Gerald N. Zimmermann

Slow potentials of the brain related to speech processing in normal speakers and stutterers

Abstract A CNV paradigm was employed to study slow potential shifts of the brain in stutterers and normal speakers during speech and nonspeech verbal tasks and in a nonverbal manual task. Slow potential shifts were recorded from Cz. iF 3 ′ and iF 4 ′ . It was found that vertex CNVs were similar in normal speakers and stutterers in the verbal and nonverbal tasks. Differences between groups were found at the lateral electrode derivations in both verbal tasks. Preceding speech, four of five of the normal speakers showed a larger shift in the left hemisphere than in the right, while only 22% of the stutterers showed a left greater than right asymmetry. In the verbal condition in which the subjects did not speak (the “expectancy” condition) similar relationships were found. It is concluded that the left and right inferior frontal areas of nonstutterers and stutterers performed differently even when the stutterers were not approaching a moment of stuttering, i.e. , when both groups appeared to have equal speech performance, and also when they were making identical decisions about the “expectancy” of having difficulty speaking a word although no overt motor speech was required. The interhemispheric differences in the two groups are not attributed to increased “anxiety” since the vertex CNV was unaffected.

Lip and jaw interaction during speech: Responses to perturbation of lower‐lip movement prior to bilabial closure

Electrical stimulation was used to produce unexpected, involuntary depression of the lower lip in three normal young adults. Stimulation was timed to begin 500 to 40 ms prior to voice offset in [aep] and (Ip]. Upper lip, lower lip, and jaw movements were measured with a strain gauge system. Movements in 104 syllables with lower-lip stimulation were compared to the preceding normal syllable. Both the jaw and upper lip compensated for the involuntary perturbations in lower-lip movement. Compensatory movements did not occur as additional, discrete gestures following stimulation onset, but appeared as an increase in the size of closing movements. Bilabial closure was produced at the typical time (within - 10 to + 20 ms of voice offset) in 68% of the perturbed syllables, but it was delayed (a mean of 61 ms) in the remaining 32%. Neither the incidence nor the magnitude of this delay appeared to be related to the jaw position at stimulation onset or to the time between stimulation onset and voice offset.

Jaw‐muscle activity during speech with the mandible fixed

Electromyographic recordings were made from the anterior temporalis, masseter, medial pterygoid, and lateral pterygoid muscles in four normal adult subjects. Discrete bursts of activity occurred in these muscles even during speech with the jaw prevented from moving by placement of a bite block between upper and lower molars. The bursts of muscle activity with the bite block were similar in frequency of occurrence, time of peak activity, and magnitude to the activity observed with the jaw free to move. A motor control system that employs a central simulation process to coordinate the lips and tongue with the jaw is not necessarily consistent with this finding. An efficient central simulation process might be expected to eliminate the discrete jaw-muscle activity with the jaw fixed as well as producing correct responses in the lips and tongue. These data are more consistent with a motor control system employing lower-level neural mechanisms to coordinate articulatory movements.

Effects of auditory maskers on whistling and voicing

The effects of high‐intensity auditory maskers on whistling and voicing pitch matching tasks were studied. Subjects whistled or produced the vowel /ah/ to match the frequency of a target tone. Nine experimental masking conditions were presented randomly: no masker, white noise, a pure tone at the target frequency, three tones above and three tones below the target frequency. Masker tones were presented at either 80 or 100 dB SPL. Mean fundamental frequency of voicing was essentially unaffected by auditory manipulations, i.e., productions obtained during masking conditions were no different than productions in quiet. For whistling, mean fundamental frequency of production appeared to be differentially affected by masker frequency. Masker frequencies lower than the target generally resulted in higher frequency of production. Similar effects were observed for both levels of masking stimuli which argues against an interpretation of the observed effects in terms of a peripheral masking of the production. Rathe...

The effects of white noise masking and low‐pass filtering on speech kinematics

The effects of reduced auditory information on spatial and temporal parameters of speech production were investigated using cinefluorographic techniques. While subjects read a series of test words embedded in carrier sentences, they received normal auditory information, auditory information regarding only the first formant of their production, or high level noise to mask all formant information. Kinematic analyses indicated that while there were some changes across conditions, these changes were not consistent either within or across the subjects. Parameters that were affected included mean displacement, vocal tract shape, interarticulator timing, and steady state duration. The results suggest that auditory information plays a role in maintaining dynamic aspects of speech kinematics. That is, while speech can be produced without auditory information, the precise action and coordination that characteristic normal production may be altered.

A reinterpretation of the effects of DAF on articulation

The effects of delayed auditory feedback as a function of the point in the articulatory cycle at which feedback occurs was analyzed. Six subjects were exposed to two delay conditions (100 and 200 ms) and a control (no delay) condition while reading an experimental passage. Strain gauges were used to monitor lower lip and jaw movements. Preliminary analyses suggest that the specific effects of DAF (increased voice duration, increased syllable duration, movement plateaus, increased durations of movements, etc.) are related to the point in the articulatory cycle that feedback occurs. The differences in the effects of 100 and 200 ms delays may be accounted for by the probabilities that feedback will occur during different points of articulatory movement under these different delays. The DAF effect will be discussed in terms of a model of the role of auditory information in speech production rather than in terms of auditory loop times.