III. The Dual Factor Model of Stuttering in Persistent and Transient Childhood Stuttering

Overview of the Dual Factor Model

There are at least two different conditions, each of which has traditionally been considered in isolation of the other, that appear to be associated with stuttering in adults. The dual factor model essentially constitutes an effort to marry our understanding of those two conditions in order to generate a broader account of stuttering in terms of neuropsychological, as well as situational and psychosocial considerations. As described by Forster and Webster (1991), the dual factor model contains three major elements. The previous sections of this thesis have attempted to address those points: (1) the brain mechanisms that underlie speech and language are lateralized in people who stutter as they are in the general population (i.e. the cerebral dominance hypothesis is not supported by the available evidence); (2) people who stutter appear to demonstrate an anomalous pattern of interhemispheric relations, possibly related to lateralized attentional mechanisms; and (3) people who stutter have a fragility in the neuromotor mechanisms underlying the organization and initiation of complex sequential responses. Neither of the two components of the dual factor model is believed to represent a condition that would be sufficient alone to produce the extreme manifestations of stuttering that may present in severe adult stutterers. This accounts, in part, for the conspicuous use of the terms "anomaly" and "fragility," to the exclusion of terms such as "deficit" or "disorder," when describing the two characteristics.

An important feature of the model is that the attentional and speech-motor components are seen as having a reciprocally interactive relationship with one another. One of the most notable characteristics of stuttering is the variability in both the frequency and severity of its most basic symptoms. Fluency may vary widely from moment to moment and from situation to situation. Anxiety and negative emotional arousal have a well-known and well-documented affect on the fluency of the person who stutters (Bloodstein, 1981; Trotter and Bergmann, 1957; Van Riper, 1937; 1982). Caruso, Chodo-Zajko, Bidinger and Sommers (1994) measured the effect on articulatory coordination of cognitive stress, and reported that stutterers' speech was more adversely affected by the stress than was the speech of nonstutterers. Other internal phenomena such as distraction (Bloodstein, 1981) and ineffective self-talk (Webster and Poulos, 1989) also appear to have a significant influence on fluency. An advantage of the model outlined above is that it addresses the variability in the expression of stuttering in terms of a dynamic relationship between attentional and speech-motor mechanisms.

On the one hand, an anomaly of interhemispheric relations could be capable of affecting fluency indirectly in a number of ways, depending on the precise nature of the anomaly. For example, a hemispheric attentional anomaly of the kind described by Kershner and Morton (1990; Kershner, 1995) in dyslexics, could lead to overactivation of the right hemisphere during situations where left hemisphere activation is required. Evidence of right hemisphere overactivation during a variety of language tasks in people who stutter was reviewed earlier. Inappropriate or excessive right hemispheric activation could influence the speech-motor mechanisms in the left hemisphere and supplementary motor area in at least two different ways. First, processing resources ("attention") would be drawn away from speech-motor mechanisms, increasing their susceptibility to neural interference. Second, overactivation in the right hemisphere could directly interfere with speech-motor mechanisms, through both intrahemispheric and interhemispheric spread of activation into the areas that mediate speech-motor control. Either of these mechanisms could operate independently, or both might be involved.

In either case, one may postulate a mechanism whereby a hemispheric attentional anomaly that led to right hemispheric overactivation could exacerbate a speech-motor fragility to the point where it resulted in an actual breakdown in fluency, or perhaps the sensation of an impending breakdown in fluency (i.e. covert stuttering). A more generalized attentional lability or a problem with the organizing mechanisms of the frontal lobe could affect speech-motor control by similar or analogous mechanisms. By leading to the activation of inappropriate information-processing systems, or overactivation of complementary mechanisms, resources could be diverted from speech-motor mechanisms or interfering activation could be generated. Either of these mechanisms could also presumably lead to a breakdown in the fragile mechanisms underlying sequential motor control in people who stutter.

