AUTISM SPECTRUM DISORDER By AMANDA B. BROWN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
2014 Amanda B. Brown
To my husband, for your steadfast love, unfailing support, and willing sacrifices To my parents, for your unconditional love, Godly guidance, and high expectations To my sisters, for always believing in me To my nieces and nephews for the joy you bring to my life To my Quinnie, for the inspiration to begin and continue this journey.
4 ACKNOWLEDGMENTS I would like to acknowledge the members of my committee who gave freely of their time and knowledge. Dr. Conroy gave me perspective about autism from outside healthcare, Dr. Rodriguez shared her ext ensive knowledge of testing pain instrumentation and Dr. Pieper Dr. Schaffer stepped up in my final semesters, and her excitement about my research re energized me. Dr. Elder, my committee chair, guided me and helped me succeed while allowing me the freedom to grow. She had a gentle, yet persuasive, way of pushing me to achieve outside of the degree requirements. She was always available and always patient. I wo uld also like to thank the many teachers over the years that have encouraged me in my educational journey; they include: Mr. Porter, Mr. Arey, Dr. Graham, Dr. Loriz, Dr. Coumeaux, Dr. Radjenovich, and Dr. Behar Horenstein. The flexibility understanding and support I received from Cyndy Jackman, Dr. Pamela Turner and Dr. Carolyn Johnson made it possible for me to work full time and complete my degree. I am indebted to each of you. I must thank my 4 Wolfson nurses also. Each and every day they provide top quality, loving care to our local children with a special sense of humor. We have shared stories of triumph and tears of sorrow. Their promises to call me Dr. Brown and their faith in me to know the answer always drives me to continue to learn degree is no exception. My parents have always set the bar high and equipped me to succeed and I owe who I am today to them. My sisters have been a source of healthy compet ition, friendship, and love; we have a bond that cannot be broken. My grandparents, even though they have left this world, continue to inspire and encourage me to this day. My nieces and nephews are remarkable and unique ; they bring a special joy to my l ife. My nephew Quinnie was my
5 inspiration to embark on this endeavor and h e continued to be a source of inspiration throughout my journey W henever I would get discouraged I would think of him and his daily struggle with autism. I started and continued t his journey because I hope to impact his life and the life of other children with autism in some meaningful way. I have a group of friends that are like family. They have tried their very hardest to understand why I returned to school yet again. They hav e been patient, empathetic and encouraging throughout this process. They have provided much needed stress relief and I cannot thank them enough. A special thank you to each of the children with autism and their families, who allowed me to film an intima te moment in their lives and share it with others in the hope that we can provide better care to other children with autism. Thank you to all of the nurses who eagerly participated. I am appreciative of the funding I received from the American Nurses Foun dation and the Society of Pediatric Nurses. Finally, I have to thank my husband, Kenny. He is the constant, calming force that balances my driven, frantic one I love him more every day and I share this degree with him.
6 TABLE OF CONTENTS P age ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 12 Purpose and Research Question ................................ ................................ .............................. 14 De finition of Terms ................................ ................................ ................................ ................ 15 Significance ................................ ................................ ................................ ............................ 15 Limitations ................................ ................................ ................................ .............................. 15 2 REVIEW OF LITERATURE ................................ ................................ ................................ 17 Autism Spectrum Disorder ................................ ................................ ................................ ..... 17 Theories of ASD Brain Function ................................ ................................ ..................... 19 Sensory Impairments ................................ ................................ ................................ ....... 20 Communication and Social Impairments ................................ ................................ ........ 21 Development of communication and social skills ................................ .................... 21 Behavioral characteristics ................................ ................................ ........................ 22 Social interaction ................................ ................................ ................................ ...... 24 Repetitive Patterns of Behavior ................................ ................................ ....................... 24 Physical Differences ................................ ................................ ................................ ........ 25 Pain in Children ................................ ................................ ................................ ...................... 25 Pain Theories ................................ ................................ ................................ ................... 25 Neuromatrix Theory of Pain ................................ ................................ .................... 25 Model of Pain Expression ................................ ................................ ........................ 26 ................................ ... 28 Prescriptive Theory of Acute Pain Management in Infants and Children ............... 29 Comparison of the four theories ................................ ................................ ............... 31 Rationale for selection of model ................................ ................................ .............. 32 Pain Assessment in Children ................................ ................................ ........................... 33 Self report measures ................................ ................................ ................................ 33 Behavioral measures ................................ ................................ ................................ 35 Pain and pai n assessment in children with special needs ................................ ......... 37 Pain and pain assessment in children with ASD ................................ ...................... 38 Video Recording in Research ................................ ................................ ................................ 40
7 Ethics in Research of Vulnerable Populations ................................ ................................ ........ 43 Respect for Autonomy and Informed Consent ................................ ................................ 44 Justice and Research Participation ................................ ................................ .................. 46 Risk Minimization ................................ ................................ ................................ ........... 47 Summary ................................ ................................ ................................ ................................ 48 3 METHODS ................................ ................................ ................................ ............................. 52 Setting ................................ ................................ ................................ ................................ ..... 52 Measurement Concepts ................................ ................................ ................................ ........... 52 Inter rater Reliability ................................ ................................ ................................ ....... 53 Construct Validity ................................ ................................ ................................ ........... 55 Sample Size Determination ................................ ................................ ............................. 57 Diagnostic Instrumentation ................................ ................................ ................................ ..... 58 Pain Assessment Instruments ................................ ................................ ................................ 59 NCCP C PV ................................ ................................ ................................ ...................... 59 Revised FLACC ................................ ................................ ................................ .............. 60 UWCH PSPNC ................................ ................................ ................................ ................ 61 Participants and Recruitment ................................ ................................ ................................ .. 61 Data Collection ................................ ................................ ................................ ....................... 63 Phase 1 Children with ASD ................................ ................................ ............................. 63 Phase 2 Pediatric Nurses ................................ ................................ ................................ 65 Data Management ................................ ................................ ................................ ................... 66 4 RESULTS ................................ ................................ ................................ ............................... 69 Participant Data ................................ ................................ ................................ ...................... 69 Inter rater Reliability ................................ ................................ ................................ .............. 69 Construct Validity ................................ ................................ ................................ ................... 70 Summary ................................ ................................ ................................ ................................ 71 5 DISCUSSION ................................ ................................ ................................ ......................... 74 Study Considerations ................................ ................................ ................................ .............. 76 Ethics and Data Security ................................ ................................ ................................ 76 Video Re cording Methodology ................................ ................................ ....................... 76 Recruitment ................................ ................................ ................................ ..................... 77 Strengths and Weaknesses ................................ ................................ ................................ ...... 78 Implications and Application to Clinical Settings ................................ ................................ .. 80 NCCPC PV ................................ ................................ ................................ ...................... 80 Revised FLACC ................................ ................................ ................................ .............. 81 UWCH PSPNC ................................ ................................ ................................ ................ 83 Future ................................ ................................ ................................ ................................ ...... 84 Conclusion ................................ ................................ ................................ .............................. 86 LIST OF REFERENCES ................................ ................................ ................................ ............... 88 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ......... 98
8 LIST OF TABLES Table p age 2 1 ASD language characteristics. ................................ ................................ .......................... 51 3 1 Pediatric pain instrument studies tests of reliability and construct validity. ..................... 67 4 1 Children with ASD age and CARS2 ST scores. ................................ ............................... 72 4 2 Spearman correlation. ................................ ................................ ................................ ........ 72
9 LIST OF FIGURES Figure page 4 1 Age range distribution of nurse participants ................................ ................................ ..... 73
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy AUTISM SPECTRUM DISORDER By Amanda B. Brown May 2014 Chair: Jennifer H. Elder Major: Nursing Sciences Characteristic impairments in Autism Spectrum Disorder (ASD) make it difficult to express pain in a way that is understood by others. Quality assessment of pain is a priority in healthcare because it leads to bet ter management and treatment. No pain assessment instruments have been tested exclusively for use in children with ASD. The purpose of this study was to test the reliability and validity of three pediatric pain assessment instruments developed for nonverb al, cognitively impaired or neurologically impaired children when utilized by pediatric nurses for children with ASD experiencing procedural pain. Six children with ASD undergoing a painful procedure as part of their routine healthcare were video recorded during the procedure and also during a non painful healthcare task. Two patients were randomly assigned to one of three pain scales: the Non Pain Checklist Postoperative Version (NCCPC PV), the Revised Faces, Legs, Activity, Cry and Pre verbal and Non verbal Children (UWCH PSPNC). Forty five practicing pediatric nurses viewed the video clips and scored the pain for the clips according to the assigned instrument.
11 Inter rater reliability was assessed using the intra class correlation coefficient (ICC 2,1) for each instrument separately. The NCCPC PV did not m e et reliability criteria (ICC=.71) but the Revised FLACC (ICC=.84) and the UWCH PS PNC (ICC=.84) were deemed reliable. Construct validity was examined by a matched t test for each instrument to assess contrasting group validity and a Spearman correlation to the other instruments to assess convergent validity. All three instruments demon strated significant contrasting group validity, with scores for painful clips being significantly higher than scores for non painful clips (p=.000 for all instruments). Each pain assessment instrument had strong and significant correlation with the other t wo instruments (p=.000 for all correlations). Based on these findings, it is recommended that the Revised FLACC and the UWCH PSPNC can be used immediately to assess pain in children with ASD in clinical settings with reasonable confidence, and that the re sults will be reliable and valid. Further investigation is needed to confirm these results and to support the use of these instruments for clinical research.
12 CHAPTER 1 INTRODUCTION Children with autism spectrum disorder (ASD) represent a significant group of the pediatric population in the United States (Rice, 2009). The most recent data reveal that as many as 1 in every 88 children in the United States has been diagnosed with ASD -a 78% increase in prevalence in six yea rs (Centers for Disease Control and Prevention, 2012). ASD is a developmental disorder of the brain that is characterized by impairments in social interaction, communication, and repetitive patterns of behavior (American Psychiatric Association, 2013). It has been proposed that children with autism experience a decreased sensitivity to pain, but recent research suggests that children with autism experience pain as much as developmentally normal children ( Daughters, Palermo, & Koh, 2007 ; Nader, Oberlander, Chambers, & Craig, 2004; Tordjmann, et al., 2009). Parents of children with autism agree, and reported in a recent survey Experts are proposing that it is n ot a decreased sensation to pain but a difference in how these children express pain (Dubois, Capdevila, Bringuier, & Pry, 2009; Terstegen, Koot, De Boer, & Tibboel, 2003; Tordjman et al., 2009). Pain expression is multi faceted and sensory, emotional, co gnitive, cultural, developmental, and contextual factors all have an influence (Craig, 2009; Finley, Kristjansdottir, & Forgeron, 2009). Assessment and treatment of pain in children is a priority in pediatric healthcare because effective pain management ca n ease and speed recovery. Lack of appropriate pain assessment is a major barrier to effective pain control and treatment (American Academy of Pediatrics & American Pain Society, 2001). Pain that is not assessed is typically undertreated or not treated a t all (Terstegen, Koot, De Boer, & Tibboel, 2003). Routine and systematic pain assessments have been shown to lead to better pain treatment outcomes (Breau, Finley, McGrath, & Camfield,
13 2002). Although routine pain assessments are the standard of care fo r pediatric patients in acute care settings, a lack of proper assessment tools presents a barrier to controlling pain and can lead to unnecessary suffering (American Academy of Pediatrics & American Pain Society; Voepel Lewis et al., 2008). A negative heal thcare experience, which may include untreated pain, can have effects on a child that extend into future encounters with the healthcare system (Souders, Freeman, Depaul, & Levy, 2002). Children with developmental delays, including autism, present challenge s to healthcare accurately, as children with neurological deficits may not be able to participate in pain assessment or self reports of pain due to their cognitive, physical or communicative limitations (Breau et a l., 2002a). A person with ASD may not be able to express pain due to social or cognitive impairments (Nader et al., 2004) and also experiences sensory abnormalities that may change their experience of painful stimuli and their expression of the pain (Ingl ese, 2008). Numerous pediatric pain assessment instruments are described in studies, but often children with known developmental delay are excluded from research that seeks to establish the reliability and validity of these instruments (Hicks, von Baeyer, Spafford, Van Korlaar, & Goodenough, 2001; Hunt, Mastroyannopoulo, Goldman, & Seers, 2003 ). There are a few pain assessment instruments that have been expressly designed to assess children who are non verbal or have cognitive impairment. Inclusion of ch ildren with ASD in these studies is highly variable (Breau, McGrath, Camfield, Rosmus, & Finley, 2000; Duivenvoorden, Tibboel, Koot, Van Dijk, & Peters, 2006; Hunt et al., 2004; Malviya, Voepel Lewis, Burke, Merkel, & Tait, 2006; Soetenga, Frank, & Pellino 1999). To date, no published research has examined the reliability and validity of existing pain assessment instruments for only children with ASD. This
14 unique population needs further research that seeks reliable and valid instruments to assess their pain so it can be successfully managed and treated. Purpose and Research Question The purpose of this study is to test the reliability and validity of three pediatric pain assessment instruments developed for non verbal, cognitively impaired or neurologica lly impaired children when utilized by pediatric nurses for children with ASD experiencing procedural pain. This descriptive study will establish baseline reliability and validity of these instruments for application to children with autism. Children wit h autism undergoing a painful medical procedure, as part of their routine healthcare, will be video recorded. Pediatric nurses at children with autism in both p ain and non pain situations using three pain assessment instruments. This study seeks to answer the research question: is the Non Communicating PV), the Revised Faces, Legs, Activity, Cry, and Con Hospital (UWCH) Pain Scale for Pre verbal and Non verbal Children (PSPNC) a reliable and valid pain assessment instrument when used by pediatric nurses to assess procedural pain in childre n with ASD? The specific aims of this study are as follows: 1. To determine the inter rater reliability of the NCCPC PV, the Revised FLACC and the UWCH PSPNC when used by pediatric nurses to assess procedural pain in children with ASD. 2. To determine the contra sting group validity of the NCCPC PV, the Revised FLACC and the UWCH PSPNC when used by pediatric nurses to assess procedural pain in children with ASD. 3. To determine the convergent validity of the NCCPC PV, the Revised FLACC and the UWCH PSPNC when used b y pediatric nurses to assess procedural pain in children with ASD.
15 Definition of Terms Autism Spectrum Disorder (ASD): A developmental disorder of the brain that is characterized by impairments in social interaction, communication and repetitive patterns of behavior (American Psychiatric Association, 2013). Direct care nurse: A healthcare provider whose primary role is to provide hands on care to patients in an acute care setting. (McGrath & Pain assessment instrument: A tool that attempts to capture the presence and may also attempt to measure the severi ty of hurting. Measurement can occur through self report, behavioral observation or biological observations (McGrath & Unruh, 2006). Significance This study is relevant to nursing practice given the rising prevalence of ASD and the importance of accurate ly assessing pain. At the present time, it is unknown if any pain assessment instrument is reliable or valid in assessing pain in children with ASD. The findings of this research may lead to the utilization of an established pain assessment instrument to assess procedural pain in children with ASD or the refinement or creation of an instrument. Interventions for relief of pain can be tested in this population once an established instrument that measures pain has been identified. Developing this assessmen t will contribute to the scarce body of knowledge on the pain experience in ASD. Once healthcare workers can begin to assess pain in children with ASD, their pain can be better managed and treated more effectively. Limitations The sample of children with ASD presents several limitations, as it will most likely have a high degree of heterogeneity with age, manifestations of ASD and the type of procedure being performed. To recruit enough subjects, the sample of children with ASD is purposive. The
16 ability of nurses to view vignettes multiple times and the fact they are not assessing the patient in real life situations may influence the generalizability of these findings to clinical settings.
17 CHAPTER 2 REVIEW OF LITERATURE This review will begin with an overview of ASD. Four theoretical approaches to the study of pain will be described and compared. A rationale for application of the presented, and pain and assessment of pain, including challenges, will be described for children, children with special needs and children with ASD. A discussion of video recording methodology in research is followed by a review of the ethics in research of vulnerable population s. Autism Spectrum Disorder ASD is a pervasive developmental disorder with impairments in social interaction, communication and repetitive patterns of behavior. The d ia gnostic and statistical manual V combines the several different diagnoses under the on e diagnosis of ASD (American Psychiatric Association, 2013 ) childhood disintegrative now a recognized genetic syndrome, was eliminated from the manual and childhood disintegrative disorder re categorized (American Psychiatric Association, 2010). Leo Kanner (1943) descr ibed a pattern of behaviors he recognized in 11 children and termed the disorder autistic disturbances of affective contact His description of the disorder is y were physically normal, although one child had a seizure disorder. The children demonstrated deficits in four main areas. First, they lacked an ability to connect or relate with people; secondly, they had language and communication difficulties, includ ing knowing words, but not
18 understanding their meaning, literal interpretations of language, and delayed echolalia; third, sensory issues were common and included texture or taste issues with food, aversion to loud noises, and a fascination with the moveme nt of objects; and fourth, the children had repetitive behaviors and a need for routine. Manifestations of the impairments in ASD are highly variable and therefore two individuals with ASD can present completely differently (Heflin & Alaimo, 2007; Notbohm, 2006; Scarpinato et al., 2010). Common co morbidities in ASD include sleep disturbances, gastrointestinal issues and seizure disorders (Scarpinato et al., 2010). Unlike the children described by Kanner (1943), many children with ASD are reported to have cognitive impairments (Heflin & Alaimo, 2007; Scarpinato et al., 2010). ASD is a neurological disorder and persons suffering from the disorder have different functional, chemical and structural components of the brain. Functionally, they may use different parts of the brain to process certain types of information, such as recognition of facial features, than those without ASD. Chemical differences include increased levels of serotonin, dopamine and endogenous opiates. Structural differences have been fou nd in the cerebrum, brain stem, limbic system and cerebellum (Heflin & Alaimo, 2007). Electroencephalography is normal in the majority of those with ASD. One study reported 35% of a sample of 46 children with ASD with epileptic paroxysmal abnormalities, 13% with seizures and 22% without seizures (Canitano, Luchetti, & Zappella, 2005). These functional, chemical and structural differences impact how the person with ASD thinks. The ASD mind has difficulty with critical thinking, executive management and s ocial pragmatics, thus contributing to their lack of ability to generalize information and difficulty in making connections between pieces of information (Notbohm, 2006).
19 Theories of ASD Brain Function Three main theories attempt to explain how the ASD b rain functions. The lack of theory of mind argues that children with ASD are unable to recognize that other people have independent mental states. They may believe that everyone thinks just like they do, or they may simply not be able to comprehend that o thers hold their own motivations for action. This lack of theory of mind is manifested in a lack of empathy, inability to enter reciprocal relationships or lack of desire to maintain relationships and may explain why children with ASD have difficulty part icipating in natural language development activities such as model imitation, joint attention and symbolic play (Barnbaum, 2008). Theory of mind provides a rationale for many of the pragmatic issues with communication experienced by those with ASD (Noens & van Berckelaer Onnes, 2005). The weak central coherence theory describes a mind that only sees information in parts and not wholes. Information is stored and retrieved separately and is specific and not generalized to other situations. Communication r equires many pieces of information from different sources to be analyzed and contextualized quickly, and so therefore a weak central coherence can be detrimental. Noens and van Berckelaer Onnes (2004 2005) argue that this theory best explains all the communication deficits -both expressive and receptive -found in ASD. It can also explain delayed echolalia or the use of neologisms seen in ASD. Children with ASD do not recognize the meaning of the ent ire message, including its context, the first time. I nstead they focus on a portion of the message, possibly a phrase and l ater may repeat that phrase, attempting to capture the meaning and context of the first message, but their recalled portion of the m essage is not enough to be meaningful to others.
20 The weak executive function theory holds that the ASD mind has impaired executive function. Executive function allows a person to plan, organize, multi task, make high level decisions and override or inhibi t automatic behaviors or impulses and permits for flexibility in thinking and learning. Without executive function, it is difficult to change after a concept has been learned (Barnbaum, 2008). Communication requires a great deal of flexibility and adapt ing over time as language continues to increa se in complexity (Prizant, Wetherby, Rubin, & Laurent 2003). P ronoun reversal is a n example of communication impairment in ASD After children d difficult for them to adapt their communication when making a request. So the children repeat what they first heard : y ou want Sensory Impairments Sensory impairments are present in as many as 80% of persons with ASD (Heflin & Alaimo, 2007). The sensory system may be hypersensitive in some areas and hyposensitive in others. Many repetitive behaviors in ASD may actually be adaptive behaviors for the sensory differences that are being experienced. Often, those wi th ASD are very sensitive to smells, usually have difficulty with auditory processing and respond best with visual communication methods. Tactile defensiveness may be present and is caused by a very low threshold for tactile stimulation. Vestibular and pr oprioceptive systems are also affected and may account for the clumsy behavior often described in individuals with ASD. Due to differences in their brain, they are unable to filter and prioritize sensory signals. For example, it may be difficult to block out background noise or they may use peripheral vision to focus on people or objects because having the eyes focus directly forward provides too much sensory information at one time (Heflin & Alaimo, 2007).
21 Communication and Social Impairments Clinical presentations of communication and social impairments vary considerably across the ASD spectrum some individuals never develop verbal language skills while others approximate normal skill levels. It is estimated that 20 50% (Tager Flusberg, Paul, & Lord, 2005 ) never develop functional communication and 25 50% do not develop verbal language (National Institute on Deafness and Other Communication Disorders, 2010; Noens & van Berckelaer Onnes, 2005). Development of communication and social skills ar e intertwined and d elay in one area impacts the development in the other. The level of development of skills in both areas is related to the level of independent functioning that the person with ASD will achieve (Prizant et al., 2003). Development of c omm unication and s ocial s kills Development of communication and social skills begins in infancy. Crying is the very first form of social interaction (Espo sito & Venuit, 2010). Through eye gaze and observation of facial expression, typically develo ping (TD) infants begin to form relationships between people and objects and recognize emotion. Infants later diagnosed with ASD prefer to look at objects over people, notice parts of objects instead of wholes, and fixate on one item instead of gaze at multiple it ems. At six to seven months, infants begin babbling and using vocal utterances to gain attention from others. Infants with ASD do not babble as much, do not seem to be as aware of language and are often thought to have hearing impairment (Heflin & Alaimo 2007; Tager Flusberg et al., 2005). In the final months of infancy, TD infants begin to gesture to in order to express needs or wants. Infants and toddlers with ASD tend to use gestures less often and in less meaningful forms of communication (Heflin & Alaimo 2007 ; Sowden, Perkins & Clegg, 2008; Tager Flusberg et al.).
