In-situ AEP amplification and wireless recording of Auditory Evoked Potentials and Otoacoustic Emissions

1
In-situ AEP amplification and wireless recording
of Auditory Evoked Potentials and Otoacoustic
Emissions

• Yuri Sokolov, PhD
• Vivosonic Inc., Toronto, ON
• Early Hearing Detection and Intervention
  ConferenceAtlanta, March 3, 2005

2
Audiology clinicians experience significant
frustrations with ABR and ASSR

• Auditory Brainstem Response (ABR) and Auditory
  Steady State Response (ASSR) are quite difficult
  to administer for many clinicians, particularly
  in post-screening assessment environments
• Noise is FRUSTRATION 1 reported by 84 of U.S.
  clinics
• Noise leads to unclear results and long test
  times up to 90-120 min, typical 45-60 min per
  test
• Long test time results in low patient throughput
• Requirement of 5 kOhm impedance is challenging
  to achieve, often by abrading the skin until
  bleeding
• Abrading the skin increases the risk of infection
  (Ferree et al., 2001. Scalp electrode impedance,
  infection risk, and EEG data quality. Clin.
  Neurophysiol., 112, p. 536-544)
• The above results in higher risks and operating
  costs

Source Tannenbaum, S US infant post-screening
market survey (The Hearing Review, Jan 2005).
3
ABR and ASSR have very low amplitudes relatively
to noise, but largely coincide with noise
frequency

• AEP signal recording, analysis, and detection is
  simply about three things
• Signal,
• Noise, and
• Signal-to-Noise Ratio (SNR)
• The three major sources of noise in AEP are
• Physiological
• Electric field, RF, and power-line
• Magnetic field

S
SNR
N
The same signal can or cannot be detected
depending on noise and SNR
4
ABR has diagnostic, screening, andthreshold-findi
ng applications

• Auditory Brainstem Response (ABR) is a transient
  response, provides valuable information on
  hearing thresholds and useful for differential
  diagnostics
• Objective
• Non-invasive
• Known generators (on the opposite from ASSR)
• Well researched over several recent decades
• Responses are looked for in the time domain in
  the form of characteristic waves
• Recommended by many established NHS protocols

• Amplitude 0.1-1 µV (millionth of V)
• Frequency range 50-3,000 Hz

Source Multiple publications by J. Hall III, M.
Hyde, C. Berlin, L. Hood, and others.
5
Click-ABR is used mostly for screening and
differential diagnostics

• Diagnostic application
• Response is generated by Acoustic Nerve and
  Brainstem
• Has characteristic wave structure
• 70-80 dB nHL click typical
• Looking for Waves I, III, V, and I-III, III-V,
  I-V intervals
• Diagnostics of Acoustic Neuroma and Auditory
  Neuropathy
• Stacked ABR may detect smaller Acoustic Neuroma
• Screening application
• 30-50 dB nHL 100 µs click stimulus
• Looking for Wave V
• Typically automated detection (e.g. AABR)

V
III
I
0
10 ms
I-III
III-V
I-V
V
0
10 ms
AABR is a registered trademark of Natus Medical
Inc. Stacked ABR is a trademark of Bio-logic
Systems Corp.
6
Tone-burst ABR is used mostly for finding
thresholds

• Established and recommended protocol
• Tone bursts instead of click stimuli typically
  500 (difficult to record), 1000, 2000, 4000 Hz
• Frequency-specific
• Levels vary to find the threshold
• Looking for Wave V threshold
• Technically similar to screening click-ABR, but
  not automated
• Detect thresholds up to 80 dB HL

V
0
10 ms
7
ASSR is a promising tool for finding hearing
thresholds

• Auditory Steady State Response (ASSR) has been
  proven to provide valuable information on hearing
  thresholds, particularly in infants
• Objective
• Non-invasive
• Frequency-specific, as tone-burst-ABR
• Not site-specific (generators are unknown)
• Typically faster than tone-burst ABR
• Accurate, particularly at higher HL, above 40 dB
  HL
• Effective at severe and profound hearing loss, up
  to 110 dB HL, while tone-burst ABR is limited to
  80 dB HL

Source Multiple publications by T. Picton, S.
John, D. Stapells, and others.
8
ASSR is a frequency-specific Evoked Potential