On the other hand, a speech-motor breakdown may be able to, indirectly, influence attentional mechanisms and overall patterns of hemispheric activation. As discussed by Van Riper (1982) and Bloodstein (1981), the act of stuttering may be associated with fear on the part of the individual who stutters. Similarly, sounds, words or situations may also become associated with fear and anxiety. It is clear that negative affect of this kind is able to influence patterns of hemispheric activation (Ahern and Schwartz, 1985; Davidson and Fox, 1982; Etcoff, 1986; Fox and Davidson, 1988; Gianotti, 1972). Thus, a disfluency or the sensation of an impending disfluency could potentially trigger the right hemisphere activation that would ultimately worsen the disfluency. According to this idea, the speech-motor and attentional components of the model stuttering exist in a positive feedback relationship with one another. Webster (1990) has described this situation as a "neural amplification system." This model provides a valuable heuristic device through which one can account for a severe and potentially crippling speech disorder in terms of a relatively subtle speech-motor anomaly and a coexistent anomaly of interhemispheric relations. Blockages and the struggle described by Wingate (1964b) may be seen as a manifestation of that neural amplification system.

Separability of the Factors

Traditionally, organic explanations of stuttering have referred to a single underlying cause for the disorder; aberrant interhemispheric relations and an impairment of motor sequencing are two such examples that have been discussed here. The dual factor model invokes two underlying causes that interact with one another to give rise to the manifestation of stuttering. It is possible, however, that the two factors described in the model reflect different manifestations of a common underlying structural or functional anomaly.

In fact, there are a number of parallels between the two components of the dual factor model that suggest the possibility of a common basis or link between the two components. For example, some authors (e.g. Verfaellie et al. 1988) consider "intention" to be one sub-component of a larger construct for attention. Intention was defined by Verfaellie et al. (1988) as the mechanism by which one selectively prepares for action - a "readiness to respond." The term "intention" has also been used to describe the function of the SMA (Deiber et al., 1991; Fried et al., 1991; Goldberg, 1985), and it has been suggested here that stuttering may involve a difficulty translating intention to action. The relationship between the concept of intention as a component of attention and the early onset readiness potential associated with the SMA (Deecke, 1987) also provides an interesting source of speculation.

The commonality goes beyond the use of similar terminology, however. Fried et al. (1991) noted that unilateral stimulation of the right SMA influenced limb movements bilaterally, while stimulation of the left SMA elicited responses almost exclusively from the right limb. The authors proposed that their results might reflect right hemisphere dominance for bilateral motor intention. They noted the similarity between their observations regarding bimanual coordination and the more bilateral representation of attentional space in the right hemisphere (Heilman and Van den Abell, 1979; 1980; Mesulam, 1981; Verfaellie et al., 1988). The fact that the right hemisphere has a more bilateral representation of attentional space than the left hemisphere has been proposed as a possible neuropsychological basis for attentional asymmetries. It is thus interesting that Fried et al., (1991) speculated that "the right supplementary motor cortex may be part of the right cerebral system dominant for attentional and intentional mechanisms directed at the body and the external space in which motor activity takes place."

The corpus callosum provides a second, and somewhat less esoteric, commonality between the two components of the dual factor model. It is possible that an anomaly in the corpus callosum could manifest itself in two separate functional neuropsychological systems. According to this idea, the two components of the dual factor model represent distinct functional systems, but arise out of a common underlying structural anomaly. This possibility is particularly appealing in light of the well-documented roles of the corpus callosum in both the functioning of the SMA (Fried et al., 1991; Preilowski, 1972; 1975) and interhemispheric relations (Dimond, 1976; 1979; Levy, 1985; Moscovitch, 1986).

While it was not a primary purpose of the research, the design employed here does offer an opportunity to address the issue of the separability of the two factors in the dual factor model. This is because, as will be described below, it was expected that transient childhood stutters would demonstrate only one of the two factors described in the model. This would, of course, not be possible if the two conditions were merely different manifestations of a single underlying etiology. This idea will be discussed further in the Discussion.