22 Toddlers and pre schoolers engage in three types of behaviors that are essential in the development of communication, social, and language skills. Motor imitation of others begins before language skills. In order to imitate actions correctly, precisely, or in the right context, slow to imitate or may imitate behaviors inaccurately because they m iss the meaning of behaviors or attribute intentions inaccurately (Heflin & Alaimo, 2007; Tager Flusberg et al., 2005). Joint attention is the ability to engage in interaction with others. Children can see what attention in that action or activity. The social interaction includes sharing of emotions and reciprocal exchange of information. Joint attention engagement is a predictor of future skills (Aylott, 2000; Bolick, 2008; Noens & van Berkelaer Onnes, 2005; Prizant et al., 2003; Tager Flusberg et al., 2005). Children with ASD are less 2008 ; Chiang, 2008; He fl in & Alaimo, 2007). Symbolic or object play in childhood helps to develop symbol representation and is critical to development of language and social skills. Pretend play with objects develops naturally and becomes more complex over time for TD children. Children with ASD are far less likely to participate in symbolic play (He fl in & Alaimo, 2007; Noens & van Berckelaer Onnes, 2005; Prizant et al., 2003). Behavioral characteristics Children with ASD have difficulties with the pragmatics of communication ( Heflin & Alaimo, 2007; Notbohm, 2006). Vocalic communicat ion includes the nuances of language including the inflection of tone, non literal language, sarcasm, puns and idioms. Those with ASD may be unable to understand messages in a vocalic context. Body language, facial
23 expressions and gesturing make up kine sthetic communication and are also often not understood by children with ASD because these non verbal aspects of communication are not recognized as important signals. Proxemic communication, the concept of personal space and boundaries, also presents ch allenges (He fl in & Alaimo 2007 ; Notbohm 2006 ). Atypical communication styles such as echolalia, contact gestures, pronoun reversals and neologisms (refer to Table 2 1 for definitions) likely develop in ASD because there is a limited understanding of the meanings and intentions of symbolic forms of language ( Heflin & Alaimo 2007 ; Noens & van Berckelaer Onnes, 2005). Their speech tends to be very formal (Tager Flusberg et al., 2005) and e ven in high functioning children with ASD, 51 81% have impairments in pragmatic aspects of communication (Volden & Phillips, 2010). Overall, they have difficulties sending and receiving messages and t his affects their ability to understand and participate in the social world (Hel f in & Alaimo 2007 ; Notbohm 2006 ; Scarpina to et al., 2010). is ignored or the wrong response is given. The use of incorrect pragmatics or odd qualitative aspects of language can contribute to a breakdown (Tager Flus berg et al., 2005). TD children will attempt to repair communication breakdowns and t hose with ASD may not (Meadan, Halle, Watkins, & Chadsey, 2006) but, t o repair communication, a person must possess a certain level of social and communication skills to r ecognize a breakdown has occurred and those with ASD often lack these skills (He fl in & Alaimo, 2007). The frustration of being unable to communicate with others can lead to behavioral outbursts in some children. Children with ASD use challenging behavior s to communicate so as to have some form of control over their environment and as a resource to regulate their emotions and behavior (Bronsard, Botbol, & Tordjman, 2010;
24 Ch ia ng, 2008; Halle, Brady, & Drasgow, 2004; Noens & van Berckelaer Onnes, 2004; Priza nt et al., 2003). Social i nteraction Communication is a link to the social world although c hildren with ASD use communication as a tool for getting their needs met, but not typically for social engagement (Chiang & Lin, 2008). They may not anticipate engagement at all and tend to declare or express their needs or wants without any expectation for others to engage (Noens & van Berckelaer Onnes, 2004). These communication difficulties along with lack of understanding of emotion, a tendency to be disorg anized and problems with generalization contribute to their problems of interacting socially. Interacting with others builds social capital and skills, and those with ASD have fewer tendencies to interact. TD children learn social interaction via a hidde n curriculum through trial and error. They are not able to decipher the messages from others about what behavior is appropriate and what is not. Children with ASD do not have the skills to access this hidden curriculum (Bolick, 2008; Heflin & Alaimo, 2007 ; Notbohm, 2006) and have difficulty in learning social skills because of their communication issues, rigid thinking and difficulties with gen eralizing (Heflin & Alaimo, 2007 ). Repetitive Patterns of Behavior Repetitive patterns of behavior are manifeste d in ASD. This may be an unusual fascination or obsession with one subject, a need for things to always be the same or motor movements. The complete focus on one task often limits the ability to multitask and causes issues with transitioning to a new tas k. Stereotypies are motor movements, such as hand flapping, that seem to have no purpose. If these movements are pleasurable or soothing to the individual, they are designated as self stimulating behaviors and may be used by the person with ASD as a mean s of coping (Heflin & Alaimo, 2007; Notbohm, 2006; Scarpinato et al., 2010).
25 Physical Differences Persons with ASD have physical differences in their neurological system, which dictates how the mind processes and responds to information. The physical differences in ASD lead to a different way of thinking and this different way of viewing the world and self is manifested in sensory differences, social and communication impairments and repetitive patterns of behavior (Heflin & Alaimo, 2007 ). Pain in Children Children with ASD have many unique features from TD children, but these children will face some of the same experiences, including pain. Children most commonly experience acute pain related to injury, illness, or medical procedures (American Academy of Pediatrics & American Pain Society, 2001). Pain is often under estimated and under treated in chil dren for a variety of reasons. Incorrect beliefs still persist that children do not experience pain like adults or if they do, they are unlikely to be affected or have long term consequences from the experience of pain. Providers may fear side effects of pain medication. The biggest barrier to pain assessment methods (American Academy of Pediatrics & American Pain Society, 2001; Pain Theorie s Neuromatrix T heory of P ain dimensional experience. There are four parts to the conceptual nervous system: the body self neuromatrix; the cyclical processing and synthesis that ma kes the neurosignature; the neural hub that converts the neurosignature to awareness; and the action that is activated by the neuromatrix to produce the pattern of movements that will bring about the goal. The neuromatrix is the body self that is
26 made of a large network of neurons, the thalamus, the cortex and the limbic system. The design of the neuromatrix is influenced at first by genetic factors and then shaped by sensory factors. The neuromatrix receives information from cognitive related brain areas, sensory signaling systems and emotion related brain areas. The neuromatrix shapes all inputs to the system (Melzack, 2005). The neurosignature is the cyclical processing and synthesis of impulses that take place throughout the neuromatrix. It is a cont inuous flow of data input, data processing and data output. Triggers to the neurosignature may cause changes or stimulate patterns, and may include sensory inputs or cognitive events. The neurosignature ultimately produces awareness and action (Melzack, 2005). The neurosignature sends pattern signals to the sentient neural hub in the brain, where information is processed and cognitive and physical reactions are produced. The neuromatrix receives information from cognitive related areas in the brain, sens ory signaling systems and emotion related brain areas. It produces outputs such as awareness (pain perception), action programs and stress regulation programs. The brain chooses an action based on experience (Melzack, 2005). Model of Pain Expression model of non verbal communication, which describes the process of non verbal communication that begins with an internal experience. This is expressed in a behavior that allows an observer to draw conclusions about the experience (Rosenthal, 2005). The Model of Pain Expression involves three events: the experience, the encoding an d the decoding. Each phase is influenced by physiological, psychological and social influences.
27 The first phase, the experience, begins with a pain stimulus. The experience is affected by intrinsic factors, such as age, developmental level, personality variables and coping styles and extrinsic factors, such as the presence of analgesics, the sensory consequences of the experience, or other environmental influences such as social expectations. There is a threshold level of pain that must be met for the pain to be expressed in the face, and this is dependent upon the individual, not necessarily the severity of pain (Prkachin & Craig, 1995). The second phase of this model is encoding. Once the threshold level is reached, a central motor program that lead s to a change in facial expression is activated. Pain is encoded in the facial expression. These facial movements are specific to pain and differ from other emotional states. The set of movements, or expression, is usually very short in duration, lasting less than five seconds. This encoding can be affected by intrinsic factors and display rules. Display rules, or socio cultural influences, can cause a person to try to lessen, exaggerate or mask their expression of pain (Prkachin & Craig, 1995). The f inal phase in the model is decoding. After the signal of a pain specific facial expression is broadcast, an observer must decode the expression and this involves detection and discrimination, attachment of meaning to what was observed, and a behavioral re action or response. Decoding is a critical phase; it is not guaranteed that a pain expression will be detected and decoded. A filter that is made up of observer bias can affect the decoding. Each observer has a gain function, or a different level of fac ial expression that must be reached for them to determine that the broadcaster is in pain. The observer acts once the gain function is reached. Actions can include giving help, imparting sympathy, or ignoring the communication (Prkachin & Craig, 1995).
28 Communications M odel for U nderstanding C hildren s P ain verbal communication model. This mo del of pain does not solely focus on facial expression or non verbal communication of pain, but includes both verbal and non verbal expressions of pain. The model has four steps: the internal experience, the encoding in expressive behaviors, the decoding, and action. The first step begins with a pain stimulus that triggers an internal experience of pain, which is very complex and personal. Multiple areas of the brain and nervous system contribute to the experience, as well as cognitive and social factors It is affected by intrapersonal influences and contextual influences. Interaction between biological stages of life, personal experiences with pain and socialization to pain influence the perception of pain (Hadjistavropoulous & Craig, 2002). In child ren, the perception of pain is affected by intrinsic, formative and situational factors. Intrinsic factors may include genetic program, biological maturation, psychological capabilities and affective mechanisms, while formative factors are family and cult ural influences, and situational factors are transient states and environmental settings (Craig et a l., 1996). The process of encoding pain involves both verbal and non verbal programs. Non verbal programs, e.g. facial expression of pain, are more of an automatic process. A person has less voluntary control over these processes. Verbal programs are a higher mental process. An individual is more likely to be able to exert control over these cues. Self reporting of pain is often cons idered the gold standard, but this model emphasizes the importance of assessing both automatic and higher mental processes encoding of pain for the most accurate decoding (Hadjistavropoulous & Craig, 2002). Children express pain cues vocally (cry or scream ), non
29 vocally (facial expressions, limb movement, or posturing) or verbally (dependent upon language development) (Craig et al., 1996). The third step is decoding of pain by the observer. Encoding of pain has no value if it is not decoded properly by t bias. The cues produced by higher order mental processing are easier to decode than non verbal cues, while the clarity of the message is affected by past experience of the decoder, t he trust the decoder has in the person sending the cues, and other decoder personal characteristics. Characteristics of the sender may trigger prejudicial attitudes in the observer and cause incorrect decoding (Hadjistavropoulous & Craig, 2002). As adult attention in order to detect and discriminate pain cues from other non pain behaviors in the child and the signals have to be interpreted and assigned a value by the decoder (Craig et al., 1996). The fourth phase of the model is action. The adult decoder can choose to intervene, to e to environmental influences, such as availability of medication or inability to stop treatment, in non pharmacological. Withdrawal of the painful stimulus m ay not be an option or a possibility. Adults may continue to impose pain if they believe the ongoing treatment will benefit the child or if they have brutal intentions (Craig et al., 1996). Prescriptive T heory of A cute P ain M anagement in I nfants and C hildren The Prescriptive Theory of Acute Pain Management in Infants and Children (Huth & Moore, 1998) consists of three constructs (initial assessment, therapeutic interventions and reassessment) that produce an outcome. These constructs are actions that are to be repeated until the outcome, satisfactory pain reduction, is met. The theory was developed from a set of acute
30 pain management guidelines for children, with the goal of providing a framework for research in applicability. In the initial assessment phase, the nurse is to evaluate several concepts ( past pain history, current pain history, develop mental level, coping strategies and cultural background ) in order to make a judgment about the presence of pain and to help determine the best intervention. Past pain history can influence the current expression of pain and also lead to clues about effectives of interventions. Children process and cope with pain differently depending on their developmental level and i t should be borne in mind that c ultural influences affect communication of pain and coping strategies (Huth & Moore, 1998). Based on the initial assessment, the nurse chooses one or more therapeutic interventions includ ing child parent teaching, opioid an algesics, pharmacologic adjuvant, and non pharmacologic adjuvant. Education can help relieve anxiety regarding pain. Opioid treatment by itself may not be completely effective in reducing pain and so adjuvants can be used and b oth non opioid analgesics and non pharmacological interventions have been shown to be effective for relieving pain They are also options when opioid therapy is contraindicated. The outcomes of these interventions should be pain relief that is satisfactory to the child, parent and nurse (Huth & Moore, 1998). Re assessment includes regular assessment of pain, behavior, physiological states and side effects. These regular assessments lead to the identification of pain relief that is not adequate, behavioral distress, unacceptable ph ysiological measures and side effects and should occur frequently -at least every hour until the pain is controlled. The reassessments contribute to 1998)
31 C omparison of the four theories The four theories described above hold a number of common characteristics. All acknowledge and describe pain as a multi faceted, complex, internal experience and each theory, with the exception of the Pres criptive Theory of Acute Pain, begins its process with a painful stimulus. After the stimulus, each describes some form of internal processing that occurs and results in an outcome. For the Neuromatrix Theory of Pain, the outcomes are awareness and action, while facial e x pression is the outcome in the M odel of P ain E xpression and, for the verbal cues are the outcome to the internal processing. The Neuromatrix Theory of Pain does not go beyond the outcome s of the individual experiencing pain. The other two models go on to describe a decoding process by an observer that is affected by many factors that are inherent to the decoder riptive Theory of Acute Pain in Infants and Children does not describe any outcomes, but only details a list of factors that may affect pain expression. The focus is on a nurse, not the child, who is assessing pain, intervening and reassessing. Each theory has unique characteristics. The Neuromatrix Theory of Pain, for instance, is very focused very much on the biological components of the pain experience and is very detailed and specific on the anatomy and physiology of the pain experience. The Model of Pain Expression focuses on just one output of pain encoding facial expression and also describes threshold and gain levels, variable across individuals, that must be reached to express pain in the face or act on pain perceived in another. The Communicati ons Model for Understanding communication and encompasses multiple means of pain expression and multiple factors of pain perception. It includes the most comprehensive desc ription of the decision to act on pain
32 perceived in others. The Prescriptive Theory of Acute Pain Management in Infants and Children includes a circular pathway for assessment, intervention, and reassessment but lacks details on the ways in which concepts interact, as well as on the criteria used by the nurse when making assessments. It limits action of pain to intervention, and therefore only considers one possible course. Rationale for s election of m odel The Communications Model for Understanding Childr model for this research. The Neuromatrix Model of Pain was not selected primarily because it does not include any concepts on the assessment of pain and therefore would not be helpful in the research problem. In additi on, this theory, although it accounts for many influences of pain expression, does not place specific emphasis on developmental, communicative and social factors that influence pain expression. The Prescriptive Theory of Acute Pain Management in Infants a nd Children does not provide enough detail or depth to support the proposed research, while the Model of Pain Expression is limited as it is focused solely on one pain expression -facial movement. Facial expression in pain is not dependent on development level or communication skills, and may be a good method for studying pain in ASD, but it is only one aspect of pain expression and assessment. From a pragmatic perspective, it has been proven most useful in research where video recordings and/or photograp hs can be examined in detail and may not be of use in clinical environments. As a result, therefore, the Communications Model for Understanding Children s Pain will serve best for this research because it places equal emphasis on the pain expression and p ain assessment. Thi s model is multi faceted and focuses on many factors of pain expression and pain assessment. It also provides a step for the decision to act following assessment of pain.
33 Pain Assessment in C hildren Pain, pain is a challenge. Indirect measurement can occur through self report, behavio ral observation have not demonstrated strong correlation to other measures of pain; these may only change in instances of sharp, acute pain and can be attributed to a large variety of other stress events. Use of only physiologic changes for pain assessment is not appropriate, therefore, but they may be used in conjunction with observations of pain behaviors or self reports of pain (American Academy of Pediatrics & Amer Spagrud, 2007). Self report measures Pain is recognized as a very subjective event (International Association for the Study of Pain, 1994) and, therefore, many hold self report to be the gold standard of pain assessment for a child to be able to self report pain, the child must developmentally understand how to use the scale, not be overly distressed, be able to co mmunicate, have the cognitive ability to grasp the scale and not be restricted excessively by medical interventions (such as bandages or sedation) (von Baeyer & Spagrud, 2007). Self report instruments exist in both verbal and non verbal formats, but even non verbal formats require the child to possess a certain level of cognitive and communication function. Self report can be used in conjunction with behavioral observations or physiologic changes to assess pain (American Academy of Pediatrics & American
34 More than 30 instruments exist for the measurement of self reported pain in children, but only six have well documented reliability and validity. Of these six scales, not one was able to establish reliability and val idity across all age groups and different types of pain. The Pieces of Hurt, Faces Pain Scale, Faces Pain Scale Revised, Oucher Photographic, Oucher Numeric Rating Scale, Wong Baker FACES Pain Scale, and Visual Analogue Scales (VAS) are the self report pa in instruments reported to have sound psychometrics (Stinson, Kavanagh, Yamada, Gill, & Stevens, 2006). Two of these scales are discussed in further detail. VAS are used for children to rate their pain intensity. They consist of a line (either horizontal or vertical), in which the endpoints represent two extremes (no pain versus extreme pain). Existing scales vary. The length is different, with some having l ines demarcating intervals between the anchor points, and the language is also variable. All scales use some type of metric to assign a numerical value to the pain -usually the distance in millimeters or centimeters from the no pain anchor point to the po int where the child has rated the pain. Overall, these scales have proven valid and reliable in developmentally normal children aged eight years and older, although some have been validated in children as young as four years (Stinson et al., 2006). The Faces Pain Scale Revised (FPS R) (Hicks et al., 2001) was developed in an effort to make the original Faces Pain Scale more conducive towards a common metric for measuring pain and improve upon criticisms. The original scale consisted of seven faces rangi ng from a smiling face (no pain) to a face frowning with tears (worse pain). Seven faces made use of the scale hard for researchers and clinical staff in comparison with other scales that most often range from 0 to 5 or 0 10. The use of the smiling and c rying faces was also criticized as children may infer that they must be smiling so as not to be in pain or crying if in pain. The revised scale
35 consisted of six faces, correlating with a 0,1,2,3,4,5 or 0,2,4,6,8,10 scale and presented a neutral face as no pain and a face with a grimace, but no tears as the most pain. The scale was validated in two groups of developmentally and cognitively normal children aged 4 12 years. The first group comprised children undergoing ear piercing and the second group cons isted of children admitted to the hospital and experiencing pain. Concurrent validity was established by correlating the scores on the FPS R to scores the children gave using an established VAS or the established Color Analogue Scale (CAS) In the ear pi ercing group, a strong positive correlation was found between the FPS R and VAS (r=.93, P < .001). In the hospital group, the FPS R correlations with both the VAS and CAS were strong (r > .70). The FPS R is simple to use, has been translated into many lang uages and is the recommended tool for children ages 4 12 (Stinson et al., 2006). Behavioral measures For many children self report of pain is not appropriate. Behavioral measures of pain can either be used alone or in conjunction with self report (Americ an Academy of Pediatrics & American Pain Society, 2001; von Baeyer & Spagrud, 2007). Research for these measures is typically correlational, with occurrence of a behavior correlated with the occurrence of pain. Well anchored events, such as short medical procedures, are normally used for the studies (Blount & Loiselle, 2009). Several types of behavioral tools exist for pain assessment. Behavior checklists consist of a number of behaviors that the observer determines if they are present or not present. Behaviors are not weighted, nor is intensity of behaviors scored on the checklists. B ehavior rating scales attempt to rate behaviors on intensity and sum scores from different categories to determine pain measurement. Global rating scales depend solely on subjective assessment of pain in the child, and often VAS or FPS R is used by the observer to
36 Hospital of Eastern Ontario Pain Scale (CHEOPS) are tw o examples of behavior rating scales. They are described below. The FLACC scale has an observer rate pain behaviors in five categories (face, legs, activity, cry, consolability) as 0, 1 or 2. The observer than adds the scores to obtain a pain measure that ranges from 0 to 10. Each behavior category contains descriptive data for each score. For example, a score of 0 in the face category is described as no particular expression or smile, while a 1 is occasional grimace or frown, withdrawn, or disinterested, and a 2 is frequent to constant quivering chin or clenched jaw (Merkel, Voepel Lewis, & Malviya, 2002). The study was validated in a developmentally normal sample of children ages two months to seven years undergoing surgery. FLACC scores prior to analg esia were significantly higher when compared with scores following analgesia at 10 minutes, 30 minutes and 60 minutes (p < 0.001 assessment of pain (r=0.41, p < 0.005) and when compared with an existing pain instrument (r=.80; p<0.001) (Merkel, Voepel Lewis, Shayevitz, & Malviya, 1997). The CHEOPS uses six categories of behaviors: cry, facial, child verbal, torso, touch and legs. Each category contains certain behaviors that, if demonstrated by the child, point towards the cumulative score given. Behaviors are scored either 0,1, or 2. The observer selects as many behaviors as t he child is currently exhibiting in each category. T otal scores range from 4 to 13. The original sample included children from 1 to 7 years and demonstrated strong validity (McGrath et al., 1985). Further research has found the instrument to be reliable and valid in developmentally normal children 4 months to 17 years. The CHEOPS sca le has, however, been criticized for the use of a non traditional metric (4 14 versus 0 10) (von Baeyer & Spagrud, 2007).
