IOPI Northwest

Development of Objective Standards of Nonspeech Oral Strength and Performance: An Advocate’s Views

Erich S. Luschei

The reader of this chapter should be warned in advance that the author (henceforth “I”) is not an expert in the area of objective measurement of oral strength and performance, and that, except for a chapter in this volume, co-authored with Donald Robin and Lori Somodi, I have no publications on the subject. I have become very interested in the topic, however, and strongly feel that our ability to help people with speech disorders can be greatly facilitated by developing standards based upon instrumentation and procedures that are available and applicable to clinics as well as research laboratories. Objective measurements and norms based on large numbers of observations would help in refining diagnostic categories, monitoring progression of a disorder, and assessing the effect of intervention therapy, as well as ultimately helping in understanding the basic process of speech production. Let me note first of all that the need for objective measurements of articulator strength and coordination has been recognized by many previous investigators in speech-language pathology. Palmer and Osborn (1940) attempted to measure tongue strength by measuring the pressure the tongue could exert on a hard rubber ball placed in the mouths of various types of speakers. While their results unfortunately were flawed technically in several ways, the basic approach has the potential for being quite useful. More recently, Dworkin (1980), Dworkin and Aronson (1986), Dworkin, Aronson, and Mulder (1980), Dworkin and Culatta (1985), and Posen (1972) published results of studies of tongue strength, measured as isometric force, in both adult and child normal and disordered speakers. Barlow and Abbs (1983) also developed an instrumentation system for measuring strength and performance of the tongue, lips, and mandible. Barlow and Netsell (1989) have shown how a sophisticated instrumentation system of this type can be used in a clinical setting to understand further the problems of dysarthric speakers.

To be perfectly honest, my knowledge of these studies came “after the fact.” My own personal interest in methods to develop normative standards for oral strength gradually arose from technical questions posed by several colleagues, and from my own fascination with instrumentation systems for measuring physiological variables. Working with Donald Robin and Lori Somodi, I helped develop a prototype of a portable and inexpensive device to measure objectively tongue strength and endurance (see Robin, Somodi, & Luschei, chap. 13, this volume). My experience with this project is not, however, the motivation for this chapter. Rather, it is the experience encountered with the review of our project for the Clinical Dysarthria Conference that stimulated my thinking. Although Donald Robin, Lori Somodi, and I were granted the opportunity to present our findings at the conference, the comments of the reviewer who evaluated our submission for the conference came as quite a shock. To quote, “My opinion about this proposal is that it is motivated by a real clinical concern for objectivity and accessibility. . . , but that it begs the question of the relevance of the measures to speech disorders.” The reviewer continued, “They [the authors] should be aware that participants at the conference . . . will demand more of a rationale for this technique than the typical blind acceptance that non-speech tasks are a reasonable part of a diagnostic battery.” The reviewer then suggested examination of criticisms raised in a recent review of maximum performance tasks written by Kent, Kent, and Rosenbek (1987).

After careful study of the Kent et al. (1987) paper, and after hearing many remarks from scientists and clinicians at the conference who are involved in the study and/or treatment of dysarthria, I have begun to appreciate the honest candor of the reviewer’s remarks, and the experience from which they arose. There are at least three basic reasons why, I think, many people in the field are reticent to embrace a proposal for the development of standards of oral strength and performance.

1. Studies on this subject have not, so far, produced methods or standards that are perceived as generally useful clinically.

2. Advocates of objective measurements and standards have perhaps “oversold” the concept and created an impression that the goal is to replace the clinician’s experience and skill with machines and numbers.

3. For many people, it is not obvious how measures of oral strength, resistance to fatigue (endurance), and maximal performance (speed) in nonspeech tasks are related to the requirements of speech articulation.

I cannot, in the space allotted, address these reasons in any depth, but I would like to comment on them. My reasons are simple: I am optimistic about the potential worth of objective measurements of simple nonspeech motor tasks, including those that require maximum performance, to clinicians who have to understand and treat dysarthria. Here are some things to think about.