Developmental Implications of the Dual Factor Model

Overview

One of the primary purposes of the present research was to investigate the relationship between transient childhood stuttering and stuttering that persists into adulthood. As discussed earlier, a substantial proportion of children who stutter will eventually stop stuttering before reaching adulthood. The exact rate of recovery is unclear from the literature, but may be as high as 80% (Andrews and Harris, 1964; Andrews et al., 1983). One principle that has emerged from the literature is that the likelihood of spontaneously recovering from stuttering decreases with age, and becomes very small after puberty or early adolescence (Andrews et al., 1983). This fact suggests that recovery from stuttering may be linked to one or more of the many important developmental changes that occur in childhood and early adolescence. However, very few attempts have been made to investigate neurodevelopmental influences on stuttering (Greiner and Fitzgerald, 1992; Sommers et al., 1975).

The neuropsychological model of stuttering described above was developed on the basis of research with adults, and was intended to refer only to persistent stuttering. However, it would be of both theoretical and clinical interest to attempt to extend the model to childhood stuttering, and to the relationship between transient childhood stuttering and stuttering that persists into adulthood. It would be difficult to address that issue by studying children who stutter directly because there is currently no reliable means of identifying those children whose stuttering will persist into adulthood.

The present research was intended to be a step in that direction, by allowing for a comparison of adults who stutter with adults who stuttered as children but who no longer do. The research used this approach to test the hypothesis that an anomalous pattern of hemispheric activation is a predisposing factor to stuttering in children. Evidence will be reviewed below suggesting that the speech-motor system may be "fragile" in all young children, in that the neural systems underlying it (particularly frontal lobe motor systems and the corpus callosum) do not reach maturity until late childhood or early adolescence. Thus, one of the two factors which, according to the model, are necessary for the expression of stuttering, may be met in all children as a part of the normal developmental course. The other factor, an anomaly of interhemispheric relations, would thus predispose the child to stuttering during childhood. According to this idea, these children would stutter temporarily but would grow out of it as the relevant speech-motor control mechanisms reached maturity.

A smaller population of children may possess a speech-motor system that will never operate as efficiently as it does in normal adults; this is the group of individuals in whom the stuttering will persist into adulthood. This could presumably relate to a developmental anomaly; they may simply fail to achieve the normative developmental endpoint with respect to the speech-motor control mechanisms that are involved in stuttering. Another possibility is that the inability of these individuals to achieve that normative endpoint may reflect a mild form of focal structural damage to the neural mechanisms that underlie the speech-motor processes involved. A third possibility is that these individuals simply fall at one extreme of the continuum of normal variability with respect to those speech-motor control mechanisms.

Speech-Motor Control

The proposition that the expression of stuttering in some children is linked to the development of speech-motor control mechanisms has face validity for a number of reasons. Perhaps most importantly, it is clear that many aspects of speech and nonspeech motor control continue to develop throughout childhood (Connolly and Stratton, 1968; Denkla, 1973; 1974; Fagard and Pezé, 1992; Kolb and Fantie, 1989; Peters, 1983; Qvarnström, Jaroma and Laine, 1994; Robbins and Klee, 1987; Sharkey and Folkins, 1985; Watkin and Fromm, 1984). Moreover, both the corpus callosum and the frontal lobes continue to develop throughout childhood both structurally (Goldman-Rakic, Isseroff, Shwartz and Bugbee, 1983; Witelson and Kigar, 1988; Yakolev and Lecours, 1967) and functionally (Jeeves, Silver and Milne, 1988; Kolb and Fantie, 1989; Lazarus and Todor, 1987; O'Leary, 1980; Passler, Isaac and Hynd, 1985; Salami, 1978; Shute and Huertas, 1990). Not only are these structures clearly important to the development of motor control, but they are critical for the aspects of speech-motor control implicated in stuttering in this paper. The supplementary motor area is a frontal lobe structure that involves extensive interhemispheric connections via the corpus callosum. The corpus callosum itself may be important for the precise integration and sequencing of the bilaterally paired speech apparatus. The presence and timing of transient childhood stuttering suggests that it could be related to the development of these neural mechanisms.

The following discussion will focus on the development of speech-motor control mechanisms, and will rely on the assumption that the developmental trends evident in a number of nonspeech motor processes are relevant to the speech-motor control mechanisms involved in stuttering.

