Hearing Research, 51 (1991) 33-48 0 1991 Elsevier Science Publishers B.V. (Biomedical

HEARES

33 Division)

0378-5955/91/$03.50

01482

Development

of auditory brainstem evoked potentials in newborn A three-channel Lissajous’ trajectory study *

H. Hafner,

H. Pratt, Z. Joachims

‘, M. Feinsod

infants:

2 and S. Blazer 3

Evoked Potentials Laboratory Technion - Israel Institute of Technology, Haifa, Departments of ’ Otorhinolaryngology and 2 Neurosurgery, and 3 Neonatal Intensive Care Unit, Rambam Medical Center, Haifa, Israel (Received

5 March

1990; accepted

12 July 1990)

Auditory brainstem evoked potentials (ABEP) were recorded from 50 newborns (35-43 weeks gestational age), using three orthogonal differential electrode pairs, in addition to the widely used vertex-mastoid derivation. Potentials were evoked by alternating polarity, 75 dBnHL clicks presented monaurally at a rate of 10/s. From the records of the three orthogonal electrode pairs (nasion-inion; vertex-spinous cervical process VII; left-right mastoids), Three-channel Lissajous’ trajectories (3CLT) were derived and analyzed. 3CLT point-by-point, as well as segmental descriptors were compared with peak latencies of the vertex-mastoid derivation. Point-by-point 3CLT descriptors included apex amplitude, latency and orientation. Segmental descriptors included planar segment beginning latencies, duration and orientation. The interpretation of these results in relation to developmental aspects of the auditory system, as well as to the question of ABEP generators, is enhanced by using 3CLT descriptors of ABEP, which are more comprehensive than their single-channel counterparts. 3CLT apices correlated well with the Vertex-Mastoid defined peaks. Both peak and apex latency changes indicate that at the developmental stages surveyed in this study, development takes place in the more central portions of the pathway, whereas the peripheral portion is already relatively mature. The results also indicate a maturational change in the relative contributions of constituent generators of ABEP components.

Newborns;

Auditory

brainstem

evoked potentials;

3CLT

Introduction ABEPs have been reported to reflect the neurological status (Hakamada et al., 1981; Steven, 1984; Kileny and Robertson, 1985; Ken-Dror et al., 1987) and audiometric prognosis (Despland and Galambos, 1980; Marshall et al., 1980; Stein et al., 1983) of neonates. Most studies of neonatal

Correspondence to: Hillel Pratt, Evoked Potentials Laboratory Behavioral Biology, Gutwirth Building, Technion - Israel Institute of Technology, Haifa 32000, Israel. * A preliminary report has been presented at the Biennial Meeting of the International Electric Response Audiometry Study Group, Tokyo, 1989.

ABEPs were conducted on single channel records, and reported decreasing absolute latencies, and inter-peak latency differences (IPLDs), during the maturational process (Rotteveel et al., 1987; KenDror et al., 1987; Cohen et al., 1987; Eyre, 1988). However, considerable disagreement exists with respect to absolute values of neonatal peak latenties in the same gestational age groups (Goldstein et al., 1979; Despland and Galambos, 1980; Fawer and Dubowitz, 1982; Salamy, 1984; Cohen et al., 1987). The variability may result from differences in electrode configurations, click intensity, rate, phase or frequency content of the stimulus. The conventional single channel records represent the potential difference between the electrodes as a function of time. The wave forms are determined by the generator activity in relation to

34

the recording electrodes. Therefore, multiple electrode configurations are needed for a complete description of generator activity (Pratt et al., 1983). In a 3-channel Lissajous’ trajectory (3CLT) data are recorded from 3 orthogonal electrode pairs, and represented in 3 dimensional voltage-space (Williston et al., 1981; Pratt et al., 1983, 1984, 1985; Paquereau et al., 1986; Jewett, 1987) as the simultaneous voltage values in the three electrode pairs. 3CLT reflects all potential changes on the volume conductor’s surface. In this voltage-voltage-voltage description, time is implied as motion along the trajectory. The shape of 3CLT is not affected by the relative position of a given set of generators and the recording electrodes, provided the electrode pairs are orthogonal (Martin et al., 1987). The purpose of this preliminary study was to compare 3CLT measures with the conventional single channel ABEP records in neonates in an attempt to resolve possible electrode placement effects on neonatal ABEP which may have caused disagreements between studies, and to enhance our understanding of electrophysiological development during neonatal maturation. Methods Fifty healthy neonates, ranging in age from 35 to 43 weeks gestational age, were studied. Gestational age was determined from menstrual data, early pregnancy tests, bimanual palpation of the uterus and/or ultrasonographic examination in the first trimester. Results were usually confirmed by at least one mid-gestation ultrasonographic examination. Post-natal assessment of gestational age was determined by physical examination, Dubowitz scoring (Dubowitz et al., 1970), and according to lens scoring (Hitner et al., 1977) when appropriate. All neonates were born in a normal nontraumatic delivery without fetal distress, were appropriate for gestational age, and scored an average Apgar of 9 over 10. All subjects were examined within 36 hours of birth during spontaneous sleep. The fifty neonates were divided into 3 groups according to their clinical gestational age assessment: 18 border line prematures (10 males and 8 females) born and tested at 35 to 37 weeks gestational age; 16 full term

