Upgrade to remove ads
Only $35.99/year

Quiz 3

Terms in this set (89)

Pass
No follow-up is indicated
Refer
Follow-up is indicated
Screening goal
Maximize the number of "refers" in ears that have hearing loss. Maximize the number of "passes" in ears with normal hearing.
Screening Task
Determine cutoff criteria that determines pass
vs. refer.
Screening Problem
There will always be some overlap between physiologic responses obtained in ears with normal hearing vs. ears with hearing loss. In this example, some ears with normal hearing have DPOAE amplitudes on the lower end. Conversely, some ears with hearing loss have DPOAE amplitudes on the higher end. No matter what cutoff you pick, there will always be some level of misidentification. An ear with hearing loss passes the screening. An ear with normal hearing refers on the screening. We pick a cutoff that minimizes both types of misidentification. Costs money, time, and stress on the part of the patient and family when a patient is not screened accurately.
4 screening outcomes
1. True positive
2. False positive
3. True negative
4. False negative
True positive
Refer when hearing is impaired
False positive
Refer when hearing is normal
True negative
Pass when hearing is normal
False negative
Pass when hearing is impaired
Sensitivity
Sums to give you all of the people screened with hearing loss who are referred. You want this number to be very high so that you are reducing Type I error.
TP/TP+FN
Specificity
The percentage of patients with normal hearing who passed the screening. You want to maximize this number.
TN/TN+FP
Sensitivity and Specificity
High sensitivity means a high proportion of individuals with hearing loss will refer. High specificity means a high proportion of individuals with normal hearing will pass. You want to pick your cutoff values to maximize both sensitivity and specificity. There is typically a tradeoff: increasing one decreases the other.
Picking a cutoff
100% specificity cutoff
All ears with normal hearing passed (specificity = 100%). But, no ears with hearing loss referred
(sensitivity = 0%). Missing chances to diagnose early.
100% sensitivity cutoff
All ears with hearing loss referred (sensitivity = 100%). But, no ears with normal hearing passed
(specificity = 0%). Requires a lot of follow up and resources.
Optimal screening cutoff
Most ears with hearing loss referred (sensitivity = 90%). Most ears with normal hearing passed (specificity = 90%). However, some ears with hearing loss passed (10%) and some ears with normal hearing referred (10%). Passing a screening does not guarantee normal hearing. Encourage parents to come back if they suspect delays in communication or other hearing concerns.
How to pick cutoff
Systematically vary your cutoff and compute sensitivity and specificity each time to make minute changes. Plot results on a Receiver Operating Characteristic (ROC) curve: x-axis: False Positive Rate (100 - specificity), y-axis: True Positive Rate (Sensitivity). Select the cutoff that maximizes sensitivity and specificity.
Optimal cutoff ROC
Cutoff that gives you closest point to upper-left of ROC curve. This point is maximum sensitivity and specificity. Shows up on ROC curve as highest true positive rate and lowest false positive rate.
Late onset HL risk factors
Early Hearing Detection and Intervention (EHDI)
Practice of screening newborns and providing diagnostic and treatment services to those who need it. All 50 states and DC have EHDI programs (either by law or voluntarily).
Joint Committee on Infant Hearing (JCIH)
Formed in 1969. Membership from audiology, otolaryngology, pediatrics, deaf education, and public health. Provides position statement on recommended practices in early hearing detection and intervention: Latest update in 2019.
1-3-6 Guidelines
ASHA Practice Portal NBHS
Comparing ROC curves
May want to compare different screening methods e.g., are TEOAEs or DPOAEs more accurate at screening for hearing loss? Compute an ROC curve for each method, then calculate the area under each ROC curve: Larger area = higher overall accuracy
TEOAE v. DPOAE v. AABR ROCs
All are good at screening for hearing loss from 1-4 kHz. AABR performs better than OAEs at 1 kHz.
OAE parameters for NBHS
Screening parameters may be set by manufacturer and not under your control.
AABR screening
Typically 35 dB nHL clicks with alternating polarity. Statistical detection methods: Template matching: waveform compared to template
Cross-correlation: can compare sweeps SNR calculations
NBHS protocols
AABR well-baby protocol
AABR NICU protocol
OAE well-baby protocol
OAE and AABR well-baby protocol
OAE and AABR well-baby and NICU
Electrocochleography
Measurement of electrical activity of cochlea and auditory nerve. Primarily used in diagnosis of Meniere's Disease, but is also useful for other conditions.
Components of ECochG
1. Cochlear microphonic (CM)
2. Summating potential (SP)
3. Action potential (AP)
Cochlear microphonic (CM)
OHCs - flips with changes in polarity. Shearing of stereocilia causes flow of ions in and out of hair cells. This flow results in a measurable alternating current (AC). The response mirrors the stimulus, similar to what a microphone does. Measured with electrodes in the ear; CM reflects OHC activity primarily from the base of the cochlea, and look like oscillations at the beginning of the waveform.
