1. Assessment Part 2: Instrumental Assessment
CD661
1. Lecture 8 Assessment Part 2 – Instrumental Assessment

2. Instrumental Assessment
Acoustic assessment
Aerodynamic assessment
Videostroboscopy
Electroglottography
Electromyography
2. There are a number of instrumental assessments that we may perform during a voice evaluation and these include acoustic, aerodynamic, videostroboscopic assessments and electroglottography. Electromyography, EMG, is performed by the ENT but the SLP may be involved in the interpretation if he or she is expert in the area of laryngeal EMG. By far, the most common instrumental voice assessment is acoustic and videostroboscopic assessment.

3. Acoustic Assessment: Spectrograms Pitch and Intensity Measurements Acoustic Characteristics of the Phonatory Signal
3. Acoustic Assessment. There are many software programs available that allow you to measure numerous acoustic characteristics of the voice. The most common is the Pentax Computerized Speech Lab or CSL which has, in addition to other programs, the Real Time Pitch and Multidimensional Voice Profile programs. You may also be familiar with Visipitch, Dr. Speech, Wevsystems and freeware such as Praat. When we perform acoustic analysis we may perform spectrographic analysis, measure habitual pitch and pitch range, measure average intensity and intensity range, and measure acoustic voicing parameters such as noise and frequency and intensity perturbations.

4. Spectrograms All 4 show sustained /i/ Y axis is frequency
X axis is time
Intensity is represented by degree of darkness, i.e. darkest frequency bands are the loudest, lightest are the softest. Areas of white indicate a complete absence of energy for that frequency.
4. Spectrograms are really ‘pictures’ of sounds and show you the frequencies present, or absent, in a given sound as well as the intensity of those frequencies. Spectrograms show frequency on the Y axis, time on the X axis and intensity in grayscale, i.e. the stronger and more intense a frequency band the darker gray it will be. There are both broad and narrow band spectrograms. The spectrograms in this slide are broad band. Broad band is great for viewing vowel formant frequencies, while narrow band spectrograms are good for seeing harmonics. Note how the spectrogram on the top right for a ‘normal’ voice shows strong, well defined formant bands and areas w/o energy (white areas). The other spectrograms show excess noise in the acoustic signal. Spectrograms, especially narrowband, are useful for voice analysis because they are able to show us how stable and periodic the VF vibration is. The National Center for Voice and Speech recommends that a spectrographic analysis be performed first to ‘type’ voice patients for further acoustic analysis. Only vocalizations that are sufficiently periodic can be analyzed for frequency and intensity perturbations.
Boone et al. 2005

5. Narrow Band Spectrograms
5. Two examples of narrow band spectrograms. The many horizontal lines are the harmonics. Both of these spectrograms are relatively clean, i.e. there is very little noise and they are fairly periodic.
/i /

6. Frequency (pitch) Measurements
Fundamental frequency of habitual pitch during reading, counting 1-15, spontaneous conversation
Women: Hz to Hz
Men Hz to Hz
Optimal pitch – on ‘mhmm’ or ‘uh huh’
Pitch variability (inflection)
Pitch range – assess highest and lowest pitches
Have patient perform ascending and descending pitch glides on /a/ or /i/
6. Assessing the patient’s fundamental frequency, or Fo, is very important. Many times we have patients who present with a lower or higher, but more often lower, than optimal habitual pitch and this must be addressed in therapy. We assess habitual Fo during counting 1-15, during reading and sometimes during speech. Depending on the study you look at, the range of normal Fo for women is Hz – Hz and for men – Hz. However, even if your patient has a habitual Fo within normal they may still be speaking too low for their voice type. This is why we assess for optimal pitch. We record the patient producing three productions of ‘mhmm’ or ‘uh- huh’ and calculate to obtain optimal habitual pitch. Most people will produce ‘mhmm’ with a frontal tone focus and comfortable, appropriate pitch. We then compare their habitual Fo to their optimal Fo to see how similar the measurements are. If the habitual Fo is lower or much higher than the optimal, then we target the optimal appropriate Fo using resonance placement techniques such as frontal or forward tone focus. Pitch variability or inflection is calculated in number of semitones by the computer program and is assessed during a reading or spontaneous speech task but NOT during counting. Note if patient can change pitch, i.e. inflect, while speaking or do they sound monotone? We also assess pitch range. We have the patient start at a comfortable pitch and glide up as high as they and sustain the pitch for 5 seconds. E do this 3 times and take the highest pitch. We then do the same thing but gliding down to the lowest pitch they can sustain for 5 seconds. Glottal fry does NOT count as a lowest pitch ! Do not take a measurement in glottal fry. Note: in order for the patient to produce an adequate pitch glide for assessment, YOU must be able to produce an adequate model for your patient ! Practice makes perfect ! :>)

