Max et al.: Neuromotor Control in Stuttering 215 Journal of Speech, Language, and Hearing Research ��� Vol. 46 ��� 215���232 ��� February 2003 ��� ��American Speech-Language-Hearing Association 1092-4388/03/4601-0215 Ludo Max University of Connecticut Storrs, CT and Haskins Laboratories New Haven, CT Anthony J. Caruso Kent State University Kent, OH Vincent L. Gracco McGill University Montr��al, Qu��bec, Canada and Haskins Laboratories New Haven, CT Kinematic Analyses of Speech, Orofacial Nonspeech, and Finger Movements in Stuttering and Nonstuttering Adults This work investigated the hypothesis that neuromotor differences between individuals who stutter and individuals who do not stutter are not limited to the movements involved in speech production. Kinematic data were obtained from gender- and age-matched stuttering (n = 10) and nonstuttering (n = 10) adults during speech movements, orofacial nonspeech movements, and finger move- ments. All movements were performed in 4 conditions differing in sequence length and location of the target movement within the sequence. Results revealed statistically significant differences between the stuttering and nonstuttering individuals on several measures of lip and jaw closing (but not opening) move- ments during perceptually fluent speech. The magnitude of these differences varied across different levels of utterance length (larger differences during shorter utterances) and across different locations of the target movement within an utterance (larger differences close to the beginning). Results further revealed statistically significant differences between the stuttering and nonstuttering groups in finger flexion (but not extension) movement duration and peak velocity latency. Overall, findings suggest that differences between stuttering and nonstuttering individuals are not confined to the sensorimotor processes underlying speech production or even movements of the orofacial system in general. Rather, it appears that the groups show generalized differences in the duration of certain goal-directed movements across unrelated motor systems. KEY WORDS: stuttering, speech motor control, nonspeech motor control, kinematic analysis Considerable evidence has accumulated in support of the view that individuals who stutter, as a group, differ from individuals who do not stutter in at least some of the neuromotor processes in- volved in speech production (e.g., Alfonso, 1991 Caruso, Abbs, & Gracco, 1988 Caruso, Max, & McClowry, 1999 Smith et al., 1993 Zimmermann, 1980). However, it remains unclear how this evidence should be inter- preted in terms of the basic mechanisms underlying stuttering. One cannot rule out the possibility that the observed between-group differ- ences reflect stuttering speakers��� self-learned or treatment-induced com- pensatory adjustments to minimize or avoid speech breakdowns (Conture, Colton, & Gleason, 1988 McClean, Kroll, & Loftus, 1990 van Lieshout, Hulstijn, & Peters, 1996a, 1996b Zimmermann, 1980). As a result, it is often impossible to unambiguously interpret findings of speech motor differences between stuttering and nonstuttering individuals.

216 Journal of Speech, Language, and Hearing Research ��� Vol. 46 ��� 215���232 ��� February 2003 One potential way to avoid this ambiguity involves examining stuttering and nonstuttering individuals��� movements in comparable tasks across different motor systems. In other words, the issue could be addressed by investigating whether between-group differences similar to those typically observed for orofacial speech movements are also present when movements of the same motor system are performed for nonspeech pur- poses and when movements are performed with an un- related motor system (e.g., the fingers). This approach could provide important insights into the nature of stut- tering by answering the question of whether individuals who stutter and individuals who do not stutter differ in the neuromotor organization of (a) speech movements only, (b) orofacial movements in general regardless of their speech or nonspeech function, or (c) movements in gen- eral regardless of the involved effector system. The question of whether differences between stut- tering and nonstuttering individuals are limited to the sensorimotor processes involved in speech production is an intriguing one in light of numerous publications sug- gesting that at least some of the central nervous system���s underlying substrates and control principles are simi- lar for speech and nonspeech movements (Evarts, 1982 Franz, Zelaznik, & Smith, 1992 Gracco, 