neonates (8 males and 8 females) born and tested at 38 to 41 weeks gestational age; 16 post term neonates (9 males and 7 females) born and tested at 42 to 43 weeks gestational age. Potentials were recorded from silver disc (9 mm diameter) electrodes arranged in 4 differential derivations: Nasion-Inion labeled ‘X’; Left-Right mastoids, ‘ Y’: Vertex-Cervical spinous process VII, ‘Z’; and the widely used Vertex-Mastoid ipsilateral to the stimulus, which we labeled ‘A’. A grounding electrode was placed on the back of the hand. The interelectrode impedance was 5 Ku or less. Potentials from each of the 4 channels were differentially amplified ( x 200000) at a band pass of 30- 3000Hz (-3 dB points, 6 dB/octave slopes). The amplified filtered potentials (Grass P511J) were averaged using 256 addresses and a dwell time of 50 ps per channel. Stimuli were clicks generated by transducing 100 ps square electric pulses in TDH-49 earphones. Potentials following 8000 monaural, 75 dB nHL, alternating polarity, clicks presented at a rate of 10/s were averaged to produce each trace. The averaged 4-channel data were magnetically stored for further analysis. Using the ‘A’ channel records and decreasing stimulus intensities, hearing threshold evaluation was conducted to verify normal hearing. Off-line analysis started with digital filtering (inverse FFT with appropriate Hamming windowing to avoid distortions) of all 4 channels to the commonly used bandpass of lOO-3000Hz. This bandpass was chosen to facilitate comparisons with adult 3CLT. Moreover, when 3CLT to wideband waveforms was compared with that of filtered waveforms, the difference seemed to be a DC shift in the entire trajectory relative to analysis origin, rather than any specific shape changes. Analysis included manual peak latency measurement in the ‘A’ (vertex-mastoid) channel, and the calculation of 3CLT measures from the ‘ X’, ‘ Y’ and ‘Z’ channels. 3CLT analysis was conducted by machine scoring algorithms as follows: First, the ‘X’ channel data were mathematically decomposed to their components in the Z direction and in the direction orthogonal to the Z/Y plane. This latter component was called ‘adjusted X’ and replaced the original ‘X’ in all subsequent analyses. ‘X’ adjustment was by 23”. as determined by

35

measurements from an anatomical atlas. 3CLTs were determined from the filtered ‘x’ (adjusted), ‘Y’ and ‘Z’ data. Analysis of 3CLT related to point-by-point attributes, as well as segmental properties of the trajectory in voltage space. Further details of our methodology were included in earlier reports (Har’El and Pratt, 1984; Pratt et al., 1985). The local rate of bending of the trajectory in voltage space was assessed by Curvature (Har’El and Pratt, 1984). Points where Curvature reached maximal values within a range of at least 6 points, were of special interest in our analysis. The point-

AUDITORY

by-point amplitude attributes of 3CLT were measured by Trajectory Amplitude, which was defined as the distance (in CLV) of each point along the trajectory from the origin of voltage space (zero potential in all channels). Local Trajectory Amplitude maxima were noted, and when such a peak coincided (within 2 data points) with a Curvature maximum, the point of coincidence was called an apex. An apex was thus a point along the trajectory where bending, as well as the absolute potential value, in any direction, were both maximal. Points with maxima in either Trajectory Amplitude or Curvature that did not coincide were not

BRAINSTEM EVOKED NEONA PAL SINGLE-CHANNEL 75dBnHL, IO/set

Left

POTENTIALS

clicks

GES TA TIONAL AGE

Right

36 WEEKS +

O.I2$4

_I +

.I

015uv

40 WEEKS + ozpv

_I

Zmsec

43 WEEKS

Fig. 1. ABEP single channel records from three representative subjects of the three gestational age &oups, in response to left and right ear stimulation.

36

analyzed further. Apices were labeled in alphabetical order of their appearance after the stimulus. Segmental analysis related to planarity of segments along the trajectory. Once apices were defined, the extent of planarity near each apex was verified. This was done by fitting a plane to the data points in increasing ranges around each apex. Fitting was conducted using a least-squares method, and the planar segment’s boundary was reached when the root-mean-square voltage (distance in voltage space) relative to the best-fit plane crossed a limit of 3.5 nV (Pratt et al., 1985). This criterion was chosen to represent deviations from the plane which were within 5% of an average distance between adjacent apices. These criteria were initially tried on 15 neonates and compared very well with manual determination of planar segments. Planar segments were described in terms of their sight vector orientation (A, B and C of the planar equation), position (D of the planar equation), the latencies of their beginning points, as well as the latency of the apex included within each planar segment. The orientation of a planar segment is described by the ‘sight vector’, which is the orthogonal line to the best-fit plane passing through the origin. Because of the non-linear nature of the cosines used to define planar segment orientation, the average orientation of planar segments, across subjects, cannot be calculated by averaging the individual orientations. In order to calculate the average orientation of planar segments, the coordinates of the ‘sight vector’ intersection with each best-fit plane were determined for each subject. The ‘X’, ‘Y’ and ‘2’ coordinates, as well as the ‘sight vector’ length (D), which are all linear, were each averaged across subjects. The average orientation of the ‘sight vector’ was calculated by dividof ing the average ‘X’, ‘ Y’ and ‘ Z’ coordinates ‘sight vector’ intersection with the plane, by the average length (D) of the ‘sight vector’. These divisions produced the ‘A’, ‘B’ and ‘C’ values, respectively, of the average orientation of each planar segment. Intersubject variability of planar segment orientation was calculated in terms of angles (‘included angles’) by which individual planar segments deviated from the average orientation. A cone representing the 90% confidence interval for sight vectors was constructed.

The 3CLT measures evaluated in this study were peak latency, planar segment orientation, beginning latency and duration, apex latency, amplitude and orientation in voltage space. Statistical evaluation of the effects of age group on apex latencies and amplitudes, as well as planar segment durations, beginning points and angles with the analysis axes included analysis of variance, with Newman-Keul’s post-hoc procedures. Correlations of ABEP measures with gestational age were assessed by linear regression. Probalities below 0.05 were considered significant. Results Single channel, ‘A’ records from 3 subjects representing the age groups studied (35-37 weeks, 38-41 weeks and 42-43 weeks) are presented in Fig. 1. In general, the records were comparable to those of adults, with peaks I, III, V and even II, IV and VI identifiable. Table I includes peak latency values for these 3 gestational age groups. 3CLTs derived from the ‘X’, ‘Y’ and ‘Z’ data, in response to left and right ear stimulation from the same 3 subjects of Fig. 1 are displayed in Figs. 2-4, with the point-by point plots of Trajectory Amplitude. Geometrical analysis enabled an objective segmentation of 3CLT into 9 planar segments, each including an apex. Apices were labeled in alphabetical order according to their temporal sequence. Planar segments were labeled according to their apices. Planar segments could be divided into 2 alternative sets: The first, included seg-