• Auditory Steady State Response (ASSR) is a
  tone-like response present as long as stimulus is
  presented.
• Elicited by amplitude (AM) or Frequency (FM) or
  combined AMFM modulation of carrier frequencies.
• Audiometric carrier frequencies 500, 1000, 2000,
  4000 Hz
• Modulation
• 40 Hz sensitive to sleep
• 80-110 Hz insensitive to sleep
• Responses are looked for in the frequency domain
  at modulation frequencies, not carrier
  frequencies
• Thresholds for carrier frequencies

Multiple-frequency ASSR responses

• Amplitudes 10-50 nV (billionth of V)
• Frequencies AM and/or FM

82 84 86 88 92 96 Hz
Source Multiple publications by T. Picton, S.
John, D. Stapells, and others.
9
Noises are introduced by multiple sourcesin most
clinical environments

• Physiological
• EEG increases in sleep
• ECG does not decrease in sleep
• EOG, EMG decrease in sleep
• Electric and magnetic
• Power line noise 50 or 60 Hz and their harmonics
• Electric field noise
• Magnetic field noise
• Radio-frequency (RF) interferences

10
Multiple sources introduce physiological noises
in AEP recording
Noise on the scalp Frequency range, Hz Amplitude
EEG awake 3-40 5 -10 µV
EEG sleep 3-16 2 400 µV
Electrooculogram (EOG) 0.5-10 10-500 µV
Electrocardiogram (ECG) 0.5-50 80 µV 2 mV
Electromyogram (EMG) 30-500 10 µV - 2 mV
Source Cutmore, James (1999). Identifying and
reducing noise in physiological recordings. Int.
J. Physiol., V. 32, No. 2, pp. 129-150.
11
ECG noises may be stronger in infants than in
adults

• The heart is positioned more centrally aligned
  with the sagittal plain
• The heart is much larger relatively to the body
• The heart is closer to the head
• The heart-beat rate is twice higher than in
  adults
• Temporary post-natal heart conditions may
  increase ECG noise frequency - up to 100 Hz

12
Filtering after the first stage of amplification
introduces distortion in conventional AEP
amplifiers
Saturation
Distorted signal
Distorted signal
High gain in the 1st stage results in saturation
by the unfiltered, often EEG noise, i.e. reaching
the maximum voltage of the 1st stages dynamic
range. Saturation distorts the signal The 1st
stage output contains periods of the maximum
voltage, and these periods become interruptions
in EP signal after band-pass filtering (BPF).
Low gain reduces EP amplitude and
signal-to-noise ratio (SNR) at the amplifier
output. Both saturation and low gain complicate
signal detection.
13
Electric field noises are introduced through
wires and cables acting like antennas

• Introduced by
• Electronic equipment
• Electric wiring
• Improper grounding
• Typical strength of electric fields in North
  American clinics, average 5.5 V/m, range 1-200
  V/m (5 Hz 2 kHz band)
• Noise amplitude up to 10 mV (1 mV a
  thousandth of V)
• Can be reduced by
• Shielding of input-circuit wires
• Shielding of wires and circuits
• Proper grounding

Unshielded lead wires
Electric fields
Differential AEP amplifier
Source www.niehs.nih.gov/emfrapid/html/QA-Work
place.html - Web site of Environmental Health
Science, NIH, U.S. Government.
14
Magnetic field noises are introduced through
wires and cables acting like antennas

• Introduced by
• Transformers
• Electric motors and wiring
• Looped wires and cables
• Typical strength of magnetic fields in North
  American clinics average 1.7 mG, range 0.1-200
  mG (milliGauss) (5 Hz 2 kHz band)
• Noise amplitude up to 10 mV
• Can be reduced by
• Reducing wire/cable length
• Positioning, NOT moving
• Reducing loop area
• Twisting wires
• Very thick shielding (steel)

Looped lead wires and cables
Magnetic fields
Loop area
Differential AEP amplifier
Source www.niehs.nih.gov/emfrapid/html/QA-Work
place.html - Web site of Environmental Health
Science, NIH, U.S. Government.
15
Long lead wires and cables introduce large
electro-magnetic field noises in a conventional
amplifier
Amp amplifier A/D analog-to-digital
conversion DSP digital signal processing
16
RF noise may strongly interfere with EP recording

