Speech Audiometry

What is Speech Audiometry?

Speech audiometry is a method used by audiologists to:

• Evaluate how well a patient can hear speech (detection) and understand speech (recognition).

• Validate puretone thresholds.

• Evaluate candidacy for hearing aids or cochlear implants.

• Monitor changes in hearing over time.

• Contribute to the diagnosis of hearing disorders.

Equipment

Speech audiometry is completed using a diagnostic audiometer with speech presented via:

• Live Voice (microphone):

  • Flexible for patients needing modifications (e.g., limited vocabulary).

  • Dialects and accents can impact performance.

• Standardized Recorded Speech:

  • Consistent, but less flexible.

  • Improves comparisons across repeated testing.

  • Speech characteristics (dialects, accents) can affect a patient's performance; audiology voice is preferred.

Speech Threshold Measures

These measure the softest sound/stimuli a person can hear.

Speech Recognition Threshold (SRT)

The lowest level of speech that is understood or recognized at least 50% of the time, typically found using an "up 5 dB, down 10 dB" procedure.

Speech Recognition Threshold (SRT): materials

• Utilizes SPONDEE word lists.

  1. Spondee words are two-syllable words with equal stress on each syllable (e.g., "hotdog", "airplane").

  2. Presented as a closed set: patients are familiarized with the words before testing begins.

Speech Recognition Threshold (SRT): clinical application

1. Cross-check for puretone thresholds.

2. Measurement of communication disability.

3. Reference for suprathreshold testing.

Speech Recognition Threshold (SRT): patient instruction + procedure

• Patient Instructions: "I am now going to find the softest level that you can recognize some common words. Please repeat the words that you hear, even if you have to guess."

• Procedure: Familiarize patient with spondee list, present words at a comfortable listening level, then use the "down-10-up-5" bracketing procedure. Other methods include block presentation and Downs and Minard (1996).

SRT / PTA Agreement

This serves as a cross-check to validate puretone thresholds.

• Puretone Average (PTA):

  • Calculated from Air Conduction (AC) thresholds.

  • 3-frequency PTA: Average of thresholds at

  • 500,1000,2000extHz

  • 500,1000,2000extHz.

• Agreement Criteria:

  • Speech thresholds are usually equal to or slightly better than PTA.

  • SRT and PTA should be within 10dBHL. 

  • Agreement within 10dB is considered good; outside 10dB suggests a discrepancy that requires reinstruction or further investigation.

 Speech Detection Threshold (SDT)

Also known as Speech Awareness Threshold (SAT), it is the lowest level of speech that is detected at least 50% of the time.

• When Used: SDT is employed when SRT cannot be obtained, such as with:

  • Severe to profound hearing loss.

  • Pediatric patients (who might be shy or require physical responses).

  • Patients with speech or language limitations (e.g., primary language is not the language of the test, and the audiologist is not familiar with nuanced differences).

• Speech materials and familiarization are not critical for SDT.

Suprathreshold Measures (Word Recognition Score (WRS))

Evaluates how well a patient can recognize speech above threshold (at an elevated, audible level). This helps determine hearing aid candidacy.

Word Recognition Score (WRS): materials

• Utilizes phonetically balanced (PB) single-syllable words (CVC - consonant-vowel-consonant, e.g., "cat", "dog").

  • PB words represent phonemes commonly used in the English language.

  • Presented as an open set: patients are not familiarized with the words prior to testing.

• Carrier Phrase: A phrase preceding each word (e.g., "Say the word…", "You will say…") to alert the patient.

• Word Lists: Can use 50-word or 25-word PB lists.

• Measurement: Determines the percentage of PB words the patient correctly identifies.

Word Recognition Score (WRS): purpose

• Estimate how a patient’s hearing loss affects speech understanding at normal conversational levels (typically

• 50 to60 dBHL

• 50 to 60 dBHL) or at a comfortable listening level.

• Compare performance between the two ears.

• Compare performance across time.

• Provide a relevant stimulus to assess patient performance with a hearing aid.

Word Recognition Score (WRS): instructions

• Patient Instructions: "I am now going to measure your recognition of words that will be presented at an audible level. You will hear an initial phrase, ‘Say the word,’ followed by the word that you are to repeat. Please repeat the word that you hear, even if you need to guess." These words do not get softer.

• Presentation Level: Presented at one level relative to either the SRT, UCL (Uncomfortable Loudness Level), or PTA.

  • Typically, +40dB above the patient's SRT. For example, if SRT = t10dBHL, WRS would be presented at 50dBHL (SRT + WRS tests are common in quiet environments.)

Word Recognition Score (WRS): clinical insights

Hearing aids might not significantly benefit individuals with very poor word recognition ability. Prolonged hearing loss can lead to the loss of neural pathways for speech understanding, as the brain may "forget" how to process speech sounds over approximately 10 years.

Speech banana

• Hearing thresholds plotted on an audiogram provide guidance on which speech sounds a person may not be hearing (e.g.,

• /F,TH,S/

• /F,TH,S/).

• Any speech sounds that fall above a person’s hearing thresholds are likely being missed, which affects the clarity and understanding of speech.

Speech in Noise (SIN) Testing

• Context: While previous tests are performed in quiet, SIN tests are designed to assess speech recognition in the presence of background noise.

• Purpose: More representative of real-world listening environments, where a smaller difference between the noise and target sound makes speech harder to hear.

• Signal to Noise Ratio (SNR):

  • Signal = 50dB HL, Noise = 45dB HL

    • 50 – 45 = +5dB SNR ((Positive SNR means the signal is above the noise)).

    • Signal = 40dB HL, Noise = 55dB HL

    • 40 – 55 = -15dB SNR (((Negative SNR means the signal is below the noise)).