/or subjective parent questionnaires such as the Infant- Toddler Meaningful Auditory Integration Scale (IT-MAIS) (Zimmerman- Phillips, McConkey Robbins, & Observer, 2001; Zimmerman-Phillips, Osberger, & Robbins, 1997) and the Parents‟ Evaluation of Aural/Oral Performance of Children (PEACH) (Ching & Hill, 2007). These questionnaires are completed by the infant‟s parents and those who work closely with the infant. Unfortunately, these methods do not yield objective information regarding the speech perception abilities of these infants and none of these measures have standardized or validated results. These questionnaires cannot provide specific information regarding the infant‟s ability to discriminate one phoneme from another. Because these questionnaires address an infant‟s perception in his/her everyday environment, there is an inability to control auditory parameters of the stimuli, such as intensity, duration, speech versus non-speech, or to allow discrete comparisons of one auditory stimulus as compared to another. Therefore, there is a critical need for the development of a standard measure for speech perception in infants. Currently, standardized speech perception tasks involve real words and therefore, cannot be used until the child has acquired language at levels appropriate for a typically developing two to three year old child. Objective assessment of an infant‟s ability to

3

discriminate speech when he/she has a significant hearing loss could yield additional information about the infant‟s overall perceptual development as well as the performance of his/her amplification. An objective measure of infant speech perception abilities could help to verify that an infant‟s amplification provides the sound necessary for the development of spoken language. Results obtained using infant speech perception measures could also help to direct habilitation efforts as well as monitor those efforts.

4

CHAPTER 2: REVIEW OF THE LITERATURE Speech perception skills in infants who have normal hearing and are typically developing have been studied in depth for decades. Most of this research has been done in the field of linguistics to assess the development of speech perception and infants‟ ability to detect particular acoustic cues, comparing these abilities with those of adults (e.g., Aslin, 1981; Bull, Eilers, & Oller, 1984; Eilers, Bull, Oller, & Lewis, 1984; Eilers, Morse, Gavin, & Oller, 1981). Various methods have been used to assess speech perception abilities in normal hearing infants. Researchers have been interested in learning about the speech perception abilities of prelinguistic children as a way to learn about early language development and have compared the abilities of children and adults to determine when particular abilities develop. Methods used to assess speech perception in prelinguistic children include high amplitude sucking (HAS) (e.g.,Eimas, Siqueland, Jusczyk, & Vigorito, 1971; Karzon, 1982), visual habituation (VH) (e.g., Bornstein & Benasich, 1986; Houston, Pisoni, Iler Kirk, Ying, & Miyamoto, 2003; Kaplan & Werner, 1986), Observer-Based Psychoacoustic Procedure (OPP) (e.g., Olsho, 1987), and Conditioned Head Turn (CHT), also referred to as Visual Reinforcement Infant Speech Discrimination (VRISD) (e.g., Eilers, Wilson, & Moore, 1977; Werker, Polka, & Pegg, 1997). These methods were primarily developed to allow investigators to examine early linguistic development, but they have also been used to measure thresholds. The methods were constructed to

5

permit investigation into the perceptual abilities of infants at different ages and stages of development. The following sections examine each of these techniques and provide examples of how they have been used in research (also see Jusczyk & Luce, 2002, for a review). High Amplitude Sucking – The high amplitude sucking (HAS) habituation procedure (e.g., Eimas et al. 1971) is a method that can be used with very young infants, even newborns, to investigate babies‟ responses to changes in sound. With this method, babies hear a repeating speech sound while sucking on a pacifier that is connected electronically to a computer so that the rate of sucking behavior can be measured. Babies tend to start sucking at a high rate while hearing the repeating sound. As their rate of sucking decreases (i.e., indicating that the baby has habituated to the sound), infants in the experimental group will hear a new speech sound. The infants in the control group continue to hear the initial sound. If an increase in sucking rate is seen in the experimental group relative to the control group, it is said that discrimination has occurred. This is a demonstration of a reflexive response; no conditioning to the task is needed. HAS has been used fairly successfully to provide valuable information about speech perception in very young infants. Data obtained via HAS is group data, therefore information about the abilities of individual infants cannot be obtained. HAS cannot address questions of audibility nor can it provide information as to how difficult a particular contrast may be for a particular infant. In addition, the attrition rate for the HAS procedure is usually

