University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln Special Education and Communication Disorders Department of Special Education and Faculty Publications Communication Disorders 2-1-2001 Relative Kinematics of the Rib Cage and Abdomen during Speech and Nonspeech Behaviors of 15-Month-Old Children Christopher A. Moore University of Washington, Seattle, [email protected] Tammy J. Caulfield University of Washington, Seattle Jordan R. Green University of Nebraska-Lincoln, [email protected] Follow this and additional works at:http://digitalcommons.unl.edu/specedfacpub Part of theSpecial Education and Teaching Commons Moore, Christopher A.; Caulfield, Tammy J.; and Green, Jordan R., "Relative Kinematics of the Rib Cage and Abdomen during Speech and Nonspeech Behaviors of 15-Month-Old Children" (2001).Special Education and Communication Disorders Faculty Publications. Paper 26. http://digitalcommons.unl.edu/specedfacpub/26 This Article is brought to you for free and open access by the Department of Special Education and Communication Disorders at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Special Education and Communication Disorders Faculty Publications by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. Published in Journal of Speech, Language & Hearing Research 44:1 (February 2001), pp. 80–94; doi: 10.1044/1092- 4388(2001/008) Copyright © 2001 American Speech-Language-Hearing Association. Used by permission. Submitted March 28, 2000; accepted October 26, 2000. Relative Kinematics of the Rib Cage and Abdomen during Speech and Nonspeech Behaviors of 15-Month-Old Children Christopher A. Moore, University of Washington, Seattle Tammy J. Caulfield, University of Pittsburgh Jordan R. Green, University of Wisconsin, Madison Corresponding author — C. A. Moore, Department of Speech and Hearing Sciences, University of Washington, 1417 NE 42nd Street, Seattle, WA 98105-6246. Email: [email protected] Abstract Speech motor control emerges in the neurophysiologic context of widely distributed, powerful coordinative mech- anisms, including those mediating respiratory function. It is unknown, however, whether developing children are able to exploit the capabilities of neural circuits controlling homeostasis for the production of speech and voice. Speech and rest breathing were investigated in eleven 15-month-old children using inductance plethysmography (Respitrace). Rib cage and abdominal kinematics were studied using a time-varying correlational index of thora- coabdominal coupling (i.e., reflecting the synchrony of movement of the rib cage and abdomen) as well as sim- ple classification of the moment-to-moment kinematic relationship of these two functional components (i.e., con- current expansion or compression, or oppositional movement). Results revealed markedly different patterns of movement for rest breathing and speech breathing, although within types of vocalization (nonspeech vocaliza- tion, babbling, true word production) no differences were apparent. Whereas rest breathing was characterized by tight coupling of rib cage and abdominal movement (average correlation coefficients usually exceeded .90), speech breathing exhibited weak coupling (the correlation coefficient ranged widely, but averaged about .60). Fur- thermore, speech production by these toddlers included the occurrence of both rib cage and abdominal paradox- ing, which are observed infrequently in adult speakers. These results fail to support the suggestion that speech emerges from the extant coordinative organization of rest breathing. Rather, even in its earliest stages breathing for speech and voice exhibits kinematic properties distinct from those of other observed behaviors. Keywords: speech development, respiration, motor control, kinematics, babbling The neurophysiologic development of speech has motor systems (e.g., locomotion; Grillner, 1982). The frequently been described as deriving its coordinative paucity of physiologic observations of speech develop- framework from existing motor control mechanisms ment renders arguments regarding the validity of these (e.g., Grillner, 1982). A wide range of candidate coordi- mechanisms moot. Sufficient empirical support has yet native mechanisms has been discussed with respect to to be built. The coordinative organization for speech speech production, some depending on well-known pat- may arise from extant motor control mechanisms, in- terns of oromotor organization (e.g., chewing) and oth- cluding those of respiration (Feldman & Smith, 1995; ers relying on related organizational structures in other Smith, Ellenberger, Ballanyi, Richter, & Feldman, 1991; 80 RespiRatoRy Kinematics duRing speech and nonspeech of 15-month-old childRen 81 von Euler, 1982), mastication (Davis & MacNeilage, iors, including speech, may invoke the adaptation and 1995), and nonspeech vocalization (Chiao, Larson, Ya- refinement of established coordinative relationships to jima, & Ko, 1994; Larson, Yajima, & Ko, 1994). Alterna- achieve new behavioral goals. tively speech motor control mechanisms may develop This idea has motivated several models of speech autonomously, independent of numerous separate, but production that explicitly incorporate the motor organi- related, behaviors. zation of established central pattern generators (CPGs; The functional components of the respiratory