Summating potential (SP)
IHCs - seen with alternating polarity. Basilar membrane displacement is not equal in both directions. Causes a shift in the baseline of the CM, moving the whole response up or down. Considered a DC (direct current) shift. CM can obscure both the SP and AP, so we typically cancel out the CM in order to measure SP and AP. In ECochG, typically exhibited as a leading "shoulder" on the AP: the first hump that comes before the AP.
Action potential (AP)
Peripheral portion of the AN - highest peak on the ABR after CM/SP. Synchronous firing of many auditory nerve fibers. Short-latency response (~1-2 ms), so it comes from distal (peripheral) portion of nerve - closer to the cochlea. Also referred to as: Compound action potential (CAP), N1, Wave I of the ABR
Mechanoelectric transduction
Hair cell at rest is negatively charged. Calcium influx causes the hair cells to become LESS negative and depolarize.
ECochG electrode placement
Electrode goes close to site of generation: Ear canal (TipTrode), Closet to TM (Tymptrode), Through TM (transtympanic)
ECochG stimuli
High intensity (e.g., 90 dB nHL) to elicit robust responses. Clicks or tone bursts can be used; tone bursts may be more sensitive for detecting Meniere's (Ferraro 2010). No masking needed since electrode is very close to generator site.
ECochG stimulus rate
Slow rate (e.g., 5.1 to 7.1/s) to elicit robust AP. If focus is on SP, can increase the rate to reduce the AP.
ECochG time window
Typically short (5 ms) - this is a very short-latency response
ECochG filter setting
10 to 1500 Hz. SP is a DC shift close to 0 Hz. Want to reduce high-frequency electromagnetic interference.
Stimulus polarity
CM is affected by stimulus polarity; SP and AP are not
CM obtained with condensation vs. rarefaction clicks are 180° out of phase. Use a single polarity stimulus to record the CM. Use alternating polarity to cancel CM and leave SP and AP. Can also make separate recordings using condensation and rarefaction then add together to cancel out CM.
ECochG electrodes
Trade-off between size of response and patient comfort/invasiveness. Most invasive --> Largest response.
Least invasive --> Smallest response.
Note that larger responses will require fewer sweeps, so shorter test time.
TIPtrodes
Can be placed by an audiologist. Requires clear canals. Requires gentle scrubbing of canal with NuPrep. Conductive gel goes on electrodes.
TympTrodes
Can be placed by an audiologist. Requires clear canals. Conductive gel goes on electrodes. Better amplitude than TipTrodes but less than transtympanic electrodes. High impedances are possible.
Transtympanic electrodes
Require physician to put needle electrode through TM that rests on the cochlear promontory.
Electrode placement
Horizontal montage is best. Remember dipole orientation. Typical montage:
Non-inverting - Ipsi canal/TM, Inverting - Contra canal/TM, Ground - Forehead
* This montage will result in AP pointing downward rather than upward.
What is artifact?
When the electrical signal sent to the insert earphones gets picked up by the electrodes. Especially problematic for measuring short-latency responses like the ECochG (esp. CM).
Verifying artifact presence
Make a measurement by crimping the earphone tubing (bottom of image). An electrical signal will be sent to the earphones, but no acoustic signal will reach the ear. If you still measure a response, this strongly suggests stimulus artifact is present.
Avoiding artifact
Ensure insert earphone boxes aren't touching electrode leads. Use electrically shielded insert earphones (see image). Use alternating polarity (if you only care about SP and AP).
Meniere's disease
Click ECochG analysis
Amplitude and latency are measured. Amplitude can be referenced from peak to trough (left picture), or baseline to peak (right picture). Normal mean amplitude: SP = 0.4 μV, AP = 2.0 μV
AP latency = 1.3 - 1.7 ms
Diagnosing Meniere's
Endolymphatic hydrops cause increased SP amplitude: Increased endolymph changes transduction properties of the cochlea, resulting in larger DC bias and thus larger SP amplitude. ECochG analysis involves calcucation of the SP/AP ratio: Expressed as a decimal or percentage. Large ratio suggests endolymphatic hydrops.
SP/AP Ratio
>0.45 is considered abnormal. Sensitivity = 55-65%
Specificity = >90%
Sensitivity will depend in part on whether pt is symptomatic at time of measurement. Sensitivity may be as high as 90% if you can measure during symptoms.
Click recording parameters ECochG
Tone burst parameters ECochG
ECochG troubleshooting
Endolymphatic hydrops
Hydrops causes excess of endolymph which throws off the motion of the cochlea therefore affecting the SP/AP ratio.