The proposal here is that the developmentally immature speech-motor control mechanisms of young children might be relatively "fragile" compared to those of adults. One possibility is that the cortical networks that mediate speech-motor control may be poorly differentiated in children as compared to adults (Peters, 1983), resulting is a susceptibility to interference or overflow from activity in adjacent cortical areas. One way to assess this idea is the measurement of motor overflow or associated movements. Associated movements are "movements accompanying a motor or intended motor function but not necessary for its performance" (Connolly and Stratton, 1968, p. 49). Associated movements often occur in a muscle group that is contralateral and homologous to the one involved in the intended movement, but this is not necessarily the case. Most authors believe that associated movements reflect the overflow of activation from the cortical region representing the intended movement to adjacent or homologous regions (Cohen, Taft, Mahadeviah and Birch, 1967; Fog and Fog, 1963; Kinsbourne and Hicks, 1978; Lazarus and Todor, 1987; 1991; Wolff, Gunnoe and Cohen, 1983). This is generally consistent with the observation that associated movements occur most frequently and to the greatest degree in muscles which are mediated by regions either physically adjacent to those that mediate the intended movement (i.e. the adjacent finger) or contralateral and homologous to them (Connolly and Stratton, 1968; Fog and Fog, 1963; Kinsbourne and Hicks, 1978). The degree to which associated movements can be inhibited may thus represent the degree of differentiation of the relevant cortical areas. The frequency of associated movements decreases with age, and stabilizes in late childhood (Cohen et al., 1967; Connolly and Stratton, 1968; Lazarus and Todor, 1987; 1991). In fact, Wolff et al. (1983) suggested that associated movements provide a better indicator of a child's "neurodevelopmental state" than chronological age.

The improvement throughout childhood in the ability to suppress associated movements parallels the development of the corpus callosum, in terms of overall size (Witelson and Kigar, 1988) myelination (Yakolev and Lecours, 1967), and speed of conduction (Salami, 1978). The development of the corpus callosum may be just as important for the suppression of associated movements ipsilaterally as contralaterally. Dennis (1976) has argued that the corpus callosum may be essential for the differentiation of the cerebral cortex within the hemispheres. She reported that two individuals with congenital callosal agenesis exhibited greatly elevated levels of ipsilateral associated movements (i.e. movements of adjacent fingers) compared to hydrocephalic and normal controls. Dennis (1976) argued that an inhibitory action of the corpus callosum during development may be essential for progressive sensory and motor differentiation within each hemisphere. Young children have an immature corpus callosum, and, perhaps more importantly, may not have had time for the progressive differentiation within the cerebral hemispheres to take place.

Studies of bimanual coordination in children have shown developmental changes that parallel the trend evident with associated movements. Fagard and coworkers (Fagard, 1987; Fagard and Pezé, 1992; Fagard, Morioka and Wolff, 1985) have used the bimanual crank-turning task developed by Preilowski (1972; 1975) to assess bimanual coordination in children (see also Jeeves et al., 1988). These authors have consistently reported that the ability to manipulate the two limbs independently continues to improve until around 10 years of age. Since the age differences interact with a variety of factors (e.g. visual feedback, ratio of required output from the two limbs) it is clear that the trend does not simply reflect a trivial age-related improvement. In fact, Jeeves et al. (1988) reported that the pattern of performance with and without visual feedback in 6-year old children was more similar to that of surgical and congenital acallosals than 10-year old children or adults.

The bimanual coordination performance of young fluent children is in a number of respects reminiscent of the bimanual coordination deficits evident in adults who stutter. As discussed earlier, Greiner et al., (1986) and Webster (1988) reported an elevated tendency for adult stutterers to make unintentional mirror-reversals of letters with their nondominant hand during bimanual handwriting performance. Greiner and Fitzgerald (1992) had child stutterers and fluent controls perform a number of variations of the same bimanual handwriting task. Three different age groups of children performed the task: 6 to 8, 9 to 11 and 12 to 15. In the fluent controls, mirror-reversals were common in the youngest age group, but were diminished by over 90% in the 9 to 11 year old group, and were almost totally absent among the oldest group. These data are significant on their own, in that they concur with the developmental trends observed for associated movements and bimanual coordination (Cohen et al., 1967; Connolly and Stratton, 1968; Fagard, 1987; Fagard and Pezé, 1992; Fagard, Morioka and Wolff, 1985; Jeeves, Silver and Milne, 1988; Lazarus and Todor, 1987; 1991), and with the known time-course of callosal development (O'Leary, 1980; Salami, 1978; Witelson and Kigar, 1988; Yakolev and Lecours, 1967).