37 Pain and pain assessment in children with special needs Pain assessment in the non verbal or developmentally delayed child can prove challenging due to their communication deficits and cognitive impairments. These children are often unable to participate in self report measures of pain (Voepel Lewis et al., 2008) and are susceptible to experiencing pain from multiple sources. In addit ion to pain experienced due to their medical condition, they still experience pain encountered by developmentally normal children, such as headaches or menstrual cramps (International Association for the Study of Pain, 2005). The belief that these childr en experience reduced sensitivity to pain is one barrier to identifying pain in this population. Self injurious behavior occurs in a large number of intellectually impaired individuals and some have argued that this is evidence of a reduced sensitivity to pain. Breau and colleagues (2003) found this theory to be false, with those in the study who exhibited self injurious behavior not showing a reduced pain reaction in other painful situations. They concluded that pain insensitivity does not contribute to self injurious behavior. In addition, self injurious behavior may be a form of pain attenuation (Breau et al., 2003; Heflin & Alaimo, 2007). Caregivers should not assume that differences in the expression of pain are indicative of differences in sensitivi ty to pain (Terstegen, Koot, De Boer, Tibboel, 2003). Mounting evidence shows that cognitively impaired children react to pain and are able to express pain through behaviors (Dubois, Capdevila, Bringuier, & Pry, 2009). These children may demonstrate unique pain behaviors (International Association for the Study of Pain, 2005; Voepel Lewis et al., 2008). Despite the individual differences among cognitively impaired children, there is growing evidence that they share a set of common pain behaviors (Hunt et a l., 2007). Some of these pain behaviors are the same as those seen in the general population (International Association for the Study of Pain, 2005). These include:
38 changes in facial expression, changes in movement or posture, vocal cues, changes in chil Dubois et al., 2009; Duivenvoorden et al., 2006; Hunt et al., 2003; Terstegen et al., 2003). Assessment of pain in this population is crucial, as these childre communicate clearly through language requires their caregivers to depend completely upon pain should be respected and sought (Internat ional Association for the Study of Pain, 2005). Familiarity with the individual can contribute to better recognition and assessment of pain (Breau, et al., 2002b; Hunt et al., 2003). The use of a standardized behavioral rating scale may decrease the need to know the patient and provide an accurate assessment of pain (Breau et al., 2002a; Hunt et al., 2004). To that end, several pain assessment instruments, including the UWCH PSPNC; the Revised FLACC; and the NCCPC PV have been developed and these are desc ribed in more detail in Chapter 3. Pain and pain assessment in children with ASD Children with ASD, similar to children with other cognitive impairments, have been thought to experience decreased sensitivity to pain. These theories are not supported by a ny research, but based on anecdotal observations and clinical judgments (Daughters et al., 2007; Messmer, Nader, & Craig, 2008; Nader et al., 2004). Although limited, there is research that shows that children with ASD experience and have the ability to ex press pain. In a study of children with ASD undergoing venipuncture, Messmer and colleagues (2008) found that facial expression of pain or non Nader et al. (2004), meanwhile, found tha t children with ASD demonstrated a substantial facial expression response to venipuncture, with no difference in the facial reaction between children with ASD and children without ASD A study of children undergoing routine dental cleaning by
39 Daughters et al. (2007) found that children with ASD had higher pain scores compared with children without ASD Tordjman and colleagues (2009), furthermore, reported that, although children with ASD displayed significantly (p<.0001) less pain behaviors compared with n ormal children during venipuncture, the children with ASD had a significantly increased heart rate endorphin levels (p<.01) than the normal control children. This suggests that, even though they may not express pai n, they do experience it. According to Inglese (2008), parents of children with ASD report that they believe their children with ASD experience pain. The social, communication and sensory impairments experienced by children with ASD can affect their expr ession of pain (Messmer et al., 2008; Nader et al., 2004), which complicates the assessment of pain in children with ASD. Ingelese (2008), in a survey of 88 parents of children with ASD, notes that 70% reported difficulty determining when their child is i n pain and 86% reported that their child responds differently to pain than typically developing children. This difficulty in determining pain was also found when the facial response of pain did not correlate to the parental ratings of pain in children wit h ASD undergoing venipuncture. On the assess and manage pain i n their child with ASD (Inglese 2008 ). There has been some investigation into pain assessment instruments in children with special needs that has included children with ASD, but typically this has entailed a small portion of the sample size. In Malviya NCCPC, meanwhile, which included testing two samples, a mere eight out of 32 children had
40 ASD in th e first sample, while the second sample included nine children with ASD out of a total of 33. Dubois and colleagues (2009), furthermore, compared various pain scales for the cognitively impaired and of the sample of 30, only 1 child had ASD. Some studies do not indicate the origin of impairment in their samples, and so it is possible that more children with ASD have been studied. Video Recording in Research Development of pain assessment instruments often includes the use of video recording. Observation i s a critical component of nursing and social sciences and it is also a powerful research data source (Haidet, Tate, Divirgilio Thomas, Kolanowski, & Happ, 2009; Up ar ela Acosta & Tuna, 2011). Observation requires the observer to see, hear, document and int erpret events (Caldwell & Atwal, 2005) and can be made in naturalistic or laboratory settings. While a laboratory setting offers complete environmental control, a naturalistic setting enables a researcher to study human behavior better as it occurs. The r esearcher level of involvement can vary while observing, ranging from covert and uninvolved to covert and fully engaged. The researcher can also make the role of research known and then choose either to participate or observe. This non participant obser ver role can occur either in live action or by video review (Caldwell & Atwal 2005 ; Stangor, 2011; Up ar ela Acosta & Tuna 2011 ). H istorically t echnological advancements have influenced how researchers collect and analyze data and will continue to do so ( Derry et al., 2010; Gibbs, Friese, & Mangabiera, 2002). M any researchers were fearful following introduction of audiotape technology that the tapes would somehow inhibit their ability to experience the observation process fully, but important knowledge in social sciences has emerged as a result of using audiotapes (DuFon, 2002). Video recordings provide a very detailed record of an event and their use to document life events is
41 becoming socially widespread (Derry et al 2010 ) and its use in nursing rese arch continues to increase (Haidet et al., 2009). A key advantage in using video in research stems from its ability to provide a recording of an event with a level of detail that is beyond what many humans could document or recall. Verbal and non verbal behaviors including interactional behaviors are captured on video (Cohen & Crabtree, 2006; DuFon, 2002; Hall, 2010) and v ideo can also be stored and reviewed multiple times thus providing several benefits. Firstly, the researcher when coding behaviors can rewind and replay as many times as necessary to capture all events and m ultiple researchers can view the same event at different times something that is not possible for live observations. Multiple researchers viewing the data can also help to estab lish inter rater reliability or bring new perspectives to an event while s haring the video beyond the original research may accelerate knowledge in the field (Caldwell & Atwal, 2005 ; Derry et al., 2010; Gibbs et al., 2002). T here are however, a number of d rawback s to using video in research The presence of a video camera may for instance, and may be a deterrent to participat ion (Haidet et al., 2009; Stangor, 2011). Most researchers are not trained video technicians, so issues with visual or audio quality may emerge (Caldwell & At w al, 2005; Derry et al., 2010; DuFon, 2002). Furthermore, p articipants cannot be guaranteed anonymity they may be recognized in the video and m asking or pixilation of faces may not be a feasibl e option depending on the focus of the research. Additional concerns of confidentiality arise if the video is shared with other researchers or placed in a database for analysis (Broyles, Tate & Happ, 2005; Derry et al. 2010 ; Gibbs et al., 2002).
42 D esp ite claims that videos are objective data collection instruments that capture complete details the method is flawed, as video clips are not objective data they are filmed or edited with a human perspective. The angle of the lens or closeness of the subje ct can affect the inferences a researcher gathers from the video (Hall, 2000). The researcher watching a video is only experiencing it as a secondary experience and t hey do not have the power to shift their gaze, listen for signals not picked up in the au dio or direct their attention more closely to certain events. They can only experience what is captured through the lens and i t is only what happened in that limited amount of time. Contextual factors which can provide very important information may be missing and t his is especially concerning with secondary analysis of videos by researchers not involved in the original observations (DuFon, 2002; Haidet et al., 2009; Hall 2000 ). R esearchers need to consider how to minimize the disadvantages of using vi deo data. The y should know for instance, the research question and define the focus and aims of the research prior to the study as t his will help narrow the choice of timing and events to film (Cohen & Crabtree, 200 6 Uparela Acosta & Tuna, 2011). The de cision to use video as a data source should be based on the study aims and also consider if the type of behavior being observed can be captured accurately on video. The information needed by the instrument measuring the behavior should be able to be prese nt on the video and t he researcher should establish the timing of the video events. This can be interval timing (a certain number of minutes per day) or event timing (record whole events) (Haidet et al., 2009). T he equipment and room set up should be tes ted to ensure that good quality video and audio are captured so as to avoid technological issues, and t he video should be shot with a wide angle and with the most minimal manipulation by the researcher (DuFon, 2002; Hall, 2000).
43 Efforts should be made to m inimize risks to participants. Informed consent should be obtained from persons who may appear in the video, even if they are not the focus of the research. Informed consent should contain information about how the video will be stored, who will have acc ess, how long it will be stored and if it will be used in any educational or scientific presentations. Specific information about the intention to share the video in scientific databases or to use in future research should be contained in the consent form A plan should be made for the video data to be properly stored and secured with access limited only to those who need it for the research. Even if complete anonymity in a video is not possible, efforts should be made to minimize identifying details as much as possible e.g. blocking the sound of names or diagnosis in the video, respecting modesty or not displaying names of institutions or room numbers (Broyles et al., 2008; Caldwell & Atwal, 2005; Derry et al., 2010). Ethics in Research of Vulnerable Populations A person or group of persons is considered vulnerable in research if they have a reduced capacity for understanding or they lack legal or social power (Hirtz & Fitzsimmons, 200 2 ; Matutina, 2009; Shamoo & Resnik, 2009). These characteristics le ave them at risk of harm and susceptible to coercion (Beauchamp & Childress, 2009). Historically, vulnerable populations including children, mentally handicapped people, prisoners and the socially disadvantaged have been victims of unethical and often ti mes harmful research (Hirtz & Fitzsimmons 2002 ; Matutina 2009 ; Shamoo & Resnik 2009 ). S trict protections for research have been enacted i n an effort to protect these populations from further harm and this has result ed in the exclusion of most of these populations from research. Th is ha s led to negative consequences for instance, l ack of healthcare research in children left practitioners with a difficult decision not to treat either due to lack of evidence or to generalize data gathered in adults to th eir pediatric patients to the extent that a dverse events or
44 less effective therapies have occurred. In addition, the overall welfare of these populations was affected because no new knowledge that may have led to improvements in their well being was gain ed. These groups were also denied potential benefits from participating in research such as access to care and new treatments (Beauchamp & Childress, 2009; Hirtz & Fitzsimmons, 200 2 ; Kopelman, 2002; Shamoo & Resnik, 2009). Since the late 1990 s r ecogniti on of these outcomes of overprotection has led to a push to include children and other vulnerable populations in research, unless there is a scientific rationale for their exclusion, (Chen, Miller, & Rosenstein, 2003; Lamert & Glacken, 2011; Shamoo & Resni k 2009 ). Ethical concerns still arise with the inclusion of vulnerable populations in research, especially children. Hence, i nformed consent and assent are ethical issues in research with children that will presented in the context of the principal respe ct for autonomy. The balance of protection from harm versus the benefit from research participation is presented from the perspective of the ethical principle of justice and s trategies to minimize risks and protect children in research will also be discu ssed. Respect for A utonomy and Informed C onsent Autonomy refers to the right of a competent individual to make decisions regarding their own person, including the decision to participate in research (Beauchamp & Childress, 2009). Informed consent is the formalization of the decision to take part in research and was designed to protect research participants and ensure they are fully aware of the risks and benefits of the study (Barnbaum, 2008; Beauchamp & Childress 2009 ). A person must be competent to g ive informed consent and t hose who are unable to make decisions regarding their overall well being, to evaluate risks and benefits or who lack legal authority are considered to be not competent. Competency requires a certain level of cognitive functioning that enables a person to understand consequences and think rationally (Hirtz & Fitzsimmons, 200 2; Kopelman, 2002; Matutina,
45 2009; Shamoo & Resnik, 2009) yet p eople with ASD face challenges in this regard due to the cognitive differences of the disorder. It is difficult for a person with ASD, even if they are high functioning, to envision alternative situations and to evaluate how those situations may make them feel in the future (Barnbaum 2008 ). If a person lacks the competence to consent to research, t hen a proxy may be sought to gain consent. This proxy should be someone who will try to protect the interests of the person they represent. For most children, parents are their legal proxy (Hirtz & Fitzsimmons, 200 Shamoo & Resnik, 2009) and such prox ies can base their decisions on substituted judgment, pure autonomy or best interest. The use of substituted judgment and pure autonomy stipulate that the person being represented was at one time competent. Children were never competent, so best interest is the rationale that should be used for decision making (Barnbaum, 2008; Beauchamp & Childress, 2009). When acting in the best interest of another, the person assesses the risks and direct and indirect benefits of the situation for them. Direct benef its are sometimes termed therapeutic benefits and they usually entail a treatment of some form that the participant will receive while i ndirect benefits are usually non therapeutic benefits. The participant is less likely to receive a benefit from partic ipating in research, but may indirectly benefit from the increased knowledge gained about their population or by learning the value of contributing to society. There is usually very low risk for participants in studies with only indirect benefits. Some p ropose that even if the risks are very low, that it is not in the best interest of the child to participate if they will not receive any direct benefits while o thers counter with the argument that indirect benefits should be included in a full risk benef it assessment and
46 being as it limits the discovery of knowledge (Barnbaum; Kopelman, 2002; Shamoo & Resnik 2009 ). Assent is gaining permission for participation from a person that is not competent to provide informed consent. It should include the procedures/treatments involved in the study and the possible outcomes ; emphasize the choice to be involved ; and the option to withdraw at any time. A ssent design shou ld be tailored towards the mental capacity, age, and developmental level of the person and t he researcher should confirm that the person understands what is being asked (Lambert & Glacken, 2011). Assent should be obtained whenever possible considering the Children with developmental delays and those with ASD may be less able to participate in an assent process (Hirtz & Fitzsimmons, 2009; Kopelman, 2002; Shamoo & Resnik, 2009) and i n some cases assent will not be possible (Chen et al., 2003). The researcher should seek to establish an assent process that is as inclusive as possible and look for potential competency for assent and tailor a process that will fit th is potential (Barnbaum, 2008; Lambert & Glacken 2011 ). Justice and Research Participation Justice is the fair distribution of resources and burdens and, i n research context means that no group should bear unequal benefits or burdens of research (Beaucham p & Childress, 2009). Ensuring justice in research with vulnerable populations requires finding a balance between protection and gaining new knowledge (Hirtz & Fitzsimmons, 200 2 ; Lambert & Glacken, 2011). A group of people should not be used in research j ust because they are a convenient, available sample if the potential knowledge gained will not benefit this group. This is exploitation and has historically led to abuse of participants (Hirtz & Fitzsimmons ; 2002 ). They should also not be excluded for pu rposes of convenience, such as having a homogenous
47 sample to analyze (Kopelman, 2002). The inclusion or exclusion of a group of people should be based on scientific rationale (Chen et al., 2003). Participation in research does have benefits, and these shou ld not be denied to members of vulnerable groups in the name of protection. Knowledge about these groups may only be gained by their participation in research. Generalizing data from other groups will lead to flawed, even dangerous conclusions. Research can lead to new information that is able to improve the well being and overall health of a group and all groups should have this opportunity (Hirtz & Fitzsimmons 2002 ; Shamoo & Resnik, 2009). The National Institutes of Health mandate that children be in cluded in research they fund unless the topic is irrelevant to children, there are legal or regulatory restrictions that prohibit their participation, the information is already known, the condition is very rare in children or the data are not sufficient to support their inclusion (Kopelman 2002 ). Risk Minimization There are several safeguards a researcher should use in order to minimize risks to the participants and prevent exploitation of vulnerable groups. The study should first seek t o gain information that will contribute meaningfully to the body of knowledge and have a scientifically sound design (Chen et al., 2003). The choice to conduct research with a vulnerable group should be based on scientific rationale that means the research cannot be conducted with another population, the knowledge gained from this research will benefit this population, and the knowledge will promote the health and well being of this population. A vulnerable population should never be used simply because the y are an available, convenient, or cost saving option (Chen et al. 2003 ; Kopelman, 2002; Shamoo & Resnik, 2009). The researcher should have an Institutional Review Board that examine s the research proposal and look s for a risk benefit ratio that is positi ve for the participants. For research with children, the board should follow the four part stepwise plan provided by the Office of Human Subjects Research (OHSR) in Title 45, Part
48 46, Subpart D (Chen et al. 2003 ; U.S. Department of Health and Human Servi ces, 2005; Shamoo & Resnik 2009 ) as this outlines increased protections as risks to participants develop. A board reviewing research with children should have at least one or two members that have some experience with children (Lambert & Glacken, 2011; S hamoo & Resnik 2009 ; Twycross, 2009). Informed consent, especially in vulnerable populations, should include a clear and realistic outcome expectation for any treatment involved in research (Barnbaum, 2008; Shamoo & Resnik 2009 ). Finally, in populations not competent to provide consent, assent should be 2008 ; Chen et al. 2003 ; Kopelman 2002 ; Lambert & Glacken 2011 ). Summary People with ASD have brains that are structurally, functionally and chemically different from TD people. These differences in the brain cause cognitive impairments, sensory issues, social and communication impairments, and repetitive patterns of behaviors. Behavior differences are evident early in life an d affect the overall development of the child (American Psychiatric Association, 2013; Heflin & Alaimo, 2007; Scarpinato et al., 2010). the study of pain expression and assessment in ASD very well. The social and communication impairments in ASD, along with atypical motor behaviors, may make decoding pain signals difficult. This model allows for a number of ways for the child to express pain, which can accommodate both typical and atypical pain behaviors. The framework for decoding pain allows the observer to understand pain based on the behaviors the child chooses to use to express pain, s preconceptions and beliefs regarding the individual.
49 No pain assessment instrument exists expressly for the assessment of pain in children with ASD. Little research has examined the reliability and validity of existing pain measurement instruments in ch ildren with ASD and no research has studied this group exclusively. Failure to assess pain consistently leads to under treatment of pain and the experience of untreated pain in children has lasting effects on their future interactions within healthcare (A merican Academy of Pediatrics & American Pain Society, 2001; Voepel Lewis et al., 2008; Souders et al., 2002). Research into pain in children with cognitive impairments has shown that pain assessment instruments can be successfully created for those unabl e to participate in self report of pain (Breau et al., 2002a; Malviya et al., 2006; Soetenga et al., 1999) and that a common set of pain behaviors may exist in even highly heterogeneous populations (Breau et al., 2002b; Dubois et al., 2009; Duivenvoorden e t al., 2006; Hunt et al., 2003, 2007; Terstegen et al., 2003). Video recording in research is a valuable tool for data collection. While there are limitations to the use of this type of data, there are also nonetheless distinct advantages. In the present research, video recording allowed for reliability testing for pain assessment instruments. Video is an appropriate medium for testing behavioral checklists, including behavior based pain assessment instruments (Haidet et al., 2009). Special consideration should be given to protect the interests of vulnerable population s in research. These protections should include a balance of risks and benefits for the group involved. The extra time and effort needed to protect vulnerable populations should not preclude these groups from benefitting from scientific discover. Instrument creation is time consuming and difficult, in that, prior to attempting to develop new instruments, existing instruments should be tested in different populations or modified as needed and retested (Hulley, Cummings, Browner, Grady, & Newman, 2007).
50 There is evidence that behavioral based pain assessment instruments have been utilized successfully in different populations or settings from the original design (Blount & Loiselle, 2009; Malv iya et al., 2006). Testing the reliability and validity of three existing pain assessment instruments in children with ASD is thus justified and will contribute to the body of knowledge regarding ASD and pain expression. The assessment and management of p ain in the acute care setting is an indicator for nursing quality and a national accreditation standard (National Database of Nursing Quality Indicators, 2012; Wells, Pasero, & McCaffery, 2008). In order to manage and treat pain effectively, nurses need in struments that are reliable and valid, even when pain expression may be atypical, such as in the child with ASD.
51 Table 2 1. ASD language characteristics Term Definition Echolalia Children repeat what has been said to them either immediately or after some period of time. Contact gestures Children use other people as a tool to get what they need or want. The gesture is not symbolic. Pronoun reversals Children use first (I, me) and second pronouns (you, he, she) incorrectly. Neologisms Children assign meaning to a word or phrase that is not the socially accepted meaning.
