Electroencephalography and clinical Neurophysiology , 1983, 55:619 636 619

Elsevier Scientific Publishers Ireland, Ltd.

Clinical Section

B R A I N S T E M A U D I T O R Y , P A T T E R N - R E V E R S A L VISUAL, AND S H O R T - L A T E N C Y
S O M A T O S E N S O R Y E V O K E D P O T E N T I A L S : L A T E N C I E S IN R E L A T I O N T O AGE, SEX,
A N D BRAIN AND B O D Y S I Z E !

T R U E T T ALLISON, C H A R L E S C. W O O D and W I L L I A M R. G O F F
Neuropsychology Laboratory, Veterans Administration Medical Center, West Haven, Conn. 06516, and Department of Neurology, Yale
University School of Medicine, New Haven, Conn. 06510 (U.S.A.)

(Accepted for publication: February 16, 1983)

Evoked potentials (EPs) generated in the audi-
tory, visual and somatosensory systems can be
recorded from the head and neck and are useful in
the assessment of neurological disorders (e.g., Hal-
liday 1978; Starr et al. 1978; Celesia 1982; Cour-
ion et al. 1982; Stohr et al. 1982). Reliable identifi-
cation of EP abnormalities requires statistical
characterization of EPs in an appropriate popula-
tion of neurologically normal subjects. From
normative results already published it is clear that
age, sex and body size influence the characteristics
of auditory (AEP), visual (VEP) and somato-
sensory (SEP) components (e.g., Asselman et al.
1975; Celesia and Daly 1977; Kritchevsky and
Wiederholt 1978; Stockard et al. 1978, 1979; Kjaer
1980; Shaw and Cant 1980; Shearer and Dustman
1980; H u m e et al. 1982). Most studies have con-
centrated on the 20-50 year age range since most
diagnostic problems involve adult-onset symptoms
suggestive of central nervous system dysfunction.
However, there has been no comprehensive de-
scription of normative results for the 3 modalities
of stimulation over a wide range of ages in a large
sample of males and females. The present study
attempts to fill that need and to assess the factors
1 This work was supported by the Medical Research Service of
the Veterans Administration, by USPHS Grant MH-05286 and
by N I H Special Research Resources Grant RR-3 to the Health
Sciences Computing Facility, U.C.L.A., Los Angeles, Calif.,
U.S.A.
which are likely to be important in the characteri-
zation of statistical norms. Preliminary results have
been reported (Allison et al. 1979).
M e t h o d and Material
Subjects were 130 males and 156 females rang-
ing in age from 4 to 95 years. Methods of stimula-
tion and recording were chosen to reflect typical
practice at the time this study was begun. Audi-
tory stimuli were 75 dB SL, 10/sec, monaural
clicks alternating in polarity and produced by 0.1
msec pulses delivered to Clark 12507-G earphones.
Thresholds were determined by the staircase
method and were considered accurate to + 2 dB.
A masking white noise (45 dB HL) was delivered
to the unstimulated ear. Averaged potentials to
1024 clicks were obtained from a C z to ipsilateral
earlobe derivation with filter settings of 30-3000
Hz ( - 3 dB points) at a sampling rate of 0.04
m s e c / p o i n t . Checkerboard-reversal monocular
visual stimuli were back-projected onto a screen
via a g a l v a n o m e t e r / m i r r o r apparatus. Each check
subtended 50' of visual field; the full-field checker-
board subtended 16 ° . Central fixation was moni-
tored by closed-circuit TV. Intensity of the light
squares was 1750_+ 50 c d / m 2 and of the dark
squares 45 c d / m 2. Averaged potentials to
128 2 / s e c reversals were obtained from O~-Fz and
O2-Fz derivations with filter settings of 1-300 Hz
at a sampling rate of 1 msec/point; VEP P100

0013-4649/83/0000-0000/$03.00 © 1983 Elsevier Scientific Publishers Ireland, Ltd.