1990 Grillner, 1982 Ostry & Cooke, 1987 Ostry & Flanagan, 1989 Ostry, Keller, & Parush, 1983 Smith, McFarland, & Weber, 1986 Smith & Zelaznik, 1990). Unfortunately, the results from many studies on nonspeech motor con- trol in individuals who stutter���including several early studies completed in the first half of the 20th century (for a review of these older studies, see Bloodstein, 1995)���are vague and inconclusive. In addition, integra- tion with the data obtained for articulatory movements continues to be precluded by the absence of comparable data across motor systems. The results from more recent studies suggest that nonspeech motor differences between stuttering and nonstuttering individuals are task dependent. On the one hand, differences have been observed in behavioral measures of finger movement sequencing accuracy, ini- tiation time, and execution time (Forster & Webster, 2001 Webster, 1986, 1989a, 1989b, 1990) manual reac- tion times (Bishop, Williams, & Cooper, 1991 McFarlane & Prins, 1978 Webster & Ryan, 1991) and bimanual coordination (Forster & Webster, 2001 Zelaznik, Smith, Franz, & Ho, 1997). On the other hand, no differences were found for manual rhythmic timing or isometric force generation tasks (Hulstijn, Summers, van Lieshout, & Peters, 1992 Max & Yudman, 2003 Zelaznik, Smith, & Franz, 1994). Most important, however, with the excep- tion of the work by Zelaznik et al. (1997), none of these studies included kinematic analyses of nonspeech move- ments. Moreover, data regarding kinematic character- istics of movements across speech and nonspeech motor systems in a single group of stuttering individuals have been completely unavailable to date. This situation is particularly unfortunate given that paradigms involving both speech and nonspeech motor tasks may yield evidence in support of the hypothesis that previously reported speech kinematic differences between stuttering and nonstuttering individuals do indeed reflect underlying sensorimotor difficulties in the stuttering group. Although generalized motor differences may still reflect a compensation strategy or an interac- tion between a compensation strategy and the underly- ing problem, the nonspeech movement differences would suggest that even if compensatory behavior contributed to the results, it occurred to minimize a sensorimotor problem. Thus, studies of nonspeech motor control may also yield critical insights regarding the level at which individuals who stutter experience difficulty during the speech production process. That is, additional evidence indicating that differences between individuals who stut- ter and individuals who do not stutter are not confined to movements of the speech system would support the hypothesis that stuttering is likely to be a disorder that is associated with aberrant sensorimotor processes, or possibly aberrant sensorimotor and linguistic processes, rather than a disorder that is essentially linguistic in nature. The aim of the present work was three-fold. A first goal was to investigate whether differences exist in the articulatory kinematics of perceptually fluent speech pro- duced by individuals who do not stutter versus individu- als who do stutter but who have not participated in stut- tering treatment for several years. A second goal was to determine whether the same nonstuttering and stutter- ing individuals show kinematic differences in orofacial nonspeech movements and finger movements. A third goal was to determine whether any such between-group dif- ferences in kinematic parameters of speech, orofacial nonspeech, and finger movements are similar for (a) ar- ticulatory closing/finger flexion movements versus articu- latory opening/finger extension movements, (b) longer versus shorter movement sequences, and (c) target move- ments near the beginning versus end of a sequence. Including movement type (closing/flexion vs. open- ing/extension) in this investigation was motivated by the fact that previous kinematic studies of stuttering individuals��� fluent speech have examined either only closing movements or only opening movements (Alfonso, 1991 Caruso et al., 1988 De Nil, 1995 J��ncke, Kai- ser, Bauer, & Kalveram, 1995 McClean et al., 1990 Ward, 1997 Zimmermann, 1980). However, it has been suggested that closing and opening movements are fun- damentally different actions operating under different constraints and that closing movements are more dif- ficult to control than opening movements (Gracco, 1988,