52 CHAPTER 3 METHODS The purpose of this study was to test the reliability and validity of three pediatric pain assessment instruments developed for non verbal, cognitively impaired or neurologically impaired children when utilized by pediatric nurses for children with ASD exp eriencing procedural pain. This study was supported by a grant through the American Nurses Foundation funded by the Society of Pediatric Nurses. Institutional Review Board approval was obtained from the University of Florida and Baptist Medical Center. Setting part of a non profit five hospital system, Baptist Health, is a regional pediatric referral hospital located in Northeast Florida. The hospital provides service to the children and families of Jacksonville, Flo rida, as well as five surrounding counties in Florida and Georgia. Wolfson has 199 inpatient pediatric beds, a pediatric emergency room and surgical services, and serves as a training facility for University of Florida pediatric residents, while sub specia Health is recognized as a Magnet Health System by the American Nurses Credentialing Center World Report ( Baptist Health, 2013). Measurement Concepts The end product of an instrument represents the true score plus or minus the amount of error in measurement which can be either systematic or random. Systematic errors are related to the validity of the instru ment, while random errors are related to reliability. Reliability of an instrument refers to how consistently an instrument measures its intended concept. In measuring different concepts related to health, reliability represents the amount of trust that a difference in
53 the score of an instrument represents a true difference in health status (McDowell, 2006; Portney & Watkins, 2009). There are four types of reliability. Rater reliability includes inter rater and intra rater reliability the former referr ing to the consistency of scores across a group and the latter consistency of scores for an individual. Test retest reliability examines stability of scores over time while p arallel or alternative forms reliability involves using two different versions o f an instrument to test the same concept and can be used in place of test retest reliability if participants are likely to remember their answers. Internal consistency meanwhile, is the stability of results across items (Portney & Watkins 2009 ; Trochim, 2006). Validity is how well the instrument measures the concept it is designed to measure. Validity can be divided into three types. Content validity has experts or participants examine the instrument and decide if the items in the instrument reflect th e concept that is being measured while c riterion validity compares the instrument being tested to a gold standard or well established reference test and hi gh correlations between the two measures are sought. Concurrent and predictive validity are associ ated with criterion validity. Construct validity is used for instruments that measure abstract or theoretical concepts (Hulley et al., 2007; McDowell, 2006; Portney & Watkins, 2009) and t ypically results from an accumulation of evidence over time, not ju st a single assessment (Lynn, 2011). Inter rater Reliability Inter rater reliability should be tested in instruments that are meant to be used in clinical situations to establish that the instrument is able to provide consistent scores (Portney & Watkins, 2009). It is recommended that standardized pain assessment instruments be used in pediatric clinical practice to evaluate and treat pain effectively (American Academy of Pediatrics & American Pain Society, 2001). It is common to use a Pearson correlatio n to analyze for inter rater reliability, but this measure may not yield the most useful data. Correlation data may
54 provide an inflated sense of reliability because it does not capture the amount of agreement in scores, but instead the covariance of the d ata. The intra class correlation coefficient is a better measure that is able to capture the amount of agreement and degree of correspondence among ratings. There are six models of the intra class correlation coefficient and it is the second model that is recommended for testing inter rater reliability where there is a desire to generalize the information. This model is based on the assumption that raters and subjects are randomly chosen from a sample, but it is robust enough to be used with non random s amples for feasibility. Non random samples should be representative of the population in which generalization is desired (Lynn, 2011; McDowell, 2006; Portney & Watkins 2009 ). This test results in an intra class correlation coefficient with a range of 00 1.00 where .00 is interpreted as all measurement values that can be attributed to error, while 1.00 means no error exists in the measurement (Portney & Watkins, 2009). The value for a correlation coefficient that represents acceptable reliability can vary based on the intended use for the instrument. In general, instruments used in clinical settings or for clinical decision making should have a higher minimum acceptable correlation coefficient (McDowell, 2006; Portney & Watkins 2009 ). Portney & Watk ins (2009) suggest the following guidelines for interpretation of correlation coefficients: <.50 poor reliability, .50 .75 moderate reliability, >.75 good reliability, >.90 for instruments that will be used in clinical decisions. Lynn (2011) adds that a co efficient of .70 may be considered acceptable in the testing of new instruments or of instruments in new populations, while .80 is the goal for a mature instrument. Both point out that intra class coefficients will probably be lower than traditional Pears on correlation coefficients. In the literature reviewed (see Table 3 1 ), the minimum accepted intra class coefficient reported was .70, the highest >.80. Results reported and found acceptable varied from .71 .90. Considering
55 the recommendations and curre nt literature, the following values were set as acceptable guidelines for reliability testing in each of the three pain instruments. A non significant p value (p>.05) for the F ratio statistic and an intra class correlation coefficient of > .75 would be ac ceptable to establish reliability. The correlation coefficient meets at least the minimum recommendations and practice standards. Construct Validity Pain is an abstract concept and construct validity is used to support pain instrumentation. Criterion va lidity, where scores are compared to a gold standard instrument, does not generally apply to concepts that cannot be directly measured because gold standards usually do not exist. Several approaches are norm ally used to support construct validity of an i nstrument (McDowell, 2006). This researcher examined two aspects of construct validity for three pain assessment instruments : contrasting group validity and convergent validity. The former selects one group that should score high on the instrument and one group that should score low (Devon et al., 2007; Lynn, 2011). For this study, video clips of children in pain and compared with video clips of the same children not experiencing pain were used for contrasting groups. The resulting mean pain scores of the groups were expected to be significantly different (Devon et al. 2007 ; Lynn 2011 ). In the literature, this contrasting groups approach is often utilized, but ha s many different labels, including construct validity, discriminant validity, sensitivity and extreme group validity (refer to Table 3 1). Divergent or discriminant validity should be reserved for use when the scores of one instrument are compared with the scores of an instrument that measures a different concept (Hulley et al, 2007; McDowell 2006 ). An example in this study would be pain scores compared to scores of level of contentment. Contrasting group validity and extreme group validity can be used interchangeably (Lynn 20 11 ). In the studies in Table 3 1, analysis of variance ( ANOVA ) was typically used to analyze for differences in mean scores if pain or non
56 pain scores were collected for more than one incident per participant. If only one incident of pain and one of non pain were scored, then a matched t test was used for analysis unless the data were non parametric T he Wilcoxon rank sum test was then used. Of the 14 studies reviewed, only one set their significance level, p <.05, for contrasting group validity a priori (Merkel et al., 1997). All studies found a statistically significant difference between their pain and non pain events with t he p values rang ing from p=.0000 to p=.01 (refer to Table 3 1). For this study, each child with ASD was scored for only one pai n event and one non pain event and a matched t test was run to compare the means between the pain event scores and non pain event scores for each instrument. If sample size is greater than 30 when utilizing a matched t test, normality can be assumed to ha ve been reached based on the tenets of the central limit theorem (Field, 2009; Portney & Watkins, 2009). Based on the hypothesis that non pain events should score significantly lower than pain events and the literature reviewed, significance was declared at p < .01. C onvergent validity was analyzed in a bid t o test the construct validity of the three instruments further. Convergent validity is the comparison of scores with another instrument that is proposed to measure the same concept. If the instruments measure the same concept, the scores should be similar (McDowell, 2006; Portney & Watkins, 2009; Trochim, 2006). In this research, the three pain instruments were compared with each other to see if similar results emerge in the scoring of painful events. Eleven of the 14 studies (refer to Table 3 1) used this method of convergent validity, although terminology varies with convergent, concurrent, criterion and construct all being used. Concurrent and criterion validity are norm ally used to validate instruments with direct measures while c onvergent validity is one method used to support the construct validity of instruments that measure abstract concepts (McDowell 2006 ;
57 Portney & Watkins 2009 another pain correlation to analyze the data and correlation. Hunt and colleagues (2004) used ANOVA to analyze convergent validity although, o nce again, only Merkel and colleagues (1997) specified a significance level (p<.05) prior to analyzing data. Reported p values for correlations ranged from p <.0001 to p < .05 for the studies. In th e present study, P instruments if the data were parametric and were non parametric. Statistical significance was declared if p < .05. Statistical an alysis of the data was facilitated by Statistical Package for the Social Sciences. Sample Size Determination Sample size was estimated based on the three statistical tests that were to be run on the data: t test, Pearson correlation and an intra class co rrelation. A minimum of 106 observations class correlation. At an effect size of 5 in a one tailed t test Finally, 23 observations of painful events were required 2012). Another way to estimate sample size in instrument development is to use a standard of more than five observations per variable (Devon et al, 2007). The UWCH PSPNC consists of 30 variables, the greatest number of variables for all three instruments. Using a standard of more than five observations per variable (or a min imum of 6), 180 observations would be needed for each instrument. Use of this standard exceeded the sample size estimates for each of the statistics.
58 Diagnostic Instrumentation The Childhood Autism Rating Scale (CARS) 2 is an instrument that helps distinguish between children with ASD and those who have developmental delays related to some other cause. Its use is also intended for comparisons of the intensity of autism related behaviors in individuals or groups. The CARS 2 Standard version (CARS2 ST), meanwhile, is intended for use in children two years of age and older. The CARS 2 High Functioning version (CARS2 HF) is to be used in children six years or older with IQs greater than 80. Both versions are administered in approximately 15 minutes. T he tools consist of 15 scales that are scored by an ratings, the observer may score 1.5, 2.5, or 3.5. Scores for each scale are totaled. The score range is from 15 60 with a score of 30 or more indicating the child has mild to moderate symptoms of ASD (Schopler, Van Bourgondien, Wellman, & Love 2010). Initial test ing of CARS was performed on a sample of 537 children and the tool was deemed reliable. Internal consistency was established by determining a reliability coefficient al pha for the entire tool (alpha= .94). Inter rater reliability was tested in 280 cases b y comparing the scores of each scale between two independent observers. The average inter rater reliability reported was 0.71. Validity was tested by comparing CARS scores with the clinical rati ngs of ASD by the observers (r= .84, p < .001) and with indepe ndent clinical assessments by child ps ychiatrists or psychologist (r= .80, p < .001) (Schopler, Reichler, Devellis, & Daly, 1980). Multiple other research studies have independently established the reliability and validity of CARS (Mayes et al., 2009). Th e CARS2 ST was tested in 1,034 people ages 2 to 36 years. The instrument demonstrated reliability with in ternal consistency estimate of .93 for total scores. A selection of
59 239 cases was tested for inter rater reliability with a correlation of 0.95 for t otal scores. Validity was tested by verifying that the average CARS2 ST total score (=38.5, SD=8.4) of the sample of patients clinically diagnosed with ASD still supported the screening cutoff score of 30, which is one standard deviation below the mean. The CARS2 ST was also compared with the ASD Diagnostic Observation Schedule (ADOS) and the Social Responsiveness Scale (SRC). Total score correlation for the ADOS and CARS2 ST was .77 in a sample of 37 patients. Total score correlation of the SRC to the CARS2 ST i n a sample of 237 patients was .38. The researcher completed the training materials and received certification in administration of the CARS from Western Psychological Services. The tool was purchased from this company for use in the research study. Pain Assessment I nstruments NCCPC PV The NCCPC PV was developed to assess pain in children with intellectual disabilities following surgical procedures or for pain due to procedures in the hospital. Observations of behaviors are made for six subs cales: vocal, social, facial, activity, body and limbs, and physiological. Each subscale has the observer rate from two to six behaviors on a scale with 0 meaning not at all, 1 just a little, 2 fairly often, 3 very often, or not applicable. The scores fo r each sub scale are calculated and added together for one cumulative score. In all, there are 27 behaviors to rate. The observer does not need to be familiar with the patient or gather any pain instrument (Breau et al., 2002a). The instrument was tested for reliability and validity in 24 children with disabilities undergoing surgery and was found to have inter rater reliability for to tal scores with correlation of .82 prior to surgery and .78 after surgery. They also established construct validity by showing that scores following surgery were significantly higher than before surgery (p= .01). Convergent
60 validity was established by having nurses score pain on a visual analog scale at the same ti me researchers scored pain on the NCCPC PV, with a statistica lly significant correlation (p< .05) (Breau et al., 2002a). The NCCPC PV tools and instructions are copyrighted, but made available freely for research funded by not for profit agencies (Breau, Mc Grath, Finley, & Camfield, 2004). Revised FLACC The FLACC, which was designed to assess pain in children with cognitive impairments, enables observers score pain based on the following five categories: face, legs, activity, cry and consolability, and ea ch category has a description of activities with an associated score ranging from 0 to 2. The score in each category is added together to give a pain score. The Revised FLACC added more descriptions of behaviors that are typical to those exhibited by chi ldren with cognitive impairments when they experience pain, and also contains a category in each section to allow a caregiver to add individual behaviors that the person may exhibit when they are in pain (Malviya et al., 2006). Reliability and validity fo r this instrument were tested in a sample of 52 children with cognitive impairment undergoing elective surgery. Inter rater reliability of the total scores had an intra class correlation coefficient of .90, which wa s statistically significant (p< .001). N ursing Assessment of Pain Intensity scores were calculated for the same pain event by a different nurse observer to test for criterion validity. The correlation between these scores and the Revised FLACC was 0.87 (p< .01). Supporting construct validity, s cores were significantly lower for observations that followed pai n medication administration (p< .001) (Malviya et al., 2006). Permission for use of this scale is has been granted by publishers John Wiley and Sons.
61 UWCH PSPNC The UWCH PSPNC was developed t o help assess pain in children who are not yet verbal and those who are cognitively unable to communicate verbally. This pain assessment tool has five categories: vocal/cry, facial, behavioral, body movement/posture, and sleep, with each category containi ng a rating of 0 5. Descriptors are provided for ratings of 0, 1, 3, and 5 for each category. An observer can score a patient a 2 or a 4 if their behaviors fall between those descriptors and an overall score of 0 5 is given based on the ratings of behavi ors in each subset (Soetenga et al., 1999). The instrument was tested for reliability and validity with 59 pre verbal and 15 non verbal patients who were undergoing procedural pain or had pain medication ordered. The score established inter rater reliab ility with correlations of .92 (p< .001). Criterion validity tests compared scores on this instrument to the Wong Baker Faces Scale scores for the same p ain event; the correlation was .62 (p< .001). Mean pain ratings for scores prior to pain medication and following pain medication were compared by ANOVA to test the construct validity and a significant di fference was found, F=80.82 (p< .001). None of the statistical tests addressed the scoring of only non verbal patients (Soetenga et al., 1999). Permissio n to use this tool was granted by Janetti Publications. Participants and Recruitment Two groups of subjects were recruited for this study. The target population for the first sample consisted of children with ASD who were anticipated to experience procedu ral pain. A purposive sample of children was recruited from local pediatric offices or the emergency department, general care units, pediatric post anesthesia care unit, the pediatric intensive care unit and the outpatient laboratory at WCH. Children elig ible for inclusion in the study were 4 12 years of age, had a score of 30 or greater on CARS 2 (Schopler et al. 2010) and were undergoing
62 a painful procedure as part of their routine healthcare. Children whose parent or guardian felt they were currently experiencing pain were excluded as were c hildren who did not speak English and c hildren who were hearing impaired, visually impaired, or whose cognition or development was impaired by a diagnosis other than ASD. Recruitment of subjects was through flyers and referral from healthcare providers within WCH and in formed consent was obtained from the parent or legal guardian. Due to the targeted age range and the limitations in cognitive, communication and social ability in children with ASD, exemption from obtaining assent from the children was granted by the IRB. At the conclusion of their participation, each child received a $75 gift card to a retail store in appreciation of their time. P ediatric acute care nurses were t he target population for the s econd group of subjects. This was a convenience sample recruited from nurses currently employed at WCH who provide direct care to patients. Nurses included in the study had a minimum of two years practice experience in pediatric acute care and were curr ently practicing as a direct care nurse. Nurses who provided any care during the video recording of the subjects in group one were excluded. Recruitment was through individual contact with potential participants and word of mouth and i nformed consent was obtained from each nurse. At the conclusion of their participation, each nurse received a $25 gift card to a retail store in appreciation of their time. Nine children with ASD were to be recruited for this study and e ach nurse was to be assigned three pa ired vignettes per pain assessment instrument and test all three instruments. A minimum sample size of 30 nurses was needed t o reach a total of 180 observations per instrument, given that each participant has six observations per instrument. Recruitment for participants of children with ASD lasted six months. The researcher was able to make contact through flyers with a number of potential participants, but none of those
63 needed a qualifying procedure during the recruitment period. The children admitted t o the hospital often did not require any qualifying procedures and e fforts were made to reach out to local physician offices with ten days being spent in the emergency department waiting for potential participants. Some of the physicians suggested that i n their experience parents of children with ASD may choose not to immunize, thus eliminating a common childhood source of procedural pain. The decision was made to end recruitment of children with ASD after 6 participants were enrolled as t he sample size of 180 observations was not dependent on the ratio of children with ASD to nurse observers. Considering the difficulties of recruiting the children with ASD and the option to achieve the same statistical power by increasing the number of nurse participant s, the decision was made to end the recruitment of children with ASD. To achieve 180 observations with six participants with ASD, the number of nurse participants was increased from 30 to 45. There were n o difficulties in recruitment of nurse participants. Data Collection Phase 1 Children with ASD Healthcare providers at WCH and in the Jacksonville community helped identify eligible patients and notified the researcher after receiving permission from the parent or legal guardian. The researcher also encountered eligible patients while participating in their healthcare. A HIPPA exemption waiver was granted from the IRB for patients the researcher may have learned about incidentally while not providing direct care -for example, during interdisciplinary rounds or discharge planning. The researcher also assessed the child for eligibility into the study and, if determined eligible, obtained informed consent from the parent or legal guardian. In all, five patients referred to the researcher were not eligible to participate. One was excluded because th ey did not speak English, three were excluded due to age and one was excluded because they were currently
64 in pain. After obtaining informed consent, the researcher completed the CARS2 ST by parental interview and direct observation of the child for approxi mately 20 minutes; all patients screened with the CARS2 and any co morbid diagnoses were collected. The social security number of either the parent or child was also collecte d for reporting to the Internal Revenue Service due to the value of the gift card. Parents, family members and other healthcare workers who appeared in the video recording gave their consent after the purpose of the study was reviewed and they were informe d about the data security measures and destruction plan. No family members or healthcare workers declined to appear in the video recordings. Prior to the procedure, nurses were instructed to perform a non painful task such as auscultating the lungs or cha nging a gown. They were then instructed to proceed as usual with the procedure, using any standard of care practices to reduce pain and anxiety, including pharmacological interventions. The camera was hand held by the researcher so as to be able to move qu The researcher began video recording the participant, after which each nurse performed a non painful task with the participant. The researcher continued to record while the procedure took place, immediately fol lowing the non painful task. All six procedures captured were venipuncture procedures. All six participants had some form of topical analgesia applied, as is standard practice, three participants had child life available for distraction, and five had paren tal presence. after they had been recorded: one clip from the non painful task the nurse performed and one clip during the painful procedure, ranging from 45 to 85 seconds in leng th. The six pairs of clips were randomly
65 assigned to one of the three pain assessment instruments: hence the NCCPC PV, the Revised FLACC, and the UWCH PSPNC were each randomly assigned two pairs of video clips. As all the procedures were venipuncture, there was no need to account for procedure type when making assignments to the pain assessment instruments. Once assigned, the researcher used a table of random numbers to determine the order in which the pain versus non pain clip would be viewed and all p articipating nurses viewed the clips in the same order. A module was designed using Microsoft PowerPoint, which explained how to use each instrument and contained the video clips. Phase 2 Pediatric Nurses After determining study eligibility, informed conse nt was obtained from each nurse participant. Ten nurses were determined ineligible -and two due to their not currently providing direct patient care. Age range, years of experience as a nurse and years of exp erience as a pediatric nurse were collected from each nurse. The researcher then started the module for nurses reviewing the instructions verbally. Nurses were able to ask questions about how to use the instrument during the study. Only general questions about the instrument itself were answered and questions that may have influenced the pain scoring process were not answered. The nurse watched one video clip and then scored the clip. Nurses were able to re watch the video clip if desired, but no one made this request. In all, 45 nurses viewed 12 clips of six different participants. Although nurses were not told which clips represented pain and which represented non pain events, it was impossible to blind the nurses truly to the task or
66 Data Management Informed consents were separated from data collected from the participant and stored in a locked cabinet. Social security numbers were collected on a separate sheet of paper without other identifying data and, as with informed consents, stored in a locked cabinet. Social security numbers will be submitted to the IRB at the University of Florida by the researc her in person and no electronic transfer of data will take place. Video data were recorded by an electronic device onto an SD card that could not be encrypted. In order to protect the privacy of the participants, the video data were transferred to an enc rypted external hard drive prior to leaving the facility. The SD card data was deleted and the card reformatted after the transfer. At the conclusion of the study, the SD card will be submitted to the University of Florida College of Nursing IT Department to be properly destroyed and/or purged. The SD card and video recorder were transported in a locked travel bag and secured in a locked cabinet and locked office when not in the direct possession of the researcher. The video data were played directly from the encrypted hard drive and were not stored locally on any devices. The nurse participants viewed the data in private locations. The demographic information and the pain score results were all de identified and are stored in a password protected file on a password protected device. The primary investigator is the only individual who has access to the device and files.
67 Table 3 1 Pediatric pain instrument studies tests of reliability and construct validity. Instrument First Author, Year Reliability Construct Validity Convergent Validity Contrasting Groups NCCPC ( Breau et al. 2000 ) Test retest Pearson correlation between score and another Termed: concurrent validity Comparison of scores during a pain event to scores during a calm event by ANOVA Termed: sensitivity to pain NCCPC Revised ( Breau 2002 ) N/A Pearson correlation between score and another Termed: concurrent validity Comparison of scores during a pain event to non event scores by matched t test s Termed: discriminative validity NCCPC PV ( Breau et al. 2002 a ) Intra class coefficient model 2 Pearson correlation between score and another Termed: convergent validity Comparison of scores pre pain medication to post pain medication by paired t test Termed: construct validity Pediatric Pain Profile ( Hunt et al. 2004 ) Intra class coefficient model 1 ANOVA of scores to Termed: concurrent validity Comparison of scores during a pain event to non event scores by paired t test s Termed: construct validity PAICP (Boldingh, Jacobs van der Bruggen, Lankhorst, & Bouter 2004) Test retest N/A Comparison of scores during a pain event to non event scores by paired t test s Termed: construct validity PEPPS ( Schultz et al. 1999 ) Pearson correlation N/A Comparison of scores pre pain medication to post pain medication b y paired t test Termed: contrasting groups/construct validity COMFORT ( Bear & Ward Smith 2006 ) One sample t test N/A N/A FACES ( Hicks et al. 2001 ) N/A Correlation between score and two other pain Termed: convergent validity N/A Pediatric Pain Profile ( Hunt et al. 2007 ) Intra class coefficient model 3 Correlation between score and another pain Termed: c oncurrent validity Comparison of scores in high pain group to those in low pain group by independent t tests Termed: e xtreme group validity
68 Table 3 1. Continued Instrument First Author, Year Reliability Construct Validity Convergent Validity Contrasting Groups CPS ( Suraseranivongse et al. 2001 ) Intra class coefficient m odel unspecified Spearman correlation between score and other Termed: concurrent validity Comparison of scores during a pain event to non event scores by Wilcoxon rank sum test Termed: construct validity UWCH PSPNC ( Soetenga et al. 1999 ) Pearson correlation Correlation between score and another pain t test between two scores Termed: criterion validity Comparison of scores during a pain event to non event scores by ANOVA Termed: construct validity FLACC ( Me rkel et al. 1997 ) Pearson correlation and percent agreement Correlation between score and another pain Termed: construct validity Comparison of scores pre pain medication to post pain medication by ANOVA Termed: construct validity FLAC C ( Voepel Lewis et al. 2002 ) Test retest between score and another Termed: criterion validity Comparison of scores during a pain event to non event scores by Wilcoxon rank sum test Termed: construct validity Revised FLACC ( Malviya et al. 2006 ) Spearman correlation and intra class coefficient model unspecified Correlation between score and two other pain Termed: criterion validity Comparison of scores during a pain event with non event sco res by Wilcoxon rank sum test Termed: construct validity
69 CHAPTER 4 RESULTS The purpose of this study was to determine the reliability and validity of three pain assessment instruments when used to assess pain in children with ASD, a population in which these instruments have not been examined. The NCCPC PV, Revised FLACC and UWCH PSPNC were each examined for inter rater reliability and construct validity. Construct validity was tested using two methods: contrasting group validity and convergent validity Participant Data The first group of participants, six in total, included children with ASD undergoing a painful medical procedure as part of their routine healthcare. All participants were male and the median age was 9.67 years. The average CARS2 ST sc ore was 43.0. Co morbid diagnoses were different: hypoglycemia, irritable bowel disease, pneumonia, vomiting, osteomyelitis and one child needed routine well check blood work. All the children in the study underwent venipuncture, either for lab work or int ravenous catheter placement. Table 4 1 contains information about this participant group. The second group of participants included pediatric acute care nurses who were involved in direct patient care. Of the 45 nurses, 44 were female and t he range of ye ars of pediatric nursing experience was 2 39 years, with the average experience 11.83 years. N urses w ere classified by category: 21 30 years (37.8%), 31 40 years (26.7%), 41 50 years (22.2%), and >50 years (13.3%) (Refer to Figure 4 1). Inter rater Reliability Inter rater reliability was assessed by using the intra class correlation coefficient (2,1). This model is based on the assumption that raters and subjects are randomly selected, but is robust enough to be used with non random samples for feasi bility (Lynn, 2011; McDowell,
70 2006; Portney & Watkins, 2009). The NCCPC PV, the Revised FLACC and the UWCH PSPNC were each examined separately with e ach instrument ha ving four cases ( two pairs of video clips from two children with ASD) assigned to it. Eac h case had 45 observations by nurses, yielding 180 observations per instrument. An intra class correlation of >.75 was set as the acceptable value to deem the instrument reliable. The NCCPC PV had an intra class a class correlation was .84, while the intra class correlation of the UWCH PSPNC was also 84. The NCCPC PV did not meet the criteria t hat would deem the instrument reliable for intra class correlation Both the Revised FLACC and the UWCH PSPNC met the standard set for the intra class coefficient and therefore demonstrate strong levels of agreement across the nurses. Hence t hese two instruments were accepted as reliable while the NCCPC PV was not. Construct Validity Construct validity was examined by two tests : a matched t test for each instrument to assess contrasting group validity ; and a Spearman correlation to assess convergent validity. For the matched t test, each pain assessment instrument had 180 obse rvations by nurses from the video clips of children with ASD and t hese were equally divided into 90 observations of a non painful task and 90 observations of a painful procedure per instrument. No test for normality of data was required for matched t test s as there were more than 30 data points (Field, 2009; Portney & Watkins, 2009). Non pain clips were compared to the pain clips for each instrument, with a significance level of p<.01 set a priori The results for the matched t test s for the NCCPC PV de monstrated that pain scores were significantly greater during the clips of painful procedures (M=28.16, SE=1.19) than during non painful tasks (M=17.70, S=1.24) (t (89) 12.73, p=.000). The Revised FLACC also demonstrated pain scores that were significantl y higher from observation of painful procedures clip (M=4.73,
71 SE=.15) than during the observation of the non painful task clip (M=0.76, SE=0.13) (t (89) 20.42, p=.000). Finally, the UWCH PSPNC demonstrated similar results. Clips of painful events (M=2.21, SE=0.17) were scored significantly higher than clips of non painful events (M=0.46, SE=0.06) (t (89) 9.02, p=.000). These results demonstrate that all three pain assessments had good contrasting group validity and this supports the overall construct valid ity of these instruments for use in this new population. Convergent validity was assessed using a Spearman correlation and e ach of the three pain with the other two instruments to see if there would be a correlation since all instruments conceptually measure pain. A Spearman correlation was chosen because the data were non parametric; none of the histograms followed a normal distribution. Refer to Table 4 2 for Spearman correlation data. In brief, eac h pain assessment instrument had strong and significant correlation with the other two instruments and the p value for all the correlations was p=.000. These results indicated that convergent validity is present for all three instruments in this populatio n and can be used in combination with the findings from the contrasting groups validity tests to establish that the NCCPC PV, the Revised FLACC, and the UWCH PSPNC have construct validity when used by pediatric nurses to assess procedural pain in children with ASD. Summary The results from the examination of reliability and validity show that the Revised FLACC and the UWCH PSPNC both demonstrate good reliability and strong construct validity when used by pediatric nurses to assess procedural pain in children with ASD. The NCCPC PV was not able to demonstrate reliability when used in the same context as the other two instruments but did demonstrate strong construct validity. Further discussion on the implications of the results follows in Chapter 5.