TABLE

I

NORMATIVE

SINGLE

CHANNEL

MEAN

CIES (MS) FOR THE 3 GESTATIONAL WEEKS)

35-37

38-41

42-43

PEAK LATEN_

AGE GROUPS

I

II

III

IV

V

1.71 (*0.2)

2.91 (kO.2)

4.53 (kO.3)

5.65 (+0.3)

6.83

8.40

(kO.3)

(kO.3)

1.70 (f0.2)

2.89 (fO.2)

4.42 (*0.3)

5.55 (*0.3)

6.67 (kO.3)

(*0,4)

1.66 (kO.2)

2.86 (fO.2)

4.35 (+0.3)

5.48 (kO.3)

6.57

8.10

(*0.3)

(*0,4)

For each group the upper line represents the lower line the Standard Deviations.

(IN

VI

8.26

the mean values, and

X= Nr-Ct

Ill

v

YriJ l

2 = cz-CVil

I

I

I!

I_ Q3rV

TRAJ. AMP, 6 Fig. 2.3CLT of ABEP from a representative, 36 weeks gestational age, subject in response to left and right ear stimulation (top). Tbe respective voltage-time records from which the 3CLTs were derived, as well as point-by-point plots of Trajectory Amplitude (TRAIAMP.), are also shown (bottom).

NEONA TAL 3CL T 40 WEEKS GESTA K’ONAL ?dBnHL,

f&b?G

Left

2

:

P+l

x--... , k.

-..:

..__---.

Lr

\ t

‘2

.. (;,

. 1

: i

:- kit 2.4 -

._ ; ., . .

; :.

!;

h

?2_.___-...__

: ‘\..

i

I’.....

clicks

Right

z c

AGE

\:

Y,.“..

“k, ._ ‘...,

Y.

: :

;,:a

y .,.”

‘:i 1

:’

“..__

‘A ..

.*.

x= Nz- cz

Y=Mr-M2

2 = cz- CVII

TRAJ.

AMP.

3 mrrc Fig. 3. 3CLT of ABEP from a representative, 40 weeks gestational age, subject in response to left and right ear stimulation respective voltage-time records from which the 3CLT were derived, as well as point-by-point plots of Trajectory (TRAJ.AMP.), are also shown (bottom).

(top), The Amplitude

39

43 WEEKS GESTA 77UNAf AGE

?SdBnS’L,lU/sec clicks

x= Nr- Ct

Y=MI-M2

2s Cz-Cull

TRAJ.

AMP.

+ OPpV I_

6 mscc Fig. 4. 3CLT of ABEP from a representative, 43 weeks gestational age, subject in response to left and right ear stimulation respective voltage-time records from which the 3CLT were derived, as well as point-by-point plots of Trajectory (TFUJ.AMP.),

are also shown (bottom).

(top). The Amplitude

40 TABLE

II

NORMATIVE GESTATIONAL

POINT-BY-POINT DESCRIPTORS AGE GROUP (16 NEONATES).

a1

a2

1.2*0.2 0.2kO.l 50 95

LATa AMPa AOVa LOVa

1.7*0.2 0.3 f 0.2 28 65

b

OF

NEONATAL

cl

2.7 f 0.3 0.2*0.2 78 141

3CLT

c2

3.4k0.2 0.4+0.1 35 65

4.4kO.3 0.3 +0.1 53 91

COMPONENTS

OF

d

el

5.5 +0.3 0.2 * 0.1 51 102

5.8kO.3 0.2 + 0.1 56 100

THE

42 TO 43 WEEKS

f

e2 6.6 f 0.4 0.4 f 0.2 24 40

8.OkO.4 0.2+0.1 58 95

Left ear Aa Ba Ca

- 0.16 -0.80 0.56

- 0.18 - 0.90 0.25

-0.85 0.42 -0.28

0.18 0.96 -0.18

- 0.28 -0.85 0.42

~ 0.23 - 0.76 0.60

-0.18 - 0.91 0.36

- 0.40 -0.12 0.90

~ 0.63 0.72 0.27

0.15 0.94 0.30

- 0.08 0.98 0.18

0.06 - 0.99 - 0.08

-0.40 - 0.78 - 0.48

0.57 0.05 0.81

-0.22 0.62 0.75

0.01 0.99 -0.15

- 0.54 0.02 0.84

- 0.22 -0.31 0.93

Right ear Aa Ba Ca

LATa and AMPa represent apex latency (in ms) and amplitude (in PV), respectively; AOVa stands for the average intersubject orientation variability for the apex, measured in average deviation of individual apices from the average orientation (in degrees of included angle); and LOVa signifies the 90% upper limit of apex orientation variability, also in degrees of included angle; Aa, Ba and Orientation Ca denote the respective apex orientation coefficients (cosines with analysis axes ‘X’, ‘Y’ and ‘2’. respectively). coefficients are listed separately for responses to left and right ear stimulation. Standard deviations are preceded by f

ments that approximately corresponded in their apex latencies to the simultaneously recorded vertex-mastoid peaks (I to VI). These correspond-

TABLE

ing 3CLT apices and planer segments were labeled ’ a2’, ‘b’, ‘c2’, ‘d’, ‘e2’ and ‘f. The second set, included transition planar segments which pre-

III

NORMATIVE GESTATIONAL

POINT-BY-POINT DESCRIPTORS AGE GROUP (16 NEONATES).