620
latency was taken as the average latency at the two
derivations. Somatosensory stimuli were 5/sec, 0.5
msec duration constant-current square wave pulses
delivered to the median nerve at the wrist at an
intensity producing a just-detectable thumb twitch.
Averaged potentials to 512 pulses were obtained
with filter settings of 30-3000 Hz at a sampling
rate of 0.08 msec/point. Four derivations were
used: ipsilateral midclavicle to Fz; C7 and C2
spinous processes to Fz; and contralateral parietal
scalp (P3 or P4) to linked ears. Subjects were seated
in an electrically and acoustically shielded record-
ing chamber maintained at 22-24°C. Three aver-
aged potentials were obtained to stimulation of
each eye, median nerve and ear.
No subject had a history of relevant neuro-
logical or diabetic disease, or of drug abuse. 'Nor-
mal' was more difficult to define in older subjects;
some had cardiovascular disease, hearing loss re-
garded as normal for age (correctable to 75 dB
SL), or slight ocular opacities (but with acuities
better than 20/100 in all cases). Subjects were
thought to be a representative sample of the neuro-
logically normal population, but no attempt was
made to select an unusually healthy sample. Since
it can be assumed that patients exhibit similar
non-specific changes to the same degree, the use of
only the healthiest normal subjects could lead to
an increase in false-positive results.
Only peak latencies were analyzed. As noted by
many others, the amplitudes of these potentials
vary widely in the normal population, and onset
latencies are often difficult to determine. Peaks
occasionally fragmented into two or more sub-
peaks; in such cases latency was taken as the
midpoint of the peaks. AEP and SEP interpeak
latencies using the peripheral nerve volleys as
benchmarks, some commonly used central inter-
peak latencies, and the absolute difference in
latency between components evoked by left and
right stimulation (LR difference) were also de-
termined. Statistical analysis was performed using
the BMDP series of computer programs (Dixon
1981). With the exception of LR differences, all
graphic and statistical summaries were based on
analysis of left (ear, eye, median nerve) data to
avoid the use of two correlated measures from the
same subject; comparisons of left and right data
T. ALLISON ET AL.
are noted below. 'Significant' will refer to two-
tailed tests of significance with P < 0.01.
Peak and interpeak latencies as a function of
age with linear regression lines representing the
mean and upper limit of normality (ULN) were
plotted, separately for age and sex groups where
appropriate. U L N was defined as + 3 standard
deviations (S.D.) above the regression line, using
the standard error of estimate as the estimate of
variability. Differences in latency related to age,
sex and height were assessed by analysis of covari-
ance. These regression methods assume normality,
linearity and homoscedasticity. Visual inspection
and analysis of residuals indicated no appreciable
departure from normality and linearity. Bartlett's
test was used to assess the assumption of homo-
scedasticity. Of 104 tests of AEPs and SEPs only
18 were significant, and of these 14 were of SEP
peak latencies in adults, in whom height dif-
ferences were a large source of variance and were
removed as described below. VEP latency variabil-
ity changed significantly with age; there were other
grounds as well for treating age groups separately.
Thus, we regard the assumptions underlying the
following analyses as being adequately met.
Results
Representative AEPs, VEPs and SEPs are il-
lustrated in Fig. 1. Components were labeled
according to 'the polarity/latency convention
(Donchin et al. 1977). The variability across repli-
cations illustrated in Fig. 1 is representative of
most subjects but was greater in (usually older)
subjects having significant scalp or neck muscle
activity. While the neural origin of components is
not of primary importance here, the following
consensus will be useful in interpreting the results:
AEP P2-P6 (waves I - V of Jewett and Williston
197 l) primarily reflect potentials generated respec-
tively in acoustic nerve, cochlear nucleus, superior
olive, lateral lemniscus and inferior colliculus (e.g.,
Jewett and Williston 1971; Starr et al. 1978; Stock-
ard et al. 1978); VEP P100 reflects activity of
visual cortex (e.g., Halliday 1978); SEP N10 is the
median nerve compound action potential at the
level of the brachial plexus (e.g., Cracco and Cracco
N O R M A T I V E A U D I T O R Y , VISUAL A N D S O M A T O S E N S O R Y EPs 621

P26
P6 P15
SEP P22 , ( ~ "

N13b N1 ~ .~ 2 0
Cz-A i

I I i I I Fz-C2 ~F-~'-- ~LY--.J' N13a~ '// \~:~-P\

0 4 8 12 16

P100

2- z Fz-S i

I I I I I I I
0 100 200 5 10 15 20 25 30
msec msec
Fig. 1. Representative potentials recorded from derivations shown; relative positivity is plotted upward at electrode location given
first. Only components analyzed in this study are labeled. A i = ipsilateral earlobe; S i = ipsilateral shoulder (mid-clavicle); Pc =
contralateral parietal scalp (P3 or P4).