Max et al.: Neuromotor Control in Stuttering 217 1994). Thus, studies including both closing and open- ing movements may provide important new insights into the neuromotor mechanisms involved in stuttering by documenting whether these two types of movements have differential effects on any kinematic differences between stuttering and nonstuttering individuals. Including movement sequences of different length and with the target movement occurring at different locations within a sequence was motivated by previous studies that suggested that stuttering is more likely to occur on longer versus shorter words (Brown, 1945 Soderberg, 1966 Taylor, 1966), on longer or more com- plex versus shorter or less complex utterances (Brundage & Ratner, 1989 Jayaram, 1984 Logan & Conture, 1995 Tornick & Bloodstein, 1976 Wells, 1979), and on words located near the beginning versus the end of an utter- ance or clause (Brown, 1945 Jayaram, 1984 Koopmans, Slis, & Rietveld, 1991 Soderberg, 1967 Taylor, 1966). In order to gain an understanding of the impact of such variables on the frequency and loci of stuttering, it is necessary to describe their influence at the level of the neuromotor processes involved in speech movement con- trol. An example of a step in this direction can be found in a recent study suggesting that increased syntactic complexity is associated with increased variability of lower lip movements in adults who stutter (Kleinow & Smith, 2000). It should be pointed out, however, that the exact nature of the relationship between variables such as utterance length or complexity and either stut- tering frequency or speech motor performance is com- plex and poorly understood. First, a number of studies have failed to find a direct relationship between stut- tering frequency and utterance length or complexity (Silverman & Ratner, 1997 Wall, Starkweather, & Cairns, 1981), and some studies that did find such a relationship were not controlled for the greater prob- ability of at least one moment of stuttering occurring in utterances that contained more syllables or words. Sec- ond, some studies using speech motor paradigms have failed to find a Group �� Utterance Length interaction (Maske-Cash & Curlee, 1995 Reich, Till, & Goldsmith, 1981 van Lieshout, Hulstijn, & Peters, 1991) or have found it only for subgroups of stuttering individuals (Dembowski & Watson, 1991 Watson et al., 1991 Watson, Pool, Devous, Freeman, & Finitzo, 1992). Clearly, fur- ther data are needed to clarify these relationships at both the behavioral and neurophysiological levels. Method Participants Participants were 10 stuttering and 10 nonstuttering adults, ranging from 27 to 45 years of age (M = 34.3, SD = 5.7) for the stuttering group and from 26 to 46 years of age (M = 33.9, SD = 5.9) for the nonstuttering group. Each group included 7 men and 3 women, all native speakers of American English. Detailed individual par- ticipant data are provided in Table 1. Stuttering individuals reported that the onset of stut- tering occurred during childhood, that they had no diag- nosed neurological or communication disorders other than stuttering, and that they had not received stuttering treat- ment for at least 8 years preceding data collection. Nine stuttering participants reported being right-handed and 1 reported being left-handed. All individuals in the nonstuttering group (a) were individually matched for age (�� 3 years), gender, and self-reported handedness with a participant in the stuttering group, and (b) reported no diagnosed neurological or communication disorders. All Table 1. Individual participant data for gender, age, and number of right (R), left (L), and either (E) responses for the Handedness subscale of the Lateral Preference Inventory (LPI). For stuttering participants, Stuttering Severity Instrument���Third Edition (SSI-3) severity classification, percentage of stuttered syllables (%SS) during a conversational speech sample, and number of years since last participation in stuttering treatment are also included. Stuttering group Nonstuttering group Years since Participant Gender Age SSI-3 (%SS) treatment LPI Participant Gender Age LPI S1 M 35 Mild (5.00) 12 4R 0L 0E N1 M 33 4R 0L 0E S2 M 28 Mild (8.67) 14 4R 0L 0E N2 M 29 2R 0L 2E S3 M 29 Mild (7.00) no treatment 4R 0L 0E N3 M 30 3R 0L 1E S4 M 27 Moderate (8.33) 16 3R 1L 0E N4 M 26 3R 0L 1E S5 M 45 Mild (6.67) 8 4R 0L 0E N5 M 46 4R 0L 0E S6 M 40 Mild (11.33) 17 0R 4L 0E N6 M 37 1R 3L 0E S7 M 37 Moderate (15.33) 9 4R 0L 0E N7 M 37 4R 0L 0E S8 F 36 Moderate (5.67) 14 4R 0L 0E N8 F 38 4R 0L 0E S9 F 35 Moderate (17.67) 16 4R 0L 0E N9 F 34 2R 2L 0E S10 F 31 Severe (7.33) 15 4R 0L 0E N10 F 29 4R 0L 0E