sys- Grillner, 1982; Wolff, 1991). These models rely on dis- tem afford a unique opportunity for empirical evalu- crete brainstem nuclei, empirically observed in ani- ation of models of speech development. In addition to mals (Feldman & Smith, 1995), as the essential coordi- being easily and noninvasively accessible, this system native mechanism underlying rhythmic behaviors such can be simply, accurately, and completely modeled by as walking, chewing, and respiration. With respect to the dynamic changes in abdominal and rib cage vol- speech development specifically, Fawcus has stated this umes (Hixon, Goldman, & Mead, 1973; Hoit, Hixon, argument most strongly: “The harnessing of basic move- Watson, & Morgan, 1990). By quantifying the man- ments by the CNS is the essential problem in...the normal ner in which these two functional components are al- development of speech” (Fawcus, 1969, p. 558). Darley, tered to achieve changing behavioral goals (e.g., gas Aronson, and Brown (1975), drawing on their extraor- exchange, modulation of alveolar pressure, abdominal dinary wealth of clinical intuition, elaborated this idea: fixation), changing control structures, strategies, and “As the baby progresses to soft and then to solid food … mechanisms can be observed. For example, one plausi- jaw movements are modified for chewing. … The motor ble hypothesis is that cortical inputs for speech produc- control of these nutritional movements must be adapted tion bypass the relatively sophisticated sensorimotor for speech production.... It is readily apparent that the integration afforded by the brainstem-level respiratory pursing of the lips for sucking may be adapted to pro- pattern generator underlying rest breathing (Feldman duce the phoneme /oo/ … opening of the jaws is nec- & Smith, 1995). It can be reasoned that, because the be- essary for the phoneme /o/ …” (p. 65). Consistent with havioral targets for speech and homeostatic breathing these models of speech development, speech breathing are so different, and because the mechanisms involved and rest breathing would also be understood as sharing are comparatively simple compared to other speech a common neural network. Grillner (1982) has suggested processes, speech motor control must rely on a dis- that centrally patterned muscle synergies are fraction- tinct coordinative mechanism. Observations of respira- ated into smaller, tightly constrained units, which in tory dynamics may reveal the presence or absence of turn may be independently controlled, perhaps by cor- these control-system redundancies during speech de- tical centers. These functional units are putatively de- velopment. This question of shared control structures rived from parts of the CPGs for respiration, mastica- has been the focus of previous investigations of devel- tion, and swallowing (Grillner, 1982), each of which can opment of speech motor control (Moore & Ruark, 1996; constrain speech subsystems in a way that simplifies the Ruark & Moore, 1997). control problems for speech development. An impor- tant immediate goal would then be to compare the co- ordinative characteristics of these behaviors with devel- oping speech. Speech as a Successor to Earlier Emerging The many similarities of babbling and speech moti- Behaviors vate the conceptualization of speech as a successor to The notion that speech develops as a successor to babbling; similarly, babbling can be seen as succeed- earlier developing mechanisms (e.g., homeostatic ven- ing other closely related, earlier emerging, centrally pat- tilation) has the appeal of parsimony and evolutionary terned behaviors. Rhythmic babbling may employ con- precedent (MacNeilage & Davis, 2000). Adaptation and trol mechanisms that underlie rhythmic patterns of exploitation of fully functioning mechanisms enable the respiration, phonation, and supralaryngeal articula- child to apply established skills to the external demands tory movement (e.g., mandibular movement; Davis & of emergent speech communication. It would seem most MacNeilage, 1995; MacNeilage & Davis, 2000). Mature efficient during motor learning to constrain the many speech may reflect the neurophysiologic basis of these degrees of freedom across speech subsystems by em- structures in the subtle rhythms that appear across lan- ploying fully formed, intact coordinative systems (Ber- guages (e.g., English has alternating strong-weak sylla- nstein, 1967). The intersystem coordination required for ble patterns; Spanish is syllable-timed; Kent, Mitchell, & speech production (e.g., among respiratory, phonatory, Sancier, 1991). These linguistic universals of phonologic and articulatory systems) might lead the toddler to re- development and rhythm lend further support to the cruit whatever organizational units are present, thereby idea that the neurophysiologic basis of emergent speech reducing the relatively large error anticipated dur- may be derived from the shaping and melding of en- ing motor learning. Development of new motor behav- dogenous control mechanisms. 