OBJECTIVE MEASUREMENT IS A BASIC APPROACH OF SCIENCE

Many, if not most, of the things we know in science and the health professions that help people and allow us to understand very complex processes have derived, in the beginning, from the application of objective measurements made in the currency of commonly accepted ”units.” Such measurements typically have been compiled, often by contributions from scientists all over the world, until a large data base is available. At this stage, a data base of this type has a certain empirical value, but it does not really explain the process in question. However, it does allow a person to take an individual sample of the process (e.g., a patient) and make a good guess whether it is like other samples with respect to this particular database. We can often use this information to make good “clinical” decisions even if we don’t know everything there is to know about the process. Consider an example. Suppose someone claims that a piece of shiny yellow metal is pure gold. How does one know this person is telling the truth, without being an expert chemist? A very useful test would be to weigh the metal and measure its volume to calculate its density. We know the density of pure gold because we have a book compiled by chemists who had very elaborate ways of determining what pure gold is. If the piece of metal has the same density as gold, then it might be gold. However, if its density is 3 standard deviations from the mean density of gold, then a very good guess would be that the metal is not gold. A good clinical decision would be to decline the offer to purchase the piece of metal.

When we can recognize really deviant samples, we can study those specimens with the question “Why?,” and our knowledge of the subject may advance. Progress in science has often followed a very whimsical path, but many success stories have started with objective measures of simple aspects of the phenomenon in question. I am quite confident that careful measurements of simple oral behaviors may one day make significant contributions to our understanding of motor speech control. Because it hasn’t happened yet is no reason for being pessimistic.

Let me present another familiar example, partly because it reinforces my point about the potential usefulness of standardized measurements, and partly because it also illustrates the point that measurements and normal standards are an aid to the clinical process, not a substitute for it. Consider the problem of a patient who can’t seem to understand what is said. Could it be the patient doesn’t hear? If we want an answer to that question, we can obtain a very good estimate of the person’s ability to hear because of the invention of the audiometer, which tests the limits of a person’s ability to detect very simple, physically specified, stimuli. Should a clinician order a hearing test just because we have that wonderful ability? He or she could, I suppose, but let common sense prevail. Is there reason to believe the person understands the clinician’s language? Is there a possibility that the person had a stroke? Can the person hear a clap of the hands? When other reasonable explanations for a problem can be eliminated, and the clinician is still unsure of the cause of the problem, then measurement and standards become very useful. However, qualitative tests of speech and nonspeech oral behaviors that clinicians have used for years continue to be important and useful even if objective methods and standards exist.

MEASUREMENT OF ORAL MAXIMAL PERFORMANCE: AN IMPERFECT BEGINNING

Scientists interested in speech and its disorders have been making simple measures of motor performance for many years. This body of work has recently been reviewed and evaluated by Kent et al. (1987). I would recommend the paper to anyone who has not already studied it. Quite a few measures are considered, but there is a subset of measures that has been studied by a number of investigators: maximum phonation duration; pitch range; maximum expiratory pressure; maximum sound pressure level (a shout); strength of the tongue, lips, and jaw; and maximum syllable repetition rate. I won’t try to review this review, but I would like to comment on some of its general conclusions.

Both intra-subject and inter-subject variability is large. To quote Kent et al. (1987):

The issue of variability should not be underestimated. For some measures at some ages (e.g., maximum phonation duration for young children), the range of normal values is nearly an order of magnitude. It is abundantly clear from the literature on measures of maximum performance that instructions to the subject can strongly affect the data. (p. 382)

Intra-subject variability should not be large. If it is, and the source of the variability cannot be found and eliminated, then it seems unlikely that much use can be made of the measurements. I am in complete agreement with the comment on instructions. They should be given in a highly standardized way, and should be detailed in the published results. Motivation and experience with the task also affect maximum performance. It would be helpful, perhaps, to have investigators with an interest in maximum performance agree on a standard procedure. Other investigators in various disciplines have met to develop common procedures. Let me suggest an example of how one might proceed in a measure of maximum performance. Ask the subject, clearly and simply, to perform the task as “long as possible,” or as “hard as you can.” Make two measurements, recording both, but interpreting the highest as the maximum. Do not give feedback between these responses, or comment on performance. Then ask the subject to perform the task once more and “try as hard as you can to go longer (harder, faster).” One could call the third performance the “motivated” performance and report the typical percentage change in the measurement.

