Neuroanatomical Summary
Reactive Inhibition Model of Developmental Stuttering
Stages of Stuttering Development
Types of Spontaneous Recovery
Recovery as a Consequence of Therapy
Challenges to and Tests of the Theory
Implications for Therapy Approaches
Veils of Stuttering

The information in this article is original, copyrighted material. References, brief quotations, and discussion of the Reactive Inhibition model are welcomed and encouraged as long as the source is cited and acknowledged.

A Reactive Inhibition Model of Stuttering Development & Behavior:
A Neuropsychological Hypothesis Based on Clinical Evidence and
Recent Research

Darrell M. Dodge, M.A., CCC-SLP
Littleton, Colorado, USA

June, 2011 - Version 04-11

Q: Why doesn't the amygdala appear to be activated in PET scans of the brains of adults who stutter?
Read an
answer here.


The hypothesis presented in this paper primarily concerns the brain functions underlying developmental stuttering. It was developed to explain the observed and reported behaviors of stuttering, including those resulting from ameliorative stuttering treatment of various types, as reported in research literature and clinically observed. The hypothesis proposes that stuttering behavior originates from increasingly conditioned reactions of the extended amygdaloid circuits to developmental disfluencies, which may be associated with several types of temporary or permanent speech-language impairments in childhood. The core behavior of developmental stuttering is described by this hypothesis as an inhibition of the speech-language motor system, resulting in disruption or freezing when some component(s) of the speech mechanism return(s) an impaired (unexpected) response. When the person experiencing this inhibition attempts to overcome the loss of speech control that results, compensatory movements (which include forcing and tension of the speech musculature) cause chaotic motor movements that result in additional compensative reactions. Each element of this behavior is thought to be influenced by genetic and developmental associations. For example, the original impairment to which the stutterer reacts may be genetic in nature, but may be a temporary characteristic associated with development. Likewise, repetition of the behavior during childhood results in the reactive forcing and resulting chaotic motor movements being classically conditioned and associated with the inhibitory response. The strength, duration, and/or target of the reaction and of this inhibitory response itself may likewise result from a genetic predisposition. In addition, the effect of the inhibitory attempt on brain function may also be influenced by genetic characteristics and developmental factors that govern the vulnerability of the neural system and circuits comprising the speech-language system to inhibition. An example of such a characteristic may be the fractional anisotropy (FA) in white matter tracts that has been measured in both children (Chang et. al. 2008) and adults (Sommer 2002), which has been enriched by Lu C. et al 2009.   After describing the neurological details of the hypothesis, the paper continues with a description of proposed stages of stuttering development, applicability of the hypothesis to natural fluency recovery and recovery as a consequence of therapy. a discussion of possible ways of disproving the hypothesis, and implications for stuttering therapy.


Essential details of the reactive inhibition model of stuttering set forth in this paper were first proposed in the author's posting on the STUTT-L listserv in November, 1995. Since then, the model has been refined and enhanced the model to include additional information obtained from recent research in stuttering, speech-language pathology, audiology, and neuro-psychology.

It should be emphasized that this model is a working hypothesis, and is not intended to be considered a statement of fact.  It was formulated as an attempt to put many of the puzzling and seemingly inconsistent characteristics of stuttering into a logical context that could eventually be scientifically proven or disproven. The model is based on the published research of others, his academic work at the University of Colorado (1995-99), substantive communication with hundreds of people who stutter (PWS), personal experience with numerous stuttering therapy approaches, and extensive clinical observations in diagnosing, treating, and studying the behavior of over 450 children, adolescents and adults with fluency disorders (developmental, psychological and neurological stuttering and cluttering) since 1998.

It is important to point out that from the very beginning (the posting referenced above) the model has been developed to resolve the "conflict" between "nature" and "nurture" hypotheses of stuttering behavior and development. It is assumed that there is an important poly-genetic component to stuttering and that this has a key influence on the bio-psychology of stuttering.

Development of this model has required considerable self-inquiry on the author's part and this has involved questioning many original conventional assumptions about stuttering. It is possible that one reason why the etiology of stuttering has been so resistant to discovery is that investigators are afflicted with preconceived notions that prevent the development of useful new hypotheses. Among these are four assumptions that are continually encountered in stuttering research, within the community of people who stutter, and in society at large:

  1. That stuttering must be primarily or solely explained by a "medical model" that requires the presence of one or more inborn neurological deficits or "flaws."
  2. That if there were no neurological deficit or flaw "causing" stuttering, then people who stutter should be able to learn how to consciously control their speech disorder or recover completely from stuttering following speech or psycho-therapy.
  3. That stuttering is a disorder that can be understood by primarily studying the activity of the cortical and neocortical areas of the brain or can be explained primarily or solely by a deficit or excess of one or more neurotransmitters.
  4. That stuttering occurs primarily because of functional anomalies of language production and the coordination of motor speech control that are based in the cortical areas of the brain and are explained by the "neurological deficits" of assumption 1.

Each of these assumptions is informed by an obsolete model of human cognitive function and behavior that has been considerably revised by experimental neurological research. Unfortunately, most of this research has not yet been applied to the field of fluency disorders and speech-language science.

The first assumption over-generalizes our knowledge of disease processes to encompass all of human behavior and is behind our obsession with drugs as a "cure-all" and genetics as an explanation for all behavior. 

The second assumption is an emotional issue that underlies or motivates much of the thinking behind current research. It is difficult to move beyond because this manifestation of "New Age Guilt" is supported by the other assumptions, as well as the common human confabulations that: 1) intellectual control over all brain processes is possible, and 2) that cognitive or "spiritual" changes are totally responsible for recovery processes that are actually better explained by physically-based de-conditioning or extinction processes.

The third assumption is understandable given the prevailing sentimental model of brain function, that defines the subcortical brain as a sort of vestigial "less than human" structure and the "higher" cortical or neocortical brain as what humans (as opposed to "lower" animals) primarily use to think and act.

The fourth area (the focus on stuttering as reflecting a functional language disorder or a disorder in speech coordination) is a rocky field that has been plowed again and again by researchers to no avail. Recent research purporting to find structural brain anomalies in stutterers (e.g., Foundas 2001) does not adequately control for the effects of life-long stuttering on the brain.

It is not difficult to see how these assumptions could persist. The function of the human brain is difficult even for neurologists to comprehend and explain, resulting in over-simplistic popular explanations and the promulgation of simplistic concepts by lay people. The ease with which even scientists can oversimplify brain function is demonstrated by Perkins (2000), who argues that stuttering cannot be assigned to one brain area and then promptly traces all stuttering to the cerebellum, which requires him to misrepresent the function of that brain structure. 

Definition of Stuttering

The reactive inhibition model described in this article assumes that stuttering behavior is a largely self-inflicted disorder of speech, communication, and awareness that comprises a reaction to neurological characteristics or weaknesses in the speech-language system that cause disruptions in the coordination of speech motor and language information that is required to produce fluent speech. It assumes that stuttering does not primarily involve a neurological flaw or a disease process, but is rather a manifestation of normal and universal human processes of reactivity and self-protection, reflected in the processes of fear and threat conditioning. In other words, stuttering involves a difference in degree -- rather than a distortion -- from a variety of "normal" functions.  Further, it assumes that most of the observed physical characteristics of stuttering are the result of that process of reactivity and self-protection, and are not in themselves the cause of stuttering. And it also draws on the increasing realization that every human being is subject to similar types of unconscious and autonomic reactivity, causing problems ranging from low self-esteem, to loss of spontaneity, to lack of self-realization. Such reactivity can be manifested in various ways (Goleman 1995).  The premise of the present hypothesis is that there are a number of genetically transmitted and environmentally triggered neurological characteristics that make certain humans more vulnerable to autonomic disruptions of speech motor-language processing (see, for example, Guitar (2003), Sommer (2001), Chang (2008)).

This definition is not intended to "blame" or "hold stutterers responsible" for their stuttering, as has been claimed in some quarters. Rather, it is required to explain the fact that there is often a lag (sometimes a significant one) between the beginnings of language usage and the onset of stuttering behaviors, indicating that stuttering is triggered by the stutterer's personal response to some disordered aspect(s) of speech communication, internal (neurological) or external (social) to him/herself. This response is almost certainly beyond the conscious control of the stutterer, at least with respect to the outcome of disfluency. In fact, a common internal experience of stuttering communicated by many people who stutter often begins with a sense that an anticipated speech gesture is impossible to perform, and results in a loss of control of speech when the person who stutters attempts to speak in spite of this premonition.

The clinical evidence that stuttering often begins in the 3rd or 4th year of life, at the very time when speech begins to become a primarily automatic rather than deliberate activity, indicates that stuttering may be a disorder of the automatic (or intrinsic) speech system.

It may be argued by some advocates of stutterers that if such a hypothesis was eventually proven and stuttering was found to be a "normal" (albeit extreme) human process that is not totally "physical" or congenital in nature, then people who stutter would lose the societal support that can facilitate their recovery. But the conditioned and reactive aspects of stuttering behavior, in light of recent research showing the physical basis of other behaviors such as post traumatic stress and anxiety disorders, actually increase its significance as a neuropsychological and social-communication disorder worthy of research investment and accommodation by society. A primary benefit from pursuit of associated research would be to provide much-needed input to understanding and improving the efficacy of stuttering therapy as well as understanding which youngsters may be more vulnerable to the development of chronic stuttering.

As detailed below, this model is somewhat unique in that it is derived not only from the evidence available from stuttering behavior, but from the process of recovery from stuttering. It seeks to explain the major features of this process, including the seeming impenetrability of "chronic" stuttering, characteristic responses to the various types of therapy, the situational nature of stuttering, the tendency to relapse, and the persistence of stuttering responses and reactivity, even in many people who have regained virtual fluency.

