IHESis ’— V LIB RA RY Michigan 5“!th University ‘ \ flat This is to certify that the thesis entitled The Effect of Surface EMG Biofeedback on Muscular Tension and Stuttering Behavior presented by Mangala Sadasivan has been accepted towards fulfillment of the requirements for M.A. Audiology G de co in , gr Speech Sc1ences T)‘ x FGLVE, (3L . (bx—k0 Major professor Date \{fumww 33) l°\2l 0-7 639 - _- - - vw-h— - [Will WNW/ill? W: 25¢ Per day per item RETURNING LIBRARY MATERIALS: Place in book return to remove charge from circulation records THE EFFECTS OF EMG BIOFEEDBHCK TRAINING ON MUSCULAR.TENSION AND STUTTERING BEHAVIOR By MANGALA.GOWRI SADASIVAN A.THESIS Submitted to Michigan State University in partial fulfillment of the requirenents for the degree of MESTER.OF{ARTS Department of Audiology and Speech Sciences 1981 \ematw ABSTRACT 'IHE EFFECIS (F M BIOFEEIIBACK TRAINING (N MUSCUIAR ‘IENSIG‘I AND STUTI'ERING BEHAVIOR By Mangala Gcwri Sadasivan 'lhis study uses a single subject research design to shcw the effects of biofeedback training on muscular tension and stuttering behavior. Such data can assist in developing a ncre efficient bio- feedback therapy program for the treatment of stuttering. Four stutterers, were assigned randomly to two groups: S-F, treatment of specific site followed by frontalis training; or F-S, treatment of frontalis training first followed by specific site nuscular training. A lZ-sessicn therapy program (six 20-minute sessions per site) was inplemented to train the subjects to control muscular tension using visual and/or auditory feedback. The data collected consisted of EM; readings for non-speech and speech tasks and the dysfluency percentage during pre-therapy, post-therapy, pretest and pcsttest within sessions for each muscle site tested. The results frcm this investigation indicated that a reduction of excessive muscle tension in the speech musculature is a basic factor in reducing stuttering behavior. ACKNCWIEIXZEMEN’IS I wish to express my appreciation to Dr. Paul Cooke, my thesis director, and Dr. Richard chpel and Dr. Leo Deal, the members of my thesis committee for their invaluable assistance and criticism in the preparation of this thesis. I also wish to thank Dr. Michael Chial and Mr. George Gamble for their expertise in setting up the instrumentation for the experiment. In addition, I am indebted to the four subjects who participated in this experiment, for their time, co-cperaticn and patience which made this experiment possible. Finally, I wish to extend a personal thank you to my family and frieids for their support, encouragement and understanding during my entire Master's program. ii TABLE OF CONTENTS Page LIST OF FIGURES ............................................. iv GBPtP—r l. INTKJDUCI'ION: REVIEWOFIJ'IERA’IURE ............... l 2. METHOD ............................................ 7 Selection of Miecles ... 7 Electrode Placement 10 Data Acquisition 14 3. RESULTS ........................................... 15 Subject F-Sl 16 Subject F-SZ ..... 19 Subject S-Fl .. 22 Sdoject S-F2 .. ..... 25 4. DISCUSSION ........................................ 29 iii Figure 1 0 LIST OF FIGURES Page Schematic drawing illustrating muscles of the face observed during the diagnostic session. 8 Schematic drawing illustrating muscles of the neck and larynx observed during the diagnostic session.- 9 Block diagram of the instrumentation used during this experiment. 11 Subject F-Sl's EM; data and the dysfluency percentage for pre-therapy, post-therapy, pretest and pcsttest within sessions for frontalis muscle training with- out biofeedback . The lines represent the biofeed- back training interval. 17 Subject F—Sl's EM; data and the dysfluency percentage for pre-therapy , post-therapy, pretest and pcsttest within sessions for crbicularis cris muscle training without biofeedback. The lines represent the bio- feedback training interval. 18 Subject F-SZ's EM; data and the dysfluency percentage for pre-therapy , post-therapy, pretest and pcsttest within sessions for frontalis muscle training with- out biofeedback. 'Ihe lines represent the biofeed- back training interval. 20 Subject F-SZ's EM; data and the dysfluency percentage for pre-therapy, post therapy, pretest and pcsttest within sessions for crbicularis cris muscle training without biofeedback. 'Ihe lines represent the bio- feedback training interval. 21 Subject S—Fl's EM; data and the dysfluency percentage for pre-therapy, post-therapy, pretest and pcsttest within sessions for geniohycid muscle training with- out biofeedback. The lines represent the biofeed- back training interval. 23 iv Figure Page 9. Subject S-fl's EM; data and the dysfluency per— centage for pre-therapy, post-therapy, pretest and posttest within sessions for frontalis muscle training without biofwdback. The lines re- present the bicfeedback training interval. ----- 24 10. Subject S—FZ's EM; data and the dysfluency per- centage for pre-therapy, post-therapy , pretest and pcsttest within sessions for orbicularis cris muscle training without biofeedback. 'Ihe lines represent the biofeedback training interval. 26 11. Subject S-FZ's EMS data and the dysfluency per- centage for pre-therapy , post-therapy, pretest and pcsttest within sessions for frontalis muscle training without biofeedback. The lines represent the biofeedback training interval. -—--- 27 Chapter I INITOIIJCI‘ICN: REVIEW OF LITERATURE Numerous references in the research literature on stuttering are directed toward the hypothesis that stuttering is the result of ex- cessive physiological tension. Many researchers are of the opinion that the stuttering block is accompanied by a spasm cf the laryngeal muscle (Van Riper, 1971; Schwartz, 1975). Based on this hypothesis, Hanna, Wilfling, and McNeill (1975) , in a single subject biofeedback treatment for stuttering, showed that laryngeal tension and stuttering were in- timately related. Shrum (1967) measured surface electrical activity of facial, neck and chest muscles in stutterers. He observed a relationship between increases in EMG signal amplitude prior to speech and the sub- sequent stuttering. Blocdstein (1975) claims that for most stutterers the act of blocking is associated with some feeling of strain or tension. 