1976; Jones 1977); the N l 2 and N13 complexes
are respectively the presynaptic spinal afferent vol-
ley and postsynaptic spinal and lower medullary
activity (e.g., Jones 1977; Allison and Hume 1981;
Desmedt and Cheron 1981a); the N14-P18 se-
quence reflects activity in more rostral levels of
somatosensory afferent pathways, while N20 and
later components primarily reflect activity of
somatosensory cortex (Hume and Cant 1978;
Kritchevsky and Wiederholt 1978; Allison and
Hume 1981; Desmedt and Cheron 1981b; Allison
1982; Kimura and Yamada 1982; Stohr et al.
1982).
Fig. 2 shows AEP latencies as a function of age.
The results may be summarized as follows: (1) the
peak latencies of all components except P12 in-
creased significantly with age; (2) some interpeak
latencies also increased significantly with age; (3)
most latencies showed a steeper increase with age
in males than in females, but the slope difference
was significant only for P2-P5; (4) all peak laten-
cies except P12, and all interpeak latencies except
P2-P3 and P2-P12, were significantly longer in
males than in females; (5) all latencies were essen-
tially identical ( < 1% difference) to left and right
ear stimulation and showed no systematic left-right
difference.
Fig. 3 shows VEP P100 peak latency as a func-
tion of age: (1) Latency decreased to the adult
level by about age 20, remained stable until up-
 622
T. ALLISON ET AL

. +.o~
. . . . . . . . . .~: . . . . . . . =. , ~. . . . . . ~:1 . . . . . . . . . . . .
: ~o~co cJ~ ~
g o ~ t =Tw ,13 o

i t: ~- ~= ~ =
o

• 'ta.. t

+" ~ ,-,- ~ ~ =.,--, _r.: • ..,~=" = '"

o

• .tz u_t I:: ~ ta..[ I~: IE ~

: ' zu- ~ -* t I~:

-t~
, ta_ t
• =u_
Ic

u.q _ u.. ~:

~*

• ~ ~ i ~ ~!

: ~: :~,,u_~_ ,,~

; o2 rq ~ ~ .' :

~A

I I
• • . . . . + . . . . +>_ , , . + , . + . . . . + . . .~-+ . . . . + . . . . . + . . . . :~ . . . . +

111

:
• .
' . . . , . , .

0
.<

•
= +~ + z+
o<,,=.+ z+
2~ + =+ + +.
+" ~ =~t~ ~ 1 "/ ~,....,~, t / ~"", :

.: o ~: ~ ,, ,, ~ .... : ....

+. ,.. o
tfi~_ =~"" • • . . ~ . . :
.t.m
• ~o.~ -/. ~ ~ ¢"?.'~. ~c,i ,-2.

o I~ ~- ~ o ~,,o0
0 r--. ~- ~ o o. . "= ~oo~o
"

+~ : :z c.,i t.m :z c-,/t.~ _.~'.~

c~ t.m -= = ~ - , .,.¢,
+t.,~

I : ta..
: :~
• " u_ t.~
.;~_ ,-,- +.~.

+"
:
~
~
+. o
.u_ lieu.. ~:
!~
u_ :t.,a
÷ ~
ta_
: ~i~ ..o
~t.._ +ca
: ~i :'*'~

+.
: ~;~ :
:~ ~ _,u_
ta. +.~

°:I I
o

: . .+

~'~
. . . . +

~
. .~ .+

~
. . . . .+

m
. . . . +

o
.~- . .. + .

m
. +

c~
. . . . +

m
. . . . +

o
, .~_ ,+

m
. . . . . . +

L~
t . . . . ~

0
. . . . + ,
i.