Now let me confess that we have not used this motivated third trial in our preliminary study (Robin, Somodi, & Luschei, chap. 13, this volume). It is a post hoc thought, and it is just a guess whether it is useful. The point is that human variables such as motivation will always be there, but we can try to neutralize them by standardization, or specifically manipulate them to assess their influence. Experiments could be done to determine which standardized instructions produced the least variability or had the best controls for motivation. Refining procedures in science sometimes takes a long time and requires help from many people.

Inter-subject variability may simply be part of the data. There is no doubt that we can weigh people accurately, but the range of weights that might be encountered among a group of middle-age males in the state of Iowa is impressive. Such variability greatly weakens the ability to detect “abnormality” if factors cannot be identified that account for much of the variability. It seems to me that these circumstances should lead us to search for variables that can be combined to form a mathematical model that is able to account for much of the intersubject variability. For example, one could ask people about their daily food intake and activity pattern as well as weigh them. A 300-pound 6-foot Iowan male would not be unusual in most circumstances, but if it was known that such Iowan males ate 3,000 calories a day and read magazines for exercise, one would have a very different clinical impression when it was noted that a particular 300- pound male ran laps and ate diet entrees.

Kent et al. (1987) suggest that maximum performance measures might not be relevant to speech because “speaking under ordinary circumstances does not tax the performance capabilities of the speech system” (p. 382). These authors subsequently modify this suggestion, however, by noting the need for timing of articulation, which can be affected by speed of movement. They note that “the temporal sequencing of speech may be performed at rates that allow for a small margin of error— a margin that is
occasionally exceeded by many talkers” (p. 382). Let me consider the first comment, that strength and power exceed that required for ordinary speech. This view arises because, I think, we usually measure strength of articulatory muscles in a highly artificial way (i.e., under static isometric conditions). The tongue of normal people, either adults or children (and interestingly they do not differ much in this regard) can produce an isometric anterior push of 15—20 N (Dworkin et al., 1980; Dworkin & Culatta, 1985; Posen, 1972). We never push with the tongue in this manner during speech, so how can this measurement be relevant? Normal speech movements require, particularly of the tongue, very high rates of muscle fiber shortening. The tongue is a muscular hydrostat. Contracting muscle fibers cause compartments within the tongue to constrict in one dimension, and the fluid that is displaced has to move to other regions of the tongue where muscle fibers are relaxed. This is a “clever” way of making muscle fibers, whose only active response is to contract, move things in one of several different directions without using bones. We don’t often think of it, but it is an old trick, “invented” to move tentacles and elephant trunks as well as tongues. Because of this feature, however, the tongue muscle has to operate against a large viscous load. The situation is like opening a door with a dashpot, a device that keeps the door from banging shut when it is released. Although it does not take much force to open the door slowly or hold it open, it requires a very large force to move the door rapidly. Thus, there is reason to believe that a significant fraction of the strength we measure in the tongue during a static isometric behavior may be used during normal speech to get the tongue to its proper positions in a timely manner. This conjecture has been given support by the observations of Dworkin et al. (1980). They found very strong negative correlations between maximum isometric tongue force and the severity of articulation defects and maximum rates of /pa/, ta/, and /ka/ in patients with amyotrophic lateral sclerosis.

Jaw muscles are another matter. The power of these muscles so far exceeds the requirements for speech that I doubt that they ever impose limitations on speech movements. I personally regard having a person produce a maximal bite on a device placed between the teeth as a very hazardous procedure, both to the teeth and to the temporomandibular joint. I would strongly recommend against further study of maximal bite forces. The normal values seen would be, in any case, extremely variable.

