---

Speech Audiometry — A Comprehensive Account (50 Marks)

---

1. INTRODUCTION

"Speech recognition requires greater synchronous neural firing than is necessary for appreciation of pure tones. Speech audiometry tests the clarity with which one hears." — Harrison's Principles of Internal Medicine, 21st Ed. (p. 1039)

• Correlating audiometric thresholds with real-world disability
• Differentiating cochlear from retrocochlear (VIII nerve/CNS) lesions
• Hearing aid evaluation and fitting
• Medico-legal assessment of hearing disability
• Monitoring disease progression and treatment response

---

2. HISTORICAL PERSPECTIVE

---

3. PREREQUISITES AND EQUIPMENT

Test Environment

• Sound-treated booth: ambient noise < 30 dB SPL (ANSI S3.1)
• Calibrated audiometer (TDH-39 or TDH-49 earphones; insert earphones ER-3A)

Transducers

• Supra-aural earphones — standard; bilateral testing
• Insert earphones — superior inter-aural attenuation (~50 dB); reduce collapsed ear canal artifact
• Bone oscillator — for bone-conduction speech testing
• Sound-field loudspeakers — used when earphones cannot be worn (children, hearing aid evaluation)

Stimuli

1. Intelligible — clearly recorded/enunciated
2. Phonetically balanced — phoneme frequencies match conversational speech
3. Familiar — common vocabulary (especially for non-native speakers)
4. Standardized — recorded (not monitored live voice) for reproducibility

---

4. SPEECH AUDIOMETRY TESTS — CLASSIFICATION

SPEECH AUDIOMETRY
       │
  ┌────┴──────────────┐
  │                   │
Threshold Tests    Suprathreshold Tests
  │                   │
  ├─ SDT/SAT          ├─ WRS/PB Score (SDS)
  ├─ SRT              ├─ PI-PB Function
  └─ MCL/UCL          ├─ Rollover Index
                      ├─ SSI
                      ├─ SPIN
                      ├─ HINT
                      └─ Synthetic Sentence ID (SSI-CCM/ICM)

---

5. SPEECH DETECTION THRESHOLD (SDT) / SPEECH AWARENESS THRESHOLD (SAT)

• Definition: Lowest intensity (dB HL) at which the patient can detect the presence of speech (without necessarily understanding it) — 50% of the time.
• Uses continuous discourse or spondee words
• Normal: approximately equal to the PTA average (500, 1000, 2000 Hz)
• Clinical use: Used in very young children, severely intellectually disabled patients, or those with profound loss where SRT cannot be obtained
• SDT is typically 8–9 dB lower than SRT (since detection precedes recognition)

---

6. SPEECH RECEPTION THRESHOLD (SRT)

Definition

Spondee Words (examples)

Airplane, baseball, birthday, cowboy, doorstep, eardrum, farewell, football, hotdog, ice cream, mushroom, northwest, outside, playground, railroad, sunset, toothbrush, whitewash, woodwork, workshop

Procedure — Step-by-Step

STEP 1: Familiarization
   └─ Present spondee word list at comfortable level (audible)
   └─ Patient repeats or points to picture

STEP 2: Starting Level
   └─ Begin 30–40 dB above estimated SRT
   └─ Present 2–3 spondees at this level (to confirm audibility)

STEP 3: Descend in 10 dB Steps
   └─ Until patient fails 2 consecutive words

STEP 4: Ascend in 2 dB Steps (modified Hughson-Westlake)
   └─ Present 3 words at each level
   └─ SRT = level at which 2/3 or 50% correct

STEP 5: Masking if needed
   └─ Contralateral masking with narrow-band noise
   └─ Applied when SRT difference > 40 dB (supra-aural) 
        or > 65 dB (insert earphones)

SRT–PTA Correlation

---

7. MOST COMFORTABLE LOUDNESS (MCL) AND UNCOMFORTABLE LOUDNESS LEVEL (UCL)

• Narrow dynamic range → Loudness recruitment (cochlear pathology)
• Guides hearing aid gain settings (amplification must not exceed UCL)
• Used in TILL (Treble Increase at Low Levels) and BILL fittings