82 mooRe, caulfield, & gReen in Journal of Speech, language & hearing reSearch 44 (2001) The interaction of these CPGs during speech develop- creases in length with increased exertion (Bunn & Mead, ment can be applied to speech production by the super- 1971; von Euler, 1982). Depending on external and in- imposition of rhythmic vocalization (including rhythmic ternal demands, then, one of these two control systems jaw movement) on the rhythmic pattern of rest breath- can be seen to dominate the coordinative organization ing. A child may exploit these patterned movements as for breathing. Finally, the observable musculoskele- she or he begins to produce speech by establishing a lo- tal actions characteristic of adult speech breathing and cus of control, presumably cortical, to coordinate these rest breathing are measurably distinct (Estenne, Zoc- rhythms. The resulting behavior would be the prod- chi, Ward, & Macklem, 1990; Hixon, Goldman, & Mead, uct of the endogenous rhythms underlying the precur- 1973; Hoit, Plassman, Lansing, & Hixon, 1988). Mus- sor behaviors. Support for this idea is derived from ob- cle activation patterns underlying speech breathing are servations of mutual entrainment of biologic oscillators, quite unlike those observed during rest breathing. Ac- which is widely observed across motor behavior (Kelso, tive expiratory effort characterizes speech breathing, Holt, Rubin, & Kugler, 1981). whereas the expiratory phase of rest breathing is pri- marily passive. In summary, models of speech development can be The Emergence of Speech Independent of seen to differ widely in their reliance on extant motor Other Motor Processes behaviors. Mature speech and nonspeech oromotor be- haviors exhibit distinct control modes and timing pat- Alternative representations of speech development terns, although their essential properties include shared are economically less appealing, even though these sug- anatomic and organizational structures. The empirical gestions have better empirical support. These alterna- problem is one of distilling the generalized, behavior- tives require that speech emerge as a behavior that is dependent control structures from the observed move- entirely distinct and separate from others. Support for ment dynamics of variable individual events. Detec- this view is found in comparative studies of speech and tion of these control structures will allow us to track and nonspeech movements by adults (Moore, 1993; Moore, describe the emergence of the control mechanisms for Smith, & Ringel, 1988; Wohlert & Goffman, 1994) and speech. Finally, the developmental process incorporat- from studies of mandibular movement in 15-month-olds ing rhythmic, patterned movements with the contextual (Moore & Ruark, 1996) and lip movement in 2-year-olds and external demands of speech will be revealed. (Ruark & Moore, 1997). Speech production has been shown in these studies to be sufficiently different from earlier emerging behaviors that it has become increas- Respiratory Kinematics in Speech and Rest ingly difficult to hypothesize any benefit to speech from Breathing extant mechanisms. With respect to respiratory control of speech and The respiratory system is most accessible in infants nonspeech behaviors, distinct coordinative mechanisms and adults through observations of rib cage and ab- may be suggested by the gross differences in the physi- dominal motion. Respiratory kinematics in adults have ologic goals of each. The competition of these behaviors been quantified most reliably for rest breathing, which entails task prioritization and functional compromise. has been shown to be dominated by synchronized dis- During speech breathing blood gas levels and extracel- placement of the rib cage and abdomen (Fugl-Meyer, lular pH likely depart significantly from homeostatic 1974; Hixon, 1973; Hixon et al., 1973; Konno & Mead, levels, with speakers hyperventilating during normal 1967; Sharp, Goldberg, Druz, & Danon, 1975). During speech (Bunn & Mead, 1971). This divergence from ho- rest breathing in adults these two functional compo- meostasis might seem to entail different control systems nents appear coupled, exhibiting synchronous expan- in meeting the distinct goals of these competing behav- sion and compression consequential to diaphragmatic iors. Von Euler (1982) noted that the goal of breathing activation patterns. Inspiration is correlated with volu- during speech production is distinct from that of ho- metric increases in both the rib cage and the abdomen; meostasis (i.e., achievement of target alveolar pressure expiration is correlated with decreases in each (Hixon, versus maintenance of blood gas levels). This devia- 1973). These two components do not contribute equally tion from homeostasis is easily tolerated under normal to air volume exchange during rest breathing, how- speaking conditions, but is obviously limiting during ever. Rest breathing is most commonly characterized conditions of increased metabolic need (e.g., aerobic ex- by greater relative displacement of the rib cage than the ercise; Bunn & Mead, 1971). abdomen, although equal contributions by each com- The competitive relationship of speech and ho- ponent or a predominance of abdominal displacement meostatic demands is also elucidated by the observa- can also be observed (Hixon, 1973; Hixon et al., 1973; tion that ventilation demands may supersede those of Hoit & Hixon, 1986; Sharp et al., 1975). Because the re- speech during vigorous exercise; speech phrasing de- spiratory system is open (i.e., with respect to the atmo- RespiRatoRy Kinematics duRing speech and nonspeech of 15-month-old childRen 83 sphere) during rest