'nnis is usually localized in the speech musculature, i.e., the muscles of articulation, mention, or respiration, but in some cases this tension may also be manifest, elsewhere in the body, for example, the arms, legs, or shoulders . Stuttering therapy has for a long time focused on re— ducing this physiological tension by using relaxation techniques. In more recent times , biofeedback has been introduced as an effective means of reducing sure of that physiological tension. Bicfadback is a method of learning to control body processes that ordinarily cannot be regulated volitionally. The concept of biofeedback is based upon three basic principles. First, an individual is able to assume control over a neurological or biological function if that function can be monitored or 2 amplified by electronic instruments and fed back to him. Second, every change in the plnysiological state is accarpanied by an appro- priatechangeinthementalerotional state andviceversa. 'Ihethird principle is that a meditative state of deep relaxation is conducive to voluntary cantrol of body processes because it allows the individual to became aware of internal imagery and sensations. 'lhe amount of literature pertaining specifically to the use of electramycgramic (D’s) biofeedback in the field of speech, particularly stuttering, is very limited. For this reason, articles dealing with other clinical problems related to tension reduction procedures using EMS biofeedback will also be reviewed. Studies have shown that EM; ‘ biofeedback may be used successfully for the treatment of several dis- orders including high blood pressure, digestive disturbances, irregular heart beat, migraine and tension headaches, and chronic anxiety. Canter, Kandc and Knott (1975) did a carparative study using anxiety reurosis as the dependent variable and two different methods of training in deep muscle relaxation. 'Ihese included biofeedback procedures arnd progressive muscle relaxation techniques. 'Ihey concluded from their experiments that EM; feedback was generally superior in producing larger reductions in muscle activity, with concomitant relief in anxiety symptoms. Acosta et a1. (1978) applied EM; biofeedback to the relanation training of schizophrenic, neurotic and tension headache patients . lbsults from their study indicated that patients did exlnibit a reduction of their frontalis muscle tension level across the ten successive biofeedback training sessions. 'Ilney further noted that learning to reduce tension through EM; biofeedback was applicable to individuals with diverse backgrounds . In addition, Raskin et a1. (1973) and Townsend and House (1975) also found 3 EM; biofeedback effective with tlne treatment of chronic anxiety patients. The frontalis muscle has been frequently used as a popular site for electrode placement in various experiments. Canter et a1. (1975) ex- plained that the frontalis muscle was chosen as the target for feedback training in their experiment "since this muscle has been slncwn to reflect the general muscle tension levels in anxious patients". Alexander et al. (1975) compared the efficacy of auditory and visual feedback in EMS bio— feedback-assisted relaxation training of the frontalis muscle. The authors found that biofeedback training with auditory feedback produced significant decreases in frontalis ENE, whereas training with visual feedback was in- effective. Haynes, Miseley and McGowan (1975) assessed the carparative effectiveess of frontalis EMS biofeedback and relaxation instructions in reducing frontalis EM; levels. Results from the one-session design deronstrated that the greatest decretent in frontalis Em level was attained when subjects received biofeedback as opposed to other techniques. The authors indicated that previous research suggest that frontalis activity may be correlated with muscle tension levels in other areas of the body. Many reserachers have shown that reduction of tension of specific speech muscle groups using biofeedback techniques is accompanied by decreases in stuttering frequency. 'Jhe muscle groups chosen for electrode placerent were specific for each individual case and included the orbicularis oris superior (lip), the anterior belly of the digastric (chin) . laryngeal muscles, the frontalis muscle (forehead), masseter muscle (jaw), and other muscles of the face, head, and neck. Guitar (1975) trained three adult male stutterers to reduce resting muscle activity by using analog EM; feedback from four sites over different 4 muscle groups. He then trained his srbjects to reduce muscle activity prior to uttering selected sentences. A functional analysis of the re- lationship between the decrease in stuttering frequency on the initial [inmates and the reduction of electrical activity at each muscle site in- dicated that fine srbjects had different responses to the training. One subject's greatest decrease in stuttering frequency was associated with muscle activity training at the lip site; the second subject's greatest decrease was associated with training at a laryngeal site; and the third srbject's decreases were related to both lip and laryngeal site training. Based on these results , the author designed a practical management program for a fourth stutterer. Feedback training to reduce EM; activity when paired with speech resulted in a reduction of stuttering in two monitored situations, conversations and telephone calls. Probes indicated that stuttering continued to be markedly reduced in all situations nine months after treatment. Alexander (1975) designed an EM; fwdback program to allow stuttering snbjects the opportunity to perceive normally subliminal physiological events. The program trained subjects to utilize feedback of muscle po- tentials to control the amount and occurrence of tension in specific muscles. 