72 T able 4 1. Children with ASD age and CARS2 ST scores Minimum Maximum Mean Std Deviation Age 4 12 9.67 3.67 CARS2 ST 32.5 43.0 39.83 3.89 Table 4 2. Spearman correlation NCCPC PV Revised FLACC UWCH PSPNC NCCPC PV p=.000 p=.000 Revised FLACC p=.000 p=.000 UWCH PSPNC p=.000 p=.000
73 Figure 4 1. Age range distribution of nurse participants
74 CHAPTER 5 DISCUSSION The purpose of this research project is to test the reliability and validity of three pediatric pain assessment instruments developed for non verbal, cognitively impaired or neurologically impaired children when utilized by pediatric nurses for children with ASD experiencing pr ocedural pain. The resear ch has sought to ascertain whether the NCCPC PV the Revised FLACC or the UWCH PSPNC was a reliable and valid pain assessment instrument when used by pediatric nurses to assess procedural pain in children with ASD and t he specific aims were to determine th e inter rater reliability, contrasting group validity and convergent validity for each instrument. The researcher designed and implemented this study in order to fill a gap in the current knowledge regarding pain assessment in children with ASD. In gener al, the entire pain experience in children with ASD has suffered from a significant lack of research and there is much to explore and learn. The recognition of the importance of accurate pain assessment to the effective treatment and management of pain (A m erican A cademy P ediatrics & A merican P ain S ociety 2001; Breau et al., 2002 a ; Terstegen et al., 2003; Wells et al., 2008) helped to focus this research in a field that is wide open. The use of existing pain instruments applied in a new population was a th oughtful choice designed to be impactful to children with ASD very quickly. Closing the bench to bedside gap in healthcare research should be a priority for all healthcare researchers. If an existing instrument can be applied to a new population successfu lly, it saves time and resources that can be focused on moving forward other areas where deficits of knowledge exist (Hulley & Cummings, 2007). There is also evidence that existing behavior based pain assessment instruments have been used successfully in different populations or settings from their original design (Blount & Louiselle, 2009; Malavyia et al., 2006).
75 The measurement of pain is always indirect (McGrath & It can be achieved either through self report, behavioral indicators or physiological indicators. While there are over 30 self report instruments for children in the literature, even the non verbal instruments require a certa in level of cognitive function, communication skills, social understanding and the ability to think abstractly (von Bayer & Spagurd, 2007). Children with ASD demonstrate deficiencies in each of these areas (American Psychological Association, 2013). Altho ugh pain assessment in children with special needs faces barriers such as unique pain behaviors, researchers have been able to identify a common core set of behaviors that does not seem to be impacted by cognitive functions, communication ability or social interaction. These include changes in facial expression, changes in body posture, changes in mood or behavior, and vocal cues (Breau et al., 2000; Dubois et al., 2009; Duivenvoorden et al., 2006; Hunt et al., 2004; Terstegen et al., 2003). In the limit ed research available on expression of pain in children with ASD, there is evidence that these children do experience and express pain (Daughters et al., 2007; Inglese, 2008; Messmer et al., 2008; Nader et al., 2004; Tordjman et al., 2009). The Communic support this research because it places equal emphasis on the expression and assessment of pain. This research is interested not only in the expression of pain in children with ASD, bu t whether or not that expression is able to be interpreted by nurses. The model also gives an in depth overview and the many factors, intrinsic and extrinsic, that affect the encoding and decoding of pain between two humans (Craig et al., 1996). This level of detail and the encompassment of both sides of pain communication not only support this study, but will allow an entire program of research to springboard off the theory.
76 Study Considerations Ethics and Data Security The ethics of working with vulnera ble populations in research have been discussed in Chapter 2, but the logistics of executing a research study with a vulnerable population using digital video data warrant discussion as they impacted this study. Two IRBs reviewed and approved this study. T he approval from the first board was obtained in June 2012, pending the approval of the second board, and the second board granted approval in August 2012. After experiencing the secure research database, the researcher became aware that, in order to play video modules from this database, they would have to be first downloaded onto a local device. This would not be acceptable for the protection of the participants in the study as the researcher would have no method of ensuring the data were destroyed proper ly each time. The risk of a confidentiality breach was too great to the participants and so the ensuing search for a more secure solution that the researcher felt was feasible, while balancing the concerns for secure data from the Information Technology Of fice and the IRB, took until March 2013 to resolve. This was six months of lost recruitment time. The difficulties faced when trying to perform research with vulnerable populations, or data that are stored electronically, may discourage other researchers f rom pursing studies in these areas. This becomes an issue of justice for the vulnerable populations (Hirtz & Fitzsimmons, 2009; Lambert & Glackens, 2011). The fragility of these groups warrants some extra precautions in their protections, but the field mus t be cautious not to cause a disservice to these groups by making research not feasible. Video Recording M ethodology The use of video recordings of the children with ASD was vetted to ensure maximum protection. This research has a well defined purpose with specific event criteria that were captured by video and t he choice of video of children experiencing pain to study pain is well
77 documented in the literature (Hunt et al., 2007; Malviya et al., 2006; Schultz et al., 1999; Suraseranivong se et al., 2001; Voepel Lewis et al., 2008). Video is an appropriate medium to use to test behavioral checklists, which include these behavioral based pain instruments (Haidet et al., 2009). The use of video allowed a much larger group of nurses to witne ss and rate the same event, which helped to determine the inter rater reliability of the three pain assessment instruments that could be used in practice to make clinical judgment decisions. The communication and social deficits in ASD reduce d the likeliho od th at the children would be influenced by the presence of the researcher or video equipment and the researcher did not feel that either affected the children These factors support the selection of video data Recruitment Participant recruitment is oft en the primary stumbling block to researchers and this study was no exception. Given the volume of pediatric patients treated at WCH annually, the researcher felt that recruitment of nine participants would be feasible. After six months, despite vigorous recruitment efforts, the recruitment of participants into the children with ASD group was ended with six participants. Since the power of the statistical tests rested on the number of total observations, the number of nurse observers was increased to acc ount for the early end to recruitment in the first group. This was accomplished without comprising the power of the statistical tests. The researcher distributed over 250 flyers to parents or legal guardians of children with ASD and even though the parents /guardians reported they were eager to participate and would contact the researcher if the need for a medical procedure arose, not one participant and agreed to request the permission of the parent/guardian to give contact information to the researcher if appropriate, but no patients were recruited in this manner either. After discussion with colleagues, and assessment of personal experience, the researcher pr oposes that most
78 children with ASD do not have an additional chronic illness that would require a medical procedure. Furthermore, from personal experience and reports from physician colleagues, many parents of children with ASD chose not vaccinate their ch ildren, eliminating a procedure that even healthy children undergo. In contrast to the difficulties of recruiting children with ASD, the nurses in this study were available and eager to participate. Strengths and Weaknesses The sample of children with ASD was very homogenous: the children were all school aged males undergoing venipuncture with a relatively small range of scores on the CARS2 ST (32.5 43.0). For the purpose of this initial study, this author views this as a str ength. With only six participants in this group, a homogenous sample may have helped to contribute to the clarity of the results. In future studies, a larger and more heterogeneous sample should be sought including females, a range of functional levels of ASD, and more procedure variation to see if these results can be generalized. A key strength of this study lies in the large number of observations (540) by practicing pediatric acute care nurses. Often in research, the observations are made only by one or two researchers but the instruments in this study are designed to be used by nurses who provide patient care and will use the results as a decision to treat and manage pain. The inclusion of clinicians in the research of reliability and validity of instru ments increases the clinical applicability. The video recording of the events allowed 45 nurses to view and score the same events, lending strength to tests of reliability and validity, and further work should test if these same results can be obtained i n a naturalistic setting. The nurses were unfamiliar with the child, save for knowing that they had ASD, and this should be an endorsement that these scales are able to be achieve valid results even when the patient is unknown to the observer. This is impo rtant
79 because it has been reported that a key element in assessing unique pain behaviors is know the child (Breau et al., 2002b; Hunt et al., 2003). Knowing the child is not feasible in many acute care settings -the child in pain needs pain to be recogni zed and treated whether or not they have encountered the nurse before. The intention of the researcher had been to blind the nurses as to whether the clip was a pain or non pain clip, but this was not possible. The non pain tasks and venipuncture procedu res were apparent on the video clips, and digitally altering the video would have blocked some area told which clips were intended to be pain or non pain and th e clip viewing order of pain versus non pain was randomly arranged. In many studies where live observation is taking place, blinding is also not possible. The ability to see what procedures were taking place (i.e. venipuncture or auscultation) may have in fluenced the nurses. One of the concerns with using behavioral based pain assessment instruments in children is that the procedure or intervention will cause the child anxiety and distress and the manifestations of these behaviors may be translated incorr ectly as pain. This effect may be greater in the child with ASD, who may have sensory issues with touch or the noise of the acute care setting or become agitated due to the unfamiliarity of the situation. To counteract this effect, the researcher had the same healthcare worker perform a non painful task just prior to the painful procedure. This task took place in the same location as the procedure, with the same people present. If the environment or healthcare worker were going to cause a reaction that may mimic pain behaviors, these behaviors would be captured here and could be contrasted with the behaviors the child exhibited during the painful procedure. The results of the matched t tests comparing the non pain and pain clip scores support strongly the conclusion that there was a
80 significant difference in non pain versus pain scores (p=.000 for all three instruments). This suggests that instruments were able to differentiate between pain behaviors and behaviors from other causes. Implications and Applica tion to Clinical Settings NCCPC PV The NCCPC PV was designed and tested in intellectually disabled children for use during the post operative period. The scoring system is based on 27 behaviors that are scored based on the frequency of occurrence during t he observation period. Researchers unfamiliar with the child observed the child for a period of 10 minutes before and after surgery (Breau et al., 2002 a). This researcher also had observers who were unfamiliar with the child, but the pain being scored was procedural, and not post operative. The type of procedure, venipuncture, is likely to entail a much shorter pain experience than post operative pain. Notably, the clips of observation were 45 85 seconds, nowhere near approximating the suggested 10 minute observation period as designed in the original research. The researcher feels this limited time of observation may have contributed to the findings related to reliability in the scoring. This instrument did not meet the standards set for reliability by t he researcher (ICC=.71). Although by some standards, the ICC could be deemed acceptable, especially when testing an instrument in a new population (Lynn, 2011). The instrument was able to show strong construct validity even when applied to a new populati on, different type of pain, and much shorter observation period. There was significant difference when comparing pain clip scores with non pain clip scores: pain clip scores were higher ( t (89) 12.73, p=.000). This helps to support the notion that the rat ed behaviors demonstrated pain and not emotions such as anxiety or distress. The strong p=.000) also lends support that the instrument does indeed measure the concept of pain.
81 Although this instrument shows promising results for use in varying types of pain in children with ASD, the researcher questions the feasibility of this instrument in practice. The instrument instructions recommend a ten minute observation period that, while not continuous, suggests that the observer be close to the patient during that time. It is stated that shorter observations may be used, but that cut off scores for pain may not apply. The likelihood of an acute care nurse having the time to b e in the presence of one patient for ten minutes, while not performing any tasks or interventions with the patient, is extremely low. Pain is now considered the fifth vital sign and many organizations are recommending assessment of pain at a minimum of ev ery four hours (National Database Nursing Quality Indicators, 2012; Wells et al., 2008). In addition to the lengthy observation time, the tool itself is long and the observer must rate 27 behaviors in five sub categories, total the scores for subcategorie s, and sum the subcategories to achieve a final rating individually. One final concern is unfamiliarity of healthcare personnel with the range of scores (0 81). Many pain assessment instruments in use have a scale of 0 5, 5 10, or 0 100. This score rang e is odd, and the cut off scores for moderate to severe pain (>10) are on the low end of the total score range (Breau et al., 2002 a). This may be confusing for healthcare providers and may inhibit the recognition of pain. Revised FLACC The Revised FLACC w as created to address pain assessment in children with cognitive impairments and was tested in a post operative setting. Behavior descriptors were added to the FLACC that were found unique to children with CI. The Revised FLACC also allows for a parent or caregiver to make additions to the descriptors and add unique pain behaviors. The study sample did include children with ASD (16%) (Malviya et al., 2006). In this research study, the procedural pain differed from post operative pain in the original study Children with ASD were the only children included. Scores in this study with the Revised FLACC were deemed reliable
82 (ICC=.84) among observers and were also significantly higher when associated with video clips of painful procedures versus the clips of no n painful procedures ( t (89) 20.42, p=.000). Furthermore, when compared with scores of the NCCPC measure pain and these results supp ort that it does. A key design measure of the Revised FLACC is that caregivers are able to add unique pain behaviors, although this is not required. It is interesting to note that only in 25% of the original research cases did caregivers elect to add unique pain behaviors. In th e present study, the researcher did not offer the parent the option to add unique pain behaviors. Since the majority of unmodified instrument. Although individuali zed pain assessment instruments seem like an ideal, it is highly desirable to have an instrument that is standardized and able to capture the pain behaviors of a child that is completely unknown to the observer. In the world of electronic medical records, individualizing a patient assessment tool presents logistical problems. It is no longer as simple as writing the unique pain behaviors down and making copies there are real obstacles to individualizing the record and this supports further the needs for u niform pain assessment instruments. It should be noted that nurses in this study use the FLACC often to assess pain in non verbal or pre verbal children and it is one of three standard pain assessment instruments utilized at WCH. The instrument itself is v ery feasible for use in clinical practice. Five behaviors are scored based on brief descriptors from 0 to 2, after which the scores are added to give a familiar range of 0 10 (Malviya et al., 2006 ). The observer can make the observations in a relatively s hort period of time and the scoring on the tool is not time consuming, although more investigation is needed to support the findings of this research as a reliable and valid
83 instrument to assess pain in children with ASD. At this point, the author recommen ds the use of this instrument for pain assessment in children with ASD in the clinical setting. UWCH PSPNC The UWCH PSPNC was designed for use in infants, pre schoolers and non verbal children. The sample included infants and pre verbal children (66%), and non verbal children (44%). It is not stated if any of the children had ASD. The instrument was used to assess pain in a variety of situations including procedural pain. Direct care nurses, in addition to researchers, were also involved in the original te sting of this instrument (Soetenga et al., 1999) and this research utilized all direct care nurses as observers in a sample of children with ASD. These children had communication impairments, but were not necessarily non verbal. All the children in the cu rrent study experienced procedural pain, inter rater reliability among the 45 nurses was good (ICC=.84) and non pain clip scores were significantly lower than pain clip scores ( t (89) 9.02, p=.000). In addition, the convergent validity found significant co rrelations to the NCCPC this instrument does measure pain. The instrument is designed around five behaviors that the observer scores, based on descriptors on a scale of 0 5. The observer then uses these sub set scores to determine an overall rating of pain from 0 5. Most of the scoring could be accomplished through short observations, but one category, sleep, requires at least some extended period of observation or co ntribution about the sleep patterns from a caregiver (Soetenga et al., 1999) The instrument does not take long to complete after the observation is complete and is appealing for use in a clinical setting because it is designed to assess pain in a wide var iety of patients. A different instrument is not needed for pain assessment in the infant versus pain assessment in a cognitively impaired 12 year old. This would create more standardization in the clinical setting and be less burdensome to
84 healthcare worke rs as the number of instruments could be reduced. One concern with this tool is that, although there are objective descriptors of pain, the overall rating is a subjectively derived score, although in this study and the original work, the tool had good reli ability. Further investigation of the use of this instrument in children with ASD should continue, but the instrument could be used in clinical settings to assess pain in children with ASD. The construct validity of these three different pain assessment i nstruments supports the findings of other research that, even though some children may exhibit unique pain behaviors, there may be a core set of behaviors demonstrated regardless of impairments that can lead to accurate assessment of pain (Breau et al., 20 02 b ; Dubois et al. 2009; Duivenvoorden, et al. 2006; Hunt et al., 2003; Terstegen et al. 2003) Further testing is needed to confirm these results, but any attempt made to assess pain objectively is a step forward for these patients. Pain that is not as sessed will go untreated (Terstegen et al., 2003). Future The researcher feels strongly that the use of the Revised FLACC or the UWCH PSPNC in children with ASD in the clinical setting can begin immediately, with some confidence that the scores obtained wi ll measure pain, differentiate from other behaviors and be reliable. Further investigation is needed to confirm these results and expand the applicability of these instruments as research tools. This study lays a foundation for work that can be built upon in several directions. One such direction is a focus on diversifying and further exploring children with ASD. The use of a larger and more diverse sample of children with ASD is needed to support these results further. An exploration into how functional level affects pain expression and assessment
85 al., 1996). A larger variety of pain experiences should also be examined, as this study was limited to procedural an d all participants underwent venipuncture. In addition to focusing on the children with ASD, attention can be directed towards the (Craig et al., 1996) Do prior held beliefs abou t pain in ASD influence the scores by nurse observers? Age, years of experience, educational preparation, or familiarity with ASD are among the other factors that may influence the interpretation of pain signals. Testing these results in a natural setting is also needed. Nurse satisfaction with the pain assessment instrument and some measures of feasibility, such as average time to observe and complete the tool, is an interesting avenue to explore further. Comparing the satisfaction of nurses with the ins truments and the feasibility with the psychometric measurements would be useful information. The gap between research and implementation needs to be significantly closed and considering the thoughts of end users in instrument development may help accelerat e changes in practice. Disseminating knowledge in research is one of the most important aspects in the research process. This author has professional experience with evaluating research, implementing practice changes based on the findings of research, eva luating outcomes of the changes and disseminating information. This researcher plans firstly to share the successful findings of her research at national clinical conferences as t he clinicians who attend these conferences are often seeking new evidence to change. The next step would be to publish the findings in a journal with an audience for pediatric acute care clinicians as t his is the target audience for this knowledge. During this time, th e researcher would be working towards change in her own institution and s uccessful
86 implementation of this change would provide further opportunities to present the findings of this research at conferences or in journals from a new angle. Sharing the process of implementation of research, including barriers, helps other institutions envision how they would implement change successfully. Conclusion Pain assessment is not an easy concept to study. Pain is an abstract, complex and subjective experien 2004). Children with communication impairments or language delays, such as those seen in ASD present even more of a challenge to the assessment of pain (Voepel Lewis et al., 2008). Overall, pain and pain assessment research in delayed, cognitively or neurologically impaired children do not seem to be a priority in healthcare. Routine, systematic pain assessments lead to better treatment of pain (American Academy of Pediatrics & Am erican Psychiatric Society, 2001; Breau et al., 2002a; Terstegen et al., 2003; Wells et al., 2008). Pain in ASD is grossly understudied and much of the focus has been on proving that children with ASD do experience pain. This research examined the inter rater reliability, the convergent validity and contrasting group validity of three existing pain assessment instruments for use by pediatric nurses assessing procedural pain in children with ASD. The findings have revealed strong construct validity in the new population for the NCCPC PV, Revised FLACC and UWCH PSPNC. These instruments were able to demonstrate the ability to measure pain and distinguish pain from non painful healthcare tasks that may induce similar behavior. These findings are encouraging, in that pain in children with ASD may be able to be recognized and assessed objectively by nurses unfamiliar with the child. The Revised FLACC and UWCH PSPNC also demonstrated good reliability and both instruments are feasible for application in the busy clinical setting. The researcher
87 recommends the use of either of these instruments to assess pain in children with ASD in a clinical setting, while recognizing that further exploration is needed.