breathing, either subcomponent (the ing (Hixon, 1973; Hodge & Rochet, 1989; Hoit, 1994; rib cage or the abdomen) can be used independently to Hoit & Hixon, 1986), changing in its frequency of oc- modulate alveolar pressure and airflow (Hixon, 1973; currence during postnatal development and even dur- Konno & Mead, 1967). ing sleep stages. It occurs most frequently during REM Speech breathing kinematics have been character- (rapid eye movement) sleep (e.g., Goldman, Williams, ized as significantly more variable than patterns as- Soo Hoo, Trang, & Gaultier, 1995). Abdominal para- sociated with resting respiration in adults (Hixon et doxing during speech (i.e., abdominal volume increas- al., 1973) and in children less that 3 years old (Boliek, ing during expiration or decreasing during inspira- Hixon, Watson, & Morgan, 1996, 1997). Generation of tion) is more common than rib cage paradoxing (i.e., rib subglottal pressure for speech production is achieved cage volume increasing during expiration or decreasing by a net compression of the lungs, which can be accom- during inspiration) and is seen most often at high lung plished by a range of rib cage and abdominal contribu- volumes (Hodge & Rochet, 1989; Hoit & Hixon, 1986). tions. The changing relaxation pressures of the system Estenne and colleagues (1990) have suggested that ab- at different lung volumes dictate how target pressures dominal paradoxing in speech breathing “prevents can be achieved most efficiently (Hixon, 1973), although pressure from dissipating and prevents shortening of a range of combined forces can be employed to modu- the diaphragm so therefore optimizes inspiratory mus- late alveolar pressure. The effects of trade-offs among cle function” (p. 2081). There is general agreement that such factors as elastic forces, muscle efficiency, latency paradoxing occurs during speech, but its frequency of of aerodynamic response to muscle activation, and mo- occurrence is uncertain. tor control complexity are unknown. Estenne and col- Boliek and colleagues investigated breathing during leagues (1990) have shown in adults that speech breath- a wide variety of behaviors on the part of 40 infants be- ing, like rest breathing, exhibits a predominance of rib tween two age ranges: 5 weeks to 1 year (1996) and 1 cage displacement compared to the relatively smaller year to 3 years (1997). These investigators demonstrated displacement of the abdomen. Stathopoulos, Hoit, clear differences between breathing patterns associated Hixon, Watson, and Solomon (1991) similarly have with vocalization and those associated with rest breath- shown that abdominal volumes are relatively lower and ing (e.g., initiation of vocalization at lung volumes that rib cage volumes are higher during speech production exceeded the predicted end-inspiratory level for tidal than during rest breathing. This finding is in agreement breathing) and showed a remarkable degree of vari- with earlier observations (Hixon, 1973) that abdominal ability across behavior types. The patterns observed in- muscles are more active during speech production than cluded paradoxical movement of the abdomen and rib during rest. cage, which was observed about 20% of the time in chil- One explanation for these differing configurations dren 1 year old or younger. These investigators specu- is that activation of abdominal muscles during speech lated that infants experiment with a wide range of spa- breathing enhances the effects of diaphragm activity, tiotemporal patterns of breathing movements during promoting the rapid inspiratory phases that are char- vocalization. acteristic of speech (Sharp et al., 1975). Hoit and col- leagues (1988) observed abdominal muscle activity during speech and rest breathing, showing that abdom- Methodologic Considerations in Studying inal muscle activity increases during speech breath- Respiratory Kinematics in Toddlers ing compared to rest breathing. Again, the explanation One major obstacle in studying the development of for differences in abdominal activity may be that more speech and nonspeech breathing in toddlers is main- forceful activity is required during speech to keep “con- taining instrumental calibration (Boliek et al., 1996, stituent parts in favorable mechanical circumstances to 1997). This problem has usually been overcome by fre- meet the inspiratory and expiratory requirements of the quent recalibration and experimental control of record- speech breathing cycle” (Hoit et al., 1988). Speech and ing conditions and experimental tasks. However, the rest breathing are further distinguished by the modu- calibration procedures and movement restraint tech- lation of expiratory effort seen during speech to under- niques themselves may prove to be too intrusive to al- lie patterns of stress and intonation (Murdoch, Chenery, low observation of naturally occurring behavior. Alter- Stokes, & Hardcastle, 1991). natively, for observations of very young children in the Paradoxical movement of the rib cage and abdomen early developmental periods of speech, it may be possi- also distinguishes speech breathing from rest breath- ble and desirable to use uncalibrated abdominal and rib ing. Rib cage or abdominal paradoxing is characterized cage signals. by the uncoupling of rib cage and abdominal movement Essential elements of speech breathing include the to the extent that