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 Conference
Atlanta, March 3, 2005
1
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 90120 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).
2
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:
S
• Signal,
• Noise, and
• Signal-to-Noise Ratio (SNR)
SNR
N
The same signal can or cannot be
detected depending on noise and SNR
The three major sources of noise in AEP are:
• Physiological
• Electric field, RF, and power-line
• Magnetic field
3
ABR has diagnostic, screening, and
threshold-finding 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.
4
Click-ABR is used mostly for screening and
differential diagnostics
V
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
III
I
0
• Stacked ABR® may detect smaller
Acoustic Neuroma
I-III
10 ms
III-V
I-V
V
Screening application
• 30-50 dB nHL 100 μs click stimulus
• Looking for Wave V
• Typically automated detection (e.g.
AABR®)
AABR® is a registered trademark of Natus Medical Inc.
Stacked ABR® is a trademark of Bio-logic Systems Corp.
0
10 ms
5
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
6
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.
7
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
AM+FM 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
Multiple-frequency
• Thresholds – for carrier frequencies
ASSR responses
• Amplitudes: 10-50 nV (billionth of V)
• Frequencies – AM and/or FM
Source: Multiple publications by T. Picton, S. John, D. Stapells, and others.
82 84
86
88
92
96 Hz
8
Noises are introduced by multiple sources
in 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
9
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.
10
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
11
Filtering after the first stage of amplification
introduces distortion in conventional AEP amplifiers
Saturation
Distorted
signal
Noise
EP
Amp 1
BPF
Distorted
signal
Amp 2
High gain in the 1st stage results in saturation by the unfiltered, often
EEG noise, i.e. reaching the maximum voltage of the 1st stage’s
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.
12
Electric field noises are introduced through wires
and cables acting like antennas
Introduced by:
•Electronic equipment
•Electric wiring
•Improper grounding
Unshielded
lead wires
Electric
fields
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
* Source: www.niehs.nih.gov/emfrapid/html/Q&A-Workplace.html - Web site of
Environmental Health Science, NIH, U.S. Government.
Differential
AEP amplifier
13
Magnetic field noises are introduced through
wires and cables acting like antennas
Introduced by:
• Transformers
• Electric motors and wiring
• Looped wires and cables
Looped
lead wires
and cables
Magnetic
fields
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)
Loop area
* Source: www.niehs.nih.gov/emfrapid/html/Q&A-Workplace.html - Web site of
Environmental Health Science, NIH, U.S. Government.
Differential
AEP amplifier
14
Long lead wires and cables introduce large electromagnetic field noises in a conventional amplifier
EP
EMI
Amp
Ground lead
Other leads
A/D
“Garbage” IN
DSP
“Garbage” OUT
Amp – amplifier
A/D – analog-to-digital conversion
DSP – digital signal processing
15
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.
16
Power line noise is not only 60 Hz and comes from
both electric field and AC power lines
Power Line noise comes from
AC outlet
• 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) –
AMP
introduced through electronic circuits
USB
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
PC
Power-line
noise in AEP
amplifiers
60 120 180 240 300 Hz
17
Low A/D resolution can significantly affect AEP
recording due to insufficient dynamic range
Dynamic Range of A/D Conversion
160.0
y = 6.0206x
R2 = 1
140.0
Dynamic Range [dB]
120.0
y = 6.0206x - 6.0206
R2 = 1
100.0
80.0
Integrity™
60.0
Typical
40.0
Low
20.0
0.0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
-20.0
A/D Resolution - Number of Bits
18
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
10 – 50
ASSR
70 - 110
(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
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)
19
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
20
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
21
In-situ amplification largely eliminates electromagnetic field-induced noises
EMI
EP
A/D
DSP
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.
22
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:
< 350 nV in 10-10,000 Hz
Noise
EP
<10 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.
23
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
24
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.
25
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.
26
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)
3200 clicks
Stimulus: Click, 30 dB nHL, 21.1/sec, ipsi
27
In “ideal” electro-magnetically shielded room, the benefit of
in-situ amplification and filtering is less pronounced
ABR in a shielded room (<1 V/m, 0.1 mG)
VivoSonic Baseline Grand Average
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).
Bio-Logic Baseline
Grand Average
Recording:
Montage Fz/A1
10.66 ms window
Band-pass filter: 30-1500 Hz for BioLogic Navigator Pro
30-1200 Hz for Amplitrode
28
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
Bio-Logic (5.5 mG;12 V/m)
Correlation coefficient = 0.43
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).
Viv osonic (5.5 mG;12 V/m)
Correlation coefficient = 0.81
Recording:
Montage Fz/A1
10.66 ms window
Band-pass filter: 30-1500 Hz for BioLogic Navigator Pro
30-1200 Hz for Amplitrode
29
Wireless recording of OAE and AEP provides mobility
and additional noise reduction
Wireless communication with PC
• No cable to the PC
o No noise coming back into the EP and OAE
amplifiers from AC power supply
o No cable-related hassles
• Mobility
o Testing can be controlled form anywhere within
the reach of Bluetooth®
o The patient or a baby’s mother can move around
– without the need to disconnect electrodes,
connectors, or transducers
o Adult and senior patients can take a relieving
break.
o In the Operating Room, testing can be done
from a distance, without cables getting in the
way.
• Battery operation
o No AC-power-related noise in the amplifier
circuits
30
Bluetooth® is a wireless communications protocol
• Wireless communications
protocol
• Digital signal in GHz range
• Noise-like, broadband (no fixed
carrier frequency – unlike FMradio)
• Low energy – below 0.1 mG
• Limited area – 30 feet (10m)
Bluetooth®
• Encoded – secure for medical
information
• FDA-approved for various
medical applications
31
Integrity™ is the world’s first and only
wireless OAE, ABR, and ASSR system
Amplitrode™
VivoLink™ interface module
• Generates DPOAE, TEOAE, ABR,
and ASSR stimuli
• Conditions stimuli for
•
•
ER-3A
ER-3A
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®
B-71
Vivo
Link™
Integrity™ computer program
controls the OAE, ABR, and
ASSR, functions
• Protocol setting – modular
• Test control – modular
• Data management - integrated
Integrity™
Wireless
Bluetooth
communication
32
Questions?
Thank you for
your interest!
33