I
,~DNalVl
NORMATIVE AUDITORY, VISUAL AND SOMATOSENSORY EPs 623

...N • .] - . ~ ] '+ . . . . + . . +: • . .+ . . . . . + .~ • .m... . +

~'+~
. . .

:~÷÷
Z ~ N
< d z.~.~ ~. ~ :
:m ~_~ ~ :~
+~ ,,mo~ ~ o
~o. +oo
:
: u_ ~qq
o~ ii ii =,=, .: ~
u_

+-~

g o ~z u - u_ ~,~ ~..
: ~cn =% +" o

:m ~-~.~
) --u_ z ~ -
<co
i ~- ~:,Z.~
.'N ~ ~ =~_ =

~- ~ :
z~-
:m ~u_ - u,_ u _ ~ m :ua
~ ..~ ~ . ~

~ o ~u_

:u-,
~ t m
u-u_ .
• N
~-u.. c-
:m I~:~u u _ •m
+c--i
u. ~_.
~ a
:ua
~ ..
: ~ t,,
'~: ~ ~.- G_ r~ •
' ~4 :u_ ~ ~ ~:~- '

• " + " ' '+ . . . . ' . . . . >-" '+ . . . . . + . . . . +~-" " "+ " ' " : 111 ~ ~ ~.~, ~1~ ~

<

=T ";

u_

...a~ • u_

0 ,..4 ~
u_: u_ ~ z . o

= ~ ~ * ,~. ==~ •

u_ x: : ,~, +, ='-

~-~ ~_~z . ~-
~ ~.~-~ ~ =
~-~ol~- ~-' ~. ~- :~',
~ > ~ ~ ~ ' ~..
~

~_4~_ •
"~ ~ B .~'~

• . .+ . . . . + .~ ,+>< + • .>-+ . . . . . . . + . ,>- .+ . . >_ , . + , . . :

~o" I ......
,A o8 ~ :1: o 03 I:q ~° o8
ADN]lV]
624 T. ALLISON ET AL.
~-.+ .... + .... + .... + .... + .... + .... + .... + .... + .... + .... + .... + .... + .... + .... *..*X + .... + ~ ~

130 ; PIO0

120

110 ~-F F ]09 IM M[MMFM ..-'F ;

>-. : 11o5 ~ FM. I zJ 'M*MFM
• .F.M M F * MMlt ' I-M . . t / FFF ~ Y
0 • M*~.*M ~FM~IM M M M M .MMMM M ~ M ~
Z IOO + M-M. F M M FF MMMMMF MI_EI~T-f~M F F F +
LIJ • FF *F.~FM M M F~M FF F * F ~ F F F *
• FM*FFF'-. MFMF * M* "I FF F F •
< Y FM FM F - FF •
• FF F FMF [:E ~F* FF[:E~'F F FF F i F
_ I 9O. + F F *'MMFM. FFMFFF F * F F. M IF +
M FF FF F r,FFI F F
• FF F I •

80. + N MEAN SD R SL MFsL REGRESSION EQ +

4-19 61 99.3 5.8 - . 3 3 S NS L=-.4717*A+105
20-59 M 65 96.0 4-4 .26 L= .0949"A+92.3
F 83 92.5 4-2 - . 0 2 NS NS L=_.OO55*A+92.8
70. + 60-95 M 33 107.3 6.9 -66 L= .5637"A+67.8 +
F 42 101.7 7.7 .24 S NS L= .2549"A+83.8
• . . + . . . . X . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . + . . . . +.
5. 15 25 35 45 55 65 75 85 95
I0 20 30 40 50 60 70 80 90
AGE
Fig. 3. VEP PI00 latency; details as in Fig. 2.