al

a2 1.7 f 0.2 0.4 f 0.2 40 66

OF

NEONATAL

b

cl

2.6 f 0.3 0.2 + 0.3 69 122

3.7*0.4 0.4+0.1 46 120

LATa AMPa AOVa LOVa

1.1*0.2 0.3 + 0.1 67 133

Left ear Aa Ba Ca

- 0.06 - 0.99 0.12

- 0.21 - 0.97 0.06

0.36 0.93 - 0.04

Right ear Aa Ba Ca

- 0.25 0.73 - 0.64

-0.16 0.98 0.06

-0.38 - 0.92 - 0.01

3CLT

d

c2

0.46 0.87 0.14

-0.44 -0.87 -0.23

COMPONENTS

4.4 f 0.4 0.3 * 0.1 41 76

5.5 f 0.3 0.2 + 0.1 56 103

OF

el 6.1 f 0.3 0.2 f 0.1 51 81

THE

38 TO 41 WEEKS

e2 6.8kO.4 0.5 f 0.1 31 61

f 8.2 f 0.4 0.3 f 0.1 53 94

0.26 - 0.69 0.67

0.05 - 0.66 0.74

0.20 -0.81 0.54

- 0.65 0.19 0.73

- 0.70 0.70 -0.13

0.61 0.15 0.77

- 0.32 0.90 0.30

- 0.03 0.98 0.19

- 0.61 - 0.05 0.79

- 0.07 - 0.87 -0.50

LATa and AMPa represent apex latency (in ms) and amplitude (in PV), respectively; AOVa stands for the average intersubject orientation variability for the apex, measured in average deviation of individual apices from the average orientation (in degrees of included angle); and LOVa signifies the 90% upper limit of apex orientation variability, also in degrees of included angle; Aa, Ba and Orientatior Ca denote the respective apex orientation coefficients (cosines with analysis axes ‘x’, ‘Y’ and ‘Z’, respectively). coefficients are listed separately for responses to left and right ear stimulation. Standard deviations are preceded by f

41 TABLE

IV

NORMATIVE GESTATIONAL

POINT-BY-POINT DESCRIPTORS AGE GROUP (18 NEONATES)

al

a2 1.8kO.3 0.3 f 0.1 23 43

b 2.850.3 0.2*0.1 53 103

OF

NEONATAL

cl

3CLT

c2

3.5 + 0.3 0.4*0.1 31 75

COMPONENTS

d

4.5 f 0.3 0.3kO.l 43 72

OF

el 5.6+0.2 0.2 kO.1 38 71

THE

35 TO

e2

6.2+0.2 0.2 + 0.1 62 115

37 WEEKS

f

LATa AMPa AOVa LOVa

1.1*0.2 0.2*0.1 60 119

6.9+0.2 0.5 +0.2 23 39

8.3kO.3 0.3 +0.1 48 95

Left ear Aa Ba Ca

0.07 - 0.27 0.95

-0.12 - 0.98 0.15

0.14 0.92 -0.35

0.01 0.98 -0.19

0.40 -0.51 0.76

0.23 -0.81 0.52

0.52 -0.65 - 0.55

- 0.60 -0.17 -0.78

- 0.78 0.56 0.28

Right ear Aa Ba Ca

- 0.08 0.45 0.89

-0.16 0.91 0.38

0.32 - 0.78 -0.53

- 0.20 - 0.78 - 0.60

0.20 -0.17 0.96

- 0.42 0.41 0.81

- 0.37 0.38 0.85

- 0.60 0.05 0.80

- 0.70 - 0.50 0.54

LATa and AMPa represent apex latency (in ms) and amplitude (in /.tV,), respectively; AOVa stands for the average intersubject orientation variability for the apex, measured in average deviation of individual apices from the average orientation (in degrees of included angle); and LOVa signifies the 90% upper limit of apex orientation variability, also in degrees of included angle; Aa, Ba and Orientation Ca denote the respective apex orientation coefficients (cosines with analysis axes ‘x’, ‘Y’ and ‘Z’, respectively). coefficients are listed separately for responses to left and right ear stimulation. Standard deviations are preceded by +

TABLE

V

NORMATIVE SEGMENTAL DESCRIPTORS TIONAL AGE GROUP (16 NEONATES)

OF NEONATAL

3CLT

COMPONENTS

OF THE

42 TO 43 WEEKS

GESTA-

al

a2

b

cl

c2

d

el

e2

f

LATp DURp AOVp LOVp

0.9 + 0.2 0.5 f 0.1 98 135

1.5*0.3 0.5kO.2 106 155

2.5kO.3 0.4*0.1 75 130

3.2+0.3 0.4*0.1 81 128

4.1*0.3 0.5 * 0.2 64 105

5.2+0.3 0.5 f 0.2 75 126

5.6kO.3 0.5 + 0.2 73 110

6.4 f 0.4 0.5 f 0.2 97 120

7.6 f 0.4 0.6 +0.2 95 160

Left ear AP BP CP DP

- 0.26 -0.68 0.68 - 0.01

-0.34 -0.44 0.83 - 0.05

0.40 0.82 - 0.41 0.06

0.43 0.90 0.06 0.05

0.30 -0.88 0.36 0.07

-0.23 - 0.87 0.44 - 0.01

-0.31 - 0.68 0.66 0.03

-0.51 - 0.06 0.86 0.00

-0.31 0.85 0.43 - 0.01

Right ear AP BP CP DP

-0.02 0.10 0.12 - 0.02

-0.11 0.96 0.25 - 0.02

-‘0.24 0.65 0.72 - 0.06

0.13 - 0.63 - 0.77 0.02

0.31 0.35 0.88 0.03

0.60 0.66 0.46 0.04

0.45 0.86 0.22 0.04

- 0.68 -0.16 0.71 - 0.04

- 0.08 -0.95 -0.31 0.02

LATp and DURp represent planar segment beginning latency and duration, respectively; AOVp stands for the average intersubject orientation variability of the sight vector, measured in average deviation of individual sight vectors from the average orientation (in degrees of included angle); and LOVp signifies the 908 upper limit of sight vector orientation variability, also in degrees of included angle; Ap, Bp and Cp denote the respective plane orientation coefficients (cosines with analysis axes ‘x’, ‘Y’ and ‘Z’, respectively). Dp represents the plane’s position (length of sight-vector, in PV). Orientation coefficients are listed separately for responses to left and right ear stimulation. Standard deviations are preceded by f