It is arguable whether lip muscles can be considered muscular hydrostats, but they do contain a great deal of noncontractile tissue (Blair & Smith, 1986), which must impose a significant viscous load for rapid movements. The fact that slowing of facial muscles can affect speech is revealed to anyone who stands in a winter wind very long without a face mask, or goes skiing on a very cold day. Speech can be made intelligible in this condition by slowing the rate, but attempts to speak at a normal rate produce, at least in me, what would clearly be called dysarthria. This effect is, I think, just due to the temperature effect on muscle. Facial muscles are thin and exposed, and their temperatures can drop significantly below body temperature. Cold muscles can still exert force, but their maximal shortening velocity is decreased.

The point of this digression into muscle biomechanics is really very simple. It is technically simple to measure static isometric muscle strength, whereas it is very difficult to measure maximum shortening velocity. But isometric strength is correlated to shortening velocity in a very predictable way, so if we observe a muscle that is weak in an isometric condition, it will have abnormally low shortening velocities with the same load as a normal muscle (i.e., weak muscles cannot move as quickly as strong ones).

There are many reasons why a muscle may be weak. There are diseases that affect contractile mechanisms directly, and diseases of the neuromuscular junction, peripheral nerves, and motoneurons. Another significant source of weakness we seldom think about is loss or absence of sufficient excitatory input to the motoneurons to produce sustained discharge in high-threshold motor units. A person who has suffered the loss of a significant portion of the cortex that provides excitation to the tongue motoneurons, or an interruption of central pathways to these motoneurons, will exhibit a loss of maximal strength of the tongue. To be normally strong, it takes more than strong muscles — one has to be able to fire all the motoneurons to the muscles.

Another source of weakness we may miss is weakness that comes with repeated use, that is, fatigue. Kent et al. (1987) note: “Few, if any, data have been published on these or similar measures of endurance or fatigue although it could be important clinically to make these determinations, especially for clients with neurological disorders” (p. 377). I couldn’t agree more. Fortunately, my colleague Donald Robin suggested we include a measure of endurance (resistance to fatigue) when we began our preliminary study of tongue strength (Robin, Somodi, & Luschei, chap. 13, this volume), and it has provided one of the more interesting results so far. Fatigue, like simple weakness, can have many causes. We are most familiar with fatigue from the loss of ATP in muscle fibers, the “gasoline” that fuels muscular contraction. Physiologists know a great deal about this source of fatigue. But I suspect that some individuals may fatigue rapidly because they simply cannot keep their motoneurons repetitively active. There may be central as well as peripheral causes of unusual fatigability. Accordingly, one of the goals in studies of maximal performance that evaluate fatigue may be to ask “Why?” when we identify a highly deviant individual. Study of these individuals may be useful to our general understanding of fatigue of oral muscles.

Maximum expiratory pressure is another variable that has been studied by investigators, and while there is considerable variation in means reported by different investigators for equivalent groups, the overriding impression is that any normal person has much more capacity for developing air pressure for speech than the person ever normally uses. Nevertheless, some geriatric individuals have been observed whose maximum expiratory pressures are barely above those necessary for speech. Knowledge of this might be quite important for the care of these individuals. While there may be some justification for measuring maximum expiratory pressure, Kent et al. (1987) note that it does carry some health risks, such as pneumothorax and hemorrhages of the nasal mucosa or conjunctiva. Given the data and its clear indication of the reserve of the normal respiratory system, a prudent diagnostic procedure would be to have the individual gradually raise expiratory pressure to 5 kPa (about 50 cm H20). If an individual could not achieve this pressure, it would certainly indicate a significant abnormality of the respiratory system. Thus, in some cases, measurement of maximum performance may not be appropriate, but evidence that a subject can safely achieve a level for normal function may still be informative.