proximately age 60, and increased thereafter. Al- (4) Latency to left and right eye stimulation was
though various non-linear functions could be used essentially identical in the 4-19 and 60-95 age
to summarize the data, there is no theoretical groups. In the 20-59 age group latency to right
justification for choosing among them. Hence for eye stimulation was 1 and 2 msec longer in females
ease of clinical use, linear regression analysis was and males respectively, and the ULNs were corre-
applied to the 4-19, 20-59 and 60-95 year age spondingly longer (106 and 111 msec in females
groups. (2) In the 4-19 group latency decreased and males respectively).
significantly with age and showed no significant Figs. 4-6, left column, show SEP peak latencies
sex difference. In the 20-59 group latency showed as a function of age: (1) In children, latencies
no significant age-related change, but was signifi- increased rapidly with age, closely paralleling the
cantly longer in males than in females. In the increase in height (Fig. 7) and corresponding in-
60-95 group latency increased significantly with crease in length of the conduction pathway. (2)
age and was significantly longer in males than in Height and latencies reached asymptote at about
females. (3) Apparent slope differences of latency 17 years; thus the 4-17 age range ('children') and
with age were observed between males and females, the 18-95 age range ('adults') were analyzed sep-
but in no age group was the difference significant. arately. (3) In children there was no significant sex
 NORMATIVE AUDITORY, VISUAL AND SOMATOSENSORY EPs 625

f . ._÷ . . . . ÷ + . . . . . . . . . + .... + . ~

• ÷ I . . . . i-'~ . . . . . . . +++
¢',1
c6 ! ....
z + + +
-**=
~ o

,, ,, ,, +"

E E
~ z

t~.
= ~ "2".'=? E u_

~LL E ;~LL " " "

.~ ~ ZO0~C:~
~LL <C~O0
L ~222 <O0~O

+ta.~ E*
E, =
E u
=:ELL
~. z m E
~.~ ~~ - ~,~
_!~ ~- ~ ~:,~ ~ ,~
_-%-~. E .E ~'~
_ u_ ±
~ -,~
~.u
E:
= ,a

+.~ Z I ~ I~ Z ~ Z ~: ~ +'~
14=I

0
.+ . . . . + . . . . + .t.,~
+.o-,

~ 3 ~ . . . .

÷+÷

• o.~.
...... +:

* ~E U-:Z Z - E u..z z +~
~z
~co t~ +'

= ~ 'q"O.

\ I ~ : I - " ~ ~ ~ :-:.~" \ ~ " •i .~ ~ ~ " ~o

':
- -
~ • ~ =: =_ +. =

:o__O " ~ ~ 0 ~- ~: C',,I

" ,',
O ' :El'= -
~E= =E
::

• .+ . . . . . + , . , + , .+ . . . . + . . ÷ • .÷

~- ~- ~- ~ I~- ~- ~- ~- ~I~- ~ ~- ~-
ADN31V]
 626
T. ALLISON ET AL.
,G-, . . . . . . . .+ .+
• - . . . . . . . . . . . ~

: k" ~ = +++
o .,,~< ,-~:
~ =+++ =+++ •
°'¢'¢"¢ ,~
"~ "~ . . . . . . . . . . ~ . . . . . . . . . . . . . I ~"~ ~ . . . . . . . . . . . . . -. _., . ., , ~ r,,
r~' '* l ~..~.¢..¢ :

• ~-~ ~: ~ ~ ~E ~ ~ ~
" ~ ~ . . . . . ~-~'::~ ~- ~. . . . ~ ~ ~ ~ i
u_

+" ~ * ~ ~ =,.~,-...~ , ~_ ~_ ~: ~-~ . . . .

,~.~w,. ~u_E = ~ ' ~ ~:,~.u_ • • •

. . . . . E E ~ u_~. ~- ~ : ~ Elu_ . ~Eu- L--:[ _ ~ " ~ . . . .

. m = E LL m E -

~D "~ ~ ~ r~ ~ :[_U-- ~ ,,~ -- " ~ ~ ~ +

+-'~ .:~:, ~ ~- _,::., ,~ ~ ~:~_ . ~- :~

E- ~. u_ _ ELL • c~

: ~ .=~ . . . ~ . ~ . . . ~.~ ,o ~ I:L ,,~ ::

Z Eu_ u_ i",1 u_ ~u_
,._ r',
~
L,_ o
,•
: ~ oZ
•+.'~ ~- . ~ o = = o ~:~- o ~ ~: "- ~ : ~

/.. .I ~ \ E ~ Z .i ,LF-~ LL /.-- ~ "/ ~P-- '

.÷1+ 7! ~ I "°1 ~ I '~/ ~-~ +~

• "~ . . . . + . . . . . . . + . . . . + ~ " ' "+ . . . . ÷ l" ' '+ . . . . + ' "~ '+ . . . . + . . . . .