42 TABLE

VI

NORMATIVE SEGMENTAL DESCRIPTORS TIONAL AGE GROUP (16 NEONATES) al LATp DURp AOVp LOVp

a2

0.9 * 0.2 0.4kO.l 55 91

b

1.5kO.3 0.4kO.l 73 117

OF NEONATAL

Cl

2.4i0.3 0.4*0.1 57 92

3.5 * 0.4 0.4 f 0.1 66 93

3CLT COMPONENTS

OF THE

38 TO 41 WEEKS

GESTA-

c2

d

el

e2

f

4.2 f 0.4 0.4kO.l 76 118

5.2*0.3 0.5 kO.2 66 132

5.8 k 0.5 0.6 f 0.2 100 140

6.5 + 0.4 0.5 * 0.2 95 161

7.9 i 0.5 0.6 * 0.2 126 156

Left ear AP BP CP DP

0.61 ~ 0.76 0.22 0.05

0.34 - 0.94 0.02 0.01

0.80 - 0.04 -0.60 0.07

0.44 0.89 0.13 0.08

0.24 - 0.94 0.22 0.01

0.86 - 0.42 0.30 0.03

~0.52 - 0.85 0.01 ~ 0.05

-. 0.08 0.27 0.96 -- 0.01

- 0.80 0.46 - 0.40 - 0.04

Right ear AP BP CP DP

0.60 0.66 - 0.45 0.01

0.07 0.78 - 0.61 0.01

0.42 - 0.90 0.12 0.06

0.74 -0.60 0.30 0.1

0.37 0.42 0.82 0.02

~ 0.43 0.85 0.30 - 0.02

0.07 0.10 0.02 - 0.03

- 0.76 0.63 - 0.12 ~ 0.02

~ ~

0.60 0.44 0.67 0.08

LATp and DURp represent planar segment beginning latency and duration, respectively: AOVp stands for the average intersubJect orientation variability of the sight vector. measured in average deviation of individual sight vectors from the average orientation (in degrees of included angle); and LOVp signifies the 90% upper limit of sight vector orientation variability. also in degrees of included angle: Ap. Bp and Cp denote the respective plane orientation coefficients (cosines with analysis axes ‘X’. ‘Y’ and ‘2’. respectively). Dp represents the plane’s position (length of sight-vector, in nV). Orientation coefficients are listed separately for responses to left and right ear stimulation. Standard deviations are preceded by f

TABLE

VII

NORMATIVE SEGMENTAL DESCRIPTORS TIONAL AGE GROUP (18 NEONATES)

3CLT COMPONENTS

OF THE 35 TO 37 WEEKS

GESTA-

a2

b

cl

c2

d

el

e2

f

0.9 + 0.3 0.5 & 0.2 61 96

1.5 +0.3 0.5 * 0.2 76 132

2.5 +0.3 0.5 kO.3 100 144

3.3kO.4 0.5 +0.2 78 126

4.2 f 0.4 0.5 +0.3 66 105

5.3kO.3 0.6 + 0.2 104 140

5.9kO.3 0.5+0.1 102 141

6.6 * 0.3 0.5 i 0.2 8X 127

8.1 F 0.3 0.5 * 0.1 108 147

0.53 -0.82 0.18 0.02

- 0.01

0.40 0.75 PO.53 0.02

0.34 0.94 0.06 0.08

0.65 -0.15 0.74 0.06

0.0x

0.55

- 0.94 0.31 - 0.02

~ 0.64 0.54 0.02

~ 0.27 0.42 0.87 0.02

-0.71 0.64 -0.31 -0.06

0.53 0.30 0.80 0.02

0.37 0.66 0.65 0.03

0.15 -0.61 -0.78 0.02

0.35 ~0.08 0.93 0.07

- 0.06 0.10 0.04 - 0.04

-0.6X 0.01 0.73 -- 0.04

- 0.51 0.50 0.70 -- 0.03

-0.56 PO.57 0.60 0.01

al LATp DURp AOVp LOVp

OF NEONATAL

Left ear AP BP CP DP

- 0.93 -0.35 0.00

Rtght eat AP BP CP DP

~ -

0.78 0.45 0.43 0.03

LATp and DURp represent planar segment beginning latency and duration, respectively; AOVp stands for the average intersubject orientation variability of the sight vector, measured in average deviation of individual sight vectors from the average orientation (in degrees of included angle); and LOVp signifies the 90% upper limit of sight vector orientation variability, also in degrees of included angle; Ap, Bp and Cp denote the respective plane orientation coefficients (cosines with analysis axes ‘X’, ‘Y’ and ‘Z’. respectively). Dp represents the plane’s position (length of sight-vector. in PV). Orientation coefficients are listed separately for responses to left and right ear stimulation. Standard deviations are preceded by +

43

AUDITORY

BRAINSTEM

EVOKED

POTENTIALS

NEONA TES APEX/PEAK

LA TENCY CORRELA TON

75dBnHL, 27

clicks

1 :.. .. . . : . .

a2

.

.

.

.

*:

‘.:

:

. ...:.: . . . . .

. . .

a,

1..

I3

IWsec

. . .

:’ 22

13 I

8

5s

:. .1

‘1: . : . . . . I ‘: . . . . . . . .

c2

.a... 1

.

1. .

.

::::’

e2 .

.

.

I

.

.

.

.

.

.

.