7. Intensity Measurements
Use a Sound Level Meter (SLM) held 30 centimeters (about 1 foot) from the patient’s mouth.
Average Conversational Intensity: Ask the patient to talk for seconds and note the dB as they talk. I usually take at least 10 readings and then calculate average dB. Normal is dB.
Maximum Intensity – measure maximum intensity for speech by having patient yell ‘hey!’
DO NOT measure maximum intensity in presence of VF polyps, VF edema, VF hemorrhage, or VF varice/ectasia OR whenever there is excessive edema or risk of further VF damage from yelling.
7. When measuring vocal intensity we use a sound level meter held 30 centimeters from the patient. To measure average conversational intensity, have the patient talk for seconds on any topic. Obtain at least measurements of dB level and then take the average. To obtain maximum intensity, have the patient yell ’hey’ as loud as they can UNLESS they are a phonotrauma patient, then you will pass on this measurement due to the risk of increasing injury to the VFs.

8. Instrumental Acoustic Evaluation of Voicing Parameters: Perturbations
Measurements of Voicing Parameters – always measured during a sustained ‘ah’ vowel.
1) frequency and intensity perturbations – if elevated indicates > than average cycle to cycle variability in frequency and amplitude.
called ‘Jitter’ & ‘Shimmer’
Jitter – frequency perturbation
Shimmer – intensity perturbation
2) VFo - variation in fundamental frequency
8. Perturbation means ‘a change in the normal state or regular movement of something’ ( Titze (1994) defines it as the variability or irregularity in a system. A frequency perturbation, or ‘jitter’ is a short term cycle to cycle change in the frequency of vibration or, said another way, a cycle to cycle variation in the period of VF vibration. Remember VF vibration is not purely periodic but ‘quasi-periodic.’ So there is always some perturbation in the vibratory system that is natural and normal. However, pathological voices often present with increased frequency and intensity perturbations. RAP, relative average perturbation, and PPQ, pitch period perturbation, are the best and most reliable jitter measurements are available in most acoustic soft ware programs. An intensity perturbation, or ‘shimmer,’ is a short term cycle to cycle variation in the peak to peak amplitude of vibration. A reliable measurement for shimmer is APQ, amplitude perturbation quotient. VFo, which is variation in fundamental frequency, reflects the short to long term variation in Fo by measuring the relative standard deviation of Fo. Large standard deviations would indicate instability in VF vibration. Perturbation measures can only be obtained during sustained phonation on vowels.

9. Instrumental Acoustic Evaluation of Voicing Parameters: Noise Measurements
Noise to harmonics ratio - NHR
Voice Turbulence Index - VTI
Soft Phonation Index - SPI
9. Noise measurements tell us how much noise is present in the acoustic signal. Excessive breathiness results in air escaping adds turbulence that results in increased noise (aperiodic sound). The hoarseness/harshness generated by irregular VF vibration also can add noise. There are 3 types of noise measurements that are commonly performed; NHR, VTI and SPI.
Noise to harmonics (NHR) ratio tells how much noise energy is present, globally, in relation to the periodic or harmonic energy. It is the average ratio of inharmonic energy, i.e. noise, to harmonic energy in the 70 Hz – 4200 Hz frequency range and includes all noise (aperiodic) sources, including contributions from jitter, shimmer and turbulent noise. Voice Turbulence Index, VTI, is a ratio measurement of high frequency noise in the Hz range to the harmonic energy in the Hz range. It correlates with turbulence caused by incomplete or lose adduction of the VFs. It may be an acoustic correlate to breathiness. Soft Phonation Index, SPI, is the average ratio of lower frequency harmonic energy ( Hz) to higher frequency harmonic energy ( Hz). SPI can be an indicator of hw completely or tightly the VFs are adducted during phonation and if elevated indicates losely or incompletely adducted VFs. However, this does not necessarily indicate disorder. For example, patients with posterior glottal chinks, as observed in 25%-30% of women, or people phonating softly, may present with elevated SPI. Also, some people speak with an overall ‘softer attack.’