6

quite high, sometimes as high as 75% (e.g., Karzon, 1982). Also, typically no more than one speech sound contrast can be assessed per measurement period. Given these drawbacks, it is unlikely that HAS will become a part of a clinical battery to assess discrimination of speech in infants. Visual Habituation – Visual habituation (VH) is a cognitive task that measures an infants‟ interest in sounds. This task has been used with infants ages 3 to 29 months. In a VH task, visual stimuli are presented via television screens or computer monitors. The monitors are placed on either side of the infant. Images such as checkerboards are displayed and are made to change in luminescence during a trial. This purpose of this change is to keep the infants‟ attention to the screens longer. Auditory stimuli are presented via a transducer such as a loudspeaker once the child begins fixating on a monitor. As long as the infant continues to look toward a monitor, the stimuli continue to play. The stimuli stop playing when the infant habituates to the sound and looks away from the monitor. The total amount of time the infant looks toward the monitor is recorded and is measured as an index of interest in the auditory stimuli. Familiarization trials occur first, during which time the infants‟ interest in the auditory stimuli decreases; the infant becomes habituated to the sound. Presentation with a novel stimulus, theoretically, results in prolongation of infants‟ visual attention toward the monitor. It is assumed that an infant will habituate to a frequent sound more quickly than he/she will habituate to a novel sound. Visual habituation assesses an infants‟ interest in novel stimuli by measuring the difference in looking time between the more

7

common stimulus and the novel one. Performance of an experimental group (babies who heard the novel stimuli) is compared with the performance of a control group (babies who hear only one stimulus), therefore, information regarding an individual infant‟s performance cannot be obtained. VH has been used to assess infants‟ linguistic development and ability to differentiate suprasegmental information (e.g., Houston et al., 2003; Spence & Moore, 2003). This technique is a cognitive skill that measures an infant‟s interest in particular stimuli, not whether or not it can be heard and/or distinguished. VH is not necessarily assessing discrimination of speech sounds. It is important to note the attrition rate for VH tends to be variable and high, ranging from 17 to 62% (e.g., Miller & Eimas, 1996; Werker & McLeod, 1989). Methodological differences among the studies probably account for much of this variability in attrition rate. If VH is to be used in the clinical setting to assess prelinguistic cognitive perception, the attrition rate needs to reliably be less than 30% (as suggested for Visual Reinforcement Audiometry; Gravel, 1989). Also, in order to measure the duration of looking time, test sessions typically are video recorded and analysis of looking time and preference occurs after the test session is over. Without instant feedback regarding the infant‟s performance, changes in procedure that may have an immediate impact on performance cannot be implemented (i.e., re-training, adjusting stimulus intensity levels), which may result in a greater attrition rate. Clinical use of a tool that requires

8

analysis after the test session and has a high attrition rate is not practical in today‟s fast-paced clinical pediatric setting. Observer-based Psychoacoustic Procedure -- The Observer-based Psychoacoustic Procedure (OPP) (Olsho, Koch, Halpin, & Carter, 1987) was developed for use with infants 2 to 12 months of age. OPP combines features of behavioral observation audiometry (BOA) and visual reinforcement audiometry (VRA). OPP brings in the element of conditioning a response from the infant and, thus, turns this into a test of prelinguistic sensory perception. Typically, a conditioned response (i.e., a head-turn) cannot be elicited reliably until approximately 6 months of age. OPP permits investigations of auditory perception in younger infants that has otherwise not been possible. In this procedure, the infant sits on a parent‟s lap while sounds are presented through earphones to the infant. An assistant, who serves as a distracter by engaging the infant with toys, is seated on one side of the parent and infant. On the opposite side of the parent and infant is a mechanical toy visual reinforcer (the same as those used in visual reinforcement audiometry). An observer is seated in a control room and observes the infant in order to make judgments about whether or not a target sound was presented based on the infant‟s behaviors and responses. When the infant is engaged with a toy, the observer initiates a trial. Two types of trials may occur: a signal trial or a no-signal trial. Signal trials are characterized by the presence of target stimuli. Target stimuli may include the onset of auditory stimuli or a change of auditory stimuli, depending on the

9

auditory skill being examined (i.e., detection or discrimination). No-signal trials are characterized by the absence of target stimuli (e.g. absence of sound or no change in stimuli). The observer, who is blinded to trial type by use of headphone sound attenuators or masking music, must decide if they believe the trial that has just occurred is indeed a signal or a no-signal trial based on the infant‟s behavior. If the observer correctly identifies that a signal trial has occurred, the toy reinforcer is activated. In contrast, if the observer incorrectly judges that a signal has occurred, the toy reinforcer is not activated. Results of each trial are displayed by the computer software, providing the observer immediate feedback. For this procedure to be successful observers must undergo training, which was designed by the developers of this procedure to take approximately one month (Olsho et al., 1987). During training the observer becomes familiar with what types of responses constitute infant reactions to sounds and sound changes. These reactions may range from a head turn toward the stimulus to an eye blink or change in facial expression. Thus, OPP is similar to BOA in that it does not require that the child make a specific type of response (i.e., head turn), but allows for a variety of responses. However, OPP differs from BOA in that the observer is blinded to trial type and a toy reinforcer is used. One limitation of OPP is a high false alarm rate. Behaviors such as an eye blink may be difficult to “read” as a response to a sound. Olsho et al. (1987) used a computer program that provided the observer feedback should

