A Theory of Stuttering
A Theory of Stuttering
The idea that FCs reduce stuttering by drawing the speaker’s attention to auditory feedback may be plausible for conditions in which auditory stimuli are presented or auditory feedback is altered, but visual, tactile, and other non-auditory feedback were also found to reduce stuttering. In some of those studies, visual or tactile feedback was applied to merely slow down speech rate. For instance, variations in the intensity of a light or of a vibrator at the participant’s hand reflected vocal variations in speaking. This feedback was presented with different delays, and participants were instructed to speak “synchronically” with that feedback (e.g., Smolka & Adamczyk, 1992). The moderate reduction in stuttering in these studies was more the effect of slow speech than of visual or tactile feedback.
Barber (1940) showed that visual and tactile rhythmic stimulation was as effective as the clicking of a metronome when stutterers were speaking in time with. This is not astonishing because, just as in speaking timed by an acoustic metronome, you must pay attention to both, the presented rhythm and your own rhythm, to monitor their synchronicity. And since auditory feedback is the main source for monitoring one’s own speech rhythm, we can assume that participants focused on both, the stimuli and their auditory feedback, such that even visual or tactile rhythmic stimulation improved the processing and integration of auditory feedback.
Kalinowski et al. (2000) replaced auditory with visual stimulation even in choral speech. They found a reduction of approximately 80% in stuttering frequency when stutterers recited a memorized text in time with another person who silently ‘mouthed’ the text; stutterers focused their gaze on the other person’s visible speech movements. Kalinowski and Dayalu (2002) supposed that this activated auditory association areas in stutterers, with reference to Calvert et al. (1997), who had found increased auditory activation in normally fluent individuals during silent lip-reading. In fact, auditory activation may increase (normalize) in stutterers during ‘visual choral speech’ for the same reason as in usual choral speech, since it requires attention to both, the other person’s silent gestures and auditory feedback to monitor the sync.
Hudock et al. (2011) demonstrated that delayed visual feedback can reduce stuttering without slowing down the speech rate. Stutterers recited memorized material at normal and fast speech rates while viewing their own mouth and jaw movements on a screen. This visual feedback was displayed with time delays of 0, 50, 200, and 400 ms. As compared to a baseline condition without visual feedback, all visual feedback conditions evoked reductions in stuttering ranging from 27% (0 ms delay) to 62% (400 ms delay). All conditions with delayed visual feedback were more effective than non-delayed visual feedback. There was no significant effect of rate.
It is doubtful whether the reduction in stuttering in this experiment immediately resulted from the visual stimulation. Participants were instructed to look at the screen, but their auditory feedback was not masked. A spontaneous attempt to perceive a relation or to achieve a match between visual and auditory feedback possibly made them pay more attention to their auditory feedback than usual. Or, while looking at the screen, they focused on auditory feedback to not become confused by the delayed visual feedback.
Taken together, the reported instances of fluency enhancement by visual and tactile stimulation do not contradict the view that FCs draw stutterers’ attention to auditory feedback and thereby improve its processing and integration in the control of speech.
A special kind of experiments was done in the 1960s and 1970s: a stimulus was presented immediately after each stuttering event. This method was called “response contingent simulation” according to the underlying theory that stuttering is a psychological response. These experiments, some of which may appear ethically doubtful today, are relevant in our context because they are conditions in which stuttering was reduced often by auditory, but also by non-auditory stimulation (see Bloodstein & Bernstein Ratner, 2008, for an overview).
Some of those studies examined the effects of punishment for stuttering. For instance, Martin & Siegel (1966) report that they reduced stuttering “essentially to zero” by electric shocks contingent on stuttering events. However, many studies of response contingent stimulation used auditory stimuli as a punishment, e.g., a 105 dB tone, the reproof “wrong!”, or recorded laughter. In other studies, neutral or reinforcing auditory stimuli were presented after each stutter, or stutter-free speech was “penalized” (see Bloodstein & Bernstein Ratner, 2008, pp. 292–294).
Martin and Siegel (1966) who applied electric shock as a ‘punishing’ stimulus tested merely 3 participants, and with only one, the reduction of core symptoms (sound and word repetitions) was tested. With the two other participants, only the reduction of secondary behaviors like nose wrinkling was examined. Such habits can be suppressed by the will if they result in punishment.
Even this astonishing effect can be explained as an effect on attention. Participants knew that the experimenter was attentively listening to them in order to detect each stutter and reply to it with punishment. In this situation, a participant may spontaneously mirror the experimenter’s behavior and, quasi with the experimenter’s ears, listen to his/her own speech, either to avoid stuttering or to anticipate the experimenter’s reply and avoid startle. If auditory stimuli were used as a reply to stuttering, this may have additionally drawn attention to the auditory channel by evoking an auditory anticipation. So, stutter-contingent stimulation may have drawn stutterers’ attention to auditory feedback and thereby reduced stuttering.