MEASURES OF ORAL MOTOR COORDINATION

The measures discussed test, in effect, the “motors” and immediate control circuits, and are not necessarily good tests of motor speech coordination. I don’t apologize for this at all. A clinician needs to know if those motors are working normally to evaluate a person. I would expect to find many dysarthric individuals for whom simple measures of maximal performance were normal. If it is known that all peripheral mechanisms are normal, then it is reasonable to assign a speaker’s difficulty to problems within the coordinative systems of the nervous system. In this case, a standardized measure of motor coordination of the articulators that was strictly dependent upon operation of intrinsic coordinative neural circuits would be useful. One measure that has been used to assess coordination of the articulators is maximum repetition rate or diadochokinetic rate. Kent et al. (1987) provide an enlightened discussion of this measure, They conclude, basically, that it is not a very useful measure. The appeal is that it is simple to measure: you need a stopwatch, the ability to count, and an “ear” that can detect failure of production accuracy or substitutions. These perceptual judgments, however, introduce a source of variability that one hopes to avoid in the development of objective standards. For this purpose, one really needs spectral analysis of sounds see if they are, in fact, being properly produced. This analysis also allows attention to the variability of both timing and amplitude of the sounds, as illustrated by the study reported by Tatsumi, Sasanuma, Hirose, and Kiritani (1979). I personally, however, would not advocate a measure of coordination that emphasized abnormally high rates of production. Maximum rates of coordinated, practiced, movements are not limited by the nervous system. They are limited by the time it takes to alternately shorten and lengthen muscle fibers. This process is greatly limited by the internal workings of muscles. To achieve high sy1lable repetition rates, even for adequate production, subjects must drive their muscles with “everything they’ve got,” and most subjects probably change their mode of production to a state that has little to do with speech. Amplitudes of movements of slow and massive articulators, such as the jaw, may decrease dramatically. I don’t know whether all people do what I do when I produce /pa/ at high rates, but I never stop voicing; my /pa/s are produced by a steady stream of air through a static larynx. This stream continues during the brief period of lip seal to recharge the intraoral pressure to produce the plosive. Thus, only the labial system is actually moving rapidly. In this case, coordination is greatly simplified, rather than “stressed,” by the rapidity of the task. The pattern of articulator coordination is completely different from normal production of /pA/.

If I were to try to develop a “pure” standard test for motor speech coordination, I would look at the variability of sequencing of articulators, and variability of their amplitude characteristics during repetitions of a simple syllable like /pa/, /ta/, or /ka/, or, if an adult, the trisyllable /pataka/. I would have the speaker time repetitions to a metronome set at a rate of 2 per second, or what is found to be a preferred repetition rate among most normal speakers. One could use elaborate instrumentation, but to keep things simple, I would measure intraoral pressure with a good transducer system (high-frequency response >100 Hz) and a microphone placed at a standard position in front of the lips. Such a high-frequency pressure transducer would allow detection of the onset of voicing from the intraoral pressure record, which may occur in some individuals before lips part for the plosive. An extra-oral microphone alone may fail to detect such an event. From the timing and amplitude relationships of intraoral pressure and acoustic signals and their token-to-token variability, I would predict that one could learn a great deal about the inherent ability of a person to coordinate his or her articulator muscles. I would expect considerable inter-subject variability in the pattern of coordination. More informative, I think, would be variability of the pattern within an individual. Remember that the essence of highly developed athletic skill, which we attribute to excellent motor coordination, is the ability to perform a complex sequence of movements in exactly the same way each time. Consider the tennis serve, the basketball free throw, or the delivery of a bowling ball. An individual’s pattern may be unique but nonetheless very reproducible from one attempt to the next if he or she has normal motor coordination.

CONCLUSIONS OF THE REVIEW BY KENT ET AL.

The major conclusion of Kent et al. (1987) is “that the data base is generally inadequate for confident clinical applications” (p. 384). They go on to call for a “second generation of speech production measures.” Although I am not sure if their plan calls for measures of speech-like responses, with less emphasis on simple behaviors and strength, I think an important point is that they do not call for abandonment of this effort, but for its refinement. I am in agreement with Kent et al. (1987), but I also think we must not leave the impression that the “first generation” was a failure. It has been a good start, and the investigators deserve a good deal of credit. It is worth noting that there are only 20 studies published on maximum phonatory duration, probably the most studied maximum performance measure. By comparison, there were over 100 papers published in 1989 on the pineal gland! Compared to most areas of scientific investigation relevant to human health and happiness, oral strength and nonspeech performance have hardly been touched.