'nnemuscleschosenwerethose judgedtobemosttenseduringthe stuttering mment. Stbsequently, the program was designed to teach the subjects to apply the learned muscle control to speech tasks to reduce tension prior to, during and immediately following the speech event. Finally, the program was designed to facilitate transfer of the newly acquired flLent status into tlne srbject's ncnclinical environment and to maintain fluency in that environment. Ibsults indicated a reduction of the individual nonfluencies and a reduction of secondary characteristics not nonmally related to speech. Ianycrn, Barringtcn and Neaman (1976) conducted a preliminary study in which eight stutterers. spent 10 to 18 one-hour sessions learning to relax their masseter muscles using EM; visual feedback. Major reductions in stuttering were demonstrated for all subjects in reading one-syllable words while feedback was present as carpared to no feedback. Similar results were obtained with six subjects who progressed through the reading of unto-syllable words, three-syllable phrases, and four-syllable sentences. Lanyon (1977) developed further metlncdology for the use of relaxation in the modification of stuttering, based an the general hypothesis that excessive tension in speech-relevant muscles was a key carpcnent in stuttering. He showed through replication and extensim of the previous study, that the voluntary relaxation of the speech muscles could be learned with the aid of ENE visual feedback. Results indicated that subjects could successfully be taught to rely on internal cues to generalize fltent speech and relaxation behaviors to periods when feedback was absent. The reports cited above depict a very limited number of cases in which reduction of tension of speech muscle sites using EMS biofeedback techniques has been accanpanied by decreases in stuttering frequency. Pore research is nwded to determine the variables that must be controlled if favorable clinical results are to be achieved. Research Questions 1. What effect does EM; biofeedback training have on the tension level of the frontalis muscle of stutterers? 2. mat effect does Em biofeedback training have on the tension level of a specific muscle site of stutterers? 6 3. Is there a transfer effect of ENE frontalis biofeedback training from sessions with feedback to sessions without feedback? 4. Is there a transfer effect of EMS biofeedback specific site training from sessions with feedback to sessions witlncut feedback? 5. What effect does EM; relaxation training of the frontalis muscle have on stuttering behavior? 6. What effect does EM; relaxation training of the specific site have on stuttering behavior? Chapter II moo SLbjectS Four male stutterers between the ages of 18 and 35, who had previous 1y enrolled for stuttering therapy were selected following diagnostic evaluations conducted by the experimenter. Only subjects who exhibited high observable facial and/or laryngeal tension during the initial interview could volunteer to participate in the experiment. The subjects ranged in severity from mild to severe, based upon ob- servations made by the experimenter during the diagnostic and therapy sessions. The subjects were assigned randomly to one of two grows: S-F, treatment consisting of specific site training followed by frontalis training; or F—S, treatment consisting of the frontalis training first followed by specific site muscular training. Selection of Muscles TWo muscles were selected for biofeedback training during this experiment. Figures 1 and 2 provide schematic drawings of the muscles of the face, neck and larynx that were used during the diagnostic interview. These included the medial frontalis, temporalis, zygomatic, levator, orbicularis oris, masseter, mentalis, buccinator, geniohyoid and the criccthyroid. One of these muscles was chosen based upon the tension level that the experimenter observed during stuttered speech. This is referred to as the specific site placement and was selected on an in- dividual basis. mm. 1 mos nus E%;—-+TEMPORALIS ZIGOHATIC LEVATDR DRRICULARIS ORIS Figure 1. Schematic drawing illustrating muscles of the face observed during the diagnostic session. Figure 20 ...—... BUCCINATCR THYROHIOID Schematic drawing illustrating muscles of the neck and larynx observed during the diagnostic session. 10 The second muscle, which was trained with all subjects, was the frontalis. The frontalis muscle was chosen as the nan-specific site since research has suggested that tension reduction of this muscle is accatpanied by general body relaxation (Canter et a1. , 1975) . Electrode Placarent The subjects were seated in a comfortable dnair with arm rests in a sound treated chamber. Bipolar ENE activity fram eacln muscle region was recorded with a pair of mniniature skin surface electrode placed in close proximity to each other. A neutral electrode was placed on the nearestbcnetoserveasaground. Theskinwaspreparedbyfirst ribbing the area with a disposable skin cleanser swab formulated to remove in- sulating elements from tlne skin surface to minimize skin resistivity. The skin was allowed to dry while the electrodes were carefully filled with Beckman electrode paste and attached with adlnesive collars over the muscle to be tested. Instrumentation Figure 3 presents a block diagram of the instrumentation used in sensing the raw EM; signal, anplifying and transforming the signal for visual and auditory feedback purposes, and converting the signal to a nunerical equivalent for analytical purposes . The muscle being tested produced an electrical signal whose arplitude is proportional at each instant to its tension. This signal was picked w by EM; electrodes, preamplified (Hewlett Packard 1510B) , and then transferred to an operational arplifier, which increased the signal level by a factor of 10. The anplified signal was then rectified and integrated using a contour ll SURFACE EMG ‘ l [ OSCILLOSCOPE I ELECTRODES SPEAKER , I l [ PREAM’LIFIER AMPLIFIER l I RECTIFIER [ INTEGRA'IDR VOLTAGE I use“; 1.1%: - r Figure 3. Block diagram of the instrumentation used during this experiment. 