The description of this working model (or hypothesis) will be periodically revised and updated.  The author does not consider this hypothesis to be mature enough to be called a "theory."

Finally, it is important to point out that this model has little in common with popular notions that stuttering is caused solely by a "panic reaction," by common "flight or fight" responses, or by "assertive - aggressive" conflicts. The notion that stuttering is caused solely by an "overactive" amygdala in adults is also an oversimplification (especially because the amygdala is probably less important in older stutterers than in childhood stuttering.) Such theories require that stutterers have a unitary psychological profile that is vastly different from fluent speakers (something not demonstrated by research) and do not explain stuttering (or fluency) that occurs in the absence (or presence) of such extreme reactions, responses, and conflicts -- something commonly observed in the behavior of people who stutter.

[ Contents ] Introduction

Recent research by the team of Joseph LeDoux (LeDoux 1996, 2001) has indicated that the amygdalae (or, more properly, the amygdaloid nuclei) are a key component of brain circuits involved in mediating classical conditioning and rapid, subconscious, reflex-like responses to conditioned stimuli in mammals, including man. The amygdalae are bilateral structures associated with the "limbic system" at the base of the telencephalon. They are a part of the so-called "Papez circuit" that modulates aggressive and fear behavior, and which includes the hippocampus, the cingulate cortex, several nuclei of the thalamus, the septum, and the mammillary bodies at the base of the hypothalamus. The characteristic actions of  this circuit are compelling when these structures are viewed as possible mediators of reactive, chronic stuttering behavior in humans. The specific factors that suggest the possible involvement of the circuit in stuttering development include:

1) The sub-cortical, involuntary nature of the amygdaloid reactions, which occur with extreme rapidity in response to auditory, tactile, somato-sensory, and situational cues, (very similar to the conditioning of speech-sounds modified by excessive bodily and articulator tension and associated with specific situational memories in stuttering behavior.)

2) The important roles of auditory and proprioceptive input in the development of learned fear responses.

3) The role of the amygdala in forming other strong, pre-cortical, conditioned memories of trauma, such as those associated with post-traumatic stress disorder.

4) The involvement of the amygdala in mediating survival responses that act directly on specific cortical sensori-motor areas, such as the "freeze" response and other inhibitory control responses in mammals.

5) The parallel processing of "implicit" amygdaloid memories with "explicit" situational hippocampal memories, which associate specific motor actions with specific situations that are imbued with emotional significance, given the well-known situational nature of chronic stuttering.

6) The existence of "extinguishing" links -- in the form of neural projections -- to the amygdalae circuit from the cortical areas, which may help explain the partial effectiveness of speech therapy in many cases when speech treatment targets fear conditioning mediated by the amygdalae.

7) The possible variability of the permanence and indelibility of reactivity in different subjects (given possible anatomical differences in the number and quality of neural projections to and from the amygdalae and other subcortical structures involved in conditioning), suggesting a contributing reason for the varying severity of stuttering and the varying success of stuttering therapy.

8) The lack of a triggering or precipitating factor for the extreme nature of both core and secondary stuttering behaviors, which are not explained by weaknesses in either language processing or motor movement initiation or sustainment.

9) The "simplicity" or parsimoniousness of the model derived from these characteristics (referring to its directness and similarity to other already known brain functions), and its close correlation with previous models of stuttering as both a genetic and learned behavior. 

Also of interest, from another perspective, are the indelible nature of the conditioned reactions associated with stuttering, which can be extinguished, but can also be quickly relearned, which correlates with the modern understanding of the neurology of classical conditioning with the common experience of relapse following stuttering therapy.

[ Contents ] Neuroanatomical Summary

The Brain's "Early Response System"

The two amygdalae (one on each side) are located anterior and medial to the hippocampus at the base of the telencephalon (forebrain.) Crucial to the model of stuttering developed here are the direct subcortical projections to the amygdala from the auditory nuclei of the thalamus and other subcortical structures, higher projections from the auditory and speech centers in the neocortex, and projections from the amygdala to the hypothalamus, periaqueductal gray, reticulo-pontis caudis, basal ganglia, anterior cingulate cortex, and cerebral cortex. Also pertinent to the effects in the model are the larger number of projections from the amygdala and other subcortical structures (such as the striatum and cerebellum) to the neocortex and the smaller number of projections from the neocortex to these structures, implying a dominant subcortical role in the mediation of conditioned responses associated with survival and fear.

The amygdala mediates the emotional significance of events and has been found to be the key structure in classical conditioning (LeDoux 1996). It is the brain's "early response system," with the primary role of alerting the individual to approaching danger and preconditioning the autonomic nervous system to allow automatic decisions and action by the cortical sensori-motor system.

The subcortical projections to the amygdala are key to this activity, allowing the brain to begin responding to an auditory, visual, or internal/proprioceptive cue that has been learned to have crude emotional significance by specialized neurons in the amygdala. That means that the amydalae actually "hear" the speaker's own voice (or lack of phonation) or are alerted to a proprioceptive or somato-sensory cue communicated by various subcortical brain circuits before the cortical brain can process the signal consciously. Because this response is pre-cortical, it occurs before the details are assembled in the sensory association cortexes and well before voluntary motoric responses are planned and initiated. It is the amygdala that registers trauma or pain and readies the organism's hormonal system to provide support to voluntary responses when its survival is threatened -- such as an incipient stuttering event that has become associated with loss of control, loss of status, communication failure, experience of frustration or fear, listener penalty or other negative consequences.

Survival and Threat Responses

The present model of stuttering development and behavior views the amygdala as a trigger of a survival response to a range of high stimulus speaking events associated with both adult and childhood conditioned learning. These events could range from incipient feelings of "loss of control" and "loss of status" due to core difficulties with fluency, up through actual experiences of trauma (helplessness, ridicule and humiliation) while speaking. 

It is not within the scope of this paper to discuss in detail the importance of maintaining fluent speech to the human child. But the pleasurable behaviors associated with speech and the external rewards received for speech performance (which include material rewards and high quality human contact) make it possible that speech is one of the several behaviors for which there is a dopaminergic reward system (some others being sexual behavior, eating, and drinking). The loss of this internal reward, coupled with the loss of corresponding external rewards and listener penalty, is theorized to provide sufficient motivation for the initiation of instinctive coping responses, including automatic and subconscious inhibitory responses. This would help explain how conscious attempts to maintain speech required for social success and adaptation can exist at the same time as inhibition responses associated with survival of the organism (or "proto-self") at or below the level of core consciousness (Damasio 1999), giving rise to the illusion that there is an "assertive - aggressive" conflict associated with stuttering. 

The particular inhibitory response described by this model is theorized to involve a human version of the "freeze" response seen in other mammals. Specifically, the response is theorized to initiate inhibitory actions designed to stop the person from speaking through disruption of the time and feedback dependent neural communications among the various distributed brain regions that are active during speech. Attempts to overcome this inhibitory reaction are thought to be reflected in the abnormal hyperactivity in the left cortical supplemental motor regions of stutterers and subcortical reactivity to auditory inputs may be reflected also in the hypoactivity in temporal auditory feedback area seen in recent PET scans of the brains of people who are stuttering (Wu 1995, Fox 1996). The direct precortical links to the amygdala from the ear via the thalamus is theorized to explain why auditory monitoring of speech aggravates chronic stuttering; why increased proprioceptive monitoring (involving the hippocampus in creating a new "explicit" memory of speech movement) is an effective therapy technique; and why delayed auditory feedback (which delays the subconscious and conscious perception of sound until after speech is initiated) is a fluency enhancing situation. The recently measured increased eye-blink startle response of stutterers relative to non-stutterers (Guitar 2003) could play a role, although this response is not specific to speech and may turn out to be insignificant.

The model predicts that speech sounds, postures and movements perceived by the amygdala to indicate tension or abnormality that indicates incipient speech difficulties that have become associated with a behavior that threatens a child will cause the amygdala to initiate a reactive "freeze" response that is mediated in other brain areas. Examples would be a hard "s" or "f" indicating tension in the articulatory muscles as well as the absence or delay of sound associated with a speech gesture that is expected to be accompanied by sound (such as the /e/ following the glottal /h/ in "hello.")  As these responses become conditioned, subtle precursors to actual speech gestures or sounds may be all that is required to illicit the response in some situations. The freeze response could result in inhibition of neural circuits involving the cranial nerves that project from the medulla to the muscles of the jaw and articulators, including the larynx. And it may result in the use of ascending neural pathways from the subcortical areas to transmit inhibitory commands to the cortical areas of the left temporal lobe and the premotor and motor areas where the intention to speak and language planning information are thought to be assembled. Alternatively, normal commands may be weakened by inhibition before or during transmission.

Because the circuits in which the amygdala participates also monitor tactile, somato-sensory, and proprioceptive feedback as well as auditory feedback for emotional significance, it is likely that the reactive response could be triggered by feedback that indicates an existing or incipient problem in the physical speech system, as well. For example, the fact that the vocal folds are adducted or abducted in a manner that would make initiation of speech difficult or impossible due to the inhibition of motor commands may be conditionally associated with speech difficulties, reinforcing the reactive inhibition. Thus, this model would explain the so-called "Valsalva hypothesis," which points out that negotiation of the Valsalva maneuver, involving the forceful adduction of the vocal folds and tension throughout the thorax in anticipation of difficulty with speech, is sometimes associated with stuttering (Parry, 1994) .

It would also be expected that the reactive inhibitory response would be projected to the cortical areas as well, as has been found with other sensori-motor responses involving classical and instrumental conditioning. At least, cortical perception of this involuntary inhibition would be perceived as a loss of control, a sensation familiar to all chronic stutterers and one that has also been classically conditioned over time. Because the learning of neurons in the amygdala is permanent, such reactions can be extinguished by the counter-action of inhibitory interneurons in the circuit. But the circuits are subject to spontaneous relearning in the presence of the appropriate stimulus (LeDoux 1989). 