10

his/her false alarm rate rise above 25%. The observer then either stops the session entirely or begins retraining. After retraining, the observer‟s false alarm rate must not rise above 20% or testing is terminated. Olsho et al. (1987) reported 16% of sessions must be excluded due to a high false alarm rate on the part of the observer. The ability to assess younger infants and thereby develop a better understanding of infant speech perception is a strong advantage of OPP. One benefit of OPP over VH for use in this young population is that OPP is a task of prelinguistic auditory perception rather than a prelinguistic cognitive task. OPP allows the examination of both auditory detection and discrimination. Concerns raised when discussing the clinical utilization of OPP include observer training and its attrition rate. The length of time it takes to train observers (one month as suggested by Olsho et al., 1987) is typically not clinically feasible. Another disadvantage of OPP is its high attrition rate. A study conducted by Marean et al. (1992) investigated 2- and 3-month old infants‟ categorization of vowel sounds using the OPP procedure. Attrition rates for this study ranged from 29% (for 3 month old infants) to 53% (for 2 month old infants). This high attrition rate is a potential problem for the clinical usefulness of OPP, as it appears that “meaningful” results will be obtained on fewer than half of those infants who are tested. The attrition rates reported by Marean et al. were encountered in studies using normal hearing infants – one can anticipate an even higher attrition rate for children

11

with hearing loss. In a fast-paced clinical setting, this will result in a lot of time spent on a test that yields few results. Visual Reinforcement Infant Speech Discrimination – Visual reinforcement infant speech discrimination (VRISD) is a variation of the commonly used VRA technique for assessing hearing sensitivity in infants and toddlers. Also known as the “conditioned head turn procedure” (CHT), VRISD was first developed by Eilers, Wilson, and Moore (1977) and modified in 1990 by Nozza. It has been proposed as being potentially useful in a clinical setting (Nozza, Miller, Rossman, & Bond, 1991). It has been used with normal hearing infants from 6 to 30 months of age with success, even in a clinical setting (Gravel, 1989). VRISD is a test of prelinguistic auditory perception, as is OPP, allowing for investigators to determine not only if a sound was heard, but also if it was distinguished from another sound. VRISD results have traditionally been analyzed as group data, analyzing differences in mean group performance on a particular task or contrast. VRISD testing is based on an oddball paradigm where one speech sound serves as the background repeating stimuli, while another is presented periodically with the expectation that the child will detect this difference in sounds. One sound from each pair serves as the repeating background stimuli, while the other sound serves as the change or target stimuli. The child sits on a parent‟s lap or in a high chair in the center of the test room. An assistant, who serves as a distracter, sits directly in front of the child. Stimuli are delivered in the soundfield or through earphones. Trials are initiated by

Full document contains 239 pages
Abstract: While measures of speech perception are an important aspect of audiological assessment and validation of amplification fitting in older children and adults, no clinical method of speech perception assessment exists for infants and toddlers. Visual Reinforcement Infant Speech Discrimination (VRISD) is a tool that has been used to assess infant speech perception in studies for over 30 years and has been deemed to have potential for clinical use. Unfortunately, the foundational work to provide an appropriate protocol for VRISD's clinical use does not exist; the reliability and validity of VRISD have not been studied. In its current research form, VRISD consists of 30 trials and is designed not as an assessment of an individual child's ability to discriminate specific phoneme or consonant vowel comparisons, but as a technique to compare the mean performance of a group of children with another - by age, by native language, or to determine whether one comparison is more difficult than another. This dissertation aimed to begin investigation into the foundational work necessary to help VRISD become a clinical tool. VRISD was used to assess 15 normal-hearing infants' abilities to discriminate three speech sound contrasts (a/i, ba/pa, and da/ga). Results of infants. performance on these three contrasts indicated a hierarchy of difficulty, with the a/i contrast being the easiest and the da/ga contrast being the most difficult, as hypothesized. Results from each of the contrasts tested were analyzed in six different methods - five of which have been used in previously published VRISD studies - so that conclusions made from the different methods could be compared. The different methods of analysis sometimes led to differing conclusions as to the ability of an infant to discriminate a contrast. It was determined that the use of a criterion based on binomial probability was the best way to analyze performance in a manner that would yield reliable results as well as provide construct validity. The current study revealed that individual children have performance profiles that indicate that they do not perform consistently across 30 trials, particularly when they demonstrate mastery or correct performance consistently over the first 10-12 trials. This inconsistent performance has previously resulted in a significant number of children who have been unable to complete the task or whose performance is correct in the beginning of testing and then becomes incorrect, possibly because of habituation or boredom. There has previously been no universally accepted and well-defined criterion for establishing whether an individual child has mastered a specific discrimination task. This dissertation compares six different methods of determining VRISD performance. A criterion-based performance measure using binomial probability provides the best technique for construct validity.