these components appear to work in timing of expiratory effort with respect to speech on- opposition. This normally occurring oppositional move- set and the modulation of expiratory airflow and pres- ment is a well-documented component of speech breath- 84 mooRe, caulfield, & gReen in Journal of Speech, language & hearing reSearch 44 (2001) sure (Boliek et al., 1996, 1997; Netsell, Lotz, Peters, & ducer position constantly, repositioning it as necessary Schulte, 1994). Quantification of the coordinative frame- to maintain transducer sensitivity and isolation of the work underlying speech has typically required care- abdomen and rib cage signals. Small changes in the po- fully controlled aerodynamic measures, including di- sition of each respiband were easily accommodated rect measures of airflow and pressure. Alternatively for by signal processing and by the analytic procedures populations that will not tolerate a face mask, or cannot described below. The rib cage respiband was placed remain sufficiently still (e.g., young children), the rela- around the rib cage, underneath the arms, as high as tive roles of thoracic and abdominal muscle groups can possible; the abdominal transducer was centered ver- be described and categorized with time-series analyses tically on the umbilicus and placed posteriorly so that of their kinematic signals (e.g., concurrent downward it overlapped the ribs as little as possible, if at all. The slopes in both abdominal and rib cage circumference transducers did not overlap. Output signals from the signals can be interpreted as giving rise to expiratory transducers were low-pass filtered using analog filters airflow). with a cut-off of 30 Hz, then recorded using an FM in- The present investigation compared relative kinemat- strumentation recorder (frequency response: DC-1250 ics of the abdomen and rib cage in toddlers to provide a Hz; S/N > 50 dB) for subsequent digitization. Speech detailed description of timing and relative displacement audio signals were obtained using a wireless lapel mi- of the respiratory components in developing speech. Of crophone worn by the subject and were recorded on a particular interest was whether children, like adults, ex- separate AM channel of the same instrumentation re- hibit task-specific movement patterns for speech breath- corder. Subjects were seated in a highchair with the ad- ing. This investigation also assessed the use of time-se- justable tray positioned as close as possible to the child ries analysis as a quantitative technique that avoids the without applying pressure on the child or on the Re- limitations usually associated with calibration of respi- spitrace bands. This positioning minimized unneces- ratory kinematic signals. Relative changes in circumfer- sary movements by the child. All stimuli and objects of ence of the two functional components of the respiratory interest were directly in front of the subject; this min- system (i.e., abdomen and rib cage) over very brief peri- imized reaching and extraneous leaning movements. ods (about one second) were compared to provide a dy- One investigator was seated next to the child and de- namic index of their changing coordinative interaction. scribed the child’s activities online. This continuous description included a gloss of all utterances (i.e., vo- calizations, babbling, and true words) produced, iden- Method tification of rest breathing periods, and alerts with re- spect to extraneous movement by the child. Postural Subjects changes and reaching movements by the child usu- ally yielded artifact in the Respitrace channels, neces- Subjects were 11 15-month-old children (7 girls, 4 sitating exclusion of affected periods from the analysis. boys) who were participants in a longitudinal study of Spontaneous and imitative utterances were elicited us- speech development. Subjects were free of known neu- ing a variety of toys, books, and games. Each child’s rologic deficit, passed otoscopic and tympanometric caregiver was present to reduce any anxiety the child screening (when tolerated by each child; tympanom- might experience and to provide assistance in eliciting etry failed for four children, who were asymptomatic vocalizations. Speech utterances were recorded over for middle ear pathology by parental report), and were a period of approximately 20 minutes. Rest breathing developing normally according to parental report of data, which were observed throughout the session, achievement of gross motor, fine motor, cognitive, were identified as uninterrupted periods of rest breath- speech, and language milestones. ing of at least 10 seconds duration constituting at least three respiratory cycles. Experimental Protocol Rest breathing and speech breathing were moni- Sampling Procedures tored in children using a Respitrace system (Ambula- tory Monitoring, Inc.), a commercially available respi- The recorded data were transcribed from the original ratory plethysmograph. This system transduces the FM tape, noting periods of rest breathing, speech, and circumferences of the rib cage and the abdomen us- artifactual movement. Speech samples contaminated ing two elasticized bands, “respibands,” as transduc- by concurrent movement, crying, laughter, chewing, or ers. Respibands were placed either on bare skin or over any other artifact were omitted from the analysis. Sam- very