+'~ l . . . . . :~ ~m "c~ o~2 ~ ~.. ~. 9 o0 e~ ~o :

~" '~ ~o ~- ~- ~ A,,A~ :
++÷

I I / .......I ...........
+-~ -- ~ .... :

- ~ . ~ . ~ ~ . . . :
...... ...... ......

E LL j
• : ................. ,~.......... ~:Ii ~-~ ...... ~,
E ~ +" O
L~

; L L ~• E LL ~-~°2- ½ ~ ~ ~ ' "

: ~ ~ ._ ..=.'EL'- ~- =,=u_
~.. ~ ~ ~- ,~ ~- ~:u.. :

+: ~ ~-._ ~-

~ .

+,'~.. . . . . . . . . . . ~ . 1 ..............................................
ADN]I~I
NORMATIVE AUDITORY, VISUAL AND SOMATOSENSORY EPs 627

• .+ .... + +.~

] I . . . . . .

.
. . . . .

. . . .

" '~] ............ 1 ÷ "'

• ~- ==~ ~

+'~ .~ ~ ~ T= ~ ~_
m~

0 ~-~- ~- +" ?

: a_ ~ Z !~ ~ , +'~
o • ~ o i~=: o
. 0 ~ -- !.
Zo- I ~ ,- ~ ~ ~ ~

........ +..~
~ o ~ -

z++÷ •
~ ~.*%~..~
+co

+. o

.. o
:m
u_ ~"9.°9.'9. +.'~

.. o
m
: ~ ,z~:u_ : o

:m : 3~ 2v
2
g o

ka- :m
+m

;°~ ik
~ ~ :o
:ca

; o 0

Z
~ . , .+ . . . . + . . . . + ,>- . .+ .:<

9
+ . . . . + . . . . >- .>< , -+1+ • . ,+ . . .÷ .>~

I +. .
ADN]IV]
628 T. ALLISON ET AL.

÷ ÷
80.
M M
M MM ~ M M MM
M M MM •
70. + M~ *
• FFMFM**M*FFF~F F F FFFF~
FFM FFFFFFF FM*F F FF *F F
oo F M FF FFF FF * F M •
(I) +Y ..*MF FF FFF F F F FFFF F FFF F F ~ I-- ----M-
¢- 60.
0
t-

5o. + N MEAN SD R sLMFSL REGRESSIONEQ +

CHILDREN 54 57.6 7-8 .92 S NS H= 2.000"A+34.7
i ADULTSM 103 69-6 2.9 -.25 H=-.O384*A+71.5
UJ ,~M F F 129 64.2 2.7 -.28 S NS H=_.O408.A+66.2
÷ +
7 - 4O.
• ,,X .... +,o..+.,,.÷.o,,÷,.o.+o,o°÷...,+o,.o+..°,+ .... +,°,,+-°.°+°°,°+ .... ÷,-,-+ .... +,,°°÷°,-°÷o
5. 15 25 35 45 55 65 75 85 95
i0 20 30 40 50 60 70 80 90
AGE
F i g . 7. B o d y h e i g h t i n r e l a t i o n to a g e ; d e t a i l s a s i n F i g . 2.