“For many years, it had occasionally been noted that stutterers are likely to have less speech difficulty in the presence of loud noise—for example, near ocean surf, a waterfall, or a passing train. Kern (1932) demonstrated this effect experimentally by means of a Barany drum” (Bloodstein & Bernstein Ratner, 2008, p. 295). A Barány drum is not a musical instrument but a mechanical device powered by a spring (see photo). It was developed in 1906 by Nobel price winner Robert Barány, for auditory masking and the detection of simulated deafness.
Shane (1955) and Cherry, Sayers, and Marland (1955) were the first who used white noise of high intensity to mask auditory feedback, and they reported that, in general, virtually complete elimination of stuttering was the result. Since then, the white noise effect has been confirmed repeatedly (e.g., Garber & Martin, 1974, 1977; Maraist & Hutton, 1957; Martin & Haroldson, 1979). Naturally, the first idea to explain the effect was that auditory feedback is somehow harmful for stutterers, and they become fluent once they do not hear themselves speak (e.g., Cherry & Sayers, 1956; Maraist & Hutton, 1957; Webster & Lubker, 1968).
However, further experiments showed that noise reduced stuttering even though participants heard themselves speak. Sutton and Chase (1961) and Webster and Dorman (1970) presented noise only in the silent periods between phonation; this reduced stuttering as effectively as continuous masking did. Others tested the effect of noise with an intensity at which one’s own voice can be heard for sure. For instance, Martin et al. (1984) found a significant reduction in stuttering at a noise level of 60 dB. Maraist and Hutton (1957) presented noise at levels of 30, 50, 70, and 90 dB and found a progressive decrease in stuttering with increasing noise intensity.
These findings indicate that noise does not reduce stuttering because stutterers do not hear themselves speak. Wingate (1970) hypothesized that increased voice intensity during speaking in noise may enhance fluency, but Garber and Martin (1977) found that the effect of masking noise was greatest when stutterers were speaking with normal voice intensity.
As far as I know, a study of the effect of masking noise during speech on brain activation has not yet been conducted with stuttering participants, but there are several studies with normally fluent individuals. Paus et al. (1996) found activation in the auditory association areas (where the auditory feedback of speech is processed) during whispered speech with auditory feedback masked by noise. Christoffels, Formisano, and Schiller (2007), Zheng, Munhall, and Johnsrude (2010), and Christoffels et al. (2011) compared speaking with normal feedback, speaking with feedback masked by noise, and listening to one’s own prerecorded speech. The activation in auditory association areas when speaking with auditory feedback being completely masked was greater than with normal auditory feedback and as great as when listening to prerecorded speech.
The neuroimaging findings suggest that masking noise has a similar effect on auditory association areas as choral or metronome-paced speech and altered auditory feedback. This would not be surprising if auditory feedback was merely partially masked, e.g., with noise of lower intensity. As Stephen and Haggard (1980) noted, DAF and FAF also have a masking effect insofar as they partially overlay natural auditory feedback. The same is true for a second speaker’s voice in choral speech. Seen from this angle, white noise, like other unaccustomed acoustic stimuli, may draw the speaker’s attention to auditory feedback—as long as one’s speech is still heard despite the noise. But what if auditory feedback is masked completely?
Cherry and Sayers (1956) reported that the loudness of the noise that made stuttering disappear approached pain level, such that participants were unaware of their own speech sound. This seems to clearly contradict the idea that all FCs improve the processing and integration of auditory feedback. Furthermore, the fact that the speech of normally fluent individuals is not impaired when their auditory feedback is completely masked by noise seems to indicate that auditory feedback is not necessary at all for speech control.
This was the main argument for Borden (1979) to claim that auditory feedback is irrelevant for skilled speech and necessary only for children developing speech or adults learning new speech patterns. Many experts believed he; that has inhibited research into the role of auditory feedback in stuttering for decades.
The view that auditory feedback is irrelevant for the control of skilled speech is inconsistent with recent findings. For instance, Kalveram (1983) found that the duration of long syllables is controlled based on auditory feedback (audio-phonatory coupling). Lind et al. (2014) found that auditory feedback is the main source for the phonological (and thus also for the lexical and semantic) self-monitoring of speech. To resolve these inconsistencies, it is helpful to look at another very effective FC in which no (external) auditory feedback is available: silent mouthing (see next section).