This situation may be explained by the fact that "implicit" learning of the amygdala is accompanied by more situational ("explicit") and spatially cued learning of the hippocampus and associated rhinocortical areas, as well as by conditioned "implicit" learning in the basal ganglia (striatum) and the cerebellum. The presence of parallel mediation areas is typical of all brain functions and may help provide important clues as to how stuttering therapy works, and why stuttering and other behaviors involving conditioned responses are so difficult to completely erase.

Anatomical Implementation of the Response

A specific anatomical complexity of the physical model is the exact manner with which the amygdala and the involved circuits trigger the reactive threat response. Recent PET scans of people who stutter have shown consistently and abnormally low neuronal activity in the left hemisphere cortical areas mediating speech and language (primarily the auditory cortex), abnormally high activity in some of the prefrontal motor planning areas, and low activity in the the right cerebellum (Wu 1995). Other brain scan measurements have found somewhat conflicting activity patterns. It has been theorized by some analysts that left hemispheric hypoactivity reflects the presence of a congenital language weakness or a "neurological flaw "that makes it difficult to encode phonemes or other language elements for processing. Further, it has been theorized (Peters and Guitar 1991, Guitar 1998) that the lateralization of speech has been subverted for some reason by this weakness or flaw, causing the brain to adjust by employing less-well-adapted right cortical areas for speech functions. However, such mechanisms seem to be partly contradicted by renewed left cortical function during induced "choral" fluency, as well as the observed resolution of stuttering in some situations. Many theorists, in fact, explain the situational nature of disfluencies as being caused by a breakdown in weak cortical or subcortical function caused by "stress."  This model would provide an alternative explanation for that supposition.

The model of chronic stuttering presented here views left hemispheric cortical hypoactivity as a consequence of a reactive response triggered by the amygdala and the "Papez circuit," for the purpose of preventing speech during specially marked or significant situations in which speech difficulties are anticipated. Particularly important are memories of situations in which "original" childhood disfluencies or language difficulties elicited responses from others or created problematic or unpleasant behaviors that the child's threat response system interpreted as dangerous. External stimuli would include covert cues that may not be consciously perceived, such as facial expressions (monitored by specialized neurons in primate amygdalae) and other reactions, as well as overt criticism, ridicule, and other forms of listener penalty. These visual stimuli would then become associated with stuttering behaviors that exhibited loss of control, as well as the verbal, proprioceptive, and auditory cues of incipient stuttering, creating a rich variety of internal conditioned stimuli, which would differ with the individual.

There are a number of ways the reactive inhibitory response could be carried out, including 1) actual disabling of the cortical brain areas through some action, which could be intrinsic or remote, and 2) the hyperpolarization (resulting in inhibition) of neurons involved in associating speech and language areas through inhibitory action of the anterior cingulate cortex (ACC) or the firing of inhibitory interneurons in the medulla, the basal ganglia (specifically, the striatum) or other implicated subcortical areas, such as the brain stem and the cerebellum. With regard to the first possibility, it is unlikely that the action would be intrinsic to the cortex since the firing of inhibitory neurons there would be seen as increased neuronal activity in the PET scans. With regard to the second possibility, there is clear evidence associating ACC activity with inhibition of "inappropriate" speech (Paus 1993) and evidence of high ACC activity in recent brain scans of people who stutter (de Nil 2000) (see also News.) There is also a prior case study showing that trauma to the brain-stem can cause temporary spontaneous recovery from stuttering (Cooper 1983). And Wu and Maguire have noted the persistence of hypoactivity in the striatum of stutterers whose brains were imaged with PET scans in a variety of situations where stuttering was elicited (Wu,, 1995).

The hyperpolarization of dopamine, epinephrine or norepinephrine neurons could help explain the abnormal excess of these neurotransmitters in the brains of stutterers found by at least one published study (Rastatter 1988) and implicated by Wu and Maguire in a recent (unpublished) research report given at the 1996 ASHA convention.  Hyperpolarization results in higher than normal levels of neurotransmitter release when the neuron is finally coaxed to fire an action potential, which could temporarily overwhelm the reuptake process. A polygenetic disorder of the dopaminergic reuptake process has already been implicated by Comings (1996) in stuttering associated with Tourette syndrome.  A third possibility would be the firing of a relatively small number of inhibitory neurons in key sectors which would have an effect on a large number of cortical cells. Or perhaps the action of a specific inhibitory neurotransmitter (such as GABA.)

In fact, it is possible that a number of methods are used to implement the threat response stimulated by amygdaloid reactivity and that these accumulate and change over time. The different patterns of inhibition may explain the wide variations in stuttering behavior and severity within a class of behaviors that still shares a great number of similar conditioning features.

Other Protective Systems

A central feature of the model that has relevance to stuttering behavior is the before-mentioned amygdaloid projections to the central (periaqueductal) gray, or PAG. Stimulation of this area does more than initiate the "freeze" response. It also is known to stimulate the release of endogenous opiates (endorphins), which in turn excite the serotonergic neurons of the raphe nuclei. These project down the medulla and excite opiate-sensitive interneurons that block incoming pain signals in the dorsal horn (Pinel 1997). Thus, the same reactive response that disrupts the speech-motor system may also have an analgesic effect, possibly decreasing conscious awareness of the articulators through the loss of kinesthetic and proprioceptive monitoring.

Certainly, some of the most notable secondary symptoms of severe stuttering are the episodes of brief, coma-like unconsciousness exhibited by PWS when in the midst of a severe block. Dr. Van Riper called these "La Petit Morte" or "the little death." These appear to be designed to extricate or protect the PWS from the anticipated emotional pain of an inescapable situation, much like the response that occurs in animals who are being consumed alive or the unconsciousness that accompanies severe physical trauma. There is also some remote similarity to Petite Mal epileptic seizures, particularly the loss of attention and "rolling up" of the eyes into the head. The very fact that such a response would occur during the stuttering block indicates the behavior's considerable consequences to the PWS.

Yet another key involvement of the PAG is its participation in activating the vocal folds, particularly during automatic vocalizations. Denny and Smith (1997) implicate the periaqueductal gray (PAG) as the possible source of the signals that could disrupt or inhibit intrinsic laryngeal function. This speculation is backed by recent studies that implicate the PAG in laryngeal enervation in humans and in the emotional vocalizations of primates (Larson 1985). Such a mechanism is consistent with the underlying hypothesis presented here.

Finally, Alm (2004) has explained the freeze response in his helpful review as resulting from the co-activation of the sympathetic and parasympathetic nervous systems, which occurs when responding to certain types of threats. Further, he cites Blanchard and Blanchard (1989) as suggesting "that the freezing response is especially associated with situations involving potential or poorly understood threats, without clear information about the best way to act (page 126)." This would certainly describe the experience of a child confronting speech difficulties.

Sensori-Motor Implications

It has been noted by many that stuttering is much more than the resultant disfluencies. Smith and Kelly (1996) provide useful insights here, theorizing that the intense oscillatory events that occur during severe stuttering are evidence of a chaotic, nonlinear, multifactoral behavioral state.

The present model theorizes that a stuttering episode begins when a variety of automatic inhibitory processes are initiated, with the objective of preventing speech. It is highly speculative to proceed theoretically beyond this point, but one can develop a number of scenarios that could result from such inhibitions. For example, the consequences of such processes (as implicated in the present model) could include the disruption of communication between the cortical speech planning and auditory processing areas and automatic motor programs in the brain stem and the basal ganglia, which normally provide input to these cortical sensory and motor activities. Without the control provided by the neocortex and the feedback provided by the cerebellum (which communicates directly with the basal ganglia), the various "primitive" reflexes mediated in the bulbar area (medulla) and in the tensed articulatory muscles themselves could be left unchecked and uncoordinated. These could include so-called "primitive" reflexes normally inhibited during maturation, as well as monosynaptic stretch reflexes initiated by gamma neurons and polysynaptic golgi tendon organ reflexes that respond to relieve extrafusal muscle tension, and withdrawal reflexes involving the alpha motor neurons themselves. In the presence of abnormally high muscle tension noted in recent research (Starkweather 1995), such reflexes could have devastating consequences for speech.

This model would certainly be enhanced by the finding of subtle constitutional differences in the brains of people who stutter that would increase their vulnerability to speech inhibition. An example of such a characteristic may be the fractional anisotropy (FA) in white matter tracts that has been measured in both children (Chang et. al. 2008) and adults (Sommer 2002). These findings have recently been enriched by the Chinese team of Lu C., (2009) which has found evidence (to use the title of their paper as description): "altered effective connectivity and anomalous anatomy in the basal ganglia-thalamocortical circuit of stuttering speakers." Such weaknesses would help explain the predisposition to experience an inhibition of speech motor capability (which requires robust communications between the cortex and basal ganglia) in people whose neurological functioning otherwise appear to be normal.

Clinical Aspects of the Resulting Model

The resulting multi-phase model has several stages, beginning with the "original" disfluencies that result from "covert repairs" of developmental phonological or motor speech errors (which may be outgrown) and early experiences of control loss during speaking events, progressing to the experience of traumatic events by the person who stutters, through consolidation of conditioning in the teenage years and adulthood. The model includes several types of spontaneous recovery and several types of recovery as the result of therapy. These are described in the next section.

An interesting anecdotal speculation is made by Guitar (1997) in his assessment and treatment of the stuttering of two young boys, "Adam" and "Bob." Because of its relevance to the present model, the passage is extensively quoted here:

"Kagan and colleagues (1987) hypothesize that some children are born with a greater reactivity to uncomfortable and unfamiliar experiences. Such children respond to threatening stimuli with heightened activation of the limbic system -- a part of the brain that mediates emotions and behavior. This heightened activation causes many responses throughout the body, including increases in muscle tension, particularly in the larynx. Gray (1987) provided a neuroanatomical basis for Kagan's theoretical perspective by suggesting that limbic circuits may respond to fear and frustration with muscular contractions that produce tense and silent immobility ("freezing"), an innate response in most species that can be easily classically conditioned. Moreover, innate flight/avoidance responses may be elicited under threatening conditions and may alternate rapidly with "freezing" responses."