FUTURE EFFORTS

Two words come to mind in thinking about the future of the effort to develop good performance standards in the area of speech motor control. One is technology, and the other is cooperation. Modern instrumentation makes it possible to measure variables with an ease and reliability that our teachers never dreamed of. Applications of technology to motor speech control are well illustrated by the tongue strength instrumentation system developed by Dworkin et al. (1980), and the multipurpose oral performance instrumentation system developed by Barlow and Netsell (1989). These are well suited to the study of dysarthric speakers, and I suspect they and their refinements will become standard in research and teaching. They are not, however, “field instruments,” and widespread application of information gained by normalization of performance standards will be enhanced by development of an inexpensive, easy to use, portable instrument specifically designed for speech pathologists, or anyone who wants to objectively measure certain motor capacities in humans. These two basic approaches to instrumentation (sophisticated, expensive, and fixed location versus simple, inexpensive, and portable) are not inherently in competition. They may, in fact, complement each other in important ways.

Whatever course we take in the development of normative standards, it will be useful to cooperate with one another. For one thing, any one investigator generally will have access to limited populations (i.e., children, elderly, patients with swallowing problems). Another area of cooperation could be in developing procedures for giving instructions and controlling the variables related to practice and motivation. It is too early to call for a convention to develop an “ANSI Spec” for speech-language pathologists, but I firmly believe that day will come.

REFERENCES

Barlow, S.M., & Abbs, J.H. (1983). Force transducers for the evaluation of labial, lingual, and mandibular motor impairments. Journal of Speech and Hearing Research, 26, 616—621.

Barlow, S.M., & Netsell, R. (1989). Clinical neurophysiology for individuals with dysarthria. In K.M. Yorkston & D.R. Beukelman (Eds.), Recent advances in clinical dysarthria (pp. 53—82). Boston: College-Hill Press.

Blair, C., & Smith, A. (1986). EMG recording in human lip muscles: Can single muscles be isolated? Journal of Speech and Hearing Research, 29, 256—266.

Dworkin, J.P. (1980). Tongue strength measurement in patients with amyotrophic lateral sclerosis: Qualitative vs quantitative procedures. Archives of Physical Medicine and Rehabilitation, 61, 422—424.

Dworkin, J.P., & Aronson, A.E. (1986). Tongue strength and alternate motion rates in normal and dysarthric patients. Journal of Communication Disorders, 19, 115— 132.

Dworkin, j.P., Aronson, A.E., & Mulder, D.W. (1980). Tongue force in normals and in dysarthric patients with amyotrophic lateral sclerosis. Journal of Speech and Hearing Research, 23, 828—837.

Dworkin,J.P., & Culatta, R.A. (1985). Oral structural and neuromuscular characteristics in children with normal and disordered articulation. Journal of Speech and Hearing Disorders, 50, 150—156.

Kent, R.D., Kent, J.F., & Rosenbek, J.C. (1987). Maximum performance tests of speech production. Journal of Speech and Hearing Disorders, 52, 367—387.

Palmer, M.F., & Osborn, C.D. (1940). A study of tongue pressures of speech defective and normal speaking individuals. Journal of Speech Disorders, 5, 133—140.

Posen, A.L. (1972). The influence of maximum perioral and tongue force on the incisor teeth. Angle Orthodontist, 42, 285—309.

Tatsumi, I.F., Sasanuma, S., Hirose, H., & Kiritani, S. (1979). Acoustic properties of ataxic and parkinsonian speech in syllable repetition tasks. Annual Bulletin of the Royal Institute of Logopedics and Phoniatrics (Tokyo), 13, 99—104.