There are many hypotheses and theories on developmental stuttering. By now, some of them have definitely been filed away, for instance, stuttering as a neurosis, stuttering as a muscular spasm, or stuttering as a learned behavior. Today, it is widely consensus: Stuttering has to do with a malfunction in the brain, and this malfunction has a genetic basis. That is suggested by a lot of empirical data, However, we neither know how the mutations found in some ‘stuttering families’(see, e.g. Kraft and Yairy, 2012) are related to the peculiar patterns of activation in the stutterers’ brain, nor how these brain activation patterns make people stutter, i.e., how they make them, against their will, repeate words or parts of words, prolong speech sounds, or get stuck.
Stuttering cannot be explained by simply stating that speech control breaks down, even if a person whose speaking is suddenly blocked may subjectively perceive this as a loss of control. A breakdown of speech control could also manifest in completely other symptoms: Why does not occur an incomprehensible babbling or mumbling, why does the person not yell, or talk nonsense – or what else one may imagine as ‘speaking out of control’ with some fantasy? Why do just those specific symptoms occur that we refer to as stuttering? It is the question for the mechanism underlying the symptoms. Solving this problem should enable us to answer many other questions related to stuttering, such as:
Why does stuttering mostly and earliest onset when children start forming sentences, whereas previously, during the babbling phase and the period of one- and two-word utterances, the same children did not show any symptoms?
Why do many of the affected children spontaneously recover from stuttering after some months or years?
Why is persistent developmental stuttering much more prevalent in males than in females, whereas, in early childhood, the proportion of boys and girls who stutter is roughly balanced? Why do girls have a better chance to recover?
Why is stuttering so much influenced by situations, by the environment, or by the number and the kind of listeners? If something goes wrong in the stutterers’ brains – why does it affect speech in some situations, but not in others? The ‘demands and capacities’ approach does not provide a convincing explanation.
What relationship does exist between stuttering and the phonological, syntactic, and semantic structure of utterances? Why is a sort of speech sounds (plosives, vowels...) more frequently stuttered by one individual, another sort by another individual? Why does stuttering occur mostly at the onset of words and always at the initial part of syllables? Why are stressed syllables, long words, and words with high information load more frequently stuttered?
What about the structural and functional abnormalities found in the brain of stutterers, for example, the right shift of activation during speech, and the structural deficits in some white matter tracts? What of them is causal for stuttering, and what is a concomitant or a consequence of the disorder?
Why is stuttering strongly influenceable by alterations of auditory feedback – so that altered auditory feedback or auditory masking can considerably reduce stuttering or cause it to disappear immediately and completely – as long as alteration or masking continue, and as long as an alteration remains unfamiliar?
Why is stuttering, at least in some cases, successfully treatable by an alteration of the stutterer’s behavior, for example, by using a special speech technique? The question is not trivial because if something actually goes wrong in a stutterer’s brain and causes the disorder – why can that be corrected by a change of behavior?
Finally, the most important question a theory should answer: Can persistent developmental stuttering be cured, and if so, how? And if not: What can be done, so that symptoms occur as rarely and mildly as possible? In other words: A theory of stuttering should enable us to say what is the best therapy.
These are the questions a theory of stuttering, in my view, should answer. Packman and Attanasio (2004) claimed that also the role of genes in stuttering should be explained (p. 55). However, it is doubtful that genetic mutations anyway directly cause stuttering – the distinctive variability of the disorder would hardly be explicable with such a premise. Therefore, I assume that not stuttering itself, but only the predisposition for stuttering is genetically caused, which might be plausible prima facie. The issue of predisposition is extensively dealt with in Section 3.3.
In the present theory, two things are assumed to play a central role in stuttering: the processing of auditory feedback and the regulation of attention during speech. For a long time, auditory feedback – hearing one’s own voice and words while speaking – has been suspected to be involved in the pathomechanism of stuttering (see, e.g., Cherry and Sayers, 1956; Maraist and Hutton, 1957; Sandow, 1898; Van Riper, 1973; Yates, 1963). Stuttering is often significantly reduced, or it even completely disappears as long as a stutterer perceives his own speech as altered and unfamiliar – either by talking in an altered manner or by artificial alterations of the auditory feedback. This is, for example, the case in chorus reading, shadowing, in talking with an assumed voice or in an unfamiliar dialect, but also with auditory masking (by loud noise via headphones) or with artificially delayed or frequency-altered auditory feedback (DAF, FAF) – see Bloodstein and Bernstein Ratner (2008) for an overview. All these observations strongly suggest: Hearing one’s own speech plays a role in stuttering.
Several researchers have assumed that, at least in some of the cases mentioned above, stuttering disappears or is reduced because the stutterer’s attention is distracted from hearing his/her own speech, thus the auditory feedback (that was assumed to be anyway distorted) would to a lesser extent interfere speech control.(see, e.g., Cherry and Sayers, 1956; Van Riper, 1973). However, brain imaging studies, have shown speaking in an altered manner and/or with altered auditory feedback to be associated with greater activations in the cortical areas responsible for perceptive language processing, namely in the secondary auditory areas, particularly the superior and middle temporal gyrus (BA21/22), which does not support the distraction hypothesis.