Guitar sees the primary or "core" stuttering response in some PWS to be a freeze response, and this is in superficial agreement with the present "reactive inhibition" model:

"It may not be too great a leap of imagination to reconceptualize the essential elements of chronic stuttering--core, escape, and avoidance behaviors--as manifestations of these three responses to fear and frustration: freezing, flight, and avoidance." (Guitar, 1997, p. 282)

It is important to emphasize that one temperamental trait or characteristic of this type is not being proposed as a cause of stuttering. Nor is it being proposed that all children with an inhibitive temperament are predisposed to stutter. Given the wide range of personalities and character traits associated with people who stutter, it is more likely that there are a number of very specific neurological characteristics involved and that the effect of these characteristics may overlap, as well as interacting with the person's environment, to predispose him or her to stuttering. 

[ Contents ] A Multi-phase, Reactive Inhibition Model of Developmental Stuttering

The model of developmental stuttering comprises five stages plus four types of so-called "spontaneous recovery" and two types of "recovery as a consequence of therapy." Each of these stages is based on case studies and formally and informally observed behavior in multiple subjects. While some of these stages are associated with specific ages, there is clear clinical evidence that the early stages could occur very rapidly in some situations, based on observed behavior. Of course it should be remembered that, while based on anecdotal facts, the theory underlying these stages is not proven (something true of all theories of stuttering etiology).

It is important to understand that "developmental" stuttering is not -- as some would claim -- a totally learned behavior. It is clear from our experience with stuttering that a genetic base or predisposition is probably involved in this process. Differences in the character of this genetic basis, as well as differences in environmental stimuli probably account for the wide range of differences we see in stuttering severity, tendency for recovery, persistence, and degree of handicap.

For purposes of orientation, the stages below are associated with those identified by Guitar in his widely used textbook (1998.)

Stages of Stuttering Development

Stage One - Original Disfluencies (Borderline Stuttering)

The child (typically aged 18 months to three years) speaks spontaneously, but with an excessive number of part-word repetitions, prolongations and (possibly) "blocking" behaviors (Yairi 1996a). The precise nature of the cause(s) underlying these "original" disfluencies is not known and may be extremely varied. The behavior may indicate an abnormal number of phonological errors due to a permanent genetic predisposition, a predisposition for late development, or delayed learning in the presence of environmental over- or under-stimulation. Repetitions and other errors may reflect an excessive number of covert corrections, which are normal processing behaviors for fluent and dysfluent speakers (Postma & Kolk 1997) or they could be associated with difficulties with expressive language, including attempts to mimic extremely complex utterances associated with over-stimulating language environments. This stage is characterized by the lack of obvious or overt reaction to the disfluencies; however that is not to suggest that no underlying reactivity is present. The whole word repetitions of many early stutterers seem to be associated with upcoming sounds on which the child subsequently has difficulty or experiences blocking ("I-I-I-I-I-I-I-Ieeee.... th . . . th . . . ink" etc.)

Stage Two - Early Speech Reactivity  (Beginning Stuttering)

During this stage, (which may be virtually concurrent with Stage One) the child begins to react to spontaneous covert error corrections or feelings of being blocked with frustration, showing evidence of conscious attempts to control articulators (tension and forcing) and voluntary or involuntary interruptions of spontaneous speech. These difficulties may be exacerbated by temperamental differences in children who begin to react to original speech difficulties (found by Anderson and others (2003) to be "slower to adapt to changes in the environment, less distractible, and more irregular in biological functions" than fluent children). These children may respond more violently to original disfluencies by attempting to push through them with tension or force, resulting in an increased number of repetitions and prolongations, together with increased silent "blocking" episodes in some cases. The later phases of this stage may include conscious reactive behaviors such as "giving up" in response to the inability to speak, observed practicing of speech before utterance, or the child attempting to "pull words out" of his mouth with his hands, indicating an underlying strategy of pushing or forcing. This stage is characterized by mild reactions such as frustration or annoyance. Lack of control may be generating sufficient reactivity in the child at this stage to initiate relatively mild subcortical reactions. Strong evidence for the notion that these reactions are learned is provided by the importance of the length of time the child has been stuttering prior to intervention shown in a recent longitudinal study (Yairi 1996b). This study found that the length of time between onset and intervention was the primary statistically significant difference between "recovered" and "persistent" child stutterers in the sample.

Stage Three - Consolidation of Reactions in Maladaptive Speech Habits (Intermediate Stuttering - I)

As dysfluencies continue, there is a consolidation of the behaviors initiated in Stage Two and the appearance of reactive anger and fear in response to difficulties in speaking. With continued reactions to conditioned stimuli, the centers of reactivity are distributed more widely, and the amygdala's role becomes less important (Dolan 2000.) Repetitions, prolongations, and blocks may become more severe in intensity and duration and forcing of articulators becomes reinforced by intermittent success at avoiding loss of jaw and mouth control. Accessory behaviors become more elaborate and ritualized, including the increased use of avoidance and postponement. The child begins to show speech "rehearsal" behaviors and may show some personality modification (including a tendency to avoid other children at times), that may not have been present before. Heightened auditory monitoring of his speech for cues that predict errors may result in fear conditioning the child to react to initial sounds (for example, a hard "s" associated with an incipient stuttering response.) Conditioned proprioceptive cues may include delayed adduction of the vocal folds, as in the silent "h." These reactions occur in half the time of the cortical inputs to the amygdala, making it difficult for the child to control them. As a result, he attempts to predict the reactions, initiating a process which will later cause more problems. Severe oscillatory blocks may begin during this stage. These blocks could be triggered by the heightened tension in the muscles and their increased ballistic potential, causing tonic tremors. Several reflexes could be involved in these severe oscillatory blocks, including polysynaptic Golgi tendon organ reflexes (to release excessive tension in the articulatory muscles) and monosynaptic stretch reflexes (which govern inappropriate tension or laxness in the interfusal muscles). In the presence of extreme tension (Starkweather, 1995), the impression of speech "difficulty" and the possible absence of cortical or cerebellar control, these reflexes may interact in the presence of a non-linear, chaos state (Smith and Kelly 1996). Reactions to speech are becoming fully integrated in the child's behavior and may be dominating his self-image.

Stage Four - Traumatic Sensitization (Intermediate Stuttering - II)

Some children may be subject to one or more significant events involving public humiliation, ridicule, and "loss of status" during speaking performances characterized by helplessness or loss of control.  An increasing number of escape behaviors begin to enter the child's repertoire of reactions to stuttering. Strong hippocampal memories that associate stuttering with specific situations and listener reactions are reinforced. Speech at these times is seen and felt by the child to become a matter of survival and he fights strong freeze reactions that may make his mind go completely blank as he is speaking or planning to speak, accompanied by experiences of loss of control. At this point, strong conditioned learning responses begin to dominate the child's speech behavior. If in school, the child's grades may begin to suffer as he becomes preoccupied with strong negative self-images that seem to be verified by every unsuccessful speaking episode and reinforced by listener penalty or ridicule from peers. More extreme behaviors such as fighting and total withdrawal may begin to be seen. Oddly enough, the child's overt stuttering may decrease during this period as he tries every method at his disposal to avoid stuttering in front of his peers, significant others, and his greatest critic: his own developing superego. The stage may be characteristic of later childhood (ages 9 to 11), when socialization and peer approval are the most important influences on the child's development.

Stage Five - Consolidation of Reactive, Chronic Stuttering Behavior (Advanced Stuttering)

Stuttering behavior is further consolidated in the teen years and early adulthood. Continuing speech failures and loss of control may make speaking seem like a difficult or impossible activity at times. Because the feelings of helplessness and loss of speech control are continuing, the reactions are strongly reinforced and begin to become an integral part of the person's behavior. The person's neurological system begins to anesthetize itself from the emotional pain and unpleasantness of stuttering by internally administering opoid neurotransmitters (the endorphines). As a result, periods of lack of consciousness may be seen during speech blocks and in the flow of speech, which begins to develop a characteristic choppiness as the PWS snips out awareness of large and small stutters, rushes away from stutters and slows down as new difficulties approach. Severe blocks may be accompanied by actual periods of unconsciousness, with the eyes of the PWS "rolling up" into the head. In many, there is a perceived "loss of status" or "bottom dog" self-image that can result in anger, open displays of aggression, and depression. This can also result in intense self-criticism and excessive or compulsive striving to achieve in other areas to compensate for speech difficulties.

Readers wishing a recent review of theories focusing on the persistent reactivity of stuttering in adults are referred to an excellent recent book chapter by David Prins (1996).

[ Contents ]

Types of "Spontaneous" Recovery

Type One - Recovery in the Pre-School Years

The most common recovery occurs in the years before the "five-to-seven shift." Clinic research and experience indicates that up to 75% of children who demonstrate stuttering behaviors may simply stop having dysfluencies. These can include children whose dysfluencies are accompanied by behaviors similar to advanced adolescent stuttering. It is not known why this happens, but possibilities include: 1) naturally helpful intervention by parents, 2) discovery of an adaptive speech technique that facilitates fluency; 3) resolution of a temporary developmental imbalance of some kind. Treatment of pre-school stuttering involves facilitation of possibilities 1 and 2, possibly (but not necessarily) negating option 3.