88 LIST OF REFERENCES American Academy of Pediatrics and American Pain Society. (2001). The assessment and management of acute pain in infants, children, and adolescents. Pediatrics, 108, 793 797. doi:10.1542/peds.108.3.793 American Psychiatric Association. (2000). Diagnostic and statistical manual of mental di sorders (4th ed., Rev.). Washington, DC: Author. American Psychiatric Association. (2010). Proposed r evision: 299.00 a utistic d isorder Retrieved August 25, 2010 from http://www.dsm5.org/ProposedRevisions/Pages/proposedrevision.aspx?rid=94# Americ an Psychiatric Association (2013 ). Diagnostic and statistical manual of mental disorders (5th ed. ). Washington, DC: Author. Aylott, J. (2000). Understanding children with au tism: Exploding the myths. British Journal of Nursing, 9, 779. Retrieved from www.britishjournalofnursing.com Baptist Health. (2013). Wolfson Children's Hospital Retrieved from http://www.wolfsonchildrens.org Barnbaum, D.R. (2008). The ethics of autism: Among them, but not of them. Bloomington, IN: Indiana University Press. Bea u champ, T.L. & Childress, J.F. (2009). Principles of biom edical ethics. New York, NY: Oxford University Press. Bear, L.A. & Ward Smith, P. (2006). Inter rater reliability of the COMFORT scale. Pediatric nursing, 32, 427 434. Retrieved from http://www.pediatricnursi ng.net world. Topics in Language Disorders, 4, 375 387. doi: 10.1097/01.tld.0000341130.39275.fa Blount, R.L., & Loiselle, K.A. (2009). Behavioral assessmen t of pediatric pain. Pain Research and Management, 14 (1), 47 52. Retrieved from: http://www.pulsus.com/journals/journalhome.jsp?sCurrPg=Home Boldingh, E.J., Jacobs van der Bruggen, M.A., Lankhorst, G.J., & Bouter, L.M. (2004). Assessing pain in patients with severe cerebral palsy: Development, reliability, and validity of a pain assessment instrument for cerebral palsy. Archives of Physical and Rehabilitation Medicine, 85 758 766. doi: 10.1016/S0003 9993(03)0990 0
89 Breau, L.M., Camfield, C.S., Symons, F.J., Bodfish, J.W., MacKay, A., Finley, G.A., & McGrath, P.J. (2003). Relation between pain and self injurious behavior in nonverbal children with severe cognitive impairments. Journa l of Pediatrics, 142 (5). doi:10.1067.mpd.2003.163 Breau, L. M., Finley, G. A., McGrath, P. J., & Camfield, C. S. (2002a). Validation of the non communicating children's pain checklist postoperative version. Anesthesiology, 96, 528 535. doi:10.1097/0000054 2 200703000 00004 Breau, L. M., McGrath, P. J., Camfield, C. S., & Finley, G. A. (2002b). Psychometric properties of the non communicating children's pain checklist revised. Pain, 99, 349 357. doi:10.1016/S0304 3959(02)00179 3 Breau L. M., McGrath, P. J., Camfield, C., Rosmus, C., & Finley, G. A. (2000). Preliminary validation of an observational pain checklist for persons with cognitive impairments and inability to communicate verbally. Developmental Medicine & Child Neurology, 42, 609 616. doi:10.1111/j.1469 8749.2000.tb066.x Breau, L., McGrath, P., Finley, A., & Camfield, C. (2004). Non communicating children's pain checklist Postoperative version Retrieved from http://www.aboutkidshealth.ca/Shared/PDFs/AKH_Breau_post op.pdf Bronsard, G., Botbol, M., & Tordjman, S. (2010). Aggression in low functioning children and adolescents with autistic disorder. PLoS ONE, 5, e14358. doi: 10.1371.journal.pone.001 4358 Broyles, L.M., Tate, J.A., & Happ, M.B. (2008). Video recording in clinical research. Nursing Research, 57 59 63. doi: 10.1097/01.nnk.0000280658.81136.e4 Canitano, R., Luchetti, A., & Zappella, M. (2005). Epilepsy, electroencephalographic abnormal ities, and regression in children with autism. Journal of Child Neurology, 20 (1), 27 31. doi: 10.1177/08830738050200010401 Caldwell, K. & Atwal, A. (2005). Non participant observation: Using video tapes to collect data in nursing research. Nurse Researc her, 13, 42 54. Retrieved from http://www.nurserearcher.co.uk Centers for Disease Control and Prevention. (2012). Prevalence of autism spectrum disorders Autism and Developmental Disabilities Monitoring Network 14 site, United States, 2008. Morality and Morbidity Weekly, 61 (3), 1 19. Retrieved from http://www.cdc.gov/mmwr/pdf/ss/ss6103.pdf Chiang, H.M. (2008). Expressive communication of children with autism: The use of challenging behavior. Journal of Intellectual Disability Research, 52, 966 972. doi: 10.1111/j.1365 2788.010142.x
90 Chiang, H.M. & Lin, Y.H. (2008). Expressive communication of children with autism. Journal of Autism and Developmental D isorders, 38, 538 545. doi: 10.1007/s10803 007 0423 z Chen, D.T., Miller, F.G., & Rosenstein, D.L. (2003). Ethical aspects of research into the etiology of autism. Mental Retardation and Developmental Disabilities, 9 48 53. doi: 10.1002/mrdd.10059 Cohen D. & Crabtree, B. (2006). Qualitative research guidelines project. Retrieved from http://www.qualresorg/homeobse 3594.html Craig, K.D. (2009). The social communications model of pain. Canadian Psy chology, 50 (1),22 32. doi: 10.1037/a0014772 Craig, K. D., Lilley, C. M., & Gilbert, C. A. (1996). Barriers to optimal pain management in infants, children, and adolescents. Social barriers to optimal pain management in infants and children. Clinical Journ al of Pain, 12 (3), 232 242. Retrieved from: http://journals.ww.com/clinicalpain/pages/default.aspx Daughters, H., Palermo, T., & Koh, J. (2007). Procedural pain and distress in children with autism a pilot study. Journal of Pain, 8 (4S1), S31. doi:10.1111/j.1469 8749.2006.t612595.x (2010). Conducting video research in the learning sciences: Guidance on selection, analysis, technology, and ethics. Journal of Learning Sciences, 19 (1), 3 53. doi: 10.1080/10508400903452884 Devon, H., Block, M. E., Moyle Wright, P., Ernst, D. M., Hayden, S. J., Lazzara, D. J., & Kostas Polston, E. (2007). A psychometric t oolbox for testing validity and reliability. Journal of Nursing Scholarship, 39 (2), 155 164. doi:10.1111/j.1547 5069.2007.00161x Duivenvoorden, H. J., Tibboel, D., Koot, H. M., Van Dijk, M., & Peters, J. W. (2006). Pain assessment in profound cognitive im paired children using the checklist pain behavior; is item reduction valid? Pain, 126, 147 154. doi:10.1016/j.pain.2006.06.025 Dubois, A., Capdevila, X., Bringuier, S., & Pry, R. (2009). Pain expression in children with an intellectual disability. Europea n Journal of Pain doi:10.1016/j.ejpain.2009.10.013 DuFon, M.A. (2002). Video recording in ethnographic SLA research: Some issues of validity in data collection. Language Learning and Technology, 6 (1), 40 59. Retrieved from http://llt.msu.edu/vol6num1/dufon Esposito, G. & Venuti, P. (2010). Understanding early communication signals in autism: A Journal of Intellectual Disability Research, 54, 216 223. doi: 10.1111/j.1365 2788.2010.01252.x
91 Field, A. (2009). Discovering statistics using SPSS (3 rd ed.). London: Sage Publications Finley, G. A., Kristjansdottir, O., & Forgeron, P. (2009). Cultural influences on the assessment of children's pain. Pain Research and Management, 14 (1), 33 37. Retrieved from: http://www.pulsus.com/journals/home.jsp?sCurrPg=Home Gibbs, G.R., Friese, S., & Mangabeire, W.C. (2002). The use of new technology in qualitative research. Q ualitative Social Research, 3 (2), Art. 8. Retrieved from http://nbn resolving.de/um:nbn:de:0114 fqs020287 Hadjistavropoulous, T., & Craig, K. D. (2002). A theoretical framework for understand ing self report and observational measures of pain: a communications model. Behavior Research, 40, 551 570. doi:10.1016/s0005 7967(01)00072 9 Haidet, K.K., Tate, J., Divirgilio Thomas, D., Kolanowski, A., & Happ, M.B. (2009). Methods to improve reliabilit y of video recorded behavioral data. Research in Nursing and Health, 32 465 474. doi: 10.1002/nur.20334 Halle, J., Brady, N.C., & Drasgo, E. (2004). Enhancing socially adaptive communicative repairs of beginning communicators with disabilities. American Journal of Speech Language Pathology, 13, 43 54. doi: 1058 0360/01/1301/0043 Hall, R. (2000). Video recording as theory. In D. Lesh & A. Kelley (Eds.) Handbook of research design in mathematics and science education (pp. 647 664). Mahweh, NJ: Lawrence Erl baum. Heflin, L. J., & Alaimo, D. F. (2007). Students with autism spectrum disorder: Effective instructional practices Upper Saddle River, NJ: Pearson Education. Hicks, C. L., von Baeyer, C. L., Spafford, P. A., Van Korlaar, I., & Goodenough B. (2001). The FACES pain scale revised: Toward a common metric in pediatric pain measurement. Pain, 93 (2), 173 183. doi:10.1016/s0304 3959(01)00314 1 Hirtz, D.G. & Fitzsimmons, L.G. (2002). Regulatory and ethical issues in the conduct of clinical research involving children. Current Opinion in Pediatrics, 14, 669 675. doi: 10.1097/MOP.0b01.3e32834ee.5cc Hulley, S.B. Cummings, S.R., Browner, W.S., Grady, D.G., & Newman, T.B. (2007). Designing clinical research (3 rd ed.). Philadelphia, PA: Wolter s Kluwer. Hunt, A., Goldman, A., Seers, K., Crichton, N., Mastroyannopoulou, K., Moffat, V., & Brady, M. (2004). Clinical validation of the paedatric pain profile. Developmental Medicine & Child Neurology, 46, 9 18. doi:10.1111/j.1469.8749.2004.tb00428.x
92 Hunt, A., Mastroyannopoulo, K., Goldman, A., & Seers, K. (2003). Not knowing the problem of pain in children with severe neurological impairment. International Journal of Nursing Studies, 40 171 183. doi:10.1016/s0020 7489(02)00058 5 Hunt, A., Wisbea ch, A., Seers, K., Goldman, A., Crichton, N. Perry, L., & Mastroyannopoulou, K. (2007). Development of the paediatric pain profile: Role of video analysis and saliva cortisol in validating a tool to assess pain in children with severe neurological disabili ty. Journal of Pain and Symptom Management, 33, 276 288. doi:10.106/j.jpainsymman.2006.08.011 Huth, M.M. & Moore, S.M. (1998). Prescriptive theory of acute pain management in infants and children. Journal of Specialists in Pediatric Nursing, 3, 1, 23 32. doi: 10.1111/j.1744 6155.1998.tb00206.x Inglese, M. D. (2008). Pain expression in children with autism: A foundation for instrument development (Dissertation, University of Florida, Gainesville, U.S.). Retrieved from http://etd.fcla.edu/UF/UFE0022440/inglese_m.pdf International Association for the Study of Pain. (1994). Pain definitions: Pain. Retrieved from http://www.iasp pain.org/AM/Template.cfm?Section=General_Resource_Links&Templates International Association for the Study of Pain. (2005). Pain issues in individ uals with limited ability to communicate due to cognitive impairment. In J.E. Charlton (Ed.), Core curriculum for professional education in pain. Retrieved from: https://www.iasp pain.org/AM/Template.cfm?Section=Home&Template=/CM/ContentDisplay.cfm&Conte ntID=2087 Kanner, L. (1943). Autistic disturbances of affective contact. The Nervous Child, 2, 217 250. Retrieved from: http://neurodiversity.com/library_Kanner_1943.pdf Kopelman, L.M. (2002). Human subjects research, ethics, and research on children. In T.H. Murray & M.J. Mehlman (Eds.), Encyclopedia of ethical, legal, and policy issues in biotechnology. doi: 10.1002/0471250597.mur062 Lambert, V. & Glacken, M. (2011). Engaging with children in research: Theoretical and practical implications of negotiating informed consent/assent. Nursing Ethics, 186 781 801. d oi: 10.1177/0969733011401122 L ynn, M.R. (2011). Classical measurement theory, reliability, and validity. Instrumentation, development, testing, and revision Symposium conducted University of North Carolina at Chapel Hill Center for Lifelong Learning, Cha pel Hill, NC.
93 Malviya, S., Voepel Lewis, T., Burke, C., Merkel, S., & Tait, A. R. (2006). The revised FLACC observational pain tool: Improved reliability and validity for pain assessment in children with cognitive impairment. Pediatric Anesthesia, 16, 25 8 265. doi:10.111/j.1460 9592.2005.01773.x Matutina, R. (2009). Ethical issues in research with children and young people. Paediatric Nursing, 21 (8), 38 44. Retrieved from: http:// www.nursingchildrenandyoungpeople.rcnpublishing.co.uk Mayes, S. D., Calhoun, S. L., Murray, M. J., Morrow, J. D., Yurich, K. K., Mahr, F.,...Petersen, C. (2009). Comparison of scores on the checklist for autism spectrum disorder, childhood autism rating s cale, and Gilliam Asperger's disorder scale for children with low functioning autism, high functioning autism, Asperger's disease, ADHD, and typical development. Journal of Autism and Developmental Disorders, 39, 1682 1693. doi:10.1007/s10803 009 0812 6 M eadan, H., Halle, J.W., Watkins, R.V., & Chadsey, J.G. (2006). Examining communication repairs of 2 young children with autism spectrum disorder: The influence of environment. American Journal of Speech Language Pathology, 15, 57 71. doi: 1058 0360/06/1501 0057 McCaffery, M. (1968). Nursing practice theories related to cognition, bodily pain, and man environment interactions. Los Angeles, CA: University of California at Los Angeles McDowell, I. (2006). Measuring health: A guide to rating scales and questionnaires (3 rd ed.). New York, NY: Oxford University Press. McGrath, P.J., Johnson, G., Goodman, J.T., Schillinger, J., Dunn, J., & Chapman, J. (1985). CHEOPS: A behavioral scale for rating postoperative pain in children. In H. Fields, R. Dubner, & F Cervero (Eds), Advances in pain research and therapy. (Vol 9 ). New York, NY: Raven Press. McGrath, P. J., & Unruh, A. M. (2006). Measurement and assessment of paediatric pain. In S. McMahon & M. Koltzenburg (Eds.), Wall and Melzack's textboo k of pain (5th ed., pp. 305 315). Philadelphia, PA: Elsevier. Melzack, R. (2005). Evolution of the neuromatrix theory of pain. The Prithvi Raj Lecture: Presented at the Third World Congress of World Institute of Pain, Barcelona 2004. World Institute of Pain, 5 (2), 85 94. doi:10.1111/j.1533 2500.2005.05203.x Merkel, S. I., Voepel Lewis, T., Shayevtiz, J. R., & Malviya, S. (1997). The FLACC: A behavioral scale for scoring postoperative pain in young children. Pediatric Nursing, 23 (3), 293 295. Retrieved f rom http://www.pediatricnursing.net
94 Merkel, S.I., Voepel Lewis, T., & Malviay, S. (2002). Pain control: Pain assessment in infants and young children: The FLACC scale. American Journal of Nursing, 102 (10), 55 58. Retrieved from http://www.jstor.org/stable/3522977 Messmer, R. L., Nader, R., & Craig, K. D. (2008). Brief report: Judging pain intensity in children with autism undergoing venipuncture: The i nfluence of facial activity. Journal of Autism and Related Developmental Disorders, 38, 1391 1394. doi:10.1007/s10803 007 0511 0 Nader, R., Oberlander, T. F., Chambers, C. T., & Craig, K. D. (2004). Expression of pain in children with autism. Clinical Jour nal of Pain, 20, 88 97. Retrieved from http://journals.lww.com/clinicalpain/pages/default.aspx National Database of Nursing Quality Indicators (2012). NDNQI transforming data into qual ity care. Retrieved from http://www.nursingquality.org National Institute on Deafness and Other Communication Disorders (2010). Communication problems in children with autism (NIH Publication No 10 4315). Retri eved from http://www.nidcd.nih.gov/staticresources/health/voice/CommunicationProblemsInChildre nWithAutism.pdf Noens, I. & van Berckelaer Onn es, I. (2004). Making sense in a fragmentary world: Communication in people with autism and learning disability. Autism, 8, 197. doi: 10.1177/1362361304042723 Noens, I. & van Berckelaer Onnes, I. (2005). Captured by details: sense making, language, and c ommunication in autism. Journal of Communication Disorders, 38, 123 141. doi: 10.1016/j.jcomdis.2004.06.002 Notbohm, E. (2006). Ten things your student with autism wishes you knew Arlington, TX: Future Horizons. policy to practice. Physical Therapy, 84 (6), 560 570. Retrieved from: http://ptjournal.apta.org Portney, L.G. & Watkins, M.P. (2009). Foundations of clinical research: Applications to practice (3 rd ed.). Upper Saddle River, NJ: Pearson Prentice Hall. Prizant, B.M, Wetherby, A.M., Rubin, E., & Laurent, A.C. (2003). The SCERTS model: A transactional, family centered approach to enhancing communication and socioemotional abilities of children with autism spectrum disorder. Infants and Young Children, 16 296 316. Retrieved from http://journals.lww.com /iycjournal/pages/default.aspx Prkachin, K. M., & Craig, K. D. (1995). Expressing pain: The communication and interpretation of facial pain signals. Journal of Nonverbal Behavior, 19 (4), 191 205. doi:10.1007/BF02173080
95 Rice, C. (2009). Prevalence of auti sm spectrum disorders autism and developmental disabilities monitoring network, United States, 2006. Mortality and Morbidity Weekly, 58 (SS 10), 1 20. Retrieved from http://www.cdc.gov/mmwr/mmwr_wk/ wk_cvol.html Rosenthal, R. (2005). Conducting judgment studies: Some methodological issues. In J. Harrigan, R. Rosenthal & K. Scherer (Eds.), The new handbook of methods in nonverbal behavior research (pp. 199 236), New York, NY: Oxford Press. Scarpinato N., Bradley, J., Kurbjun, K., Bateman, X., Holtzer, B., & Ely, B. (2010). Caring for the child with an Autism Spectrum Disorder in the acute care setting. Journal for Specialists in Pediatric Nursing, 15 (3), 244 254. doi:10.1111/j.1744 6155.2010.00244.x Schopler, E., Reichler, R. J., Devellis, R. F., & Daly, K. (1980). Toward objective classification of childhood autism: Childhood autism rating scale (CARS). Journal of Autism and Development Disorders, 10 (1), 91 103. doi:10.1007/BF02408436 Schopler, E., Van Bourgondien, M.E., Wellman, G.J., & Love, S.R. (2010). Childhood autism rating scale (2 nd ed.). Torrance, CA : Western Psychological Services. Schultz, A.A., Murphy, E., Morton, J., Stempel, A., Messenger Rioux, C., & Bennett, K. (1999). Pre verbal early verbal pediatric pain scale (PEPPS): Development and early psychometric testing. Journal of Pediatric Nursing, 14 (1), 19 27. doi: 10.1016/S0882 5963(99)80056 6 Shamoo, A.E. & Resnik, D.B. (2009). Responsible conduct of research. New York, NY: Oxford University Press. Soper, D. (2012). Statistics calculator Retrieved from: http://www.danielsoper.com/statcalc3/ Soetenga, D., Frank, J., & Pellino, T. (1999). Assessment of the validity and reli ability of the University of Wisconsin Children's Hospital Pain Scale for Pre verbal and Nonverbal Children. Pediatric Nursing, 29 (6), 670. Retrieved from http://pediatricnursing.net/ Souders, M. C., Freeman, K. G., Depaul, D., & Levy, S. E. (2002). Caring for children and adolescents with autism who require challenging procedures. Pediatric Nursing, 28 (6), 555. Retrieved from http://pediatricnursing.net Sowden, H., P erkins, M., & Clegg, J. (2008). The co development of speech and gesture in children with autism. Clinical Linguistics and Phonetics, 22, 804 813. doi: 10.1080/02699200801919745 Stangor, C. (2011). Research methods for the behavioral sciences (4 th ed.) Belmont, CA: Wadsworth.
96 Stinson, J.N., Kavanagh, T., Yamada, J. Gill, N., & Stevens, B. (2006). Systematic review of the psychometric properties, interpretability and feasibility of self report pain intensity measures for use in clinical trials in child ren and adolescents. Pain, 125, 143 157. doi:10.1016/j.pain.2006.02.006 Suraseranivongse, S., Santawat, U., Kraiprasit, K., Petcharatana, S., Prakkamodom, S., & Muntraporn, N. (2001). Cross validation of a composite pain scale for preschool children withi n 24 hours of surgery. British Journal of Anaesthesia, 87, 400 405. doi: 10.1093/bja/87.3.400 Tager Flusberg, H., Paul, R., & Lord, C. (2005). Language and communication in autism. In F. Volkmer, R. Paul, A. Klin & D. Cohen (Eds), Handbook of a utism and p ervasive d evelopmental d isorders (pp. 355 364). Hoboken, NJ: John Wiley and Sons. Terstegen, C., Koot, H. M., De Boer, J. B., & Tibboel, D. (2003). Measuring pain in children with cognitive impairment: Pain response to surgical procedures. Pain, 103, 187 198. doi:10.10196/s0304 3959(02)00453 0 Tordjman, S., Anderson, G.M., Botbol, M., Tabard, S.B., Perez Diaz, F., Graignic, R., Carlier, endorphin in children and adolescents with autistic disorde r. PLoS ONE, 4 (8), e5289. doi:10.1371/journal.pone.0005289 Trochim, M.K. (2006). Research methods knowledge base Retrieved from http://www.socialresearchmethods.net/kb/index.php Twycross, A (2009). An inter professional approach to the ethics of undertaking research with children. Nurse Researcher, 16 (3), 7 20. Retrieved from: http://www.nurseresearcher.co.uk Uparela Acosta, M. & Tuna, L. (20 11). What has to be considered by observation Retrieved from http://www.tim.ethz.com/education/courses/courses_fs_2011/S17_observations_Uparela _Tuna U.S. Department of Health and Human Services, Office of Human Subjects Research. (2005). Regulation and ethical guidelines Part 46 Protection of Human Subjects Retrieved from: http://oshr.od.nih.gov
97 Voepel Lewis, T., Malviya, S., Tait, A. R., Merkel, S., Foster, R., & Krane, E. J. (2008). A comparison of the clinical utility of pain assessment tools for c hildren with cognitive impairment. Pediatric Anesthesiology, 106, 72. doi:10.1213/01.ane.0000287680.21212.d0 Voepel Lewis, T., Merkel, S., Tait, A.R., Trzcinka, A., & Malviya, S. (2002). The reliability and validity of the Face, Legs, Activity, Cry, Co nsolability observational tool as a measure of pain in children with cognitive impairment. Anesthesia and Analgesia, 95. doi:10.1213/01.ANE.0000032868.29441.37 von Baeyer, C.L., & Spagrud, L.J. (2007). Systematic review of observational (behavioral) measur es of pain for children and adolescents aged 3 to 18 years. Pain, 127, 140 150. doi:10.1016/j.pain.2006.08.014 Volden, J. & Phillips, L. (2010). Measuring pragmatic language in speakers with autism 2 and the Test of Pragmatic Language. American Journal of Speech Language Pathology, 19, 204 212. doi: 10.1044/1058 0360(201/09 0011) Wells, N., Pasero, C., & McCaffery, M. (2008). Improving the quality of care through pain assessment and management. In R.G. Hughes (Ed s .) Patient safety and quality: An evidenced based handbook for nurses. Rockville, MD: Agency for Healthcare Research a nd Quality.