10. Instrumental Acoustic Evaluation of Voicing Parameters: Tremor Indices
Vocal Tremor typically has modulations in both frequency and amplitude.
Both can be measured
Amplitude Tremor Index - degree of amplitude modulation in the tremor
Frequency Tremor Index – degree of frequency modulation
10. Vocal tremor typically has two components; a regular frequency modulation and a regular amplitude modulation. The extent of each is measurable. Amplitude Tremor Index assesses the acoustic signal for the presence and degree of amplitude modulation, while Frequency Tremor Index assesses for the presence of and degree of frequency modulation. Note that most acoustic analysis programs have been normed on ‘straight tones,’ sustained vowels with no vibrato present. Sometimes singers, or even non-singers, have a natural vibrato that is difficult to suppress. This will show up as elevated Amplitude and/or Frequency Tremor Indices and is not pathological in these cases.

11. Multidimensional Voice Profile Radialgraph – Dx Reflux Laryngitis
Frequency perturbations
Tremor Indices
11. A radialgraph generated from the Multidimensional Voice Profile which is a software program within the Computerized Speech Lab (CSL). This patient’s diagnosis was reflux laryngitis with secondary muscle tension. Note the elevated frequency and amplitude perturbations, elevated soft phonation index and elevated amplitude tremor index.
Amplitude
perturbations
Noise
Parameters

12. Instrumental Aerodynamic Evaluation – What do we measure ?
Subglottal pressure and mean transglottal airflow –
1) Subglottal pressure is measured indirectly and estimated via a measure of the intraoral pressure during repetitions of /pi/ or /pæ/ at 1.5 syllables per second (Smitheran and Hixon, 1987).
2) Transglottal airflow – measured during sustained phonation on /a/ or during the vowel portion of the cv syllable.
12. There are now a number of aerodynamic measurement systems available, such as Pentax Phonatory Aerodynamics System and Glottal Enterprise’s Aeroview System. The data are collected using a Rothenberg mask fitted with a pressure transducer and a flow transducer. The signals are captured, transduced from a mechanical signal to an electronic signal and analyzed via computer software programs. Subglottal pressure is measured in cm/H2O or kilapascals and airflow is measured in Liters/sec or ml/sec. Measurements of Ps and airflow are useful pre- and post – therapy for hypofunctional patients (paresis, paralysis, presbylarygis), for functional, i.e. muscle tension, patients who present with increased closed phase and hyper-adducted VFs and even for phonotrauma patients who present w/ hourglass or irregular glottic closure. Hypofunctional patients and phonotrauma patients with incomplete glottic closure will typically present with increased airflow rates, while functional patients will usually present with decreased airflow rates due to long closed phases and hyperadduction.

13. Rothenberg Mask Pressure transducer
13. This is a Rothenberg mask. Note that it covers the nose and mouth. On the inside of the mask, the pressure transducer is fitted with a silastic plastic tube which goes in the patient’s mouth and must sit above the tongue. So, why does indirect measurement of Ps work? Indirect measurement of Ps from oral pressure measurements works because the measurement is obtained during the occlusion for /p/. When the lips are sealed around the tube, the pressures in the respiratory and vocal tract are equal, i.e. lung pressure = subglottic pressure = oral pressure. This is because when the lips are sealed the system is occluded at the lips but the VFs are OPEN and so the entire system is connected! Thus, pressure is equal throughout the system.
Pressure transducer

14. Plastic tube Airflow transducer Pressure transducer
14. Different views of the Rothenberg mask. Notice the tube on the inside of the mask as well as the pressure and flow transducers.
Pressure transducer

15. Other Techniques for Measuring Subglottal Pressure
Esophageal balloon – A second indirect method; assumes esophageal pressure is a good estimate of subglottal or tracheal pressure; invasive; used for research
Tracheal puncture – Directly measures subglottal pressure via tracheal puncture ; invasive; used primarily for research
15. There are other methods for measuring Ps but they quite invasive and used for research purposes. They are not used clinically. The first is indirect measurement via esophageal balloon. The second is direct measurement via tracheal or cricothyroid (CT) membrane puncture. In this method, a very small catheter is placed in the subglottal space by tracheal or CT membrane puncture. The catheter is attached on the outside to a pressure transducer. Computer software displays and analyzes the signal.