12 following integrator (Ooulbourn 576-01) . Rectification is the process by whidn the amplified altnerating current (AC) is converted to an unidirectional current (DC). The integration of the signal produces a relatively smooth curve whose amplitude at any point is proportional to the peak arplitude of the pulsating DC irnput. This smoothing function can be controlled by choice of an integrater time constant, whidn for this experiment was set at 500 msec. A voltage controlled oscillator (Oculbourn 824-05) accepts the smoothed DC signal and converts it to a freqrency-modulated sinusoidal wave and a synchronous frequency-modulated sqmre wave. The former was visually displayed by means of an oscilloscope and simultaneously transmitted tlnrough a loudspeaker as a pure tone by means of an audiomixer anplifier (Coulbourn $82-24) . Both of these displays (auditory and visual) were used by the subject as biofadback signals of his muscular activity. As the muscle became more relaxed (i.e., a decrease in electrical activity), the pure tone became lover in pitch and the oscillo- scopic display indicated a smaller nurber of sinusoidal waves per unit of time. The square wave was converted to a numerical value by a digital counter (Cbulbourn Rll-Ol) which produced an output equal to the number of pulses counted within a given period. These valtes were used, in conjunction with a calibration factor derived by a method described in the "data acqui- sition" section, to determine the average mnicrovolts (uv) emitted by the muscle over a period of time. The sampling interval used in this experi- ment was 60 seconds and was manually initiated by a switch module (Coulbourn 13 Procedure The four stutterers were assigned randomly to one of on grows: S—F, treatment consisting of specific site training followed by frontalis training; or F—S, treatment consisting of the frontalis training first followed by specific site muscular training. 'Ihe experiment consisted of a pretest, six training sessions and a pcsttest for each of the two muscle sites investigated. The pretest and pcsttest involved measuring muscular activity at the particular site during nonspeech and speech tasks without biofeedback training. Each pretest or pcsttest session was corpleted within a ten- mninute interval. The nonspeech and baseline segrent consisted of three consecutive one-minute intervals . The subjects were instructed to keep their lneads still and to avoid unnecessary moverents in order to minimize experimental artefacts. The speedn segment corpcsed of the subject speaking for three one-minute segnents . Open-ended general topic questions were asked by the experimenter to stimulate the speech activity. The six training sessions for each of the muscular sites was approximately 45 minutes in length. The first and last ten minutes of each session involved a pretest and pcsttest segnent as described above. The middle twenty mninutes of the training session focused on FM: bio- feedback therapy. During the training sessions the subject was instructed on the association between the variations in the auditory and visual signals and muscular tension. The subject's task was to try to maintain a low frequency tone from the loudspeaker and/or elongate the sinusoidal wave on the oscilloscope. During the therapy sessions, each subject con- trolled the type of biofeedback received, (i.e., the subject had the option to look away from the visual display or to vary the arplitude of the l4 auditory signal) . Throughout the therapy session the experimenter used open-ended questions to elicit the maximum amoimt of continuous speech. Four one-minute sarples of EMG readings while speaking were taken at intervals of approximately five minutes. Data Acquisition The EM; microvolt readings were computed using the following cali- bration procedure. Known signals of O, 20, 40, and 60 mnicrovolts were introduced into the input of the system to produce an equivalent electronic counter reading for a 60 second period of time. The counts collected were divided by 60 to obtain tie ccunts-per-second average and these values were plotted as a function of the input signal. A straight line approxi— mation (y = mx + b) was applied to these data. The readings recorded throughout the experiment were substituted into the equation to convert electronic counter readings into amplitude values of the EM; signals. The percentage of dysfluency presented for each subject was calculated using Wingate's (1964) definition of stuttering as a guide to determine the occurrence of a stuttering block. The term stuttering focused on the disruption in the flLency of verbal expression by repetitions and pro- longaticns. Sarples of speech collected during eacln pretest and pcsttest session were transcribed, and only the first two minutes of continuous speech for each session was used in the analysis. The percentage of dysfltency was obtained by dividing the number of dysfluencies counted by the total number of words uttered during the two—minute sample of speech and multiplying the result by 100. Chapter III RESULTS The data collected on each subject during this experiment are described individually. This information is displayed in two figures , one for each muscle site investigated. Each figure presents the percentage of dysfluency and EM; data for a pretesting and a posttesting period across the six therapy sessions. T\wo different y-axis scales were required to represent the data in this fashion. The left hand scale is for the EMG values measured in microvolts, which are symbolized by an "x" for speech tasks and by an "o" for non-speech (resting) conditions. The right hand scale represents the percentage of dysfltency and are symbolized by a "A". Each pretest and posttest session incorporated EM; recordings during non- speech and speech tasks without the presense of biofeedback. Lines are drawn between the data obtained during pre- and post- testing conditions indicating that biofeedback training had occurred during this interval. In addition, pre-therapy and post-therapy readings were obtained. A note of caution must be interjected regarding the interpretation of these figures. Muscle regions differ with respect to the absolute magnitude of activity that can be presented by the underlying muscles. Thus, the total range of activity recorded from the surface electrodes will vary from region to region as well as from individual to individual. 