I. 52

38 111

Fig. 5. Scatter plots demonstrating the correlation between apex latencies and the respective peak latencies. The correlations between a2 and I, c2 and III as well as e2 and V were significant, while apex b and the second peak were not significantly correlated.

ceded planes ‘a2’, ‘~2’ and ‘e2’, and were labeled ‘al’, ‘cl’ and ‘el’. Apex measures for the 3 gestational age groups are listed in Tables II-IV, and planar segment attributes for the 3 gestational age groups are provided in Tables V-VII. Note the changes in apex latencies. The scatter plots in Fig. 5 demonstrate the correlation between the single channel peak latencies and the respective 3CLT apex latencies (these correlations were made across all age groups). There was a highly significant positive linear correlation between all component pairs (R = 0.5, P < 0.0001 for I and ‘a2’; R = 0.7, P -c 0.0001 for III and ‘~2’; and R = 0.8, P -C0.0001 for V and ‘e2’), except for apex ‘b’ and peak II. A significant effect of gestational age group on ‘al’ and ‘a2’ apex amplitudes was indicated by analysis of variance procedures (F(2,96) < 0.03; F(2,98) < O.OOOl;, respectively). However, using linear regression analysis, only the correlation be-

tween ‘f’ apex amplitude and gestational age was significant (b = - 1.2; P < 0.009). This is most probably due to the larger amplitudes of the 38-41 weeks gestational age group relative to the other two groups, except for ‘f’, where amplitudes were smallest in the 42-43 weeks gestational age group. Apex latencies were also significantly affected by neonatal gestational age group. ‘el’, ‘e2’ and ‘f were significantly affected with post-hoc procedures indicating that the change occurred between 38 and 42 weeks was significant. A significant inverse correlation was found for 3CLT apex latencies as a function of gestational age for apices ‘d’, ‘el’, ‘e2’ and ‘f (b = - 23.6 P < 0.04; b = - 49 P < 0.0001; b = - 38.1 P -e 0.006 and b = - 48.7 P < 0.002; respectively). Latencies of the ‘A’ channel peaks (III, IV, V and VI) also showed inverse correlations with gestational age (b = -28.7 P < 0.008; b = -24.2 P < 0.05; b = - 37.6 P < 0.003 and b = - 44.58 P < 0.002 respectively). When interpeak latency differences (IPLDs) in

44

the ‘A’ records were evaluated as a function of gestational age, a negative correlation was found for the V-I and III-I IPLDs (b = - 31.8 P < 0.005 and b = - 22.8 P -c0.01respectively). However, no significant correlation with age was found for the 3CLT inter apex latency differences. No significant effects on segmental duration were found, except for a single effect of gestational age group on ‘al’ planar segment duration (F(2,96) < 0.04) which was shortest in the 38-41 gestational age group. The effect of gestational age group on the latencies of planar segment beginning points was significant only for planar segments ‘cl’ ‘el’ and ‘f (F(2,96) < 0.008; F(2,77) < 0.01 and F(2,93) < 0.001; respectively) which were all shortest in the 42-43 weeks gestational age group. The orientation changes as a function of gestational age were evaluated using two measures: The effects of age group on apex and planar segment sight vector angles with each of the analysis axes; or on the ‘included angles’ between respective apex orientations or planar segment orientations in the age groups compared. ‘e2’ apex angle with the X and Z axes was significantly affected by gestational age group (F(2,98) < 0.0201 and F(2,98) < 0.0112; respectively). A comparison of the ‘included angles’ for the apices revealed a significant change only for apex ‘f’ between 35 37wks and 38-41wks gestational age (included angle between age groups was larger than LOVa for ‘f). In contrast to the apex findings, neither method revealed any significant effect of gestational age group on planar segment orientation (all included angles between groups fell within the respective LOVp limits of Tables V to VII). Discussion In this study, neonatal ABEPs were characterized using the 3CLT analysis as well as the widely used single channel derivation. Earlier studies (Pratt et a1.,1983, 1984, 1985; Sininger et al., 1987) demonstrated the reproducibility of some characteristics of 3CLT across adult subjects, enabling identification of homologous components across subjects. This study confirms those earlier reports, extending them to neonates, in whom the maturational processes have not yet ended. There

was a definitive identification of all 3CLT latency measures with intersubject variability similar to that of the corresponding ‘A’ channel measures. Thus, the 3CLT planar segments are reproducible enough, even across neonates during the maturational process of their nervous system, to enable their follow-up and quantitative evaluation. The advantage of 3CLT over single channel records is its more comprehensive representation of auditory brain stem activity, providing all the data necessary for calculating a centrally located dipole equivalent of the surface-recorded activity (Pratt et al.. 1983; Scherg, 1984). The shape of the neonatal 3CLT (Figs. 2-4) differs from that of adults (c.f. Pratt et al., 1983, 1984, 1985; Sininger et al., 1987). The adult 3CLT of ABEP includes 5 prominent loops that are planar in nature, with the first two positioned roughly along the coronal plane and the third to fifth more vertical in orientation, the fifth being the most prominent. In contrast, in the neonatal 3CLT the most prominent planar segment is the third loop (‘~2’) rather than ‘e2’ which is less developed than in adults. Moreover, the orientation of ‘e2’ is not as vertical (along the ‘Z’ axis) as that of adults. When 3CLT apex latencies in neonates were compared with the conventional single channel peak latencies, a highly significant linear correlation between all corresponding components was observed (Fig. 5). Apex ‘b’ and peak II of the single channel record were an exception. The well known low detectability of the second component in neonates may account for this nonsignificant correlation. The larger variability of segment ‘b’ (and peak II) and their lower detectability may be related to its generators (Pratt et al., 1985; KenDror et al., 1987). In some reports, the generator of the second component is assumed to be in the pontomedullary junction (Starr and Hamilton, 1976; Stockard and Rossiter, 1977). Others indicate that apex ‘b’, and the corresponding peak II, derive from activity of the auditory nerve and possible overlapped brainstem contributions (Moller et al.. 1981; Scherg and Von Cramon, 1985; Martin et al., 1986; Gardi et al. 1987). The overlapped complex generators may explain the lower detectability of the second component (relative to its ad-