---

8. WORD RECOGNITION SCORE (WRS) / SPEECH DISCRIMINATION SCORE (SDS) / PB MAX

Definition

PB Word Lists (Standard)

Standard PB Word Test Procedure

1. Set level: SRT + 30–40 dB (or MCL)
2. Present 25 or 50 monosyllabic PB words
3. Patient repeats each word
4. Score: (Correct responses / Total words) × 100 = WRS %
5. Apply masking to non-test ear if interaural difference warrants

Interpretation (Jerger Classification — cited in Scott Brown, Cummings)

(Scott Brown's Otorhinolaryngology Head & Neck Surgery, 8th Ed.; Cummings Otolaryngology, 6th Ed.)

---

9. PERFORMANCE INTENSITY FUNCTION FOR PB WORDS (PI-PB FUNCTION)

Concept

• X-axis: Presentation level (dB HL)
• Y-axis: Word Recognition Score (%)

Normal PI-PB Curve

WRS%
100│                ___________
   │            __/
 75│          _/
   │         /
 50│        /
   │       /
 25│      /
   │    _/
  0└──────────────────────────────► dB HL
   0   20   40   60   80  100  120
        ↑
    SRT (50% point)

• WRS rises with increasing intensity
• Reaches a plateau (PBmax) of 90–100%
• Remains at plateau with further intensity increase

The ROLLOVER PHENOMENON (Retrocochlear Signature)

WRS%
100│             ___
   │          __/   \___
 75│        _/          \___
   │       /                ↘  ← ROLLOVER
 50│      /
   │
  0└────────────────────────────► dB HL
         ↑PBmax                ↑PBmin

Rollover Index (RI)

RI = (PBmax − PBmin) / PBmax

---

10. SPEECH AUDIOMETRY IN DIFFERENT TYPES OF HEARING LOSS

TYPE OF LOSS        SRT           WRS/SDS          PI-PB Curve        Special Features
─────────────────────────────────────────────────────────────────────────────────────────
Conductive         Elevated       Good (90–100%)   Normal plateau      SRT-PTA agreement;
                                                                        WRS improves with
                                                                        amplification

Cochlear SNHL      Elevated       Reduced           Normal plateau;     Recruitment present;
(e.g., presbycusis (correlates    (60–80%)          PBmax reached but   Disproportionate WRS
noise-induced)     with 2-freq                      may be lower        loss relative to PTA
                   PTA)

Cochlear SNHL      Elevated       Very poor         Low PBmax           Narrow dynamic range
(Meniere's,                       (<50%)                                Fluctuating
severe)

Retrocochlear      May be normal  Disproportionately Rollover (RI>0.45)  WRS worse at high
(VIII nerve /      or mildly      poor relative to                        levels; SRT-WRS
schwannoma)        elevated       PTA                                     dissociation

CAPD (Central      Normal PTA    Poor in noise;     May be abnormal      Fails SSI-ICM, SPIN,
Auditory           and SRT       normal in quiet    under CCM           dichotic tests
Processing
Disorder)

Non-organic HL     Elevated SRT   SRT-PTA poor       Inconsistent        Stenger test;
(Functional)       worse than PTA agreement                              SRT much better
                                                                         than claimed PTA

---

11. SPEECH-IN-NOISE TESTS

11a. SPIN Test (Speech Perception in Noise)

• Developed by Kalikow, Stevens & Elliott (1977)
• Sentences with target words at end (High Predictability / Low Predictability)
• Presented in multi-talker babble noise at fixed SNR
• Score: % correct target words

11b. HINT — Hearing in Noise Test (Nilsson, 1994)

• IEEE sentences presented in speech-spectrum noise
• Speech Reception Threshold in Noise (SRTN) measured
• Normal SRTN: approximately +2 dB SNR (speech 2 dB above noise)
• Widely used for hearing aid evaluation and cochlear implant outcome

11c. HHIE / QuickSIN

• QuickSIN (Quick Speech-in-Noise): 6 sentences per list; SNR Loss computed
• SNR Loss > 7 dB = significant benefit from directional microphones