15 16 Subject F—Sl Figure 4 illustrates the data from object F-Sl when the frontalis was the training muscular site. From pre-therapy to post—therapy testing his EMG activity while resting reduced from 12 AN to 6 an. In addition, this subject consistently reduced his EM; activity for speech during the therapy session, with the exception of session 2 where the frontalis activity increased slightly from 36 to 38 in. However, any carryover from the biofeedback training for speech purposes was not apparent, since the pre-therapy value of 15 nminoreased by 12 to 27 in. during the post-therapy measurenent. Thus, the reduction of dysfltency from 28% at pre-therapy to 7% during post-therapy testing cannot be accounted for by the reduction in frontalis activity. This is supported by the .37 corre- lation that existed between EMG activity and the dysfltency percentage. Figure 5 displays data from F—Sl wlen the orbicularis cris muscle was being trained with biofeedback . Comparing pre- and post- therapy conditions, the resting muscular activity demonstrated a sliglnt increase from 13 to 16)1V., while the activity during speech reduced 6%, from 90 to 85 m. This relatively modest change is indicative of the inconsistent pre- and pcsttest valtes obtained at each therapy session. The subject demonstrated slight reductions in ENE activity during sessions 2 and 4 and a sharp decrease in orbicularis oris activity (112 to 79 nv) during session 3. However, modest increases in the electrical activity were reported during the other three training sessions . The correlation between the muscular activity and the dysfluency percentage was .21 across the six therapy sessions. Such a correlation was a result of inconsistent associations from session to session, as the values for the first three . 17 DYSFLUENCY PERCENTAGE OH me om mm om mm ow / .Hm>pouce ocecwmsu xomoooowoeo one pcomouooe mocwfi one .xomoooomoeo escape: ocecemeu oHomse mwfimucopm how chwmmom canoe: umouumoo ocm umoposo .aomeomu -pmoo .Aomuocu-opo pom couscoouoo aococHmeo one use some 02m m.Hm-m uocnocm .e unseen mHAmx OH mm om mm om mm ow "n'OHS 18 DYSFLUENCY PERCENTAGE OH me om mN om mm ov me .Hm>houcH academy“ xomoooowoeo one snowshoes mocaa one .xomnoooMOMo unocoez ocecemsu ofiumoa mwho mfiemacoeneo sow mcoemmom cease: announce one unmoved .zomeoau -umoo .xomeoco-oso how ommucooeoo accosfimmxo new one must 02m m.Hm-m somehow .m apnoea mHoo mHe253?er 4 i 9: >m¥ m? ems l9 therapy intervals deronstrated a direct association between the two variables, whereas the values for the last three sessions were inversely related. Subject F—SZ Figure 6 represents data of subject F-SZ Men the frontalis was being trained. For five of the six therapy sessions the subject demon- strated a consistent pattern. In those sessions the EM; baseline data was always higher than the ENS speech data during pretest but were re- versed during the posttest evaluation. Session 5 was the exception with the EM; baseline being lower than the ELF/I3 activity during both pretest and pcsttest conditions. Thus, while the biofeedback training did result in lower frontalis activity for the baserate condition by the end of the session, it increased the EMG activity during speech. In fact, the electrical activity during post-therapy for speech, after six therapy sessions had been completed, was 5 any higher than before biofeedback training was initiated (8 AN. corpared to 3117. ). A moderate correlation of .52 occurred between the frontalis activity for speech and the percentage of dysfluencies . The arount of dysfltency decreased across the six biofwdback sessions from 17% to 12%. Figure 7 displays data from subject F-SZ during orbicularis oris training. The non-speech data were consistent throughout the therapy sessions, with a slight increase in muscular activity at the termination of therapy when compared to the initiation of the session. The largest increment of this nature occurred during session 4 as the pretest ENG non-speech recording was 4hr. corpared to 14m. during the posttest interval. ‘ The most notable change that occurred througlnout the orbicularis 20 DYSFLUENCY PERCENTAGE 0H 2 ON mm om mm oe .Hm>houcH mchHmep MomnooomoHo new ucomoeoou mocHH one .xomoomomoHo uconqu wchHmHu oHomsE mHHmocoem How mconmom cHnqu umcuumoo ecu umopoeo .zomeonu -umoo .aomumnu-ouo pow ommucoouoo accocHmet one one some use m.Nm-m pcomocm .o unseen quHoucH mchHoHu xoonooomoHo one ucomouoop mocHH one .xoonoooono Hoozqu ochHoHu oHomoE mHHo mHHHoHsoHan How mconmom chsz umopumoo oco amououo .xoononu -umoo .xoohozu ohm How omoucoohoo accousmao cow was moot 02m m.Nm m econosm .5 ouowHo mHoo mHm25onon < I >mx 1 oH ON om oo on ow oo ooH. OHH QNH omH ovH omH "u'DHE 22 oris training was tlne sharp reduction of the EMS activity for speech. Although the EM; activity during pre- and post- testing evaluations were highly variable, the pre-therapy session was 134 in while the post— therapy reading was 102 m, a reduction of 24%. This was consistent with the decrease in stuttering behavior, from 17% before therapy to 11% after the biofeedback sessions. Across the entire therapy program, a correlation of .45 existed between the orbicularis oris activity and the percentage of dysfluency. SuEiect S—Fl Subject S—Fl was the only individual for whom a muscle other than the orbicularis cris was used for the specific site biofeedback training. The geniohyoid was the muscle selected and the data are displayed in Figure 8 . For both , the tension levels were higher for the geniohyoid during posttesting compared to pretesting. For the baseline measures, this occurred in five of six sessions (the exception being session 3); and for speech activities an increase occurred in four therapy sessions. 'Ihe effect of this during the post-therapy session was consistent. The EM; activity for baseline had increased from 7 in before the onset of therapy to a level of 20 AW after biofeedback training had terminated. The electrical activity during speech, however, was markedly lower, starting at 112 m and dropping down to 70 in! at posttesting evaluation. The stuttering behavior similarly decreased in frequency as demonstrated by the 38% to 36% reduction from pre- to post-therapy testing. Figure 9 represents data from S-Fl during frontalis biofeedback training. While the therapy consistently reduced the EMS activity by 2;." during baseline or resting conditions, the muscular levels were higher at 23 DYSFLUENCY PERCENTAGE .Ho>poch ochHopp xooooooHoHo opp pcooopoop oocHH one .xoonoooonn pcoanz ochHopp oHoooE oHoanoHcom pom ocoHoooo canHz pooppooo too poopopo .xoopocp -pooo .xooponp-opo pom ooopcoopoo accoonoao ocp oco opoo Ozm o.Hm-m poonnsm .w opsmHm o o m m QHOEiOHzmu ”Bum UHmHumom