difference in latency. Hence single mean regres- 0.76-0.80 and thus accounted for about 60% of
sion and U L N lines are plotted for each compo- the variance in latency. ULNs which depend jointly
nent. (4) In adults all latencies increased signifi- upon both age and height are difficult to present
cantly with age and showed a steeper increase with graphically, but can be calculated easily from the
age in males than in females, but the difference in multiple regression equations given in Table I.
slope was significant only for some of the earlier Figs. 4-6, right columns, show SEP interpeak
components. (5) In adults all latencies were signifi- latencies as a function of age: (1) In children there
cantly longer in males than in females. (6) In were no significant sex differences. With the ex-
children and adults, latencies were slightly ception of N13a-N20, which decreased signifi-
(0.5-1.5%) but consistently later to right than to cantly with age, most interpeak latencies increased
left median nerve stimulation. However, these dif- with age, but the increase was significant only for
ferences are near the limits of resolution in de- N10-N12 and N10-N13. (2) In adults there was a
termining latency from Figs. 4 - 6 and need not be significant age-related change only for N10-NI2a,
plotted separately. (7) The best predictors of peak N10-N20, N10-P22 and N13a-N20, the first de-
latency differed between age groups. In children, creasing with age in parallel with the decrease in
latency was highly correlated with age (r = height (Fig. 7), and the latter increasing with age.
0.57-0.89, range across components) and height (3) In adults all interpeak latencies with the excep-
( r = 0.68-0.94), but not with sex. Age and height tion of N 10-P26 and N 13a-N20 were significantly
were so highly correlated ( r = 0.92) that either longer in males than in females. (4) In children
variable alone was a good predictor. In adults, age, and adults, latencies to left and right median nerve
sex and height all were significant predictors of stimulation were essentially identical and showed
latency. Multiple regression analysis showed that no consistent left-right difference.
age and height combined provided better predic- Most LR differences in all modalities increased
tion than either variable alone; the range of multi- slightly with age, probably reflecting the generally
ple correlation coefficients across components was poorer signal-to-noise ratio and resulting increased
N O R M A T I V E A U D I T O R Y , V I S U A L A N D S O M A T O S E N S O R Y EPs 629

TABLE I T A B L E II

Regression equations for prediction of the upper limit of nor- Absolute difference in latency (msec) to left and right stimula-
mality of SEP components in adults. tion.

Component Upper limit of normal Mean S.D. ULN

NI0 L = 0 . 0 3 4 4 " A + 0 . 1 1 8 6 " H + 0.8545 + 1.866 AEP

L = 0 . 0 1 7 6 " A + 0 . 1 4 1 2 " H - 0.0345 + 1.448 P2 0.12 0.09 0.39
N12a L = 0.0282 *A + 0.1275" H + 2.074 + 1.943 P3 0.13 0.10 0.43
L = 0 . 0 1 7 5 " A + 0 . 1 3 8 5 " H + 1.485+ 1.608 P4 0.13 0.11 0.46
Nl2b L = 0 . 0 3 6 2 " A + 0 . 1 1 9 7 " H + 2.987 + 2.048 P5 0.12 0.10 0.42
L = 0 . 0 2 0 9 " A + 0.1569* H + 0.7826 + 1.691 P6 0.11 0.09 0.38
N13a L = 0 . 0 3 5 2 " A + 0 . 1 2 8 0 " H + 3 . 7 6 0 + 1.726 P7 0.20 0.16 0.68
L = 0.0202"A+ 0.1521"H + 2.344+ 1.609 P9 0.34 0.31 1.3
N13b L = 0.0346"A + 0.1433"H + 3.050 + 1.927 P12 0.69 0.55 2.3
L = 0.0199"A +0.1341"H + 3.832 + 1.642 P2-P4 0.12 0.11 0.45
N14 L = 0.0361"A + 0.1209"H + 5.104+ 1.873 P2-P6 0.13 0.10 0.43
L = 0 . 0 2 0 2 " A + 0 . 1 5 9 5 " H + 2.772 + 1.734 P4-P6 0.14 0.12 0.50
P15 L = 0 . 0 3 6 9 " A + 0 . 1 2 6 3 " H + 5.388 + 2.048
VEP
L = 0 . 0 1 9 8 " A + 0 . 1 6 8 8 " H + 2.833 + 1.798 Pl00 1.9 1.5 6.4
P16 L = 0.0323"A + 0.1149"H + 7.966 + 2.080
L = 0.0200*A + 0 . 1 6 4 2 " H + 4.703 + 1.838 SEP
P18 L = 0.0324*A+0.1345*H + 8.227 + 2.188 N10 0.21 0.18 0.75
L = 0.0239"A + 0.1737"H + 5.605 + 1.843 N 12a 0.28 0.23 0.97
N20 L = 0.0444"A + 0.1376"H + 8.541 + 2.237 N 12b 0.27 0.21 0.90
L = 0.0314"A+ 0.1609"H + 7.012+ 1.934 N13a 0.25 0.20 0.85
P22 L = 0 . 0 5 8 8 " A + 0 . 1 1 9 9 " H + 12.51 + 3.049 N13b 0.26 0.21 0.89
L = 0 . 0 4 3 6 " A + 0 . 1 9 4 2 " H + 7.730 + 3.306 Nl4 0.27 0.23 0.96
P26 L = 0 . 0 3 7 5 " A + 0 . 1 3 1 5 " H + 16.68 + 5.064 P15 0.28 0.25 1.0
L = 0 . 0 1 3 6 " A + 0 . 2 2 2 4 " H + 11.79 + 4.660 P16 0.33 0.26 1.1
P18 0.36 0.29 1.2
For each component, upper equation is for males, lower is for N20 0.37 0.28 1.2
females. Equations are of the form L = a * A + b * H + c + d, where P22 0.57 0.54 2.2
L = latency in msec; a, b = regression coefficients; A = age in P26 0.94 0.87 3.6
years; H = height in inches; c = y intercept; d = 3* (standard N10-NI3a 0.21 0.18 0.75
error of estimate about the multiple regression line). N 10-N20 0.34 0.29 1.2
N 13a-N20 0.34 0.28 1.2