• Radio-frequency (RF) noise comes from various
  sources
• Cell phones, pagers, Blackberry, wireless
  intercom
• FM-systems, FM-radio
• Wireless computer networks used in many hospitals
• PDAs (Personal Digital Assistants), Palmtops
• Medical equipment (ICUs, operating rooms, general
  offices)
• Office equipment copiers, fax-machines,
  computers
• Introduce mostly electrical noise
• Interferes at EP (low) frequencies despite RF
  frequencies are much higher in MHz and GHz
  ranges because of amplifier non-linearity
• There is no common-mode rejection (CMR) at
  frequencies 20 kHz
• Amplitude up to 10 mV (thousandth of V)

Source Kitchin et al. (2003). Input filter
prevents instrumentation-amp RF-rectification
errors. EDN, Nov 13, p. 101-102.
17
Power line noise is not only 60 Hz and comes from
both electric field and AC power lines

• Power Line noise comes from
• Electric field picked up by electrode wires
  cables
• AC power outlets when plugged into the wall
  introduced through electronic circuits, power
  supplies
• Through USB computer ports (5 V) introduced
  through electronic circuits
• Interferes with EP at a number of frequencies
  mostly 50 / 60 Hz harmonics
• 60 Hz, 120 Hz, 180 Hz, 240 Hz due to
  amplifier non-linearity
• Amplitude up to 10 mV and higher

AC outlet
AMP
PC
USB
Power-line noise in AEP amplifiers
60 120 180 240 300 Hz
18
Low A/D resolution can significantly affect AEP
recording due to insufficient dynamic range
Integrity
Typical
Low
19
Putting all things together ABR and especially
ASSR are very small signals as compared to noises
in AEP recording
Signal Frequency, Hz Amplitude, nV (dB)
AEP Signals AEP Signals AEP Signals
ASSR 70 - 110 10 50 (0)
ABR 50 - 3,000 100 - 1,000 (10-20)
MLR 3 - 300 500 - 3,000 (15-25)
LLR 1.5 - 15 200 - 16,000 (16-60)
P300 1 - 15 5,000 - 20,000 (15-65)
Noises in AEP recording Noises in AEP recording Noises in AEP recording
Electrooculogram (EOG) 0.5-10 10,000 - 500,000 (60-85)
EEG awake 3-40 5,000 - 10,000 (55-60)
EEG sleep 3-16 2,000 - 400,000 (65-90)
Electrocardiogram (ECG) 0.5-50 (up to 100) 80,000 - 2,000,000 (70-110)
Electromyogram (EMG) 30-500 10,000 - 2,000,000 (70-110)
Electric, magnetic, RF 50/60 Hz, MHz, GHz Up to 10,000,000 (up to 120)
20
Clinical ABR/ASSR testing is challenging in
practice

• Long testing time
• Best reported
• 19 minutes (Luts, Wooters, unpublished), 21
  minute (Perez-Abalo et al., 2001)
• Typical 45-60 minutes (John et al., 2003), up to
  90 120 min (Tannenbaum, 2004)
• Sensitivity to electromagnetic interferences
• Electromagnetically shielded booth required
• Sensitivity to electrode impedance
• Requires rubbing the skin
• Need for sedation in many cases
• Difficult to administer in electro-magnetically
  shielded booth

21
In-situ AEP amplification and filtering is a
novel method of noise reduction in Auditory
Evoked Potentials

• Amplifier is mounted in-situ directly on the
  ground electrode pad, with no lead
• Lead length to non-inverting () and inverting
  (-) electrodes minimized to the distance between
  electrodes
• Filtering prior to amplification
• Gains optimized for ASSR and ABR
• Impedance mismatch monitored in real time
• Risk of wrong electrode connection minimized

22
In-situ amplification largely eliminates
electro-magnetic field-induced noises
In-situ pre-amplifier, the Amplitrode, is
mounted directly on the ground electrode
eliminating ground lead. The other leads are
very short and shielded. This significantly
reduces electric and magnetic field-induced and
allows for a clearer EP signal at the amplifier
output.
23
Filtering prior to amplification allows
optimizing gain and reducing physiological and RF
noises
Higher gain 150,000 for ASSR 15,000 for
ABR Exceptionally low intrinsic noise lt 350
nV in 10-10,000 Hz lt10 nV in 0.05 Hz bands in
70-110 Hz EP signals at the Amplitrode output
have large amplitude, contain little noise, have
high SNR, and therefore, can be easier converted
from analog to digital form, recorded, and
detected.
24
In-situ amplification and wireless communications
make AEP testing efficient