e e m. m N N H H aoopocp poop poop poop poop .poop poop poop poop poop poop poop poop zooponp OH mH... ON: mNu omr mmw men on Jumom apnbl -DHQ -HnDl . H J 4 ,- I cooooococ 02m O 58% 02m X pcoopoo AccooHuHobo 4 E -o o loco -opo -opo OH We O0 ON. Om OO OOH OHH ONH OmH O: OmH um ma 21in DYSFLUENCY PERCENTAGE OH mH Om mN Om mm ow .Ho>poch mchHopp HooooooHoHo onp pcooopoop oocHH one .xoonooomoHo pooan3.wchHopp oHooos oHHopcopm pom ocoHoooo canHz pooppooo poo poopopo .zoopocp -pooo .aoopocp-opo pom omopcoupoo zucooHMoxo omp too opoo 02m o.Hm-m poonoom mHnme q\q vmillion // -opo (HLIIIIJ_ O ... OH mH ON mN om mm OV mm mm 25 the end of a trerapy session compared to the beginning. This increase of muscular activity was reflected in the pretherapy and post-therapy speech data. Before biofeedback training was initiated, the frontalis activity was 11 J17- compared to the level of 16 in after the six therapy sessions . However, the percentage of dysfluency behavior decreased over this sate period, from 33% to 28%. Therefore, it is not unexpected that the correlation between EM; speech levels and the percentage of dysfluency was .27. Subject S—FZ Figure 10 displays data of subject S-F2 during orbicularis oris bio- feedback training. The figure displays that the therapy sessions resulted in drastic reductions of muscular activity while speaking. Decreases of 40 11v within a particular session was comonplace. In addition, corparing the EMG activity for speech at all the pretest conditions illustrates a systeratic decrease of the muscular activity across time. This is re- flected in the corparison of the pre—therapy to post-therapy values which indicates a decrease during speech from 143 in to 102 m. The other two measures, EMG activity during baseline recordings and the dysfluency percentages were relatively stable througnout the course of the bio— feedback training. The data from S-F2 during frontalis muscle training is focused in Figure 11. His at levels for speech and non-speech were consistently the lowest of all the subjects. Slight decreases in the three measured variables were observed across the sessions. The baseline muscular activity was reduced after every biofeedback session corpared to the pretest condition. Such a training effect is evident by the decrease in the baseline EMG 26 DYSFIUENCY PERCENTAGE .Ho>poch wchHopp xoooooomoHo ozp pcooopoop oocHH one .HoonooomoHo pooanz wchHopp oHoooE oHpo oHpoHooano pom ocoHoooo canH3.pooppooo too poopopo .xooponp -pooo .xoopozp-opo pom owopcoopoo zucoonoxo onp ecu opoo 02m o.Nm-m puonoom .OH opoon mHmo mHHZHOUHmmo 5.3m UHmHummm o m m e e m m N N H H NAQHwHG H “mm“ Hmmu HmOH Hmmu HmQH Hmong. HMOH Hmmu. Hmmu. Hmmu. Hmmu 0mm“. \Anmmhbfiu. IHNm n. . n ... u. . n ... n. . n ... ... .. a on. :0 Q u m& Ive:— anQ - I O O I I 0 0/0 . . . q 4\m fl QIIO 0 OH 1 l mH I 4 ON 1 mN r om I. on 1 mm .. a oe.1 :ooooococ Him 0 .. nooooo uzm X me r owopcoopoo HocooHMoxo 4 J Om E NH x ON Om Oo ON. Ow Oo OOH OHH ONH OmH OH OH ’4“ on 27 DYSFLUENCY PERCENTAGE OH mH ON mN om mm ow .Ho>poch mchHopp xoonooomoHo ocp pcooopoop oocHH one .HooooooHoHn poocsz mchHopp oHoooe oHHopcopm pow occHoooo cHszz pooppooo oco poopopo .zoopocp -pooo .xoopocp-opo pom ooopcoUpoo xocoonozo onp oco opoo ozm o.Nm-m puonocm .HH opoon mHHm¥ OH mH ON mN Om mm OQ 'Atf 9N3 28 reduction from 5 an to 1 117 at pre-therapy and post-therapy intervals. TheEMG for speechwas reducedto 3nvby session 4 and renained there throughout the remaining therapy sessions. Chapter IV DISCUSSIOT Excessive physiological tension is associated with many maladaptive behaviors . As a result, researchers and clinicians have erployed various techniques to reduce or control sore of this muscular tension, including hypnosis, meditation and relaxation (Barber, 1963; Null et al., 1974; Jacobson, 1938; Wolpe, 1969) . Stuttering is a corplex speech disorder that frequently results from or causes excessive physiological tension. The most widely accepted therapeutic program which focuses won reducing physiological tension in stutterers is systeratic desensitization using relaxation as the reciprocal irnlnibition to anxiety (Adams, 1972; Burgraff, 1974; Gray and England, 1972) . These techniques are aimed at reducing general body tension. Altlnough stutterers may have higher overall tension lemls than non-stutterers during speech activities , each stutterer has a specific pattern of excessive physiological tension. That is, particular muscle grows may exhibit more obvious abnormalities than other muscle sites. Recently, muscular biofeedback has been introduced to reduce tension fromcertainareas inthebody. Ithasbeenfoundtcbean effective and efficient means of procuring relaxation (Haynes, et al., 1975; Thinking et al., 1975; Coursey, 1975) . These biofeedback procedures have been used occasionally in atterpts to reduce stutterers' muscular tension (Guitar, 1975; Alexander, 1975; Ianyon, 1976). Based upon the limnited research in this potentially useful therapy for stutterers, this experiment 29 30 was designed to explore the effectiveness of EM; biofeedback training to reduce muscular activity and stuttering behavior. Generally, the results from this investigation are consistent with previous research in that the reduction of excessive muscle tension in the speech musculature was a basic factor in reducing stuttering behavior (Guitar, 1975; Lanyon, 1977). All four subjects in the present study denonstrated a reduction in electrical activity and stuttering behavior after biofeedback training in one of the two muscle sites. For subjects F—SZ, S-Fl, and S-F2 the therapy sessions involving specific muscular sites produced the most positive effects, whereas the best results for subject F-Sl occurred during the frontalis sessions. As in all of the previous EM; biofeedback studies , the therapeutic gains varied dramatically . This is denonstrated by the pre- and post- biofeedback dysfluency per- centages from the muscle site that was most beneficial F-Sl (28% to 7%) , F—SZ (17% to 11%) S-Fl (38% to 36%), S-F2 (6% to 4%). It should be noted that F—Sl, who denonstrated the most drastic dysfluency reduction across the biofeedback sessions, was the only subject of the four for whom the frontalis training was beneficial. Examining the data