16. Inverse Filtering
Inverse Filtering of the Airflow Signal : filtering removes resonant effects of the vocal tract. Removes F1 and F2.
More accurate estimate of the airflow waveform at the level of the VFs
Can be performed on the acoustic sound pressure wave (audio signal) or on the airflow waveform.
AC flow – alternating flow, i.e., the amount of airflow that is ‘vibrated’ during VF vibration. AC Peak and AC average airflow measurements during vibration are obtained.
DC flow – airflow leakage, i.e., amount of air leakage during vibration (unvibrated air) when VFs are supposed to be closed. DC Peak and DC average measurements.
Max. Flow Declination Rate (MFDR)- The rate of decline in airflow reflects closing rate of VFs. Related to vocal intensity.
16. Airflow measurements obtained without inverse filtering give us only a gross idea of what is going on at the level of the gottis because the airflow signal (airflow waveform) is contaminated by the effects of vocal tract resonances. Inverse filtering the waveform removes the first and second formant frequencies and produced a glottal flow waveform that is much cleaner and the opening and closing moments are much more well defined. There are several measurements that are made during inverse filtering; AC peak flow, AC average flow, DC peak flow, DC average flow and maximum flow declination rate (MFDR). To inverse filter the airflow signal, the signal from the airflow transducer on the Rothenberg mask is directed into a filter that is set to filter out the first two formants of the sustained vowel in order t remove vocal tract resonances from the signal.

17. Filters: filter out F1 and F2, first two vowel formants
Audio
Flow Glottogram
17. This slide shows the audio signal, the inverse filtered glottal flow called a ‘flow glottogram’ and an inverted electroglottography signal. Look at the flow glottogram signal. The Y axis is airflow in ml/sec and the X axis is time. The peak of the wave is the point of maximum airflow during vibration and of course occurs during the open phase of vibration. Notice for this person, the airflow never goes to zero during VF closure. This results in a ‘DC offset’ meaning that even when the VFs are closed there is some air leakage, perhaps due to a posterior glottal gap or incomplete VF closure.
DC offset
Electroglottogram

18. Electroglottography (EGG)
EGG – indirectly assesses the relative contact area of the VFs during a cycle of vibration
Often done concurrently with stroboscopy
How it works – measures changes in electrical impedance and conductivity associated with opening and closing of VFs
When VFs are closed, more current flows across them than when they are open and this sis reflected in the waveform
18.To do an EGG, two small disks (electrodes) are placed on either side of the thyroid cartilage. An mild electric current passes between the disks. The way the signal between the two disks behaves shows how much the vocal folds are in contact. When the VF are closed, conductivity increases and the current passes across the VF tissue and this creates a ‘peak’ in the EGG wave, but when the VFs are open, there is greater impedance to the current and this creates the ‘valley’ in the wave.

19. EGG Electrodes
19. EGG electrodes

20. EGG Electrodes
20. Placement of EGG electrodes on the neck.

21. EGG Electrodes NCVS.org
21. Individual obtaining EGG measurements on himself.
NCVS.org

22. EGG Normal intensity phonation High intensity phonation Normal Breathy
22. Here are samples of EGG waveforms. In the top left example, ‘a’ represents the beginning of VF closure, ‘b’ represents complete VF closure, the downward slope from ‘b to c’ represents the beginning of the opening phase and ‘c – d’ represents the open phase of vibration. Note that for high intensity phonation the VFs are closed longer, take longer to open and that the open phase is very short. Note that for breathy phonation, the closure is gradual and the open phase quite long. Aperiodic vibration is presented by highly variable waves.
EGG
Hoarse
Boone et al., 2005

23. EGG Modal Falsetto Breathy
23. Simultaneous measurement of EGG with videostroboscopy. The vertical lines reflect the VF position in the videostill.
EGG
Breathy

24. Indirect Measures of Glottic Closure
Maximum Phonation Time – sustained /a/ - time duration of sustain with stopwatch – repeat 3x’s take best time
Normal female range secs.; males secs.
S/Z Ratio -
1) time duration sustain /s/ 2x’s and average
2) time duration of sustained /z/ 2x’s and average
3) divide average time for /s/ by average time for /z/ Normal is 0.80 to 1.40 for adults
Less than .80 may reflect hyperfunction , while greater than 1.4 reflects incomplete glottic closure
24. Both maximum phonation time and s/z ratio have been used to indirectly assess glottic closure. However, be aware that an individual with VF nodules may have so much secondary muscle tension that their s/z ratio is within normal limits due to excessive medial compression. Also, a trained singer or woodwind or brass instrument player may have excellent expiratory control and be able to sustain /s/ for as long as seconds and /z/ for seconds. These values would result in an elevated s/z ratio that is invalid and not indicative of incomplete glottic closure. Maximum phonation time can also be affected by respiratory diseases that affect pulmonary function and also by inadequate breath support. These measurements are best used for pre- and post- therapy comparisons. Keep in mind that the only way to truly assess VF closure is with videostroboscopy or high speed video.