Collapsed across all three age groups, the stutterers produced over three times as many mirror-reversals as the nonstutterers. Interestingly, the youngest group of stutterers did not produce significantly more reversals than the controls of the same age (though there was a trend in that direction). However, the dramatic improvement noted in the older age groups of fluent children was not present in the performance of the stutterers. They showed essentially no improvement from the youngest to the oldest age group. These findings are particularly interesting because they indicate that, while performance on the bimanual handwriting task is able to clearly differentiate the two groups at older ages, it does not differentiate young children who stutter from their fluent peers. This observation would appear to support the idea that a developmentally immature speech-motor system may be sufficient to trigger stuttering in some children. The statistically nonsignificant trend towards more mirror reversals in the youngest group of stutterers may reflect the contribution of the minority whose stuttering is more severe and destined to persist into adulthood.

Greiner and Fitzgerald (1992) concluded that interhemispheric communication remains immature through at least 15 years of age in people who stutter. However, the inability to differentiate between transient and persistent stutterers in the younger age groups in their sample creates a problem for interpretation of their data. The three age groups are almost certainly composed of differing proportions of transient and persistent stutterers. Interpretation of the developmental trend (or lack of one) across age groups of stutterers is therefore problematic. It is possible, for example, that the lack of a developmental trend in the stutterers may reflect a growing proportion of persistent stutterers in the older age groups, which could effectively cancel out any age-related improvements that might exist in transient childhood stutterers. The data do indicate, however, that as stuttering persists into later childhood and adolescence, bimanual coordination deficits become more predominant.

Three major inferences can be drawn from the data of Greiner and Fitzgerald (1992): (1) the developmental trend for performance of the task in fluent controls closely follows the established time-course for bimanual coordination and callosal development; (2) the mechanisms underlying bimanual coordination are immature in young children who stutter, but apparently no more so than in fluent children of the same age; and (3) compared with their fluent peers, adolescents and young adults who stutter do appear to be delayed or blocked developmentally with respect to those brain mechanisms.

Thus, children who stutter appear to be similar to their fluent peers, though "impaired" relative to fluent adults on the bimanual handwriting task. It is perhaps significant that similar findings have been obtained from studies directly comparing the speech-motor coordination of young children who stutter with their fluent peers. Conture, Colton and Gleason (1988), and Caruso, Conture and Colton (1988) have examined various aspects of the temporal coordination of fluent speech utterances of young children (the mean ages in the two studies were approximately four and five years of age, respectively). They identified no differences between the two groups on the measures they employed. This is particularly noteworthy in light of the fact that, in a separate study (discussed earlier), the first author of the second study mentioned above (Caruso et al., 1988) did observe differences in the stability of the temporal coordination of fluent speech in adult stutterers and nonstutterers. Similarly, Howell, Sackin and Rustin (1995) recently reported speech-motor coordination differences between children who stutter and their fluent peers in an older age range (7 - 12 yrs.). Taken together, these studies suggest what Greiner and Fitzgerald (1992) observed in bimanual coordination: The speech-motor control mechanisms of young children who stutter are indistinguishable from those of their fluent peers, but differences between the groups do emerge as the two populations age, presumably reflecting a greater proportion of persistent stutterers among the heterogeneous samples of "stutterers" with increasing age.

In light of these observations, it is perhaps significant that many "normal" preschoolers experience a brief period of "developmental disfluency" (Andrews et al., 1983; Bloodstein, 1981; Johnson et al., 1959; Yairi, 1982), and that most "stutterers" of this age will eventually recover (Andrews and Harris, 1964). Whether developmental disfluency is qualitatively different from early stuttering is unclear and a matter of some controversy (Andrews et al., 1983). However, the basic characteristics of early stuttering and developmental disfluency appear to be similar (Bloodstein, 1970; Westby, 1979; Yairi and Clifton, 1972), suggesting a continuity between them.