reveals that F-Sl was the only subject who denonstrated frontalis EM; readings between the 20-40 nv. range during speecln activities. The other three srbjects had fluctuating frontalis EM; recordings from 5-15 av. Thus, while the frontalis bio- fwdback training sessions were not beneficial for those illustrating low levels from this site, for subject F—Sl the frontalis muscle was the site that had He greatest potential for tension reduction. The dysflLency reductions during the most beneficial biofeedback site are slight (except for F-Sl) compared to other therapeutic programs. There are several factors that may account for this. Cmulative effects of 31 biofeedback training were noted from the objects. That is, there was little, if any, carryover of reduced muscular activity when the experiment changed from one muscle site to the next. Thus, any therapeutic gains that occurred during muscular training from one site were not beneficial during the training of the second muscle site. These results are con- sistent with other EM; biofeedback studies indicating that generalization from such tlnerapeutic procedures is difficult (York, 1977; Alexander et a1, 1977; Iegewie et a1. 1975) . Thus, the present behavior changes must be evaluated in terms of single-site training (i.e., six 20-minute bio- feedback session). In addition, spontaneous speech was the only speech behavior that was evaluated in this investigation. This represents the most corplex speech activity and therefore is usually the most difficult for stutterers . Otlner therapeutic programs frequently report behavioral data indicating dramatic changes in speech utterances. that are less com- plicated, such as words, short phrases, short sentences, and reading. Therefore, considering the high level of communicative skills that were under investigation and the limited amount of training (two hours), the reduction of dys fluency across the therapeutic setting was encouraging. Therewere soredata thatwerenmexpected. Itwas assured that the sninject would exhibit a greater degree of relaxation (less electrical activity) during the baseline recordings than during conversational speech , where the muscle being monitored would play an active role. However, subjects F-SZ and S-F2 demonstrated the reverse pattern on several occasions. This is illustrated in Figure 6 during the pre-test conditions of sessions 1, 2, 3, 4, and 6 and in Figure 11 during the pre-test sessions 2, 3, 5, and 6. The most logical explanation concerns the order in which the baseline and speech activities were measured. airing the pre-test 32 the baseline EMG level was recorded first and‘was fOllowed by the measurement of the EMG activity during speeCh. It is very probable that.because the baseline level was always measured first, any overall arousal level that the sUbject had entering the therapy session would get reflected.in the first value obtained. Therefore, if the individual had.been anxious during the day, the higher arousal level might have been apparent during the first few minutes of baseline recordings. By the time the EMG speech data were taken (about 10 minutes after the sUbject had entered the roomo, the individual may have become generally more relaxed, especially since that.was the general goal of the therapy sessions. There- fOre, on these few occasions fromnthese two subjects, the speech EMG data were lower than the baselineemeasure that.was obtained several minutes earlier. This explanation is the most.plausible since this reversal of Emfiivalues only occurred during pretesting and NEVER.occurred.afte£_the biofeedbadk session was completed. That is, the post-testing data always illustrated that the subject had higher FIG levels for speech compared to nonrspeech. .A:more perplexing phenomenon occurred concerning the EMG levels during speech activity. Three of the fOur subjects, to varying degrees, demonstrated greater muscular activity during conversational speech after, biofwdback training compared to pre-therapy levels. That is, the data suggest that the subjects became more tense as a result of individual sessions and after the entire six session therapy programrwas completed. subject S-Fl's data during the frontalis féedbadk training illustrate this point in Figure 9. The post-therapy EM; level was 16;“ . compared to 11 nv. level befiOre any training had.commenced. In addition, during five of the six therapy sessions (the exception is session 5) the posttest 33 EM; activity was higher than the pretest level. Since this type of data has not been illustrated in other biofeedback studies and since the snbjects reduced their dysfluency percentage across the therapy programs, factors within the experimental designs were reviewed for possible ex- planations. The biofeedback signals were controlled by the subjects. They could receive visual feedback and auditory feedback simultaneously and irncrease the amplitude of the auditory signal. Subject S-Fl, whose data most exemplified this reversal of EM; levels across biofeedback training, indeed increased the audio signal during the feedback sessions. This subject reported that he did not have control of the EM; activity unless this high amplitude was continuously used. It is feasible that sane subjects were being overstimulated with biofeedback, a situation that may have hindered relaxation. This could have been achieved by the subjects to obtain feedback fromn both auditory and visual channels simultaneously and/or by intensifying the pure tone fran the speaker. This could have raisedthegeneral tension levelofthebodysothatbytheendofthe therapy session the actual muscular activity was greater than the level preceding therapy. The data acquisition procedure used to determine the EM; levels could have contributed to higher posttesting values carpared to pre- testing measures. The data were cbtained by averaging the levels from three 60-second samples during continuous speech. These intervals were initiated by the experimenter, and throughout the tine period the EM; activity was being accmulated by a digital counter. The equipment autanatically limited counting the muscular activity after 60 seconds. Therefore, any pauses that occurred during this 60 second interval were 34 included in the continuous speech