light clothing (e.g., a T-shirt) for each subject and pling criteria required further that the audio signal for were secured only by the elasticity of the bands them- each speech sample be free of acoustic artifact (i.e., au- selves. An experimental assistant monitored the trans- dio signals were not contaminated by simultaneous ut- RespiRatoRy Kinematics duRing speech and nonspeech of 15-month-old childRen 85 terances by the experimenter or the parent). Samples per second per channel. Following digitization, the au- were collected and categorized as rest breathing (includ- dio waveforms were full-wave rectified and integrated ing at least three contiguous cycles), babbling, vocaliza- to facilitate identification of speech onset and offset dur- tion (single phonemes ≥ 1 s duration), and speech (true ing the analysis. words). Classification of babbling on the basis of the acoustic signal alone is particularly difficult in children at this stage (Oller, 1978). Accordingly, the online gloss Analysis and subsequent transcription relied heavily on con- Time-series analyses were used to identify the chang- text. Utterances that were clearly referential or explicit ing within-cycle motion of the rib cage and abdomen requests, for example, were classified as speech. Utter- during rest breathing and the other target tasks. The ances that were recognized by the parent as a part of the analyses were designed to provide two indices of respi- child’s meaningful speech repertoire also were classified ratory function: (1) a dynamic index reflecting the cou- as speech. Conversely, utterances that were nonreferen- pling of rib cage and abdominal movements, and (2) tial and not requests were classified as babbling. Mul- a four-way classification scheme reflecting moment- tisyllabic utterances were parsed into individual sylla- by-moment changes in the relative direction of move- bles before analysis; syllable order for each token was ment of each of the two components (i.e., both expand- coded as an additional sample descriptor. These classifi- ing, both compressing, abdominal compression with rib cations provided a detailed description of the behaviors cage expansion, or abdominal expansion with rib cage sampled, although all data obtained from speech, bab- compression). Using custom routines written for Mat- bling, and vocalization utterances were subsequently lab (Mathworks, 1999), a moving rectangular window, combined into a single category, “speech,” for statistical one second in length, was used to compute the simple analysis. Each sample was coded for subject identity, be- correlation of the abdomen and rib cage signals over havior type (rest breathing, speech, nonspeech, and vo- the course of each observation. The window was ad- calizations), position of each syllable in the utterance se- vanced one point (15 ms) for each correlation compu- quence (e.g., second of three syllables), and total number tation, yielding a function (“r ”) made up of the of syllables in the sample. All tokens meeting sampling moving coefficients derived from each computation of the cor- criteria within each subject’s session were included in relation. The 1-s window width, selected empirically the data corpus. from a range of 100 ms to 2 s, was judged to provide the Sampling criteria for parsing target behaviors were most appropriate temporal resolution. Window sizes intended to maximize the number of samples acquired larger than one second were overly smoothed and were while maintaining the homogeneity of each sample by insufficiently sensitive to detect changes within speech minimizing the inclusion of nontarget behaviors (e.g., events, each of which was less than one second in dura- speech breathing samples did not include leading or tion. Smaller window sizes were overly sensitive, with trailing rest breathing). Rest-breathing samples included the resultant function emphasizing very brief (e.g., less 3 to 7 breathing cycles and were analyzed as a single be- than 100 ms) or transient relationships between the two havior. Speech samples were identified using only the signals. Speech or rest breathing segments were isolated audio channel and were demarcated by acoustic onset from surrounding events. Portions of the r func- and offset. The boundaries of speech, babbling, and vo- moving tion associated with rest or speech were isolated using calization samples were identified by simultaneously computer-assisted identification of speech onset and off- viewing the digitized sample and listening to the origi- set, or inspiratory onsets. R values within each seg- nal audio recording, which ensured accurate identifica- moving ment were transformed using the Fisher Z transform tion of the beginning and end of each utterance. Multiple and were averaged. This average coefficient provided sequential utterances, operationally defined as utter- an index of coupling for each observation. ances separated by more than 500 ms, were digitized In addition to the magnitude of each coefficient on the separately. Reliability for the method used to identify correlation function, each r point was categorized speech and rest samples was confirmed by reanalysis of moving according to the slope direction of each Respitrace sig- 10% of the data by the same investigator. Correlations of nal at that point. The slope of each signal was catego- the raw counts of each kinematic category between the rized as upward, downward, or flat within the middle two separate analyses ranged from .96 to 1. 