Qualitatively, both early stuttering and developmental disfluency are different in a number of respects from the disfluency which is associated with the speech of older children and adults who stutter. First, prolongations and blockages of sounds are not typically present in either early stuttering or early normal disfluency (Johnson et al., 1959; Bloodstein, 1960a; 1960b). The disfluencies are limited largely to repetitions of syllables and whole-words in both groups. Second, the struggle and associated accessory activities that are so characteristic of adult stuttering are not present in early forms of disfluency. Many "spontaneous recoveries" occur at this age when the disfluency is considered to be stuttering (Andrews and Harris, 1964; Bloodstein, 1981). The longitudinal prevalence data of Andrews and Harris (1964) reveal a group whose period of stuttering was so early (age 2-5) and so brief (from 3 months to 1 year) that the authors referred to them as "developmental stutterers."

It is appealing to speculate that both early stuttering and early normal disfluency may reflect an underlying state of immaturity of the speech-motor mechanisms. This idea is certainly consistent with the data of Greiner and Fitzgerald (1992). It is also consistent with the observation that both the prevalence of stuttering (Andrews and Harris, 1964) and the severity of normal disfluency (Yairi, 1982) diminish as those mechanisms progress toward maturity.

Interhemispheric Relations and Attentional Mechanisms

Evidence relating to the role of development in the anomalous interhemispheric relations of people who stutter is less clear and more difficult to assess than is the case with speech-motor control. One study has attempted to examine cerebral dominance in children and adults who stutter. Sommers et al. (1975) assessed dichotic listening ear preferences in children, adolescents and adults. In agreement with the literature in general, they reported that: (1) while there was an overall right ear advantage in both stutterers and nonstutterers, (2) the stutterers had a weaker right ear preference on the task; (3) and there was a higher proportion of stutterers than nonstutterers with atypical ear preferences. Most significantly, however, the authors reported that child stutterers had significantly smaller right ear preferences than the adolescent and adult stutterers. The authors related their findings to recovery from stuttering in children, suggesting that as one hemisphere emerges as dominant, recovery from stuttering becomes more likely. This conclusion is clearly incompatible with the position taken here with respect to the cerebral dominance hypothesis. It is also incompatible with more recent evidence pertaining to the development of cerebral dominance. Converging evidence from a variety of sources indicates that hemispheric asymmetries are present from birth and do not increase in magnitude over time in normal individuals (Bryden and Allard, 1981; Dennis and Kohn, 1975; Dennis and Whitaker, 1976; Hiscock and Kinsbourne, 1977; 1980; Ingram, 1975; Kinsbourne, 1975; 1989; Molfese, Freeman and Palermo, 1975; Vargha-Khadem et al., 1985; Witelson, 1970; Woods and Carey, 1978; Woods and Teuber, 1973).

As an alternative to the conclusions of Sommers et al. (1972), one might propose that recovery is in some way related to the development of the hemispheric attentional mechanisms. On the surface, this proposition appears unlikely to be correct. As with structural hemispheric asymmetries, attentional asymmetries can be demonstrated in young children and do not appear to increase with age (Geffen, 1978; Geffen and Sexton, 1978; Hiscock and Kinsbourne, 1977; 1980; Obrzut, Mondor and Uecker, 1993). On the other hand, selective attentional processes may be influenced by development independently of lateral asymmetries (Beckie and Hiscock, 1991; Geffen and Wale, 1979). Geffen (1978; Geffen and Wale, 1979), for example, has found that the ability to overcome the right ear advantage for consonant-vowel recognition improves with age in children between six and eight years of age. She argued that control of selective attention to one ear improves with development, while structural asymmetries remain constant. Developmental changes in the ability to direct attention to one or the other ear may relate to the maturation of the corpus callosum and its role in the integration of the two hemispheres. Consistent with this idea, Molfese et al. (1975) reported that asymmetries in auditory evoked responses to speech and nonspeech stimuli decreased with age. They interpreted that finding in terms of increasing interaction between the cerebral hemispheres with maturation of the corpus callosum.