45

jacent components) in term and pre-term neonatal ABEPs (Salamy and McKean, 1976; Starr et al., 1977). The slower transmission along the immature auditory pathway could reduce the overlap of sequential generators. Gestational age group had a significant effect on ‘al’ and ‘a2’ apex amplitudes, but this effect did not result in a linear correlation. A significant correlation with gestational age was found for ‘f apex amplitude. Improved synchronization and enhanced high frequency contributions, and/or advancing myelination processes with maturation may be responsible for these effects. However, progressing myelination and enhanced high frequency contributions with maturation may not be appropriate explanations, because no significant latency decrease was observed for ‘al’ and ‘a2’ apices. Furthermore, among the cranial nerves, the eighth nerve is the first to show myelinated fibres with myelination being evident from the 14th fetal week in the vestibular nerve deriving from the semicircular canals (Yakovlev and Lecour, 1967). In the brain stem the fiber systems mediating the acoustic modality myelinate early and rapidly before birth (Yakovlev and Lecour, 1967; Lemir et al., 1975). Specifically for ABEP, Krumholz et al. (1985) did not find maturational changes in the V/I amplitude ratio range as a function of conceptional age between 30-43 weeks compared to adult values. Rotteveel et al. (1987) however, did report a persistent and significant amplitude increase, particularly between 30 and 40 weeks, but this effect was limited to peak V. A negative correlation with gestational age was observed for apex, as well as peak latencies beginning with the fourth component (‘d’ or IV). However, in the ‘A’ channel, a significant correlation was also observed for peak III. The results for the ‘A’ channel are consistent with previous reports which indicate decreasing latencies with gestational age, particularly in the later components (III, IV, v) (Goldstein et al., 1979; Despland and Galambos, 1980; Fawer and Dubowitz, 1982; Cohen et al., 1987; Rotteveel et al., 1987). The significance of this correlation for peak III and its non-significance for apex ‘c’, may be explained by a change in orientation of the equivalent generator, relative to the recording electrodes. In 3CLT such changes affect apex orientation but not

latency. In contrast, single channel records may show a marked latency difference as a result. Indeed, apex ‘e2’ angle with the X and Z axis were significantly affected by gestational age group. Similar effects were reported by Pratt et al. (1987) on such a discrepancy between single channel and 3CLT results in pathological records. Such effects may explain at least some of the variability and disagreement about neonatal peak latencies in previous studies. These effects are unlikely to result from head shape changes with maturation because orientation changes were observed for only some components (e.g., ‘e2’) but not others. Head shape changes should affect orientations of all components similarly. Calculation of IPLDs as a function of gestational age, for the ‘A’ channel results, revealed an inverse correlation for V-I and III-I. Peripheral effects that might mimick maturation due to developing sensitivity of the auditory apparatus (Jewett and Romano, 1972) are unlikely in this study because hearing threshold was determined by the ‘A’ channel records and found to be the same across gestational age groups. The inverse correlation of III-I latency with gestational age indicates a maturational process of the auditory nerve and lower brain stem. In contrast, 3CLT inter-apex latency differences did not show any correlation. Thus, the ‘A’ channel IPLD decreases may reflect changes in the relative contributions of constituent generators and not from accelerated conduction. Decreasing IPLD with gestational age in the ‘A’ channel, and its absence in 3CLT, may thus stem from changes in the relative contributions of the constituents of the compound generators of the involved components, and not from conduction acceleration with development. The prominence of planar segment ‘c2’ in neonates, relative to that of adults, is compatible with this suggestion of maturational changes in constituents of ABEP complex generators. The variability of planar segment measures (beginning point and orientation) is greater than that of the respective apex measures (latency and orientation), and their sensitivity to maturational processes was therefore smaller. Variability in planar segment measures was observed for both intersubject and side to side (within the same subject) comparisons. Apices are probably more

46

reliable measures for quantifying the activity of brainstem generators. In conclusion, both peak and apex latency changes in the later ABEP components indicate that at the developmental stages surveyed in our study (35-43 weeks gestational age), development takes place in the more central portions of the pathway, whereas the peripheral portion is already relatively mature. The orientational changes in apex ‘e2’ indicate a maturational change in the relative contributions of constituent generators of the fifth ABEP component. Aknowledgements The assistance of the Neonatal Department staff in Rambam Medical Center, graphical work by Ms. Naomi Bleich and the cooperation of our subjects’ parents are all gratefully acknowledged. References Cohen. B.A., Kovnar, E.H., Chadi, R. and Graham, M. (1987) Brainstem auditory evoked potentials in normal term and premature infants. Electromyogr. Clin. Neurophysiol. 27. 469-480. Despland, P.A. and Galambos, R. (1980) The auditory brainstern response (ABR) is a useful diagnostic tool in the intensive care nursery. Pediat. Res. 14, 1544158. Dubowitz, L.M.S.. Dubowitz, V. and Goldberg, C. (1970) Clinical assessment of gestational age in the newborn infant. J. Pediat. 77,1-10. Eyre. J.A. (1988) Neurophysiological assessment of the immature central nervous system. Br. Med. Bull. 44, 1076-1092. Fawer, CL. and Dubowitz, L.M.S. (1982) Auditory brainstem response in neurologically normal preterm and full-term newborn infants. Neuropediatrics 13, 200-206. Gardi, J.N., Sininger, Y.S., Martin, W.H., Jewett, D.L. and Morledge, D.E. (1987) The 3-channel Lissajous’ trajectory of the auditory brain-stem response. VI. Effects of lesions in the cat. Electroenceph. Clin. Neurophysiol. 68, 360-367. Goldstein, P.J., Krumholz, A., Felix, J.K.. Shannon, D. and Carr, R.F. (1979) Brainstem evoked response in neonates. Am. J. Obstet. Gynecol. 135, 622-628. Hakamada, S., Watanabe, K., Hara, K. and Miyazaki. S. (1981) The evolution of visual and auditory evoked potentials in infants with perinatal disorder. Brain Develop. 3. 339-344. Har’El. 2. and Pratt, H. (1984) Geometric analysis of shortlatency evoked potentials. Math. Biosci. 69, l-10. Hitner. H.M.. Hirsch, N.J. and Rudolph, A.J. (1977) Assessment of gestational age by examination of the anterior vascular capsule of the lens. J. Pediatr. 91, 455-458.