Type Two - Recovery in Young Adulthood

In late teens to mid twenties, the person may gradually cease to stutter following a period of personal growth and increased self-efficacy. This has been explained by subjects as a sudden ability to control the onset of stuttering, perhaps indicating a greater predisposition to cortical control over reactivity. Sporadic successes may empower the person to use a "new voice" or modified manner of speaking in situations normally associated with stuttering, generalizing the extinguishment of the reactive inhibitory response, similar to the "transfer" stage of formal stuttering therapy. This scenario is based on personal accounts, which often inaccurately attribute the recovery solely to cognitive factors.  

Type Three - Recovery Associated with Competing Survival Threats

The subject suddenly stops stuttering following or during a significant survival-threatening experience such as terminal cancer or a significant injury. In this case, the significance of stuttering as a survival threat is overwhelmed by the primary threat and complex neuronal communications ("voting") may be changed sufficiently to reduce the disabling of speech  production areas. If hormonal activity is a precursor for stuttering that sets up a subconscious state of watchfulness that precipitates stuttering, changes in such activity might also have an effect. Stuttering gradually returns when the greater survival threat is no longer present and the brain chemistry state associated with speech inhibition is restored.

Type Four - Recovery Associated with Physical Trauma

Sudden cessation of stuttering following physical trauma to the brain such as an auto accident or stroke, documented for brain-stem trauma by Cooper (1983). Stuttering may return as the lesions heal or as swelling due to fluid build-up is reduced. One case of undocumented recovery due to stroke was apparently instantaneous and long-lasting, perhaps indicating damage to the area containing neurons critical to the stuttering response.

Type Five - Gradual Recovery with Aging

Some PWS have noted that the severity of stuttering begins to diminish in their 40's and 50's and that stuttering therapy seems to become more effective. This phenomenon would seem to be consistent with a gradual recovery associated with release from inhibition, rather than recovery due to repair. Loss of neuronal tissue would not be expected to result in improved brain function if cortical processing were implicated in stuttering. However, increased cortical inhibition of the amygdaloid response, extinguishing of conditioned responses over time, or the loss of amygdaloid cells or inhibitory cells which have assumed the long-term specialized functions of inhibiting speech, would seem more consistent with this effect. The observed easing of reactivity associated with anger and other responses with age provides a paradigm for this behavior change. In addition, since few people who stutter have not experienced various types of therapy at various times during their lives, it is possible that the accumulative effects of this therapy may interact with other effects to improve fluency.

[ Contents ]

Recovery as a Consequence of Therapy

Therapy approaches such as fluency shaping have been shown to be highly effective for childhood stuttering, but are apparently less effective with increasing age. This suggests a difference between childhood and adult stuttering, but does not rule out the notion that there are some stutterers for whom it would never be lastingly effective. Very few people who continue to stutter into adulthood ever completely overcome all of the interior phenomena and external manifestations of stuttering. Nevertheless, there are many adult PWS who can be said to be "recovered," in the sense that they are substantially free of the severe behavioral effects of stuttering, including a large number of disfluencies, and the conditioning and reactivity has largely abated.

Type One - Fluency Shaping

Following enrollment in one or more intensive fluency shaping or "air flow" programs, the PWS is overtrained for fluency control and manages to maintain it following a typical relapse period of three to four months, beginning several months after therapy. In this case, the physical reactions are commonly explained to the self as an underlying tendency to stutter. For example, one author who has been essentially fluent for over four years still says that "I have often dreamed of waking up fluent, but I know it is not going to happen" (Tunbridge 1994, p. 128). Such speakers seem to maintain their fluency by reducing the significance of situational stuttering states, by continually demonstrating inhibition of stuttering; and by training themselves to extinguish or ignore the specific auditory and tactile cues that caused the original conditioned fear response. The PWS speaks with a slow rate and characteristic monotone inflection under stress, indicating use of fluency strategies to avoid activating the stuttering response. The typical observation of the PWS is of someone who is being very careful. Long pauses may occur, during which the person is apparently searching for the proper way to initiate the next sound and waiting for an internal fluency cue. The PWS must practice speech targets regularly or face the gradual re-emergence of stuttering, including the dreaded feeling of the physical reaction. In time, some of these individuals can become virtually fluent; however (as noted above) this scenario is more likely for younger PWS, in whom the reactive inhibitory response has not been overlearned.

Type Two - Stuttering Modification

The person who "recovers" using the stuttering modification approach will have success in producing fluent speech in spite of the continued presence of the subcortical reactions. These reactions will continue to be felt, but their severity will gradually be reduced as the associations tied to stuttering situations and the stuttering cues are modified. The specific mechanism of extinguishment is not known, of course, but it is likely that it could involve habituation or demonstration of lack of consequence due to stuttering. Prior stuttering was rewarded because it allowed the PWS to move past the reaction. The act of openly stuttering or using techniques such as pullouts and cancellations -- while they may not result in behavior that looks appreciably different to an observer -- completely stands the reward system on its head. The organism has been attempting to shut down speech, but has been thwarted by the compulsion of the person to speak in spite of this warning. In this case it will try again and again to achieve the desired behavior because the behavior is being intermittently reinforced. In other words, because of the reaction, the speaker is afraid of speaking and is reacting to that fear by feeling that he will have difficulty speaking. As long as this is the case, the attempt is being rewarded and will be continued by the attempt to reject or avoid the stuttering behavior. But open and voluntary stuttering or disfluency actually punishes the reactive response and the attempt to extinguish speech because the feared behavior is being actively produced.

Despite the high degree of fluency such speakers can attain, many still report the presence of physical reactions, which appear in many or most cases to be indelible, if greatly reduced in severity. Severe recovered stutterers may live like Charles Van Riper, with continuing experiences of reaction. Milder stutterers may experience the reactivity only infrequently and in specific situations. There must be continued vigilance in this case also, to keep from using avoidance to push the reactions into the background, causing them to once more become "hot."

[ Contents ] Challenges to and Tests of the Hypothesis

Stuttering has a number of paradoxical characteristics which have been seen as tests for neuromuscular or organic theories of stuttering behavior. Most of these involve the lack of consistency and constancy of the behavior. It is argued that if stuttering is organic, it would be present in equal measure all the time. This would seem logical, but is not really an adequate test since many other disorders known or accepted to be organic are variable for one reason or another, including Tourette syndrome, ADHD, epilepsy and so-called developmental apraxia of speech. Nonetheless, these paradoxes make a good foil for testing any theory of stuttering, because they do describe the disorder.

1. Situational Nature of Stuttering

The model was partly created to explain this paradox, which actually holds the key to understanding the physical basis for stuttering. Explicit hippocampal memories of unsuccessful speech and trauma situations become associated with the feeling of dysfluency and create a brain state that involves a high degree of watchfulness and expectation of failure. In this state, the implicit amygdala reactions are more likely to occur in response to internal, auditory and visual cues. It is interesting that one of the few drugs known to have a slight beneficial influence on stuttering (primarily by reducing secondary behaviors) is Haloperidol, a primary action of which is to reduce "watchful" behavior that describes the role of the amygdalae.

2. Resolution of Stuttering During Choral Reading or Speaking

Auditory input is overwhelmed by the sound of other speakers (like DAF), dampening the implicit reactive responses. Past experiences of success in this situation increase overall confidence and overwhelm explicit memories. Left cortical areas are shown to be aroused in PET scans during choral fluency, exhibiting activity very similar to fluent speakers. In addition, choral reading essentially parallels the experience of group singing, which is a fluent situation for most people who stutter.

The recent observation of Kalinowski that watching another person mouth the words to be spoken is also fluency enhancing raises the interesting possibility that external auditory or visual cues are providing a replacement signal or alternative pathway for speech-motor plans that are disrupted by the reactive inhibition and that are required to complete a speech-gesture circuit that activates speech-motor program sequences.

Mirror neurons thought to be found in area BA44 of the human brain are currently being studied to explain this phenomenon, which may engage the external loop of a hypothesized "dual pre-motor system" that by-passes an internal loop affected by the stuttering inhibition response.

3. Reduction of Stuttering with Shock Treatment

"Any theory involving the neuromuscular system . . . must explain the effects of shock on stuttering that, theoretically, should increase stuttering, not reduce it." (D. Mowrer 1996). The reduction of stuttering due to punishment by electrical shock actually provides a verification of the model of stuttering provided here. The survival action of the amygdala is in response to an expected penalty for loss of status or loss of control, which is a result of stuttering. It is not the stuttering that is being responded to, rather, the fear of rejection or loss. The punishment of shock is very different and would not be expected to elicit a similar response. In addition, it must be remembered that the organism is "taking care" of the loss of control not by stuttering, but by causing an inhibition of speech. The stuttering behavior is a consequence of attempting to override this inhibition. This is not the same as physical punishment for actual stuttering, which is a tangible consequence of the attempt to override. Stuttering is decreased because the person perceives stuttering as a more immediate threat, overriding the risk of losing status. Stuttering has its own secondary reward system as well, which is overridden by the immediate threat. For example, stuttering blocks and secondary behaviors are actually reinforced when the person is able to move past the feared loss of control. If the stuttering--however obvious it appears to others--is still less than the feared activity (total loss of control or actual inability to speak) it will still be positively reinforced. This situation can also be explained by the model in much the same way as the recoveries of PWS subject to trauma or life-threatening illnesses. The survival system is distracted by the more immediate threat of the shock. There is also a possibility that -- faced with the possibility of shock -- the person will use more avoidance behaviors (word substitutions, etc.) and this would decrease the number of apparent disfluencies.

4. "Spontaneous" Recovery from Stuttering

Discussed under "Types of Spontaneous Recovery" above.