200 ms of the 1-s r window. Again, the choice of a moving 200-ms window reflected an empirically determined de- cision regarding the sensitivity of the measure to small Digitization and Signal Processing fluctuations in each signal. Table 1 includes four possible Respitrace (two channels) and audio (one channel) respiratory movement combinations as defined by the waveforms from all acceptable samples were filtered for average slopes of the waveforms in each 1-s window: (1) anti-aliasing (f = 30 Hz) and digitized at 66.7 samples Coupled Inspiration—both waveforms positive-going; lp 86 mooRe, caulfield, & gReen in Journal of Speech, language & hearing reSearch 44 (2001) (2) Coupled Expiration—both negative-going; (3) AB↑— signals. Because the direction of airflow [i.e., inspiratory oppositional movement with a negative-going rib cage or expiratory] could be inferred only during vocaliza- signal and a positive-going abdominal signal; (4) RC↑— tion [i.e., expiratory], it was not possible to distinguish oppositional movement with positive-going rib cage sig- rib cage paradoxing [i.e., expansion of the rib cage dur- nal and a negative-going abdominal signal). For the op- ing expiration, or compression of the rib cage during in- positional movement categories, no inference of airflow spiration] from abdominal paradoxing [i.e., expansion direction was attempted, as the transducers yielded un- of the abdomen during expiration, or compression of calibrated values and airflow was not monitored. An ad- the abdomen during inspiration] during rest breathing. ditional category, annotated as Unspecified, consisted of The present description simply annotated the expansive those samples during which the average slope of either component during an oppositional event [e.g., RC↑ indi- signal was nearly zero (i.e., one or both of the signals was cated expansive rib cage displacement with compressive flat, such as might occur at the peaks and troughs of the abdominal displacement]. Ribcage and abdominal par- waveforms). Operationally, Unspecified points were de- adoxing could be identified during vocalization, how- fined as those occurrences for which the slope of the 200- ever, because no vocalized event was observed to rely ms analysis window was less than 30.3% of the standard on inspiratory flow.) deviation of the derivative of the entire signal for that Type 5 (i.e., unspecified) points were not plotted, ap- channel. The inclusion of this category greatly reduced pearing as gaps in the r function where the slope moving false indications of oppositional movement, which arose of at least one of the signals was nearly flat. when the slopes of each signal were near zero. Slight dif- The coupling strength and the relative movement of ferences or asynchronies at those points can yield oppo- the rib cage and abdomen for rest breathing and speech site slope directions, which were not judged to be true in- behaviors were compared using each of these measure- stances of paradoxing. Occurrences of samples in each of ments. The composite function of r and categoriza- moving these five categories are included in graphs of the r tion of kinematics facilitated qualitative and quantitative moving function. This graphic representation provided a three- evaluation of the dynamics between these components. dimensional composite that showed (1) strength of cou- Quantitative differences among behaviors were tested pling between abdominal and rib cage movement (i.e., by statistical analysis of r and of the proportionate moving the magnitude of r ) and (2) the relative kinematics occurrences of kinematic categories. The averaged Fisher moving (i.e., the symbol used to plot r ) over (3) time. Z transforms of the correlation coefficients and the pro- moving In Figures 1 and 2 symbols on the r waveform portionate frequencies of kinematic configurations from moving reflect categorization of changing thoracoabdominal co- each interval analyzed were subjected to descriptive and ordination, especially highlighting occurrences of op- inferential statistical analyses (one-way analysis of vari- positional movement. Coupled inspiration was inferred ance: ANOVA). from observation of concurrently rising signals (i.e., ex- pansion of each subsystem) and was plotted with light gray dots; coupled expiration was inferred from concur- Results rent decreases in both signals (i.e., compression of each subsystem resulting in expiration) and was plotted with This investigation was designed to describe and con- dark gray dots. Oppositional movements were classi- trast the coupling and relative dynamics of the rib cage fied into two types: AB↑, which was defined as decreas- and abdomen in 15-month-old children during rest ing rib cage circumference concurrent with increasing breathing and speech breathing. A total of 339 samples abdominal circumference (plotted with +s), and RC↑, were obtained for a range of behaviors from these 11 chil- which was defined as increasing rib cage circumfer- dren. Table 2 summarizes the complete data set. These ence concurrent with decreasing abdominal circumfer- signals were evaluated with relatively high temporal res- ence (plotted with ×s). (AB↑ and RC↑ are nonstandard