The development of the corpus callosum and its proposed roles in interhemispheric interaction (Molfese et al., 1975) and the unification of attentional processes (Ellenberg and Sperry, 1979; Gazzaniga, 1987; Levy, 1985) certainly leave open the possibility that attentional mechanisms play a role in the expression of stuttering in children. It is interesting, in this context, to contrast the findings of Sommers et al. (1975) with those of Greiner and Fitzgerald (1992). Relative to the performance of fluent children, the bimanual coordination deficits noted by Greiner and Fitzgerald (1992) became more prominent with increasing age. The reverse pattern was evident in the dichotic listening data of Sommers et al. (1975). Among the stutterers, the youngest children had weaker (i.e. less normative) asymmetries than the adolescents and adults. Given the developmental invariance of both structural and attentional asymmetries in the general population, it is unlikely that dichotic listening asymmetries increase with age in people who stutter. A more reasonable interpretation of the data of Sommers et al. (1975) is that they reflect differences in the composition of the three age groups of stutterers with respect to transient and persistent stutterers. The youngest group would be expected to include a substantially greater proportion of transient stutterers. That same group demonstrated a weaker dichotic listening asymmetry than the other groups. This observation, in light of the data of Greiner and Fitzgerald (1992) is consistent with the idea that the proposed anomaly of interhemispheric relations may be sufficient to induce stuttering in young children in whom the speech-motor system is immature.

One question that has not been addressed relates to the variability in the age of recovery. The developmental model proposed here is not intended to provide a complete account of the factors that mediate recovery from childhood stuttering. It is likely that a variety of environmental and psychosocial factors interact with the mechanism proposed here to determine the developmental course of an individual's stuttering. It is fundamental to the original dual factor model that this kind of individual and temporal variability is able to influence the expression of stuttering. Applied to transient stuttering, the situation becomes even more dynamic and hence more complex. There are likely to be considerable variations across individuals in the time course of speech-motor development and the nature and degree of the attentional anomaly. Learned associations with sounds, words, places and situations could influence the expression of a child's stuttering. The influence of these associations could in turn influence the time and perhaps even extent of the recovery. The child's self-image and attitude toward his stuttering could play a similar role in expediting or delaying recovery. The essential feature of the developmental model proposed here is that it proposes changes in the underlying neural basis of the disorder which ultimately permit that recovery.

Summary: An extension to the model of Forster and Webster (1991) has been proposed in an attempt to account for the high rate of recovery from stuttering in childhood. This extension to the model has three major elements: (1) at the time of speech onset, the neural mechanisms that underlie the aspects of speech-motor control that are relevant to stuttering are immature in all children, and are thus fragile and poorly differentiated compared to those of adults, (2) because of the immaturity of the young speech-motor system in all children, a hemispheric attentional anomaly may be sufficient to trigger transient childhood stuttering, (3) the stuttering would be expected to improve with maturation of the speech motor system. The evidence for this model is largely indirect and speculative. The present research was intended to be a more direct test of the model.

Rationale of the Research

The research involved a series of four tasks intended to investigate both the attentional anomaly and speech-motor control. In order to address the issue of recovery, the same three groups of participants performed all four tasks: (1) adults who stutter ("stutterers," ST); (2) adults who stuttered as children but have recovered ("exstutterers," ES); (3) a control group of individuals who have never stuttered ("nonstutterers," NS). Of the four tasks, two were intended to assess the speech-motor fragility, and two were intended to assess the anomaly of interhemispheric relations. According to the model described here to account for transient childhood stuttering, the speech-motor tasks were expected to differentiate the persistent stutterers from the other two groups. The anomaly of interhemispheric relations, which was assessed in terms of the size and direction of the performance asymmetry on two divided visual field tasks, was expected to differentiate both the stutterers and ex-stutterers from the control group.

Stated differently, the following hypotheses were made with respect to the performance of the three groups on the tasks:

 

Speech-motor control tasks: (NS = ES) > ST

Divided visual field asymmetries: NS > (ES = ST)

 

The four tasks chosen for the research will be outlined briefly below.