For example, greater activations in secondary auditory areas were observed in stutterers during speech paced by the beat of a metronome, during chorus reading, and singing, compared to stuttered speech (see Table below). Greater activations in the secondary auditory areas of the cortex were also observed in speaking with altered auditory feedback. Unfortunately, there have been only studies on this issue with nonstutterers up to now (see Table). In a meta-analysis of seven brain-imaging studies, in which fluency-enhancing means like metronome pacing or chorus reading were applied (Budde, Barron, & Fox, 2014), it became apparent that only auditory areas were consistently greater activated during induced fluency, compared to stuttered speech (read more).
Those results, in my view, do not support the hypothesis that distraction from auditory feedback would improve speech fluency – especially as greater activity in the same cortex areas was found in normal fluent speakers, when they were listening to acoustic verbal stimuli, but lower activation, when the stimuli were still presented, but the participants’ attention was distracted from (Hugdahl et al., 2003; Jäncke, Mirzazade, & Shah, 1999; Sabri et al., 2008)
Enhanced speech fluency was found to be associated with greater activations in secondary auditory areas not only in fluency-enhancing conditions, but also in other cases: Normal fluent speakers, but also recovered stutterers, as groups, showed greater activation in those cortical areas during speech than persistent stutterers (see Table). In some studies, a negative correlation between stuttering severity and the activity in secondary auditory areas was found (Braun et al., 1997; Fox et al., 2000; Ingham et al., 2004), that is: The more severe the stuttering, the lower, on average, was the activity in the secondary auditory brain areas, and vice versa. In milder stutterers, as a group, activity in secondary auditory areas during speech was greater than in more severe stutterers (Neumann et al., 2003), and, after a successful fluency-shaping therapy, participants, as a group, showed greater activity in secondary auditory brain areas than before therapy (see Table). The latter finding is surprising because to influence auditory processing is not intended in fluency-shaping therapies.
Two meta-analyses, Brown et al. (2005) and Budde, Barron, and Fox (2014), in which all before published brain-imaging studies of the activation patterns in stutterers versus nonstutterers were included, revealed that the lack of activation in the secondary auditory areas is one of the very characteristics in the brain activation patterns of stutterers. The following table gives an overview of studies in which a relationship between stuttering and the activation in the secondary auditory brain areas was found (read more). In the five lower rows of the table, studies dealing with fluency-enhancing conditions are listed.
|Lower Activation||Greater Activation||Studies|
| Braun et al., 1997
Chang et al., 2009
Fox et al., 1996
Ingham et al., 2003
Pool et al., 1991
Wu et al., 1995
Brown et al., 2005 *
Budde, Baron, & Fox, 2014 *
|Stutterers|| Recovered stutterers
(stuttered as a child)
|Ingham et al., 2003|
| Severe stuttering /
higher stutter rate
| Mild stuttering /
lower stutter rate
| Fox et al., 2000
Ingham et al., 2004
Neumann et al., 2003
| Stutterers prior to
| Stutterers after
| De Nil et al., 2003
Ingham et al., 2003
Neumann et al., 2003
|Solo reading||Chorus reading|| Fox et al., 1996
Wu et al., 1995
| Narrative speech +
with a given verb
| Overlearned speech
+ paced speech
with a metronome
|Braun et al., 1997|
| Narrative speech +
with a given verb
| Chorus reading
|Stager, Jeffries, & Braun, 2003|
|Solo reading|| Chorus reading
+ paced speech
with a metronome
|Toyomura, Fujii, & Kuriki, 2011|
|Stuttered speech||Induced fluency||Budde, Baron, & Fox, 2014 *|
| Speech with natural
| Speech with altered
| Hashimoto & Sakai, 2003 **
Takaso et al., 2010 **
Tourville, Reilly, & Guenther, 2008 **
Table 1: Brain activation in the secondary auditory areas (in which perceptive language processing is localized). * = Meta-analysis, ** = Normal fluent speakers.
All the findings mentioned above suggest that auditory feedback and stuttering are anyway linked, but the nature of the link has hardly been understood until now. One reason is that the observations are contradicting: On the one hand, stuttering disappears as long as auditory feedback is masked, delayed or otherwise altered, which seems to suggest an effect of distraction. On the other hand, enhanced speech fluency is associated with greater activations in secondary auditory brain areas, suggesting a more intensive processing of auditory feedback. In the theory presented on the following pages, I will try to solve this contradiction and to explain the relationship between stuttering and auditory feedback – a relationship that seems to play a crucial role in the pathomechanism of stuttering.
A bunch of empirical findings obtained mainly in the last decade are behavioral data suggesting a deficient attention regulation in children and adults who stutter: They tend to be less able to allocate attentional resources under dual task conditions, and to be more prone to exhibit attention disorders (see Anderson and Wagovich (2010), Eggers, De Nil, and Van den Bergh (2012), or Alm (2014) for an overview). In addition to behavioral data, findings from brain research indicate deficits in central auditory processing, among them in mechanisms underlying the automatic control of auditory attention (see Section 3.3). As far as I know, these relatively new insights have hardly been included in causal theories of stuttering until now. In the theory proposed here, attention control plays a crucial role, which will enable us both, to describe a brain mechanism that possibly causes stuttering, and to explain the astonishing variability of the disorder depending on psychological, linguistic, and environmental factors.