olution (i.e., sample rate = 66.7 Hz) to observe transient symbols referring to oppositional movements of the rib relative changes in thoracoabdominal motion and r moving cage and abdomen. Their use was necessitated by meth- values. Figures 1 and 2 illustrate the nature of these mea- odologic limitations imposed by the use of uncalibrated sures in rest breathing and in speech respectively. Table 1. Categorization of all combinations of rib cage and abdominal slopes. Net slope of abdominal circumference signal Net slope of rib cage circumference signal Increasing Decreasing “Flat” Increasing Coupled inspiratory movement Oppositional movement with RC↑ Unspecified Decreasing Oppositional movement with AB↑ Coupled expiratory movement Unspecified “Flat” Unspecified Unspecified Unspecified RespiRatoRy Kinematics duRing speech and nonspeech of 15-month-old childRen 87 Figure 1. Results of windowed correlational analysis during rest breathing by Subject A. The tight coupling and synchrony of abdominal and rib cage signals (lower two traces) is reflected by uniformly high values (i.e., approx- imately 1.0) for r (top trace) and the infrequent occurrence of RC arrow up oppositional events (indicated by moving x symbols in the r trace). The light dots, dark dots, ×, and + symbols represent, respectively, coupled inspi- moving ration, coupled expiration, RC↑, and AB↑ in this figure and in Figure 2. Figure 1 provides an example of the results of these two waveforms, and a correlation of –1 indicated that correlational and classification analyses for approxi- the signals were moving in perfect opposition within mately 7.5 s of rest breathing. Interpretation of the r the 1-s analysis window. Points plotted with dots indi- mov- function is quite straightforward; a correlation of 1 cated that the signals were both increasing or decreas- ing indicated synchronous increase and/or decrease of the ing; points plotted with × or + indicated oppositional Table 2. Composition of the complete data set (i.e., number of samples of each behavior type produced by each subject). Babbling True word Vocalization Sum of babbling, speech, Rest-breathing Subject samples samples samples and vocalization samples samples A 6 22 — 28 3 B 18 5 5 28 5 C 5 32 3 40 3 D 12 12 3 27 6 E 11 34 1 46 5 F 9 3 — 12 1 G 3 40 — 43 3 H 16 — — 16 6 I 4 1 — 5 3 J 27 — — 27 4 K 14 11 — 25 3 Total 125 160 12 297 42 88 mooRe, caulfield, & gReen in Journal of Speech, language & hearing reSearch 44 (2001) Figure 2. Results of the windowed correlational analysis of the utterance /εnt/ by Subject C revealed RC↑ (× sym- bols), then AB↑ (+ symbols), oppositional movement during speech (portion between dotted lines). The entire graph comprises approximately 2.5 s. movement (i.e., RC↑ and AB↑, respectively). The most tion, yielding a comparatively small data set. In sharp obvious relationship in this figure is the coupling of the contrast to rest breathing, speech production was usu- rib cage and abdominal traces. Coupled inspiration and ally characterized by a sharp drop in the value of r mov- expiration constituted almost 80% of the points repre- and a greater frequency of occurrence of oppositional ing sented in this figure. This stable coupling was also re- movements (i.e., 83% of the kinematic samples shown in flected in the results of the correlational analysis, which Figure 2 were categorized as paradoxical; 17% were un- revealed a consistently high r function (usually ex- specified; none showed coupled inspiration or expira- moving ceeding .90; top trace). Occurrences of AB↑ or RC↑ were tion in this utterance). The numeric data sets from which found to be absent or very infrequent (i.e., together con- Figures 1 and 2 were derived are shown in Table 3. In- stituting only 1% of samples), as reflected by the ab- cluded in this table are the raw counts of the frequency sence or small number of + and × points, respectively, in of occurrence of each kinematic category, the relative the r function. The remaining 19% of rest-breath- frequency of occurrence for each category in that sam- moving ing analysis points shown in this figure were Unclassi- ple (expressed as percentages), and the mean and stan- fied, indicating that at least one of the two signals failed dard deviation of the correlation coefficients constitut- to exhibit a slope in excess of the criterion and was clas- ing r for each analyzed section. moving sified as “flat.” These unspecified points occurred al- Analysis of the complete data set proceeded in sev- most exclusively at the peaks and troughs of each respi- eral stages, each of which represented successively ratory cycle. greater reduction. Initially, frequencies of occurrence Figure 2, which includes a single vocalization pre- of each category type and average r values within moving ceded by several seconds of rest breathing, illustrates each observation were collapsed across repetitions of a very different pattern, reflecting the dramatic differ- each task (e.g., repeated observations of rest breathing ences sometimes observed for the utterances sampled. or of vocalizations) within subjects. Analysis of vari- Only those points between speech onset and offset, as ance revealed a significant subject effect for coupling indicated on Figure 2, were analyzed for speech produc- strength [F(10) = 2.71, p = .003], as reflected by the av-