25. Rationale for Instrumental Assessment
Perceptual signs should be verified and supported by acoustic/aerodynamic methods (instrumental analysis)
Noninvasive (acoustic & aerodynamic)
Available and relatively low cost
Correspondence with the underlying physiology of voice disorders
Application to therapy goals
Pre-, mid-, and post- therapy comparisons
Objective documentation of change for insurance reimbursement purposes
Allows for comparison of vocal performance to normative data
25. There are a number of reasons why instrumental assessment is valuable. They are listed here in this slide.

26. EQUIPMENT AND MATERIALS FOR ACOUSTIC & AERODYNAMIC ANALYSIS
Computer
Acoustic and Aerodynamic Analysis Hardware and Software
A good quality microphone
A sound level meter (SLM)
Stopwatch
Ruler/Tape measure
Good quality digital recorder
26. Equipment needed for instrumental assessment. A stopwatch is needed for obtaining s/z ratio and maximum phonation time. A ruler or tape measure is needed to measure mic to mouth distance for audio recordings and for intensity measurements with the sound level meter.

27. Electromyography
Fine wire or needle electrodes are inserted into the laryngeal muscles to measure the electrical activity of muscles.
Invasive: insertion requires expertise & knowledge of head and neck anatomy and is done by the ENT
Interpretation requires expertise and practice
Used to
Confirm VF paralysis
Differentiate VF paralysis from arytenoid cartilage fixation or dislocation
Assess for excessive or abnormal muscle activity related to neurological disease
Verify site of injection for Botox tx for spasmodic dysphonia
27. Laryngeal electromyography (LEMG) is a test that gives information about the motor activity within the laryngeal. This information is not available by any other test. This activity is important because it has implications for diagnosis and for predicting recovery of function. Diagnostically, LEMG is essential to determine the neuromuscular status of the VFs. Subtle weakness in the VFs may not apparent on endoscopic examination, and may only be evident via LEMG. In cases of VF motion impairment, LEMG can differentiate VF paralysis from impairment due to scarring or arytenoid fixation.
LEMG also helps to predict recovery. Following the onset of vocal cord paralysis, recovery can occur spontaneously up to 12 months following injury. Patients with VF paralysis are often treated with a "wait and see" policy. While this policy allows some patients to avoid surgery, it also causes some patients to continue to have symptoms (without treatment) for a full year, only to find out that in their paralysis is permanent. LEMG offers an alternative to the "wait and see" policy because it allows the physician to predict whether or not the paralysis is temporary or permanent early in on.The physician can then make treatment decisions based on this information. (

28. Electromyography Excessive Activity Excessive Activity
28. Examples of abnormal EMG signals. The top signal is an audio recording of a sustained vowel. The 2nd signal is CT muscle activity. CT muscle activity should cease once phonation has stopped. Note that the CT muscle stays active for some time after phonation has stopped. This is abnormal. The 2nd audio signal is a recording of the cv syllables /ba/ and /da/. CT muscle activity is shown below. Notice how the CT muscle gradually increases in activity. This patient did not increase pitch. Such an increase in activity is abnormal.

29. Non-instrumental Elevation of Voice
Pitch – subjective judgment or ‘1,2,3’ method
Pitch range - use keyboard, pitch pipe, or ear
Intensity - subjective judgment or use SLM
Glottic competence and coordination & interaction of respiratory and phonatory systems – use MPT and s/z ratio, have pt. cough, throat clear, use hard glottal attack
Carefully and accurately describe vocal quality !
29. It is well acknowledged in the field of voice that non-instrumental assessment is generally inadequate for vice assessment. However, depending on your work setting, you may have no choice but assess voice without instrumentation.

30. Oral Mech Exams
Perform oral mechanism exams or screenings on all voice patients
The depth or extent of the exam depends on the patient
30. I’m assuming that the oral mech exam has been covered in your other classes so I will not go into detail here. The depth to which you test or screen will depend on the patient. Whenever I have patient with a neurological condition, voice related or not, or have a patient who has had head/neck surgery or radiation, I perform a comprehensive oral mech exam to assess the integrity of the cranial nerves.