segment. Some muscular activity for speecln is greater than during non-speech activity. The 60-second segments that contained the least speech would be biased toward a lower EMS reading. Frequently the subjects had difficulty at the beginning of a therapy session engaging in continuous speech. However, by the end of the 20 minute biofeedback training session, they were more likely to have continuous discourse without any further questions from the experimenter. Thus, the pretest EM; data for speech was probably abnormally low in many instances, which suggested that the individual had actually increased his muscular activity across the biofeedback training. Recomendations Several points have been raised during this experiment that should serve as a caution to researchers and clinicians who might consider EMG biofeedback as a therapeutic technique. Indiscriminate application of these procedures is inefficient, and an extended diagnostic process should be considered to determine which candidates would benefit most from the program. Clients who deronstrate obvious muscle tension would be considered for biofadback training. Several sessions might be necessary to determine a specific muscle site that would produce the most beneficial results for each individual. According to this study, three of the specific muscle sites seemed appropriate for such training, whereas the frontalis produced the most drastic effects from the other subject. In addition, the type of feedback should be examined individually, since sore subjects benefit more from visual feedback, whereas others are better attuned to the auditory mode. Also, it is cautioned that the client should not have free control to modify the feedback provided since loud auditory stinuli may interfere 35 with.the intent.of relaxation training. Ideally, an instrumentation configuration that could record.two EMG sites while the subject is being trained.by one would be beneficial. This would assist in determining whether muscular tension from one muscle was systematically having a controlling influence over the other muscles. If quantifiable EMG measurements are needed for canparison with speech output, then a synchronization between the two must occur. This can be accorplished by triggering the initiation and termination of muscular activity accumulation by the onset and offset of voicing. BIBLIOGRAPHY BIBLIOGRAPHY Acosta, F. and Yamamoto, J. "Application of electromyograph biofeedback to the relaxation training of schizophrenic, neurotic and tension headache patients. " Journal of Consulting Clinical Psychology. 46: 383-384, April, 1978. Adams, M. "The reciprocal inhibition treatment of stuttering." Journal of Communication Disorder. 5: 59-66, 1972. Alexander, A.B. "An experimental test of assumptions relating to the use of EVE biofeedback as a general relaxation training technique. " Psychophysiology. 12: 656-662, 1975. Alexander, A.B., French, C.A. and Goodman, N.J. "Training and transfer of training effects in EMG biofeedback assisted muscular relaxation." Psychophysiology. 14: 551-559, 1977. Barber, T-X- and Hanan, K.W. "Hypnotic induction and 'relaxation'" Archives of General Psychiaty. 8: 295-300, 1963. Blocdstein, O. A Handbook on Stuttering. revised ed. Chicago: National Easter Seal Society for Crippled Children and Adults, 1975. Burgraff, "The efficacy of systeratic desensitization via visual imagery as a therapeutic technique with stutterers. " British Journal of Disorders of Oommmnication. 9: 134-139, 1974. Canter, A., Kondo, C.Y. and Knott, J.R. "A comparison of ENE feedback and progressive muscle relaxation training in anxiety nemesis. " British Journal of Psychiatry. 127: 470-477, Noverber, 1975. Coursey, R.D. "ENE feedback as a relaxation technique." Journal of Consulting and Clinical Psychology. 43: 825-834, 1975. Gray and England. "Sore effects of anxiety deconditioning on stuttering frequency." Journal of Speech and Hearing Research. 15: 114-122, 1972. Guitar, B. "Reduction of stuttering frequency using analog EMG biofeedback. " Journal of Speech and HeariniResearch. 18: 672-685, 1975. Hanna, R. Wilfling, F., and McNeill, B. "A biofeedback treatment for stuttering." Journal of Speech and Hearing Disorder. 40: 270-273, Way, 1975. 36 37 Haynes, S.N., Nbseley, D., and McGowan, W.T. "Relaxation training and biofeedback in the reduction of frontalis muscle tension." Psychophysiology. 12: 5, 1975. Jacobson, E. Progressive Relaxation. 2nd ed. Chicago: University of Chicago Press, 1938. Lanyon, R. I. "Effects of biofeedback based relaxation on stuttering during reading and spontaneous speech. " Journal of Consulting Clinical Psychology. 45: 860-866, Octdner, 1977. Ianyon, R.I., Barringtcn, C.C. and Newman, A.C. "Nbdification of stuttering through 13le biofwdback: a preliminary study. " Behavior Therapy. 7: 96-103, 1976. Iegewie, H., Cleary, P. and Rackensperger, W. "HIE-recording and bio- feedback in the diagnosis and therapy of stuttering: A case study. " European Journnal of Behavioral Analysis and Pbdification. 1: 137-143, Decen'ber, 1975. Null, Gary and staff. Biofeedback, Fasting, and Meditation. New York: Pyramid Books, 1974. Paskin, M., Johnson, G., and Rondestvedt, J.W. "Chronic anxiety treated by feedback-induced muscle relaxation. " Archives of General Psychiatry. 28: 263-267, 1973. Schwartz, M. "The core of the stuttering block." Journal of Speech and Hearing Disorder. 39: 169-177, 1974. Shrum, "A study of speaking behavior of stutterers and non-stutterers by means of multidnannel electromyography. " Doctoral Dissertation, University of Iowa, 1967. Townsend, R.E., House, J.F. and Addaro, D. "A conparison of biofeedback- mediated relaxation and group therapy in the treatment of chronic anxiety." American Journal of Psychiatry. 132: 598-601, 1975. Van Riper, The Nature of Stuttering. Prentice-Hall, Inc. , Englewood Cliffs, New Jersey. Wingate, M.E. "A standard definition of stuttering." Journal of Speech and Hearing Disorders. 29: 484-489, 1964. Wolpe, J. The Practice of Behavior, Pergamon Press, 1969. York, T.J. "Ens feedback as a relaxation procedure: effects on muscle tension and focus of control." Perceptual Motor Skills. 46: 955-958. STATE UNIV. LIBRARIES VI "Iml”IWIVIHHIWWWIII‘IW 3008759171 E3 nICHIan WWII” 13112