11d. BKB (Bamford-Kowal-Bench) Sentence Lists

• Pediatric sentences; low linguistic complexity
• Used for children's speech audiometry and cochlear implant outcomes

---

12. SYNTHETIC SENTENCE IDENTIFICATION (SSI)

Concept (Jerger & Jerger, 1974)

• Third-order approximations to English sentences (syntactically correct but semantically meaningless)
  • e.g., "Small boat with a picture has become"
• Presented with a competing message (continuous discourse)
• Patient identifies sentence from a printed list of 10

Two Conditions:

SSI Score Interpretation (Stell & Maran; Cummings):

• Normal brainstem: SSI-CCM ≥ 60%
• VIII nerve lesion: SSI-CCM poor; SSI-ICM may be near normal
• Cortical lesion: SSI-ICM poor; SSI-CCM relatively preserved

---

13. DICHOTIC SPEECH TESTS (for CAPD)

---

14. MASKING IN SPEECH AUDIOMETRY

When to Mask?

• SRT testing: if test-ear SRT exceeds non-test ear PTA by ≥ 40 dB (supra-aural) or ≥ 65 dB (insert)
• WRS testing: if presentation level exceeds non-test ear PTA by ≥ 40 dB

Masking Noise for Speech Tests

• Speech-spectrum noise (broadband weighted to speech frequencies)
• White noise — acceptable but less optimal
• Multi-talker babble — sometimes used in research

Masking Level

• Initial masking: Non-test ear PTA + 10 dB
• Effective masking level (EML) used in most clinics

---

15. SPEECH AUDIOMETRY IN SPECIAL POPULATIONS

Pediatric Speech Audiometry

Elderly (Presbycusis)

• WRS disproportionately reduced relative to PTA
• "Phonemic regression" (Gaeth, 1948): decline in word recognition exceeds audiometric threshold prediction
• SSI and SPIN particularly useful

Hearing Aid and Cochlear Implant Evaluation

• Pre-fitting: WRS in quiet and noise
• Post-fitting/post-CI: HINT, AzBio sentences, BKB-SIN
• Target: WRS improvement ≥ 10–15%

Non-Organic (Functional/Malingering) Hearing Loss

• Stenger test (speech version): based on Stenger principle — if two tones presented simultaneously, only the louder is perceived
• Speech SRT significantly better than claimed PTA
• SRT-PTA discrepancy > 12 dB → suspicion raised

---

16. COMPLETE SPEECH AUDIOMETRY FLOWCHART

                        PATIENT REFERRED FOR HEARING ASSESSMENT
                                          │
                               ┌──────────▼──────────┐
                               │   Pure Tone Audiogram │
                               │   (AC + BC + MCL/UCL) │
                               └──────────┬────────────┘
                                          │
                               ┌──────────▼──────────┐
                               │    SPEECH AUDIOMETRY  │
                               └──────────┬────────────┘
                                          │
                    ┌─────────────────────┼──────────────────────┐
                    │                     │                        │
          ┌─────────▼──────┐   ┌──────────▼─────────┐  ┌────────▼───────────┐
          │ SDT (Detection) │   │  SRT (Recognition)  │  │ MCL / UCL / Dynamic│
          │ (Non-verbal pts)│   │  Spondee words 50%  │  │      Range          │
          └─────────────────┘   └──────────┬──────────┘  └────────────────────┘
                                           │
                               ┌───────────▼───────────┐
                               │  SRT–PTA Agreement?    │
                               └──┬────────────┬────────┘
                                  │            │
                             YES (±10dB)    NO (>12dB)
                                  │            │
                                  │        FUNCTIONAL HL
                                  │        (Stenger Test)
                                  │
                    ┌─────────────▼──────────────────┐
                    │  WRS / PB Score at SRT + 30–40  │
                    │  (Phonetically Balanced Words)   │
                    └──────────┬─────────────┬────────┘
                               │             │
                         WRS ≥ 90%       WRS < 60%
                               │             │
                         NORMAL/          Consider PI-PB
                         CONDUCTIVE       Function Testing
                                               │
                              ┌────────────────▼──────────────┐
                              │  PI-PB at Multiple Intensities │
                              │  Plot WRS vs. dB HL            │
                              └──────┬────────────────┬────────┘
                                     │                │
                              No Rollover          ROLLOVER
                              (RI < 0.25)          (RI > 0.45)
                                     │                │
                               COCHLEAR           RETROCOCHLEAR
                               PATHOLOGY          LESION (MRI IAC)
                                     │
                        ┌────────────▼──────────────────────┐
                        │   Speech-in-Noise Testing          │
                        │   (HINT / QuickSIN / SPIN)         │
                        │   if PTA near normal but patient   │
                        │   complains of difficulty in noise  │
                        └────────────┬──────────────────────┘
                                     │
                              ┌──────▼────────┐
                              │  CAPD Tests    │
                              │  (SSI-ICM/CCM, │
                              │  Dichotic,     │
                              │  SCAN-3)       │
                              └───────────────┘