• Reduced physiological noise
• Largely reduced electromagnetic noise
• No big boxes
• Less attention to electrode impedance
• Easy mounting on electrode pads
• No need to achieve 5 kOhm impedance
• No hassles with long lead wires and cables
• Less risk of electrode lead misconnection

25
Amplitrode monitors electrode mismatch in real
time

• Electrode impedance mismatch (EIMM) is more
  relevant than electrode impedance.
• Amplitrode measures EIMM in real time during
  testing, not only prior to it.
• Operator is notified of EIMM immediately.
• Reduces set up time.
• Measuring EIMM and very high input impedance of
  the Amplitrode eliminates the need for skin
  abrasion no need to achieve impedance below 5
  kOhm.

Ferree et al. (2001). Scalp electrode impedance,
infection risk, and EEG data quality. Clin.
Neurophysiol., 112, p. 536-544.
26
Amplitrode eliminates the risk of improper
mounting

• Amplifier is mounted on the ground electrode pad.
• The other two leads have different length.
• Electrode button release makes easy mounting and
  dismounting amplifier and clips on electrode
  pads.
• It is much easier to use even for less
  experienced practitioners.

27
In-situ AEP recording speeds up testing
ABR
800 clicks
100 clicks
400 clicks
1600 clicks
Subject Normal hearing female, 24 yrs, R
ear Place Vivosonic office, EMI 0.5
mGauss Phone ER-3A (correction for 0.9
ms) Stimulus Click, 30 dB nHL, 21.1/sec, ipsi
3200 clicks
28
In ideal electro-magnetically shielded room,
the benefit of in-situ amplification and
filtering is less pronounced
ABR in a shielded room (lt1 V/m, 0.1 mG)
Recording Montage Fz/A1 10.66 ms
window Band-pass filter 30-1500 Hz for Bio-Logic
Navigator Pro 30-1200 Hz for Amplitrode
Subject T.V., 44, normal hearing Stimulus 1000
clicks 21.1 clicks per second Artifact Rejection
disabled ER-3A Insert Headphones
Source I. Kurtz, T. Venema, 2004 (unpublished).
29
Outside a shielded room, the benefit of in-situ
amplification and filtering is very significant
ABR in moderate electric (12 V/m) and magnetic (
5.5 mG) fields
Correlation coefficient 0.81
Correlation coefficient 0.43
Recording Montage Fz/A1 10.66 ms
window Band-pass filter 30-1500 Hz for Bio-Logic
Navigator Pro 30-1200 Hz for Amplitrode
Subject T.V., 44, normal hearing Stimulus 1000
clicks 21.1 clicks per second Artifact Rejection
disabled ER-3A Insert Headphones
Source I. Kurtz, 2004 (unpublished).
30
Wireless recording of OAE and AEP provides
mobility and additional noise reduction

• Wireless communication with PC
• No cable to the PC
• No noise coming back into the EP and OAE
  amplifiers from AC power supply
• No cable-related hassles
• Mobility
• Testing can be controlled form anywhere within
  the reach of Bluetooth
• The patient or a babys mother can move around
  without the need to disconnect electrodes,
  connectors, or transducers
• Adult and senior patients can take a relieving
  break.
• In the Operating Room, testing can be done from a
  distance, without cables getting in the way.
• Battery operation
• No AC-power-related noise in the amplifier
  circuits

31
Bluetooth is a wireless communications protocol

• Wireless communications protocol
• Digital signal in GHz range
• Noise-like, broadband (no fixed carrier frequency
  unlike FM-radio)
• Low energy below 0.1 mG
• Limited area 30 feet (10m)
• Encoded secure for medical information
• FDA-approved for various medical applications

Bluetooth
32
Integrity is the worlds first and only wireless
OAE, ABR, and ASSR system

• VivoLink interface module
• Generates DPOAE, TEOAE, ABR,
  and ASSR stimuli
• Conditions stimuli for
• ER-3A Insert Phones
• B-71 Bone Conductor
• Converts EP signals from the Amplitrode into
  digital form, 16 bit
• Processes signals and communicates to the
  computer software through Bluetooth
• Integrity computer program controls the OAE,
  ABR, and ASSR, functions
• Protocol setting modular
• Test control modular
• Data management - integrated

B-71
Integrity
Wireless Bluetooth communication
33
Questions?

• Thank you for your interest!