5. Extreme Variability of the Behavior

There are a number of elements of the model which allow for extreme variability in behavior. The first is the original disfluency, speech difficulties or excessive covert repairs, which may be varied by the child's environment and by the specific number and nature of phonological errors. The second is the development of these core behaviors, leading to the first loss of control or traumatic experiences and the specific types of disfluencies involved. For example, the types of speech blocks may vary from silent blocks to oscillatory repetitions, which has an impact on the cues learned by the amygdala. The third is the specific action of the amygdaloid circuits in disabling of cortical and cerebellar areas or in the possible direct affect of musculature through inhibition of neurons in the periaqueductal gray and the medulla. This would also differ depending upon the modality involved during traumatic or threatening speech events. This would partly explain, for example, why there are some PWS who stutter more when reading and some who have more difficulty during spontaneous speech. A fourth is the degree of cortical control over the amygdalae, as determined by their number, specific route, and condition.

6. Resolution of Stuttering in a Delayed Auditory Feedback (DAF) Condition

It is invariably the case that PWS who speak under conditions of electronic delayed auditory feedback (DAF) experience a resolution of stuttering. For some, it is nearly impossible to feel that one will stutter under these conditions. Experimental investigators have also observed that the increased fluency of PWS is reported by subjects to be accompanied by an attendant feeling of "invulnerability" to stuttering (Kalinowski, 1996).

This reported "feeling of invulnerability" has led some analysts to assume that stuttering is caused by a timing delay or deficit, which is relieved by the slow speech enforced by DAF. However, no current theory explaining the DAF effect takes into account the advanced warning of incipient speech production or possible stuttering provided by the amygdala. The reactive stuttering mechanism discussed in this paper would occur just prior to conscious awareness of the sound of one's own speech due to the faster auditory pathway from the ear to the thalamus to the amygdala. That would mean that the PWS speaking into a DAF machine would have no pre-cortical or cortical perception of the sound of his own voice until a few microseconds after the actual onset of phonation. In the case of voiced vowels following unvoiced fricatives (one of the common stuttering difficulties) the PWS probably would not hear his own voice sub-cortically or cortically until the actual voicing was initiated. In both cases, there would be no amygdaloid reaction to the sound, delaying any chance of reactive inhibition until after the start of phonation. (Brain imaging studies of PWS speaking under DAF conditions are required to determine if hypoactivity of the left cortical areas and right cerebellar region is resolved.)

As mentioned above, the recent discovery by LeDoux's research team of pre-cortical auditory circuits that result in perception of sound by the amygdala before sound is perceived by the cortical areas has not (to my knowledge) been incorporated in DAF theory. Until this is accomplished, all existing DAF theory relative to stuttering is suspect and needs to be reconceptualized.

The failure to find clear evidence of abnormal function or a meaningful auditory processing delay in PWS (Postma & Kolk, 1992), has led investigators to search for other reasons why some of the fluency-inducing methods work. For example, Wingate theorized that the increased fluency of PWS under conditions of delayed auditory feedback was created by the slowing of speech rate caused by DAF (Wingate, 1970). This conclusion has been assumed to be correct for many years.

However, recent experimental results seem to verify the explanation of DAF fluency set forth here. Kalinowski’s team at East Carolina University reopened the question by showing that significant fluency enhancement occurred under DAF at both normal and fast speech rates (Kalinowski, 1996). The finding that delayed auditory feedback may indeed ameliorate stuttering frequency because of the delayed audition--not because of slowed speech rate—has even been touted as precipitating "a crisis in the study of stuttering" (Stuart, 1996).

Interest in further study of the relationship of audition to stuttering appears justified in light of the DAF results and recent positron emission tomography (PET) brain scans, which show significant changes in brain areas associated with audition during stuttering and induced fluency (Fox, 1996).

7. Lack of Electrical Brain Activity Indicating an Inhibitive Response or Reaction

Until recently, there was no evidence (from electroencephalographic or EEG measurements) of unusual brain activity that would indicate the presence of a reactive response in people who stutter. However, this was apparently due to the fact that previous EEG measurements by Watson and Freeman (1997) had been made when the subjects were "at rest" (in other words, not talking.)

Why would any investigator make measurements in the absence of the specific target behavior of interest? The answer is contained in Assumption One (discussed in the Preface, above). Because it has been assumed that stuttering was associated with some kind of deficit, most measurements of the physical and neurological characteristics of people who stutter have been made when they are not stuttering, in the hope that the underlying deficit could be isolated. Because the hypothesis of these studies was that PWS were neurologically different than other people, the presence of actual stuttering during measurements was even seen as a possible "confound," invalidating the results. (This "tyranny of the hypothesis" is one of the pitfalls of experimental research designs.)

The most recent relevant research bearing on this issue concerns the comparison of startle response of stuttering and non-stuttering adults (Guitar 2003). This research, which compared 14 individuals from each group, found that the initial and repeated startle responses of people who stutter were significantly higher than subjects who did not stutter.

In addition, Rastatter and Kalinowski (1998) have found that during speech, the EEG patterns of people who stutter are characterized by hyperactivity, at least in the beta 1 and 2 frequencies, which would seem to contradict the findings of the PET scans that the cortical region is hypoactive in the speech, language, and auditory (temporal) regions during stuttered speech.

"For all stuttering participants, results showed hyper-reactivity in both beta 1 and beta 2 bands over both the left and right temporal-parietal lobe regions under the NAF speaking condition. The epochs used to derive the EEG activity were sampled from a stuttering segment for the NAF [non-altered auditory feedback] condition."

As discussed in this paper, if this measured hyperactivity was associated with an attempt to inhibit speech, it would resolve or be absent under conditions where the speaker's voice was not heard under after speech was initiated, or if it was disguised (through frequency alteration) so that the highly specialized conditioned neurons in the amygdala weren't able to recognize the voice as the speaker's own. That is exactly what they found:

"Under conditions of DAF [delayed auditory feedback] and FAF [frequency-altered auditory feedback], at least two phenomena were observed. For some of the stuttering subjects, a strong reduction in beta 1 and 2 activity occurred in the right hemisphere temporal-parietal leads, accompanied by marked reductions in stuttering behavior for both the DAF and FAF conditions. For other subjects, a strong decrease in beta 1 and 2 activity was obtained for all electrode sites, with the right hemisphere posterior sites evidencing a larger area of under-reactivity. Again, the DAF and FAF conditions produced parallel results. Such findings indicate . . . an alteration in the electrical fields of the cortex occurred in the stuttering subjects under both DAF and FAF speaking conditions . . . "

Then, however, the researcher's assumptions begin to control the interpretation of these findings, even to the point that the problems defined must be linguistically based:

"Such findings suggest that stuttering may reflect a sensory-linguistic, motor-integration anomaly where each hemisphere generates competing linguistic messages at hyper-reactive amplitude levels. Conditions of AAF [altered auditory feedback] appear to result in inhibition of activity through out the cortex and specifically within the right temporal-parietal areas, resulting in fluent speech production in stuttering subjects" (Rastatter 1998).

An interpretation that is just as plausible is that the measured activity during speech is an inhibitory activity, which just ceases (rather than being inhibited) when the reactive inhibition response is not triggered by the amygdaloid early warning system. To be sure, there is an interpretation of the researcher's theory that is in line with that in the present paper. This is that the "competing linguistic messages" are simply a nonverbal communication of the concept of DANGER. This type of input would be more in keeping with the nonverbal "language" used by the brain's threat response system.

The complexity of the Rastatter and Kalinowski model is typical of the theories set forth for stuttering, as is the manner in which new processes (in this case, the "competing linguistic input") are commonly developed for the functional explanations. This complexity would be unnecessary if a very basic neural message (a nonverbal command to inhibit speech) is behind the process.

8. Lack of High Activity in the Amygdala in Brain Scans Using Positron Emission Tomography (PET Scans)

This investigator is not aware of a brain scan study of stuttering that has purposefully targeted the amygdala or its circuits. Several researchers have informed me personally that they have not seen notable amygdala activity in passing during PET scans partially encompassing that area. The researchers were properly circumspect, citing their lack of specialized knowledge about the amygdala. However, some commentators have falsely assumed that this means "case closed" for the amygdala's involvement in stuttering.

The recent rise in the use of brain scans has led, unfortunately, to a tendency to use a "pinball machine" metaphor to describe brain activity by some commentators.  In other words, if there is abnormally high or low brain activity in one area during a certain behavior, that area "must" be the "site" of the difficulty. Conversely, it is thought, a lack of high activity seen in PET scans means that the site is not involved in the behavior at all. PET scans are particularly good at showing activity (oxygen uptake) in large areas of the brain, particularly the cortical regions. They are not particularly good at showing activity in circuits that may be activating these areas and discriminating particular types of activity in structures (such as the baso-lateral amygdala) that may have a continuously high level of activity in ongoing brain functions.

However, it would be a very unusual situation for the amygdaloid circuits to not be involved in stuttering, when they are absolutely essential to human conditioning. One would have to prove that there is no conditioning involved in stuttering at all, another totally impossible situation since conditioning is universal and underlies all human behavior.

PET scans of the brains of humans processing fear-eliciting stimuli have indicated high amygdala activity (Dolan, 2000). However, these studies have used stimuli that present novel situations to the subject. Researchers have found that once conditioned learning has taken place, amygdaloid activity in response to subsequent similar conditioned stimuli (CS) is "time-limited" (Dolan 2000, p 634.) The implication here is that the response has been substantially delegated by the amygdala to other brain structures. The dangerous lack of threat awareness that accompanies loss of, or damage to, the amygdalae indicates that there is probably some remaining key role, which might involve a small number of neurons -- perhaps too small to be picked up by PET technology. In other words, the longer a person has been stuttering (40 years in the case of some PET scan subjects), the less likely it is that the amygdala will need to be activated in a major way to implement the reactive inhibition it originally initiated.

The task is, therefore, to identify the brain areas that are implementing the inhibition of speech in response to memories conditioned by the amygdala. This requires not just PET scans, but invasive procedures that are not ethical for use with human subjects. Results of research using other primates is available, and investigation of its applicability to stuttering has been started by Juergens, as reviewed in the summary of the State of the Science Conference: Developmental Stuttering (Juergens 2005). This research is more likely to show important differences between human and lower primate vocalizations.