98 BIOGRAPHICAL SKETCH Amanda B. Brown was born and raised in Jacksonville, Florida. She graduated from University Christian High School in 1999 and attended the University of North Florida, where she graduated summa cum laude in 2003 with a Bachelor of Science in Nursing. During her time there, she was a cheerleader and served as President for the Student Nurses Association chapter. In 2005, she re turned to school to pursue a Master of Science in Nursing at the University of Florida where she was part of the inaugural class of students to enter the Clinical Nurse Leader program. She graduated in 2007 and became a certified Clinical Nurse Leader th at same year. Amanda began her journey towards a Doctor of Philosophy in 2009 and wanted to focus on improving the acute care experience for children with autism and their families. She received an American Nurses Foundation grant to fund her research on the assessment of pain in children with autism. She received her degree in 2014. beginning her career as a nursing assistant and mov ing into a registered nurse position after graduation in 2003. In January of 2006, she became one of the first C linical N urse L eaders at the hospital. She continues to serve in this role in the neonatal intensive care unit Amanda has presented nationally on the role of the Clinical Nurse Leader and pediatric acute nursing care and has published an article on the role of the C linical Nurse Leader. She intends to continue her career
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ESV0VA7U8_HDRH9C INGEST_TIME 2014-10-03T21:26:19Z PACKAGE UFE0046442_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
TheHigh-VoiceSuperiorityEffectinPolyphonicMusicIsInfluencedby Experience:AComparisonofMusiciansWhoPlaySoprano-Range ComparedWithBass-RangeInstrumentsClineMarieMcMasterUniversityTakakoFujiokaMcMasterUniversityandRotmanResearchInstituteLelandHerringtonMcMasterUniversityLaurelJ.TrainorMcMasterUniversityandRotmanResearchInstituteWesternpolyphonicmusicistypicallycomposedofmultiplesimultaneousmelodiclinesofequalimportance, referredtoasvoices.Previousstudieshaveshownthatadultnonmusiciansareabletoencodeeachvoice inseparateparallelsensorymemorytracesduringpassivelistening.Specifically,whenpresentedwith sequencescomposedoftwosimultaneousvoices(melodies),listenersshowmismatchnegativity(MMN) responsestopitchchangesineachvoice,althoughonly50%oftrialsareunchanged.Interestingly,MMNis largerforthechangeinthehighercomparedtolowervoiceinbothmusiciansandnonmusicians.This high-voicesuperiorityeffecthasalsobeenfoundinnonmusicianadultsand7-month-oldinfantspresented withtwosimultaneoustones,suggestingthatamorerobustmemorytraceforthehigher-pitchedvoicemight beaninnateorearly-acquiredcharacteristicofhumanauditoryprocessing.Thepresentstudytestedwhether musicianswithexperienceplayingabass-rangeinstrument(e.g.,cello,doublebass)wouldshowasimilar high-voicesuperiorityeffectasmusicianswithexperienceplayingasoprano-rangeinstrument(e.g.,violin, flute).Wefoundthatmusiciansplayingsoprano-rangeinstrumentsshowedahigh-voicesuperiorityeffectin linewithpreviousstudies,butmusiciansplayingbass-rangeinstrumentsshowedsimilarMMNresponsesfor bothvoices.Theseresultssuggestthatwithyearsofexperienceplayingalower-voicedinstrument,cortical encodingoftheloweroftwosimultaneousvoicescanbeenhancedtosomeextentdespitetheearlydeveloping biasforbetterencodingofthehighervoice. Keywords: musicalexpertise,brainplasticity,auditorysceneanalysis(ASA),mismatchnegativity (MMN),polyphonicmusic Supplementalmaterials: http://dx.doi.org/10.1037/a0030858.suppAuditoryenvironmentstypicallycontainmultipleoverlapping sounds.Forexample,itisnotunusualinasocialsituationto experiencesimultaneouslyseveralconversations,music,animal noises,andvariousenvironmentalsounds.Separatingthesedifferentsoundsourcesisreferredtoasauditorysceneanalysis(Bregman,1990).Musicitselfoftencontainsmultiplesimultaneous tonesormelodies.Forexample,Westernpolyphonicmusicis composedoftwoormoresimultaneousmelodiclines(oftenreferredtoasvoices),whichcancarryequalimportanceinthe music.Toprocesssuchmusic,itiscrucialforindividualstobe abletoseparateandsimultaneouslyanalyzetheindividualmelodiesastheyunfoldovertime.Previousresearchhasshownthatfor bothadults(Fujioka,Trainor,Ross,Kakigi,&Pantev,2005; Fujioka,Trainor,&Ross,2008)andinfants(Marie&Trainor, 2012),separatememorytracesareformedinauditorycortexfor eachoftwosimultaneousvoices.Furthermore,thesestudieshave shownahigh-voicesuperiorityeffectinthatthememorytracefor thehigher-pitchedvoiceismorerobustthanthatforthelowerpitchedvoice.Thepurposeofthepresentexperimentistoinvestigatewhetherexperienceplayingahigher-pitched(sopranorange)comparedwithlower-pitched(bass-range)musical ClineMarie,DepartmentofPsychology,NeuroscienceandBehaviour,and McMasterInstituteforMusicandtheMind,McMasterUniversity,Hamilton, Ontario,Canada;TakakoFujioka,McMasterInstituteforMusicandtheMind, McMasterUniversity,andRotmanResearchInstitute,Baycrest,Toronto,Ontario, Canada;LelandHerrington,DepartmentofPsychology,NeuroscienceandBehaviour,McMasterUniversity;LaurelJ.Trainor,DepartmentofPsychology,NeuroscienceandBehaviour,andMcMasterInstituteforMusicandtheMind,McMaster University,andRotmanResearchInstitute. ThisresearchwassupportedbyagrantfromtheCanadianInstitutesof HealthResearch(toLaurelJ.Trainor).ClineMariewassupportedbya postdoctoralfellowshipfromtheNaturalSciencesandEngineeringResearch CouncilCREATEGrantinAuditoryCognitiveNeuroscience.Wethank ElaineWhiskinforherhelpwithtestingandAndreaUnrauforproof-reading. Photosandbriefbiographiesofeachauthorareavailableathttp:// dx.doi.org/10.1037/a0030858.supp CorrespondenceconcerningthisarticleshouldbeaddressedtoLaurelJ. Trainor,DepartmentofPsychology,NeuroscienceandBehaviour,McMasterUniversity,1280MainStreetWest,Hamilton,OntarioL8S4K1, Canada.E-mail:LJT@mcmaster.caPsychomusicology:Music,Mind,andBrain 2012AmericanPsychologicalAssociation 2012,Vol.22,No.2,97 0275-3987/12/$12.00DOI:10.1037/a003085897
instrumentaffectstheencodingofpolyphonicmusic,inparticular, whetherthehigh-voicesuperiorityeffectismodifiedbyintensive experienceplayingthelowestlineinamusicalensemble. Forbothmusicalandnonmusicalsounds,todeterminewhat auditoryobjectsarepresent,theauditorysystemmustperforma spectrotemporalanalysisoftheincomingsoundwavetodetermine whichcomponentsbelongtogether(e.g.,theharmonicsofasingle soundsourcesuchasamusicalinstrumentoratalker,orthe successivesoundsofaninstrumentortalker)andwhichgroupsof componentsbelongtoseparateobjects(e.g.,twodifferentinstrumentsortwodifferenttalkers).Theseprocessesareknownas auditorystreamintegrationandsegregation,respectively,andtogethertheyconstituteauditorysceneanalysis.Bregman(1990) proposedthatmuchofauditorysceneanalysisoccursautomaticallyandpreattentively,andthisiscorroboratedbynumerous event-relatedpotential(ERP)studiesbasedonelectroencephalograph(EEG)ormagnetoencephalograph(MEG)recordings(e.g., Brattico,Winkler,Ntnen,Paavilainen,&Tervaniemi,2002; Lee,Skoe,Kraus,2009;Nager,Teder-Slejrvi,Kunze,&Mnte, 2003;Ritter,Deacon,Gomes,Javitt,&Vaughan,1995;Shinozaki etal.,2000;Sussman,Ritter,&Vaughan,1999;Winkler, Paavilainen,Ntnen,1992;Yabeetal.,2001).Manyofthese studiesmeasurethemismatchnegativity(MMN)componentofthe ERP.MMNisseenbetweenapproximately150and250msafter theonsetofadeviantsoundinastreamofstandardsounds.The deviancecaninvolveachangeinanysoundfeaturesuchaspitch, duration,loudness,ortimbre,aswellasachangeinsoundcategory(e.g.,onephonemetoanother)orachangeinthepatternof thepresentedsounds,suchasareversalinorder(Ntnen, Gaillard,&Mntysalo,1978;forreviews,seeNtnen, Paavilainen,Rinne,&Alho,2007;Ntnen&Winkler,1999; Picton,Alain,Otten,Ritter,&Achim,2000).MMNisobserved evenwhenlistenersarepayingnoattentiontothesounds.Thus, MMNisthoughttoreflectanautomaticprocessofupdatingof sensorymemorytraceswhenthebrainfailstopredictthenext soundeventasinthecaseofdeviantsounds.MMNisonly producedwhendeviantsarerelativelyrare,andthesizeofthe MMNincreasesastheratioofdeviantstostandardsdecreases. FujiokaandcolleaguesinvestigatedpreattentivesoundprocessingforsimultaneousmusicalstreamsorvoicesusingMMNprotocol.MeasuringMEG,Fujiokaetal.(2005)presentedadultswith trialsoftwosimultaneousfive-notemelodies,andintroduced deviantsonthefifthnoteon50%oftrials,suchthat25%contained awrongnoteinthehigher,and25%awrongnoteinthelower, melody.MMNwaselicitedbydeviantsinbothvoices,although theoveralldevianceratewas50%,indicatingthatexpectationsand thereforeseparatememorytraceswereformedforthehigherand lowermelodies.Atthesametime,MMNwaslargerfordeviantsin thehighthanthelowvoice,reflectingabiasforbetterencodingof thehighvoice.Furthermore,Fujiokaetal.(2008)foundthesame patternofresultswithsimplifiedstimuli,inwhicheachstandard trial(50%oftrials)consistedofadyadoftwosimultaneoustones ratherthanmelodies,and25%oftrialscontainedapitchchange (deviant)inthehigher-pitchedtoneand25%containedapitch change(deviant)inthelower-pitchedtone.Again,MMNwas elicitedbydeviantsinbothstreams,evenwithanoveralldeviance rateof50%,suggestingthatseparatememorytraceswereformed foreachstreamoftones.Moreover,theresultsshowedahighvoicesuperiorityeffectaswell,withlargerMMNelicitedby deviantsforthehigher-pitchedthanlower-pitchedtone.Finally, similarresultshaverecentlybeenfoundin7-month-oldinfants (Marie&Trainor,2012).Theresultsofthislaststudyindicatethat simultaneoussoundprocessingand,moresurprisingly,thehighvoicesuperiorityeffectemergeearlyindevelopment,suggestinga possibleinnatebasis.However,littleisknownaboutwhetherthis biascanbemodifiedbyexperience-relatedneuroplasticchanges. Overthepast15years,therehasbeenagreatdealofresearch examininghowmusicaltrainingorexpertiseaffectsbrainfunctionsand,morespecifically,influencesauditoryprocessing(for reviews,seeBesson,Chobert,&Marie,2011a;Kujala& Ntnen,2010;Trainor&Corrigall,2010).Relatedtotheprocessingofsoundduringpassivelistening,thegeneralconclusion acrosstheliteratureisthatmusiciansseemtohavebetterprocessingabilitiesascomparedwithnonmusicians.Forinstance,musiciansareabletodiscriminatepitchchangesamongunfamiliarand familiartonepatternsfasterthannonmusicians,asindexedbya shorterMMNlatency(Brattico,Ntnen,&Tervaniemi,2001). TheyshowlargerMMNtochangesinmelodiccontourandintervalstructurethannonmusicians(Fujioka,Trainor,Ross,Kakigi,& Pantev,2004).Whenpitchdeviationsareintroducedinpolyphonic andsingle-voicecontexts,musiciansshowabetterencodingof thesevariations,reflectedbylargerMMN,thannonmusicians (Fujiokaetal.,2005).Outsidethedomainsofpitchprocessing, musicalexpertiseinfluencesdurationprocessing,aswellasthe abilitytodetectchangesintemporalstructureandinnumerical regularity(e.g.,Marie,Kujala,&Besson,2012;Rsseler,Altenmller,Nager,Kohlmetz,&Mnte,2001;vanZuijen,Sussman, Winkler,Ntnen,&Tervaniemi,2005;Vuustetal.,2005,Vuust, Ostergaard,Pallesen,Bailey,&Roepstorff,2009).Itiseventhe casethatcertainaspectsofspeechprocessing,suchasmeterand suprasegmentalandsegmentalpitchfeatures,arealsoenhancedby musicaltraining(forareview,seeBesson,Chobert,&Marie, 2011b).Theseresultssuggestthatthroughyearsofintensive trainingwithamusicalinstrument,musicalexpertiseisassociated withimprovedsoundprocessing. Mostrelatedtoourconcern,differenttypesofmusicalexperienceappeartoaffectthewaymusiciansprocessmusicalsounds. Forexample,Pantev,Roberts,Schulz,Engelien,Ross(2001) demonstratedthatbrainresponsestotrumpetsoundsarelargerin trumpeterscomparedwithviolinists,whereastheoppositepattern istrueforviolinsounds.Musicianswhodidnotusemusicalscores whenpracticingorplayinganinstrumentshowedgreaterbehavioraldiscriminationandanenhancedMMNtocontourchanges comparedwiththoseplayingwithascoreandwithnonmusicians (Tervaniemi,Rytkonen,Schrger,Ilmoniemi,&Ntnen,2001). Inaddition,Seppnen,Brattico,&Ntnen(2007)demonstrated thattherecanevenbedifferencesbetweenmusicians,depending ontheirpreferredpracticestrategiessuchasreadingamusical score,improvising,playingbyear,andrehearsingbylisteningto recordings.Thosewhopreferredauralpracticingwerefasterat discriminatingchangesinmelodicintervalandcontourcompared withthosewhopreferredotherpracticestrategies(seealsoVuust, Brattico,Seppnen,Ntnen,Tervaniemi,2012).Thus,different musicalexperiencescanresultindifferentcorticalreorganizations inthebrain.Hereweexaminedhowspecificmusicaltrainingcan affecttheprocessingofmelodiesinapolyphoniccontext. Theoriginofthehigh-voicesuperiorityeffectisnotyetwell understood.Thefactthatitispresentalreadyinyounginfants98MARIE,FUJIOKA,HERRINGTON,ANDTRAINOR
suggeststhatiteitherhasaninnateoriginorisreadilylearned throughexposuretoWesternmusic,inwhichthemostimportant melodylineistypicallyplacedinthehighestvoice.However,itis notcleartowhatextentthehigh-voicesuperiorityeffectismalleablebyexperiencebeyondtheinfancyperiod.Ifitismodifiable, weexpectedthatmusicianswithexperienceplayingcelloorbass linesinanensemblewouldovercomethisbiasandshowlarger and/orearlierMMNstopitchchangesinthelowerthaninthe highervoicecomparedtomusicianswithexperienceplayingviolin orflutelinesinanensemble,whowouldbeexpectedtoshowa typicalhigh-voicesuperiorityeffect.MaterialsandMethods ParticipantsTwenty-fiveadultmusiciansweretested.Threewereexcluded owingtoexcessiveartifactsintheEEGdata,leaving22musicians inthefinalsample(12maleand10femaleindividuals;mean age 29.8years).Afterprovidinginformedwrittenconsentto participate,musicianscompletedaquestionnaireforauditory screeningpurposesandtoassessmusicalandlinguisticbackground.Subjectswererequiredtobetrainedmusicianswhowere playingregularlyinamusicalgroup(e.g.,orchestra,ensemble, choral)withmorethan5yearsoftrainingontheirmaininstrument.Elevensubjects(meanage 30years, SD 12)played instrumentsinthehighervoiceregister(violin,flute,soprano singer;meantraining 21years, SD 8;seeTable1fordetails). Elevensubjects(meanage 28years, SD 11)playedinstrumentsinthelowervoiceregister(bass,cello,orbass-vocal-range singer;meantraining 15years, SD 10;seeTable1).StimuliThetwofive-notemelodies(A,B)fromFujiokaetal.(2005) wereused.Theywerecomposedusingthefirstfivediatonicscale notesoftheWesternmajorscale(e.g.,C,D,E,F,andG).Inthe keyofCmajor,MelodyAwasthesequenceC-D-F-E-Gand MelodyBwasG-F-D-C-E.MelodiesAandBwerecombinedin twoversions(seeFigure1),onewithMelodyAinthehighvoice (C5G5)andMelodyBinthelowvoice(C4G4;American notation,High-A/Low-B)andonewiththevoicesreversed(HighB/Low-A).Deviantversionswerecreatedbyeitherraisingthelast noteofMelodyAbyonetone(one-sixthoctave)orloweringthe lastnoteofMelodyBbyonetone.Thiscreatedfourdeviant versions:High-A,Low-A,High-B,andLow-B.Thesechangesdid notalterthepitchcontoursofthemelodies.Fromonetrialtothe next,thecombinedmelodiesweretransposedtooneofeightkeys sequencedintheorderC-E-C#-F-D-F#-D#-Gtoavoidpotential primingeffectsthatcouldarisefromabsoluteandnotrelativepitch processing.Forbothmelody-voicecombinations(High-A/Low-B andHigh-B/Low-A),thesequencesofstimuliwerepresentedinan oddballparadigm.Sixty-fourpercentofthetrialswerepresented withthestandardlastnote(seeFigure1).Thirty-sixpercentofthe sequenceswerepresentedwitheitherthedeviantlastnoteinthe highvoice(18%)orthelowvoice(18%).AlthoughMelodiesA Table1 MusicalBackgroundofEachParticipant Participants FirstinstrumentPracticefirst(yr)Musictheory(yr)SecondinstrumentPracticesecond(yr) Bass-rangeplayers 1F Cello 36 7 VioladaSamba 8 2M Bass 6 15 Piano 17 3M Bass 16 5 Drums 20 4M Bassvocal 34 3 Saxophone 2 5M Bass 5 1 Guitar 9 6M Uprightbass 8 6 Electricbass 11 7F Cello 15 4 Piano 10 8M Cello 15 1 Piano 21 9M Bassvocal 10 4 none 10M Cello 8 5 Piano 6 11F Cello 15 10 Piano 16 Mean 15.3 5.6 12 SD 10.5 4.1 6.3 Soprano-rangeplayers 1F Sopranovocal 30 0 Cello 5 2M Flute 35 5 none 3F Flute 15 6 Piano 20 4M Violin 29 12 Piano 6 5F Flute 27 14 Piano 6 6M Violin 27 8 Piano 20 7F Violin 16 5 Piano 15 8F Violin 10 3 Piano 10 9F Flute 11 10 Piano 4 10F Violin 15 10 Piano 15 11M Violin 14 0 Viola 3 Mean 20.8 6.6 10.4 SD 8.8 4.6 6.6 Note. M male;F female.99MUSICALEXPERIENCEALTERSSIMULTANEOUSSOUNDENCODING
andBusedthesamegroupoffivenotesforstandards,theyhad differentnotesateachtimepointinthemelodies.Theharmonyat eachtimepointwasalwaysmusicallyconsonant,eveninthe deviantversions.Thesoundfileswerecreatedfromdigitally recordedpianotimbresforeachnoteatasamplingrateof44,100 Hz.Thedurationofeachnotewas300msforatotalmelodylength of1,500ms.Successivetrialswereseparatedbya900-mssilent interval.ProcedureAllparticipantsweretestedindividually.Theprocedurewas explainedtoeachparticipant,whogaveconsenttoparticipate.The procedureswereclearedbytheMcMasterResearchEthicsBoard. Eachparticipantsatinthesound-attenuatedroom(Industrial AcousticsCompany)facingaloudspeakerandascreenplaced1m infrontoftheirhead.Duringtheexperiment,theparticipant watchedasilentmoviewithsubtitlesandwasinstructedtopay attentiontothemovieandnottothesoundsthatwerecomingfrom theloudspeaker.Theywerealsoaskedtominimizetheirmovement,includingblinkingandfacialmovements,soastoobtainthe bestsignal-to-noiseratio.Fourblocksof384trialseachwere presentedusingE-primesoftwareinpseudorandomordersuchthat trialsofthesamedeviantneverfollowedsuccessively.Eachblock lastedapproximately12min.TwoblockscontainedMelodyAin thehighervoiceandMelodyBinthelowervoice,andtwoblocks containedthereversecombination.EEGRecordingandProcessingEEGdatawererecordedatasamplingrateof1,000Hzfrom 128-channelHydroCelGSNnets(ElectricalGeodesics,Eugene, OR)referencedtoCz.Theimpedanceofallelectrodeswas 50 k duringtherecording.EEGdatawereband-passfilteredbetween0.5and20Hz(roll-off 12dB/oct)usingEEprobesoftware.Recordingswerere-referencedoff-lineusinganaverage referenceandthensegmentedinto500-msepochs( 100to400 msrelativetotheonsetofthelastnoteofthemelody).EEG responsesexceeding 70 Vinanyepochwereconsideredartifactsandexcludedfromtheaveraging.ERPDataAnalysisForeachparticipantforeachcondition(High-A/Low-B,HighB/Low-A),responsestostandardsandeachofthetwodeviants wereaveragedseparately,anddifferencewaveformswerecomputedforeachconditionandparticipantbysubtractingERPs elicitedbythestandardsfromthoseelicitedbyeachofthedeviants.Thus,fourdifferentdifferencewaveswerecreated,onefor eachofthedeviantsHigh-A,Low-A,High-BandLow-Binboth thebass-rangeandsoprano-rangemusicians.Subsequently,for statisticalanalysis,70electrodeswereselectedanddividedinto fourgroupsforeachhemisphere(leftandright)representing frontal,central,occipital,andtemporalregions(FL,FR,CL,CR, OR,OL,TL,TR;seeFigure2).Groupingtheelectrodesinthis wayenhancessignal-to-noiseratiosandenablesexaminationof theaverageresponseacrossclassicscalpregions.Fifty-eightelectrodeswereexcludedfromthegroupingsowingtothefollowing considerations:electrodesontheforeheadneartheeyestofurther reducethecontaminationofeyemovementartifacts,electrodesat theedgeofthegeodesicnettoreducecontaminationoffaceand neckmusclemovement,electrodesinthemidlinetoenablecomFigure1. Descriptionofthestimulussequencesillustratedinmusical notation.Twodifferentmelodies(AandB)areplayedsimultaneouslyin highandlowvoicecorrespondingtotwolinesinmusicalnotation.Both melodiesconsistoffivenoteswithastandardoradeviantterminalnote. IntheHigh-A/Low-Bcase,thehighvoiceisMelodyAandthelowvoice isMelodyB(top),whereasintheHigh-B/Low-Acase,themelodyvoice combinationwasreversed(bottom). Figure2. Thegroupingofelectrodesinthegeodesicnet.Of128electrodes,70wereselectedtobedividedintofourgroups(frontal,central, occipital,andtemporal)foreachhemisphere.Thewaveformsforall channelsineachregionwereaveragedtogethertorepresentEEGresponses fromthatscalpregion.Opencirclewithblacksolidline,frontalleft;open circlewithblackdottedline,frontalright;darkgrayfilledcirclewithblack solidline,centralleft;darkgrayfilledcirclewithblackdottedline,central right;lightgrayfilledcirclewithblacksolidline,occipitalleft;lightgray filledcirclewithblackdottedline,occipitalright;opencirclewithgray solidline,temporalleft;andopencirclewithgraydottedline,temporal right.100MARIE,FUJIOKA,HERRINGTON,ANDTRAINOR