Normal v. abnormal SP/AP ratio
Click v. tone bursts SP/AP ratio
Clicks on the left, tone bursts on the right.
Intraoperative monitoring example
Top: Abnormal SP/AP ratio at baseline
Middle: Reduction in SP/AP ratio during sac opening
Bottom: Normal SP/AP ratio after closing sac
Survey on utility of ECochG
Otologist and Neurotologist survey results
Auditory neuropathy (AN)
Functional outer hair cells but abnormal neural synchrony. CM should be present - use ECochG to help enhance microphonic. ABR should be abnormal or absent.
ECochG and auditory neuropathy
Starr et al. (2004) suggested that we can distinguish pre-synaptic from post-synaptic neuropathy. Pre-synaptic: Both SP and AP are abnormal, suggesting IHC pathology. Post-synaptic: SP is normal but AP is abnormal. This distinction is theoretically grounded but these are not two distinct clinical diagnoses.
AN and ECochG
ECochG can be useful for enhancing the CM relative to standard ABR measurement. Measure rarefaction and condensation: CM will flip its polarity. Sum rarefaction & condensation, or use alternating polarity: CM will cancel out, Abnormal/absent ABR indicates ANSD. To assess presence of stimulus artifact, crimp the earphone tubing and make a measurement. If no response, this confirms that previous recordings were CM. If a response still present, suggests there was at least some artifact present and that you need to troubleshoot.
First: Rarefaction
Second: Condensation
Third: Sum
Fourth: Tubing crimped
ABR and ANSD
Weighted averaging
With standard signal averaging, all sweeps contribute equally to the mean waveform. With weighted averaging, noisier sweeps contribute less to the mean waveform. Weight is inverse of the variance: Higher noise = higher variance = lower weight.
Fsp technique
Provides level of confidence in presence of response. Numerator: Variance across time points in the averaged waveform. Denominator: Variance across sweeps at a single point (sp) in time. Higher value indicates higher confidence that response is present: Fsp = 3.1 indicates 99% probability of present response. Fsp is objective but still requires visual inspection: e.g., electrical line noise and post-auricular muscle artifact can result in high Fsp values.
Fmp
Involves multiple points
Standardized ABR protocols
The US has no standardized ABR protocol which introduces variability and the possibility of mistakes. There are 3 other countries that have solid protoccols for ABR:
1. National Health Services Newborn Hearing Screening Program
2. Ontario Infant Hearing Program
3. British Columbia Early Hearing Program
Identifying ABR peaks
Three outcomes: Clear response, Response absent, and Inconclusive. When determining outcome: Does it replicate? Is the latency in expected range? Is the morphology as expected? If you increase intensity, does the amplitude increase and latency decrease? When I decrease intensity below threshold, is a response absent? Other factors to consider: Is the response amplitude sufficiently large? Is the noise amplitude sufficiently low? What is the SNR?
Residual noise
Measured in uV or nV. May be provided in the software. Can be estimated visually using two replications. Superimpose waves on top of each other by baseline.
Clear response
NHS guidelines: Response amplitude of at least 0.04 uV. SNR of at least 3:1. Lowest trough is SN10 (see figure).
Response absent
NHS guidelines:
Flat tracing with no measurable response. Not always possible. Residual noise of ≤ 0.025 uV. Standard is rigorous but provides high confidence in results.
Inconclusive
No clear response but residual noise > 0.025 uV. May be a clear response that is buried in high noise. Residual noise ≤ 0.025 uV but amplitude and/or latency of wave is questionable. Unclear if response absent.
Chirp ABR
Click is broadband but stimulates base synchronously due to traveling wave mechanics. CE-chirp is engineered to stimulate regions of cochlea simultaneously. Low frequencies presented first followed by high frequencies. CE = Claus Elberling. Helps with neural synchrony, bigger responses, etc.
Octave band chirp
Narrow band chirp
Level specific chirp
ABR v. Click amplitude and latency
Increasing confidence in peak picking
Three replications at threshold and additional measurement above threshold or use a template for expected latencies to compare.
Estimating residual noise
Superimpose waveform baselines. Find amplitude where gap between waveforms is largest. Rule of thumb: average gap is 1/3 of this amplitude. Example: Largest gap = 0.03 uV. Estimated residual noise = 0.03 * 1/3 = 0.01 uV. Real-time assessment: Height (amplitude) of perturbations in a single replication. Still double check with rule of thumb. This may overestimate residual noise values calculated by the software but allows you to determine if further averaging is needed.
Gold standard ABR
Chirps v. clicks
Chirp shows waves I, III, and V while click at the same level only shows wave V
Chirps v. tone bursts
Both chirp signals show lower thresholds. Chirps also show larger amplitudes than clicks.
Students also viewed