Juergens noted that monkeys do not stutter; however, he makes an important distinction between human and lower primate speech:

"... a bilateral lesion in the inferior motor cortex completely eliminates the capacity to speak in humans but has no effect on vocal communication in monkeys. This discrepancy can be explained by the fact that speech consists of learned vocal patterns, while monkey calls represent genetically programmed vocal patterns. The motor cortex is needed only for the production of learned, but not that of innate, motor patterns." (Juergens 2005)

Jeurgens also states that:

"... the strong influence of the emotional state on stuttering indicates that the limbic system (e.g., anterior cingulate cortex, amygdala, hypothalamus, and PAG) is also involved in stuttering control." (Juergens 2005)

Given the increasing amount of evidence for the amygdala's universal involvement in mediating conditioning, and the overwhelming clinical evidence that conditioning plays a primary role in developing and developed stuttering behavior, it is actually not up to the amygdala to prove it is involved in stuttering ...  Rather, it is up to the research community to explain why their measurement technology is not capable of sensing the role of the amygdaloid circuits in the behavior. Instead, we have a situation where a recent major thesis on stuttering which has garnered substantial interest (Alm 2005) provides an intriguing review of one possible underlying characteristic of some children who stutter (an imbalance of D1 and D2 dopamine neurons in the basal ganglia), but doesn't mention conditioning at all, and this major weakness has not even been noticed by most readers.

9. Lack of Clear Advantage of Stuttering to the Organism

An argument could be made that it just doesn't make sense that the human organism would harm itself by implementing a damaging behavior like stuttering. Despite the evidence that other such self-inflicted maladaptive behaviors are unfortunately common in humans, this deserves an interpretation here because it explains something that lies at the heart of stuttering behavior: some observed primary behaviors and all secondary stuttering behaviors are actually the result of trying to avoid or control stuttering. Thus, they are an indirect consequence of the internal reactivity. This creates a disjunction between the subcortical reaction and the cortical response to the initial reaction. It should be remembered that the initial subcortical reaction was itself actually a consequence of initial conscious attempts (reactions) to control "original" dysfluencies. It is no wonder that stuttering seems so impenetrable and mysterious to those caught in this Chinese puzzle box.

The disjunction described above would explain the continued survival reaction of the amygdala. This reaction is primitive in nature and is -- like other subcortical systems -- passed on from other mammals that lack a high degree of consciousness or voluntary control over their actions. It may remain in humans because of a biological advantage or adaptation. For example, protecting humans by stifling speech in the presence of danger, predators, or human rivals. Or even serving as a subtle, unconscious warning for hunters that would assist them in remaining quiet while they were stalking prey. Such reactions are not subject to logic. If they were responsive to the conscious brain's logic, humans would have probably disappeared long ago. Basic survival mechanisms -- like the monitoring of gas levels in the blood during sleep -- have been delegated to subcortical processes, leaving the cortical areas free to reason and initiate actions (and to create wonderful confabulations of what's going on) without distractions.

When the amygdaloid circuits react by inhibiting speech in the child whose original dysfluencies are not matching his intensions or are eliciting unexpected reactions from adults and peers, and who has one or more genetically influenced characteristics that predispose her to such inhibitions, a long battle is enjoined between instinct and consciousness. The circuits containing the amygdala probably do not respond to language or to the subtleties of its logic. They can "read," through specialized neurons, the reactions in the facial expressions of other humans to unexpected things the child is doing (LeDoux 1996). They can "read" the frustration of the child when she make speech mistakes that are unexpected. They can "read" abnormal articulator tension or abrupt or strained speech sounds and voicing that are not associated with fluency -- and respond accordingly. The circuits can also "read" the conscious mind's interpretation of rejection and ridicule and the rejection of self that can result. They may or may not be able to read the conscious mind's association of dysfluencies with these factors, but they certainly are capable of making their own associations. All they "know" is that speech is causing the problem and they go to their own particular bag of tricks to solve it, implementing speech-motor disruptions that are well below the child's conscious awareness. But the child's conscious self has its own agenda here. The child needs to speak and use language to flourish in the social realm. Because of the demands of life and the rich rewards associated with communication, the child will attempt to override the reactive inhibitions in whatever way he can manage. When the disruption of the brain's language and speech-motor centers make speech more and more difficult, he will persist in speech attempts, using the range of compensating behaviors to enable the communication process to continue. As the developing person acquires more and more compensatory stuttering behaviors that in turn need to be overcome and covered up, much of his or her life can become taken up by the attempt ... or by minimizing the importance of this situation.

That is the answer to this most puzzling of questions about stuttering: the human child must communicate to live. Whatever he has to do to continue communicating -- in whatever form he can manage -- is therefore worth it.

[ Contents ] Implications for Therapy Approaches

This model explains the efficacy and inconsistency of the large variety of therapeutic approaches for treating stuttering, including the recent encouraging successes with early intervention and treatment of very young children (2 1/2 to 3 years) using "fluency shaping" approaches. It also indicates that many modern treatment approaches (while successful in improving fluency) may not succeed in quite the manner claimed for them by their advocates.

For example, fluency shaping, voice shaping treatment approaches (including "air flow"), or so-called "speech re-education" approaches, are claimed to work simply by replacing a dysfluent speaking behavior with a new speaking behavior devoid of "bad habits" that characterize stuttering. In fact, these methods are most likely extinguishing classical conditioning by 1) eliminating internal and auditory cues of incipient stuttering and 2) punishing the amygdaloid reaction by demonstrating fluency and lack of fear due to over-learning fluency, stimulating cortical areas capable of inhibiting the learned reactivity. This is accomplished by having the client repetitively produce the highly exaggerated and heavily self-monitored speech gestures that Charles Van Riper called for (Van Riper 1973) to stimulate the proprioceptive learning function of the hippocampus and cerebellum, creating a new set of learned cues for fluent speech.

It is commonly observed that many severe stutterers do not "succeed" at fluency shaping, at least in terms of achieving the degree of control that these programs require for "success."  This has often been attributed to "lack of motivation" or "lack of readiness to change" on the part of the client. While these judgments may indeed be correct in many cases, it is highly likely -- based on this model -- that there are also many PWS for whom total extinguishing of subcortical reactivity is not really possible due to the strength of its conditioning or specific physical differences in the number and quality of neural projections from the cortex to the amygdaloid circuits. LeDoux has recently cited research indicating that rats with completely extinguished learning instantaneously relearned their old behaviors when placed in another cage (LeDoux 1996). In PWS who had severe experiences of childhood trauma during stuttering episodes, whose stuttering was not treated (or was not treated until later in childhood or in adulthood), or who lack the ability to focus on "fluency targets" and ignore habitual subcortical fear reactions, several months, or even years of fluency shaping (or, for that matter, stuttering modification) may simply not be sufficient to fully extinguish the stuttering reaction. This is made more difficult by the fact that only one or two failures of the fluency shaping techniques to produce the expected fluency may be sufficient to undo the entire fluency program

Likewise, cognitive therapy or psychological approaches are often said to work by correcting false thoughts about one's ability to speak openly in public or by showing the PWS that stuttering is caused by attempts to control speech, which interfere with the normal automatic nature of fluent speech. That dynamic is similar to the child's attempts to control core dysfluencies in the earlier phases of this model. A common feature of cognitive approaches is the notion that conscious control of speech is the problem and that the PWS simply needs to "let go" and speak in a more automatic manner. However, in the context of the present model, such directions do not account for strong amygdala reactions, and failure to include emotion is often a limitation of cognitive approaches. In fact, one of the pitfalls for stutterers undergoing therapy is so-called "spontaneous fluency," during which the PWS is able to produce fluent speech due to the relaxation of the stuttering reactions, but without using behaviors used to induce operant fluency (which typically include slow rate, conscious monitoring of proprioception and tactile feedback, and continuos voicing.) In the absence of these behaviors, the PWS will eventually experience a temporary relapse. In fact, a sober appraisal will show that all cognitive therapy approaches invariably induce operant fluency in the clinic and in a group situation, and then attempt to generalize it to various other situations, either during therapy sessions or by the client in life. If the operant fluency simply involves a distraction of some kind it will not work for long--even for those with moderate reactivity.

The more successful cognitive therapists encourage PWS to use their operant fluency in a number of situations, varying from "easy" to "difficult," practicing positive replacement thoughts and self images. This is really a behavioral "extinction" technique in which speech drill and articulation practice occurs and is probably a key (but unrecognized) component if the therapy is successful. Cognitive therapy, even if it includes coincidental speech drill, is rarely successful in permanently increasing the fluency of severe reactive stutterers.

Stuttering modification therapy -- developed by Charles Van Riper -- is a direct application of operant fluency or modified speech to extinguish conditioned learning (Van Riper 1973, Stuttering Foundation of America 1996). The course of this therapy, encompassing the behavioral steps of Identification, Desensitization, Modification, and Stabilization, is specifically targeted at a model of stuttering similar to that proposed here, if lacking the anatomical details. Because it provides the PWS with specific techniques for modifying dysfluencies, stuttering modification may be the therapy of choice for severe cases where cortical control is not able to extinguish the stuttering reaction. In fact, there is clear evidence that most successful fluency shaping therapy actually involves a form of stuttering modification. Once of the more popular recent books on stuttering describes the difficulty of using "slow speech" and fluency shaping targets in public (Tunbridge 1994). The author is only one of many who points out the necessity of the PWS desensitizing himself to using this speech, which is actual the same process used in stuttering modification therapy.

This model would, therefore, indicate that a combination of fluency shaping and stuttering modification therapy is indicated to treat most moderate to severe stuttering behavior, except for very small children in whom the reactive subcortical response is not overly conditioned. This treatment advice (if not this particular theory supporting it) is the emerging consensus among legitimate practitioners of stuttering therapy.