Jewett, D.L. and Romano, M.N. (1972) Neonatal development of auditory system potentials from the scalp of rat and cat. Brain Res. 36, 101-115. Jewett, D.L. (1987) The 3-channel Lissajous’ traJectory of the auditory brain-stem response. IX. Theoretical aspects. Electroenceph. Clin. Neurophysiol. 68, 386-408. Ken-Dror, A.. Pratt. H., Zeltzer. M., Benderley. A. (1987) Auditory brain-stem evoked potentials to clicks at different presentation rates: estimating maturation of pre-term and full-term neonates. Electroenceph. Clin. Neurophysiol. 68, 209-218. Kileny, P. and Robertson, C.M.T. (1985) Neurological aspects of infant hearing assessment. J. Otolaryngol. 14(Suppl.), 34-39. Krumholz, A.. Felix, J.K., Goldstein. P.J. and McKenzie, E. (1985). Maturation of the brainstem auditory evoked potential in premature infants. EEG. Clin. Neurophysiol. 62. 124-134. Lemir, R.J., Loeser. J.D., Leech, R.W. and Alvord, E.C. (1975) Normal and abnormal development of the human nervous system. Harper and Row, MD. Marshall, R.E., Reichert, T.J., Kerely. SM. and Davis, H. (1980) Auditory function in newborn intensive care unit patients revealed by auditory brainstem potentials. J. Pediat. 96. 731-735. Martin. W.H., Pratt, H. and Bleich. N. (1986) Three-channel LissaJous’ trajectory of human auditory brainstem evoked potentials. II. Effects of click intensity. Electroenceph. Clin. Neurophysiol. 63, 54-61. Martin. W.H., Jewett. D.L.. Randolf, M.G., Williston, J.S. and Garde, J.N. (1987) The 3-channel Lissajous’ Trajectory of the auditory brain-stem response. IV effects of electrode position in the cat. Electroenceph. Clin. Neurophysiol. 68. 341-348. Mailer. A.R.. Jannetta, P.J.. Bennett. M. and Mailer. M.B. (1981) Intracranially recorded responses from the human auditory nerve: New insights into the origin of brainstem evoked potentials (BSEPs). EEG. Clin. Neurophysiol. 52. 18-27. Paquereau, J.. Marillaud, A., lngrand, P. and Kremer-Merere, Ch. (1986) Three-dimentional curves: main parameters of brain-stem auditory-evoked responses in the normal subject. Audiology 25, 1077115. Pratt, H., Har’EI, Z. and Golos, E. (1983) Three-Channel Lissajous’ Trajectory of human auditory brainstem evoked potentials. Electroenceph. Clin. Neurophysiol. 56, 682-68X. Pratt. H.. Har’El, Z. and Golos, E. (1984) Geometrical analysis of human three channel Lissajous’ trajectory of auditory brain-stem evoked potentials. Electroenceph. Clin. Neurophysiol. 5X, 83-88. Pratt, H., Bleich, N. and Martin, W.H. (1985) Three channel Lissajous’ trajectory of human auditory brain stem evoked potentials. I. Normative measures. Electroenceph. Clin. Neurophysiol. 61, 530- 538. Pratt. H., Bleich, N. and Sussel, Z. (1987) Three-channel Lissajous’ trajectory of auditory brainstem evoked potentials in patients with neurological lesions affecting the brainstern. Preliminary impressions. Audiology 26. 247-256.

47 Rotteveel, J.J., de Graaf, R., Colon, E.J., Stegeman, D.F. and Visco, Y.M. (1987) The maturation of the central auditory conduction in preterm infants until three months post term. II. The auditory brainstem responses (ABRs). Hear. Res. 26, 21-35. Salamy, A. and McKean, C.M. (1976) Maturational changes in auditory transmission as reflected in human brain-stem potentials. Brain Res. 96, 361-366. Salamy, A. (1984) Maturation of the auditory brainstem response from birth through early childhood. J. Clin. Neurophysiol. 1, 293-329. Scherg, M. (1984) Spatio-temporal modelling of early audiometry evoked potentials. Rev. Laryng. (Bordeaux). 105, 163170. Scherg, M. and Crarnon, D.V. (1985) A new interpretation of the genarators of BAEP waves I-V: Results of a spatio-temporal dipole model. EEG Clin. Neurophysiol. 62, 290-299. Sininger, Y.S., Gardi, J.N., Morris, J.H., Martin, W.H. and Jewett, D.L. (1987) The 3-channel Lissajous’ trajectory of the auditory brainstem response. VII.Planar segments in humans. EEG Clin. Neurophysiol. 68, 368-379. Starr, A. and Hamilton, A.E. (1976) Correlation between confirmed sites of neurological lesions and abnormalities of far

field auditory brainstem responses. Electroenceph. Clin. Neurophysiol. 41, 595-60. Starr, A., Amlie, R.N., Martin, W.H. and Sanders, S. (1977) Development of auditory function in newborn infants revealed by auditory brainstem potentials. Pediatrics 60, 831-839. Stein, L., Ozdamar, O., Kraus, N. and Paton, G. (1983) Followup of infant screened by auditory brainstem response in the neonatal intensive care unit. J. Pediat. 103, 447-453. Steven, A.J. (1984) Hearing screening of high-risk newborns with brainstem auditory evoked responses. Int. Med. Care J. (Hospimedica) 9- 10, 83-88. Stockard, J.J. and Rossiter, V.S. (1977) Clinical and pathological correlates of brainstem auditory response abnormalities. Neurology 27, 316-325. Williston, J.S., Jewett, D.L. and Martin, W.H. (1981) Planarcurve analysis of three channel auditory brainstem response: a preliminary report. Brain Res. 223, 181-184. Yakovlev, P.I. and Lecour, A. (1967) The myelogenic cycles of regional maturation of the brain. In: Minkowski, A. (Ed.), Regional development of the brain in early life. Blackwells, Oxford, pp. 3-65.