---

17. AUDIOMETRIC PATTERNS — VISUAL SUMMARY

• Elevated SRT (correlating with the 2-frequency PTA at 500 and 1000 Hz)
• Moderately reduced WRS (~65–75%) due to high-frequency consonant loss
• Normal PI-PB curve shape (no rollover) — cochlear, not retrocochlear

---

18. SPEECH AUDIOMETRY INTERPRETATION ALGORITHM

┌──────────────────────────────────────────────────────────────────┐
│                   SPEECH AUDIOMETRY RESULTS                       │
│                                                                    │
│  Parameter    │  Conductive │ Cochlear SNHL │ Retrocochlear │ CAPD│
│───────────────│─────────────│───────────────│───────────────│─────│
│  SRT          │  Elevated   │  Elevated     │  Nl / Elev    │  Nl │
│  SRT-PTA      │  Good agree │  Good         │  Poor agree   │  Nl │
│  WRS (quiet)  │  Near 100%  │  Reduced      │  Very poor    │  Nl │
│  WRS (noise)  │  Good       │  Reduced      │  Poor         │ Poor│
│  PI-PB        │  No rollover│  No rollover  │  ROLLOVER     │ Var │
│  RI           │  < 0.25     │  < 0.25       │  > 0.45       │ Var │
│  UCL          │  Normal     │  Reduced      │  Normal       │  Nl │
│  Dynamic range│  Normal     │  Narrow       │  Normal       │  Nl │
└──────────────────────────────────────────────────────────────────┘

---

19. IMPORTANT SPEECH TESTS — QUICK REFERENCE TABLE

---

20. RECENT ADVANCES IN SPEECH AUDIOMETRY

20a. Automated and Computerized Speech Audiometry

• Software-based audiometers (Equinox, AURICAL) with digital playback eliminate variability of live-voice testing
• Automated WRS testing using forced-choice paradigms
• Self-administered computerized testing for large-scale screening

20b. Tele-Audiometry and Remote Speech Testing

• Cloud-based audiometers (e.g., OtoAccess, Shoebox) allow speech testing via smartphones
• Validated for SRT and WRS in remote/rural settings
• Post-COVID expansion; confirmed reliability within ±5 dB of booth testing

20c. Digits-in-Noise (DIN) Test

• Simple, language-independent
• Validated in 11+ languages (International Collegium of Rehabilitative Audiology)
• Smartphone-based screening: sensitivity ~92%, specificity ~94% for detecting HL
• Dutch/European standard (Smits et al., 2013; IJzerman et al.)