[ Contents ] References

Aggleton, J.P. (Ed) (2000). The Amygdala: A Functional Analysis, Second Edition, Oxford: Oxford University Press.

Alm, P.A. (2004). Stuttering, emotions, and heart rate during anticipatory anxiety: a critical review, Journal of Fluency Disorders 29, 123133

Alm, P.A. (2005) On the Causal Mechanisms of Stuttering (Thesis). Sweden: Lund University.

Anderson, J.D., Pellowski, M.W., Conture, E.G., Kelly, E.M. (Oct 2003). Temperamental Characteristics of Young Children Who Stutter. Journal of Speech Language Hearing Research, 46(5): 12211233.

Chang, S.E., Erickson, K.I., Ambrose, N.G., Hasegawa-Johnson, M.A., Ludlow, C.L.. (2008) Brain anatomy differences in childhood stuttering. Neuroimage. Feb. 1:39(3):1333-44.

Comings, D., Wu, S., Chiu, C., Ring, R.H., Gade, R., Ahn, C., MacMurray, J.P., Dietz, G., & Muhleman, D. (1996). Polygenic inheritance of tourette syndrome, stuttering, attention deficit hyperactivity, conduct and oppositional defiant disorder: The additive and subtractive effect of three dopaminergic genes--DRD2, DBH, and DAT1. American Journal of Medical Genetics, 67, 264-288.

Cooper, E.B. (1983). A brain-stem contusion and fluency: Vicki's story. Journal of Fluency Disorders, 8, 269-274.

Damasio, A. (1999). The Feeling of What Happens: body and emotion in the making of consciousness. New York: Harcourt, Brace and Company.

De Nil, LF, Kroll, RM, Kapur S, Houle, S, (2000), A positron emission tomography study of silent and oral single word reading in stuttering and nonstuttering adults. Journal of Speech, Language, and Hearing Research, 43: 1038-1053. 

Denny M, Smith A, (1997), Respiratory and laryngeal control in stuttering. In Curlee RF, Seigel GM (Eds) Nature and treatment of stuttering: new directions, Boston: Allyn and Bacon, 128 – 142.

Dolan, R.J., (2000). Functional neuroimaging of the human amygdala during emotional processing and learning. In Aggleton, J.P. (Ed) The Amygdala: A Functional Analysis, Second Edition, Oxford: Oxford University Press, 631 - 653.

Foundas, A.L., Bollich, A.M., Corey, D.M., Hurley, M., Heilman, K.M. (2001) Anomalous anatomy of speech-language areas in adults with persistent developmental stuttering. Neurology, 57: 207-215.

Fox, P.T., et. al., (1996). A PET study of the neural systems of stuttering. Nature, 382, 158-161.

Goleman, D. (1995). Emotional intelligence. New York: Bantam Books.

Gray, J. (1987). The psychology of fear and stress (2nd Edition). Cambridge: Cambridge University Press.

Guitar, B. (1997). Therapy for children’s stuttering and emotions, In Curlee RF, Seigel GM (Eds) Nature and treatment of stuttering: new directions, Boston: Allyn and Bacon, 280 – 291.

Guitar, B. (1998). Stuttering: An integrated approach to its nature and treatment, Second Edition. Baltimore: Williams & Wilkins.

Guitar, B. (2003). Acoustic Startle Responses and Temperament in Individuals Who Stutter. Journal of Speech, Language, and Hearing Research. 46, 233-240.

Juergens, U. (2005). Neural Mechanisms for Voice Control: Evidence from Experimental Monkey Studies, in State of the Science Conference: Developmental Stuttering, March 21-23, 2005. National Institute on Deafness and Other Communication Disorders, National Institutes of Health.

Kagan, J., Reznick, J.S., and Snidman, N. (1987). The physiology and psychology of behavioral inhibition in children. Child Development, 58, 1459-1473.

Kagan, J. (1994). Galen's prophecy. New York, Basic Books.

Kalinowski, J. and Stuart, A., (1996), Stuttering amelioration at various auditory feedback delays and speech rates, European Journal of Disorders of Communication, 31, 259-69.

Kolb, L.C. (1987). A neuropsychological hypothesis explaining posttraumatic stress disorder. American Journal of Psychiatry, 144, 989-995.

Larson C.R., (1985), The midbrain periaqueductal gray: a brainstem structure involved in vocalization, Journal of Speech and Hearing Research, 28, 241-249.

Le Doux, J. (2001). The synaptic self. New York: Viking.

LeDoux, J. (1996). The emotional brain. New York: Simon & Schuster.

LeDoux, J., Iwata, J., Ciccheti, P. & Reis, D.J. (1988). Different projections of the central amygdaloid nucleus mediate autonomic and behavioral correlates of conditioned fear. The Journal of Neuroscience, 8, 2517-2529.

LeDoux, J., Romanski, L., & Xagoraris, A. (1989). Indelibility of subcortical emotional memories. Journal of Cognitive Neuroscience, 1, 238-243.

Linden, D.J., (2007). The accidental mind. Cambridge, Massachusetts: The Belknap Press of Harvard University Press.

 Lu C, Peng D, Chen C, Ning N, Ding G, Li K, Yang Y, Lin C. Altered effective connectivity and anomalous anatomy in the basal ganglia-thalamocortical circuit of stuttering speakers.
2009 Mar 13. [Epub ahead of print] ABSTRACT ONLY

Oyler, M.E., (2003). Temperamental sensitivity in children who stutter. ISAD 2003 Online Conference, October 1, 2003

Parry, W.D. (1994). Understanding and Controlling Stuttering, a comprehensive new approach based on the Valsalva Hypothesis. San Francisco: The National Stuttering Project.

Paus, T., Petrides, M., Evans, AC. and Meyer, E. (1993). Role of the human anterior cingulate cortex in the control of oculomotor, manual, and speech responses: a positron emission tomography study. Journal of Neurophysiology, Vol 70, Issue 2 453-469

Perkins, W.H. (2000). Tongue wars: recovery from stuttering. Reno, Nevada: Athens Press.

Peters, T.J. & Guitar, B. (1991). Stuttering: An integrated approach to its nature and treatment. Baltimore: Williams & Wilkins.

Pinel, J.P.J. (1997). Biopsychology, Third Edition. Boston: Allyn and Bacon.

Postma, H. & Kolk, A. (1997). Stuttering as a covert repair phenomenon. In Nature and Treatment of Stuttering: New Directions, Curlee, R.F., & Siegel, G.M. (Eds.). Boston: Allyn and Bacon, 182-203.

Prins, D. (1996). Modifying stuttering--the stutterer's reactive behavior: Perspectives on past, present, and future. In Nature and Treatment of Stuttering: New Directions, Curlee, R.F., & Siegel, G.M. (Eds.). Boston: Allyn and Bacon, 335-355.

Rastatter, M.P., Harr, R. (1988). Measurements of plasma levels of adrenergic neurotransmitters and primary amino acids in five stuttering subjects: A preliminary report (biochemical aspects of stuttering). Journal of Fluency Disorders, 13, 127-139.

Rastatter, M.P.,  Kalinowski, J., Cranford, J.,  and Stuart, A. (1998). Topographic EEG mapping of stuttering subjects during speech tasks under normal and altered feedback conditions, East Carolina University WWW Site.

Smith, A. and Kelly, E. (1996). Stuttering: A dynamic multifactorial model. In Nature and treatment of stuttering: new directions, Curlee, R.F., & Siegel, G.M. (Eds.). Boston: Allyn and Bacon, 204-217.

Sommer, M., Koch, M.A., Paulus, W., Weiller, C., Buchel, C. (2002). Disconnection of speech-relevant brain areas in persistent developmental stuttering. The Lancet, 360, 380-383

Starkweather, C.W. (1995). A simple theory of stuttering. Journal of Fluency Disorders, 20, 91-116.

Starkweather, C.W. and Givens-Ackerman, Janet. (1997). Stuttering. Austin, Texas: PRO-ED, Inc.

Stuart, A., & Kalinowski, J., (1996). Fluent speech, fast articulatory rate, and delayed auditory feedback: creating a crisis for a scientific revolution? Perceptual and Motor Skills, 82, 211-218.

Stuttering Foundation of America (1996). Therapy for stutterers. Memphis: Stuttering Foundation of America.

Tunbridge, N. (1994). The stutterer's survival guide. Sydney: Addison & Wesley.

Van Riper, C.G. (1973). The treatment of stuttering. Englewood Cliffs: Prentice-Hall.

VanRiper, C.G. (1982). The nature of stuttering (2nd edition). Englewood Cliffs: Prentice-Hall.

Watson, B.C., Freeman, F.J., (1997). Brain imaging contributions. In Nature and Treatment of Stuttering: New Directions, Curlee, R.F., & Siegel, G.M. (Eds.). Boston: Allyn and Bacon, 143-166.

Wu, J.C., Maguire, G., Riley, G., Fallon, J., LaCasse, L., Chin, S., Klein, E., Tang, C., Cadwell, S., & Lottenberg, S. (1995). A positron emission tomography [18F]deoxyglucose study of developmental stuttering. Cognitive Neuroscience and Neuropsychology, 6, 501-505.

Yairi, E. (1996a). Disfluency characteristics of childhood stuttering. In Nature and Treatment of Stuttering: New Directions, Curlee, R.F., & Siegel, G.M. (Eds.). Boston: Allyn and Bacon, 49-78.

Yairi, E., Ambrose, N., Paden, E. & Throneberg, R. (1996b). Predictive factors of persistence and recovery: Pathways of childhood stuttering. Journal of Communication Disorders, 29, 51-77.

1995-2006 Darrell M. Dodge MA, CCC-SLP

Updated: October 08, 2012

Contact the Author

[ Top ] [ Veils of Stuttering ]