parisonoftheEEGresponseacrosshemispheres;andparietal electrodeswhereMMNamplitudeiscloseto0 V. ToanalyzeMMNamplitude,firstthemostnegativepeakinthe rightfrontalregion(FR)between150and250msafterstimulus onsetwasdeterminedfromthegrandaveragedifferencewavesfor eachofthefourconditions(High-A,Low-A,High-B,Low-B),and a50-mstimewindowwasconstructedcenteredatthislatency.For eachsubjectandeachregion,theaverageamplitudeinthis50-ms timewindowforeachconditionwasusedasthemeasureofMMN amplitude.Finally,thelatencyoftheMMNwasmeasuredasthe timeofthemostnegativepeakbetween150and250msattheFR regionforeachsubject,foreachcondition,asvisualinspection showedthelargestMMNamplitudeatthisregion.Analysesof variance(ANOVAs)wereconductedonamplitudeandlatency data.GreenhouseGeissercorrectionswereappliedwhereappropriate,andTukeyposthoctestswereconductedtodeterminethe sourceofsignificantinteractions.Results MMNAmplitudeAmplitude. Afive-wayANOVAwasconductedwithGroup (soprano,bass)asabetween-subjectsfactor;Voice(high,low), Melody(A,B),Hemisphere(left,right),andRegion(frontal:FL andFR;central:CLandCR;temporal:TLandTR;occipital:OL andOR)aswithin-subjectfactors;andMMNamplitudeasthe dependentmeasure.ThemaineffectsofGroup( F(1,20) 0.33, p .57)andVoice(F(1,20) 1.94, p .17)werenotsignificant. However,therewasasignificantinteractionbetweenGroupand Voice(F(1,20) 4.79, p .04).Posthoctestsshowedaslight trendforalargerMMNdifferencebetweensoprano-rangeand bass-rangeplayersontheLowvoice(p .23)butnotrendfora groupdifferencefortheHighvoice( p .82,seeTable2for details).AscanbeseeninTable2,differencesappeartobelargest atFRandCRregions.Moreover,aninteractionbetweenGroupby VoicebyMelodybyRegionwassignificant( F(3,60) 4.52, p .03).Nootherinteractionsweresignificant. Tobetterunderstandtheseinteractions,four-wayANOVAs wereconductedseparatelyforeachgroup,withVoice,Melody, Hemisphere,andRegionaswithin-subjectfactors.Forsopranorangeplayers,themaineffectofvoicewassignificant( F(1,10) 6.08, p .03,partialetasquared[p 2] .38),revealinglarger MMNresponsesforthehigherthanforthelowervoice(see Table2andFigure3).Noothermaineffectsorinteractions weresignificant.Forthebass-rangeplayers,themaineffectof voicewasnotsignificant( F (1,10) 0.33, p .57, p 2 .03; seeTable2andFigure3),revealingthatMMNresponseswere notlargerforthehigherthanforthelowervoice.Noothermain effectsorinteractionsweresignificant(seeFigure4).MMNLatencyAthree-wayANOVAwithGroupasabetween-subjectsfactor, VoiceandMelodyaswithin-subjectfactors,andlatencyatregion FRasthedependentmeasurerevealedamaineffectofmelody (F(1,20) 19.46, p .001],withsignificantlyshorterMMN latencytodeviantsinMelodyA(148ms)thantodeviantsin MelodyB(163ms,seeFigure5).Noothermaineffectsor interactionsweresignificant.DiscussionInWesternpolyphonicmusic,themelodyismostlycommonly placedinthehighestvoice,butitisnotknowntowhatextentthis Table2 DetailedMeanMMNAmplitude(Valuein V)forEachGroup, forEachVoiceandEachRegionbyHemisphere Soprano-range players Bass-range players Regions High voice Low voice High voice Low voice Frontalleft 0.81 0.37 0.34 0.39 Frontalright 0.33 0.18 0.55 0.44 Centralleft 0.58 0.30 0.27 0.33 Centralright 0.18 0.14 0.54 0.35 Occipitalleft0.230.130.410.28 Occipitalright0.422.214.171.124 Temporalleft0.110.100.470.45 Temporal right0.560.360.170.18 Note. AninversionofMMNpolarityintheoccipitalandtemporal regions. Figure3. Differencewaveforms(deviantstandard)in(A),thegroupof soprano-rangeplayers(n 11)and(B),andthegroupofbass-range players(n 11)fortheHighvoice(blackline)andtheLowvoice(dashed line).DeviantsinMelodyAandBwerecombinedforeachvoice.101MUSICALEXPERIENCEALTERSSIMULTANEOUSSOUNDENCODING
isaculturalconventionoraresultofinnatebiologicalconstraints onperception.PreviousresearchusingEEGandMEGindicates thatahigh-voicesuperiorityeffectisfoundbothinadults(Fujioka etal.,2005,2008)andininfants(Marie&Trainor,2012).The infantfindingssuggestthatinnateconstraintsmightplayarole. Hereweinvestigatedwhethermusicalexperiencemightalsoplay arole.Specifically,wetestedwhethertheexperienceofplayinga soprano-rangecomparedtoabass-rangeinstrumentmodifiesthe high-voicesuperiorityeffect.Theresultsindicatedthatthiseffect isindeedmalleablebyexperiencebeyondtheinfancyperiod.As expected,soprano-rangeplayersshowedahigh-voicesuperiority effectasreflectedbylargerMMNamplitudefordeviantsinthe highvoicethaninthelowvoice.Ofmostinterest,bass-range playersdidnotshowahigh-orlow-voicesuperiorityeffect.Post hocanalysesindicatedthattherewasaslighttrendforbass-range musicianstoshowhigher-amplitudeMMNforchangesinthelow voicethansoprano-rangemusicians,butnotrendforadifference betweengroupsforthehighvoice. Acrossbothgroupsofmusicians,wefoundthatMMNlatency wasshorterinresponsetodeviantsinMelodyAthaninresponse todeviantsinMelodyB,regardlessofthevoiceassignment.This effectisnotrelatedtothehigh-voicesuperiorityeffect,asMelody Awasinthehighvoiceonhalfoftheconditions,whereasMelody Bwasinthehighvoiceontheotherhalf.Themostlikely explanationisthatinthecaseofMelodyA,thedeviantnotedid notoccurinthepreviousmelodiccontext,whereasinthecaseof MelodyB,thedeviantnotewasidenticaltothethirdnoteofthe melody(seeFigure1).Thus,thedeviantnoteinMelodyBmay havebeenlessnovelthanthedeviantnoteinMelodyA,causing aslightdelayinprocessingthechange.Anotherinterpretationof thedifferentMMNlatenciesforthetwomelodiesisthatthey resultedfromthefactthatthedeviantlastnoteinMelodyAwas higherinpitchthanthelastnoteofthestandardmelody,causing anincreaseinthesizeofthefinalinterval(minorthirdtoperfect fourth),whereasinMelodyB,thedeviantlastnotewaslowerin pitch,causingadecreaseinthesizeofthefinalinterval(major thirdtomajorfourth).However,notethatthesizeofthedeviance wasidenticalinbothcases(amajorsecond),soifthisexplanation iscorrect,itimpliesthatincreasesinpitchareeasiertodetectthan decreases.Athirdpossibilityisrelatedtotonalstructure.Melody AmightbeconsideredtobemorestablethanMelodyB,asit beginsonthetonic(firstandmoststablenoteofthescale)and endsonthedominant(fifthandsecondmoststablenote).However,previousworkindicatesthatMMNislittleaffectedbytonal structure;rather,itisprimarilyaffectedbythephysicalsizeofthe changeratherthanthemeaningintermsofmusicalscaleandkey (e.g.,Ntnen,Pakarinen,Rinne,&Takegata,2004;Trainor, McDonald,&Alain,2002;Trainor&Zatorre,2009),makingthis explanationunlikely.Inanycase,itwouldbeinterestingforfuture studiestoexploretheperceptualanddevelopmentaloriginsofthis effectaswell. Thepresenceofthehigh-voicesuperiorityeffectinyoung infantssuggeststhatiteitherhasaninnateoriginorisreadily learnedthroughexposuretoWesternpolyphonicmusic,inwhich themostimportantmelodylineistypicallyplacedinthehighest voice.Thepresentresultsindicatethat,atthesametime,the high-voicesuperiorityeffectismalleabletosomeextentbyextensiveexperienceplayingthelowestvoiceinmusicalgroups.Interestingly,despitethemanyyearsofpracticewithalow-voice instrument,wedidnotobserveacompletereversalofthisbiasin musicianswhoplaybass-rangeinstruments,inwhichcase,we wouldhaveseenalow-voicesuperiorityeffect.Rather,inthese musicians,thehighandlowvoicesappearedtobeencodedequally well,suggestinganinterplaybetweenaninnatetendencyfora high-voicesuperiorityeffectandgreaterexperiencewithlowerFigure4. High-voicesuperiorityeffectinbothgroups.DifferencebetweentheHigh-voicedifferencewavesminustheLow-voicedifference waves. Figure5. AveragedifferencewaveformsforMelodyA(blackline)andforMelodyB(dashedline)withboth groupscombined.ShorterMMNlatencyforMelodyAthanforMelodyB.102MARIE,FUJIOKA,HERRINGTON,ANDTRAINOR
thanhigher-voiceparts.However,itispossiblethatwehave underestimatedtheeffectsofexperience.AscanbeseeninTable 1,itisdifficulttorecruitmusicianswhoplayonlyoneinstrument, andmanymusiciansplaythepianoasasecondinstrument.Becausepianistsneedtoplayallmelodylines,andhavemuch experienceplayingthemostimportantmelodyinthehighestvoice, asagroup,thebass-rangeinstrumentplayerslikelyhadsignificant experienceplayinghigh-voicemelodyparts.Itisalsopossiblethat bassinstrumentplayerswithmoreextensivetrainingthanthosein oursamplemighthaveshownalow-voicesuperiorityeffect. Finally,melodiccontextclearlyaffectshowmemorytracesare formedbecausethehigh-voicesuperiorityeffectappearstobe smallerwithourpolyphonicmelodycontextthaninthecontextof arepeatingdyadofcomplextones(e.g.,Fujiokaetal.,2008;Marie &Trainor,2012).Thepresentresultsthussuggestthatmusical experienceinplayingabass-rangeinstrumentcanspecifically modifythedegreeofhigh-voicedominanceeventhoughitis alreadypresentinyounginfants.ConclusionThisstudyprovidesevidencethatexperienceintheformof extensivemusicaltrainingcaninfluencethehigh-voicesuperiority effectinpolyphonicmusic.Yearsofpracticewithabass-range instrumentappearstoenhancetheencodingofthelowervoiceand tomodifythebiasforbetterencodingofthehighermelody. Whereasmusiciansplayingasoprano-rangeinstrumentdemonstratedatypicalhigh-voicesuperiorityeffect,thoseplayinga bass-rangeinstrumentshowedequalencodingofthehighandlow voices.Thepresenceofthehigh-voicesuperiorityeffectinyoung infantsandthefactthatextensiveexperienceplayingabass-range instrumentdidnotresultinalow-voicesuperiorityeffectsuggest aninnatebiastowardsuperiorencodingofthehighervoice.Atthe sametime,modificationofthehigh-voicesuperiorityeffectin bass-rangemusiciansindicatesthatthehigh-voicesuperiorityeffectismodifiabletosomeextentbyexperience.ReferencesBesson,M.,Chobert,J.,&Marie,C.(2011a).Languageandmusicinthe musicianbrain. LanguageandLinguisticsCompass,5, 617634.doi: 10.1111/j.1749-818X.2011.00302.x Besson,M.,Chobert,J.,&Marie,C.(2011b).Transferoftrainingbetween musicandspeech:Commonprocessing,attentionandmemory. FrontiersinPsychology,2,94.doi:10.3389/fpsyg.2011.00094 Brattico,E.,Ntnen,R.,&Tervaniemi,M.(2001).Contexteffectson pitchperceptioninmusiciansandnonmusicians:Evidencefromeventrelatedpotentialrecordings. MusicPerception:AnInterdisciplinary Journal,19, 199.doi:10.1525/mp.2001.19.2.199 Brattico,E.,Winkler,I.,Ntnen,R.,Paavilainen,P.,&Tervaniemi,M. (2002).Simultaneousstorageoftwocomplextemporalsoundpatternsin auditorysensorymemory. Neuroreport,13, 1747. Bregman,A.S.(1990). Auditorysceneanalysis;Theperceptualorganizationofsound.Cambridge,MA:MITPress.doi:10.1121/1.408434 Fujioka,T.,Trainor,L.J.,&Ross,B.(2008).Simultaneouspitchesare encodedseparatelyinauditorycortex:AnMMNmstudy. NeuroReport, 19, 361.doi:10.1097/WNR.0b013e3282f51d91 Fujioka,T.,Trainor,L.J.,Ross,B.,Kakigi,R.,&Pantev,C.(2004). Musicaltrainingenhancesautomaticencodingofmelodiccontourand intervalstructure. JournalofCognitiveNeuroscience,16, 1010. doi:10.1162/0898929041502706 Fujioka,T.,Trainor,L.J.,Ross,B.,Kakigi,R.,&Pantev,C.(2005). Automaticencodingofpolyphonicmelodiesinmusiciansandnonmusicians. JournalofCognitiveNeuroscience,17, 1578.doi: 10.1162/089892905774597263 Kujala,T.,&Ntnen,R.(2010).Theadaptivebrain:Aneurophysiologicalperspective. ProgressinNeurobiology,91, 5567.doi:10.1016/j .pneurobio.2010.01.006 Lee,K.M.,Skoe,E.,Kraus,N.,&Ashley,R.(2009).Selectivesubcortical enhancementofmusicalintervalsinmusicians. TheJournalofNeuroscience,29, 5832.doi:10.1523/JNEUROSCI.6133-08.2009 Marie,C.,Kujala,T.,&Besson,M.(2012).Musicalandlinguisticexpertiseinfluencepre-attentiveandattentiveprocessingofnon-speech sounds. Cortex,48, 447457.doi:10.1016/j.cortex.2010.11.006 Marie,C.,&Trainor,L.J.(2012).Developmentofsimultaneouspitch encoding:Infantsshowahighvoicesuperiorityeffect. CerebralCortex. Advanceonlinepublication.doi:10.1093/cercor/bhs050 Ntnen,R.,Gaillard,A.W.K.,&Mntysalo,S.(1978).Earlyselective attentioneffectonevokedpotentialreinterpretes. ActaPsychological, 42, 313.doi:10.1016/0001-6918(78)90006-9 Ntnen,R.,Paavilainen,P.,Rinne,T.,&Alho,K.(2007).Themismatch negativity(MMN)inbasicresearchofcentralauditoryprocessing:A review. ClinicalNeurophysiology,118, 2544.doi:10.1016/j .clinph.2007.04.026 Ntnen,R.,Pakarinen,S.,Rinne,T.,&Takegata,R.(2004).Themismatchnegativity(MMN):Towardstheoptimalparadigm. ClinicalNeurophysiology,115, 140.doi:10.1016/j.clinph.2003.04.001 Ntnen,R.,&Winkler,I.(1999).Theconceptofauditorystimulus representationincognitiveneuroscience. PsychologicalBulletin,125, 826859.doi:10.1037/0033-2909.125.6.826 Nager,W.,Teder-Slejrvi,W.,Kunze,S.,&Mnte,T.F.(2003).Preattentiveevaluationofmultipleperceptualstreamsinhumanaudition. NeuroReport:ForRapidCommunicationofNeuroscienceResearch,14, 871874.doi:10.1097/00001756-200305060-00019 Pantev,C.,Roberts,L.E.,Schulz,M.,Engelien,A.,&Ross,B.(2001). Timbre-specificenhancementofauditorycorticalrepresentationsin musicians. Neuroreport,12, 169. Picton,T.W.,Alain,C.,Otten,L.,Ritter,W.,&Achim,A.(2000). Mismatchnegativity:Differentwaterinthesameriver. Audiologyand Neuro-Otology,5, 111. Ritter,W.,Deacon,D.,Gomes,H.,Javitt,D.C.,&Vaughan,H.G.Jr. (1995).Themismatchnegativityofevent-relatedpotentialsasaprobeof transientauditorymemory:Areview. EarandHearing,16, 5267. doi:10.1097/00003446-199502000-00005 Rsseler,J.,Altenmller,E.,Nager,W.,Kohlmetz,C.,&Mnte,T.F. (2001).Event-relatedbrainpotentialstosoundomissionsdifferinmusiciansandnon-musicians. NeuroscienceLetters,308, 33.doi: 10.1016/S0304-3940(01)01977-2 Seppnen,M.,Brattico,E.,&Ntnen,R.(2007).Practicestrategiesof musiciansmodulateneuralprocessingandthelearningofsoundpatterns. NeurobiologyofLearningandMemory,87, 236.doi: 10.1016/j.nlm.2006.08.011 Shinozaki,N.,Yabe,H.,Sato,Y.,Sutoh,T.,Hiruma,T.,Nashida,T.,& Kaneko,S.(2000).Mismatchnegativity(MMN)revealssoundgrouping inthehumanbrain. Neuroreport,11, 1597.doi:10.1097/ 00001756-200006050-00001 Sussman,E.,Ritter,W.,&Vaughan,H.G.Jr.(1999).Aninvestigationof theauditorystreamingeffectusingevent-relatedbrainpotentials. Psychophysiology,36, 22.doi:10.1017/S0048577299971056 Tervaniemi,M.,Rytkonen,M.,Schrger,E.,Ilmoniemi,R.J.,&Ntnen, R.(2001).Superiorformationofcorticalmemorytracesformelodic patternsinmusicians. LearningandMemory,8, 295.doi:10.1101/ lm.39501 Trainor,L.J.,&Corrigall,K.A.(2010).Musicacquisitionandeffectsof103MUSICALEXPERIENCEALTERSSIMULTANEOUSSOUNDENCODING
musicalexperience.InM.Riess-Jones&R.R.Fay(Eds.), Springer handbookofauditoryresearch:Musicperception (pp.89).Heidelberg:Springer.doi:10.1007/978-1-4419-6114-3_4 Trainor,L.J.,McDonald,K.L.,&Alain,C.(2002).Automaticand controlledprocessingofmelodiccontourandintervalinformationmeasuredbyelectricalbrainactivity. JournalofCognitiveNeuroscience,14, 430442.doi:10.1162/089892902317361949 Trainor,L.J.,&Zatorre,R.(2009).Theneurobiologicalbasisofmusical expectations:Fromprobabilitiestoemotionalmeaning.InS.Hallen,I. Cross,&M.Thaut(Eds.), Oxfordhandbookofmusicpsychology (pp. 171).Oxford:OxfordUniversityPress. vanZuijen,T.L.,Sussman,E.,Winkler,I.,Ntnen,R.,&Tervaniemi, M.(2005).Auditoryorganizationofsoundsequencesbyatemporalor numericalregularityamismatchnegativitystudycomparingmusicians andnon-musicians. CognitiveBrainResearch,23, 270.doi: 10.1016/j.cogbrainres.2004.10.007 Vuust,P.,Brattico,E.,Seppnen,M.,Ntnen,R.,Tervaniemi,M. (2012).Practicedmusicalstyleshapesauditoryskills. AnnalsoftheNew YorkAcademyofScience,1252, 139.doi:10.1111/j.1749-6632 .2011.06409.x Vuust,P.,Ostergaard,L.,Pallesen,K.J.,Bailey,C.,&Roepstorff,A. (2009).Predictivecodingofmusic. Cortex,45, 80.doi:10.1016/j .cortex.2008.05.014 Vuust,P.,Pallesen,K.J.,Bailey,C.,VanZuijen,T.L.,Gjedde,A., Roepstor,V.A.,stergaard,L.(2005).Tomusicians,themessageisin themeterPreattentiveneuronalresponsestoincongruentrhythmare left-lateralizedinmusicians. NeuroImage,24, 560.doi:10.1016/j .neuroimage.2004.08.039 Winkler,I.,Paavilainen,P.,&Ntnen,R.(1992).Canechoicmemory storetwotracessimultaneously?Astudyofevent-relatedbrainpotentials. Psychophysiology,29, 337.doi:10.1111/j.1469-8986.1992 .tb01707.x Yabe,H.,Winkler,I.,Czigler,I.,Koyama,S.,Kakigi,R.,Suto,T.,Hiruma, T.,&Kaneko,S.(2001).Organizingsoundsequencesinthehuman brain:Theinterplayofauditorystreamingandtemporalintegration. BrainResearch,897, 222.doi:10.1016/S0006-8993(01)02224-7ReceivedApril30,2012 RevisionreceivedJuly25,2012 AcceptedOctober2,2012 CallforPapers:SocietyforMusicPerceptionandCognition(SMPC)2013ThebiennialmeetingoftheSocietyforMusicPerceptionandCognition(SMPC)willbeheldat RyersonUniversityinToronto,Canada,onAugust8-11,2013.Abstractsforpresentationsshould benolongerthan300wordsandshoulddescribethemotivation,methodology,results,and implicationstothedegreethatthisinformationisavailableatthetimeofsubmission.Empirical contributionsshouldrefertothestimuli/corpus,methodology,anddatacollected.Theoretical contributionsarealsowelcome,providedthattheconnectiontomusicperceptionandcognitionis underscoredthroughdiscussionofaims,methods,and/orresults.Abstractsforproposedsymposia arewelcomeandshouldincludeindividualabstractsaswellasabriefdescriptionofthetheme. Submissiondetailsandadditionalconferenceinformationareavailableontheconferencewebsite: http://www.ryerson.ca/smpc2013. ThesubmissiondeadlineisFebruary1,2013. ProgramCommittee: TonyaBergeson(IndianaUniversitySchoolofMedicine) JohnIversen(TheNeurosciencesInstitute) ScottLipscomb(UniversityofMinnesotaSchoolofMusic) LisaMargulis(UniversityofArkansasDepartmentofMusic) FrankRusso(RyersonUniversityDepartmentofPsychology)MichaelSchutz(McMasterInstituteforMusicandtheMind)LeighVanHandel(MichiganStateUniversityCollegeofMusic)ConferenceChairProgramChair Thefollowingsatellitemeetingsmayalsobeofinterest The3rdAnnualSeminaronCognitivelyBasedMusicInformatics(CogMIR),August7 E-mail:nvempalapsych.ryerson.ca The5thAnnualMeetingofAdvancingInterdisciplinaryResearchinSinging(AIRS),August11-12 E-mail:acohenupei.ca 104MARIE,FUJIOKA,HERRINGTON,ANDTRAINOR