20d. Matrix Speech Tests

• OLSA (German), PRESTO (English), French Matrix, BaMBA (Indian languages)
• Sentences with fixed syntactic structure; words replaced across lists
• Allows precise SNR50 measurement; minimal learning effect
• Better ecological validity than PB words

20e. Spatial Speech-in-Noise Testing

• Speech from front; noise from different azimuths
• Tests binaural unmasking and spatial release from masking
• Predicts real-world hearing disability better than standard tests

20f. Electrophysiological Correlates of Speech (Cortical Auditory Evoked Potentials — CAEP)

• P1-N1-P2 complex to speech syllables (/ba/, /da/)
• Mismatch Negativity (MMN) — objective measure of phoneme discrimination
• Useful for pre-verbal infants and cochlear implant programming
• Auditory Steady State Response (ASSR) to speech-modulated carriers

20g. AI and Machine Learning

• Neural networks trained on audiograms + WRS to predict real-world handicap
• Automatic speech recognition (ASR)-based scoring of WRS without examiner judgment
• AI-driven adaptive speech testing reduces test time by 40–60%

20h. Indian Language Speech Audiometry

• Ali Yavar Jung National Institute (AYJNISHD) developed standardized Hindi, Marathi, Tamil, Telugu, Kannada, Bengali, Oriya speech materials
• Open-set and closed-set tests available
• Critical for accurate testing in multilingual India (Zakir Hussain; Hazarika)

20i. Extended High-Frequency Speech Audiometry

• Testing at 8–16 kHz range
• Detects early noise-induced and ototoxic damage before standard-frequency loss appears
• Research tool for monitoring aminoglycoside and cisplatin toxicity

---

21. SPEECH AUDIOMETRY IN MEDICO-LEGAL DISABILITY ASSESSMENT

---

22. MASKING DILEMMA IN SPEECH AUDIOMETRY

• Use insert earphones (superior interaural attenuation ~50 dB vs 40 dB for supra-aural)
• Apply plateau method (increase masker in 5 dB steps; WRS should remain stable over a 20 dB plateau range)

---

23. SUMMARY TABLE — SPEECH TESTS BY SITE OF LESION

SITE OF LESION         KEY SPEECH FINDINGS
─────────────────────────────────────────────────────────────────────
Outer/Middle Ear       Elevated SRT; WRS normal; dynamic range normal
(Conductive)           Pure gain deficit; improves with amplification

Cochlea                Elevated SRT; WRS reduced; narrow dynamic range;
(Meniere's, NIHL,      Phonemic regression; no rollover; recruitment
Presbycusis)

VIII Nerve             WRS disproportionately poor; ROLLOVER;
(Vestibular            RI > 0.45; SSI-CCM poor
Schwannoma)

Brainstem              SSI-CCM bilaterally poor; dichotic tests
                       abnormal; binaural interaction tests fail

Cortex                 SSI-ICM poor (ipsilateral competing);
                       Dichotic tests poor contralateral to lesion;
                       PTA and SRT may be normal

CAPD                   Normal PTA, SRT, WRS in quiet; POOR in noise;
                       Dichotic, SSI, SPIN — all poor

---

24. KEY MNEMONICS

---

REFERENCES

1. Harrison's Principles of Internal Medicine, 21st Ed., p. 1039 — Speech audiometry fundamentals, SRT definition, PB words
2. Scott Brown's Otorhinolaryngology Head & Neck Surgery, 8th Ed. — PI-PB function, rollover, retrocochlear testing
3. Cummings Otolaryngology — Head and Neck Surgery, 6th Ed. — WRS interpretation, SSI, Jerger classification
4. Stell & Maran's Textbook of Head and Neck Surgery, 5th Ed. — Speech audiometry in clinical practice
5. Dhingra PL — Diseases of Ear Nose Throat & Head and Neck Surgery, 7th Ed. — SRT-PTA correlation, disability assessment, Indian context
6. Hazarika P — ENT Head Neck Surgery, 4th Ed. — Indian language materials, pediatric speech audiometry
7. Zakir Hussain — Textbook of ENT & Head and Neck Surgery — Comprehensive speech tests, clinical correlation
8. Bailey & Love's Short Practice of Surgery, 28th Ed., p. 783 — Audiogram patterns (pure-tone reference)
9. Smits C et al. (2013) — Digits-in-noise test; Int J Audiol
10. Nilsson M et al. (1994) — HINT development; J Acoust Soc Am
11. Jerger J & Jerger S (1971) — PI-PB rollover; Archives of Otolaryngology
12. AIIMS / AYJNISHD Guidelines — Indian language speech audiometry standardization

---