Subject of the theory is idiopathic, i.e. in almost all cases, developmental stuttering, but not neurogenic stuttering caused by stroke, brain injury, or medication. I do not use the term ‘psychogenic stuttering’ because it may be difficult to distinguish in some cases of late-onset stuttering, whether the disorder is caused more by a genetic predisposition or more by psychological and/or environmental factors. I think a causal theory of developmental stuttering should include cases of late-onset stuttering, becauses the symptomatology as well as neuroanatomical differences from normal fluent individuals were found to be similar as with developmental stuttering (Chang et al., 2010).
The position I start from is the assumption that developmental stuttering is a disorder of normal speech, that is: As long as stutterers are spontaneously fluent, they speak normally, and when they stutter, then their normal fluent speech is disrupted. One may object that brain activation patterns in stutterers were found to be different from those of nonstutterers also when the stutterers were fluent, and even when they were silently reading or thinking (see. e.g., Neumann et al., 2003). Do such findings not imply that stutterers never speak normally, not even when they are fluent?
To decide this issue, we should distinguish between speaking, on the one hand, and the brain processes underlying speech, on the other hand. For the sake of simplicity, let us say that ‘normal speaking’ means to string words together fluently and in a syntactically correct order. After this definition, a stutterer who is doing so speaks normally, regardless of whether his brain possibly needs to expend more effort compared to a nonstutterer’s brain. Maybe, the brains of stutterers are less suited for speech control, possibly, their structure is less optimal for this task, thus they need a greater extent or a deviant pattern of activity for the control of normal speech. I think, we get a problem if we link ‘normal speaking’ to a particular pattern of brain activity, since the brain activations in several nonstutterers during speech may differ too. Therefore, I start from the position that stutterers speak normally, as long as they are spontaneously fluent without the help of fluency-enhancing techniques or devices.
In order to understand stuttering as a disruption of normal speech, we must at first realize how normal speaking works and what role auditory feedback plays in normal speech. This issue is dealt with on the next pages.
Although it may be considered politically incorrect, I prefer to use the term ‘stutterer’. In my view, a stutterer is a person with a probably innate predisposition for stuttering, who has developed a persistent stuttering, thus stuttering symptoms in some likelihood occur with this person. From this, it does not follow that the affected person necessarily stutters at any certain moment. At the best, for example, after a successful therapy, the person never or rarely stutters, but nevertheless, he or she remains a (recovered) stutterer – with the predisposition still existing and with a risk of relapse. By contrast, the term ‘person who stutters’ (PWS), in my view, suggests someone who actually stutters. To be a recovered stutterer is possible, but a recovered PWS – a person who stutters who doesn’t stutter – is nonsense.
To refer to someone as a stutterer, in my view, is more to attribute a property to him/her than to attribute a behavior – but it does neither mean to reduce the person to this one property nor to imply assumptions about personality traits like intelligence, self-assurance, etc. Further, abbreviations like PWS, PWNS, CWS, CWNS are unhandsome in the text and do not support understanding; therefore, I avoid them.
Speech shadowing is a technique in the treatment of stuttering: The stutterer repeats connected speech immediately after hearing it (with a short time lag), thus he or she must listen to the ‘leader’ and speak at the same time. In this condition, stuttering usually disappears (Cherry & Sayers,. 1956; Kelham & McHale, 1966; Kondas, 1967; MacLaren, 1960; Marland, 1957). More about shadowing in Section 3.1 (see here).
;“...left auditory (superior temporal gyrus) activations were greater in controls than in PDS, but bilateral auditory cortex activations were greater during fluency than dysfluency, especially on the right. No auditory cortex activations were present during dysfluency. Absence of left auditory cortex activation (compared to controls) indicated stuttering trait, and absence of right auditory cortex activation indicated dysfluent state” (Budde, Barron, & Fox, 2014, p. 104) – A/N: The latter may result from the fact that stuttered speech and induced, i.e., not spontaneous fluency were compared in the within-group analysis. Induced fluency is associated with greater activation mainly in the right auditory cortex. The reason may be that, in most fluency-enhancing tasks, stutterers only read, but not produce self-formulated speech; thus the left-hemispheric language network may not fully be activated. I think, self-formulated fluent speech requires auditory activation on the language-dominant hemisphere.
“Looking at the group comparison in the meta analysis, the area that showed the largest inter-group differences [i.e., between stutterers and controls, A/N] is a part of the superior temporal sulcus situated just anterior to those areas found to have voice-selective auditory representations […]. In sum, the published literature supports a robust auditory inhibitory effect in stutterers, which is consistent with the meta-analysis results.” (Brown et al., 2005, p. 113) – See also the previous footnote and further citations. (return)