latency variability in older subjects. This trend was

significant only for SEP N12a and N12b, which
were sometimes small and difficult to quantify in
the elderly. Because the age-related increase was
small relative to the variability of these measures,
and because there were no significant sex dif-
ferences, a single age-independent U L N for all
peak latencies, and the commonly used interpeak
latencies, is given in Table II.
Discussion
Interlaboratory use o f normative data
Since latencies are dependent on the effects of
unstandardized parameters of stimulation and re-
cording, it has been necessary for each laboratory
to determine its own norms. This is to be expected
for the research-oriented laboratories in which the
tests were developed. However, with the advent of
commercially available apparatus for clinical EP
testing these procedures will likely be used in
laboratories without the resources to perform ex-
tensive normative studies. Since reservations have
been expressed about the interlaboratory transfer
of normative data (e.g., Halliday 1978; Stockard et
al. 1978; Shearer and Dustman 1980), the follow-
ing discussion will focus on this issue.
The method of recording AEPs used here is
630
common, and these norms should be similar to the
values that would be obtained elsewhere. This is
especially true of the interpeak latencies, which do
not vary appreciably over the range of intensity
(60-80 dB H L or SL) commonly used for neuro-
logical assessment. Some laboratories use clicks
less intense than the 75 dB SL level used here.
Several studies suggest that in this intensity range
the increase in latency with each 10 dB decrease in
intensity is 0.10-0.25 msec (Lieber and Sohmer
1973; Starr and Achor 1975; Stockard et al. 1979).
In 26 ears we compared AEPs to 75 and 65 dB SL
clicks; the mean latency difference for all compo-
nents (between-component differences were not
significant) was 0.16 msec, in agreement with the
values noted above. Thus in this intensity range
our norms probably could be used if a 0.16 msec
correction was applied for each 10 dB change in
intensity.
Interlaboratory comparisons of VEP P100
latency continue to be difficult due to the different
stimuli used (e.g., Halliday 1978; Allison et al.
1979; Stockard et al. 1979; Shearer and Dustman
1980). However, in the 20-59 age range our calcu-
lations show that the correlation of latency with
age in different studies is: 0.05 (Asselman et al.
1975), 0.11 (Celesia and Daly 1977), 0.08 (Stock-
ard et al. 1979), 0.34 (Shaw and Cant 1980), 0.40
(Shearer and Dustman 1980), 0.12 (Sokol et al.
1981), and 0.10 (present results). That is, most
laboratories find little or no age-related change in
this age range. All workers agree that latency
increases with age in the elderly. Differing age-re-
lated changes have been reported in younger sub-
jects and may be due in part to differences in
luminance (Shaw and Cant 1980) and check size
(Sokol et al. 1981).
The SEPs described here appear to be compara-
ble to those obtained in other laboratories, which
usually use stimulus intensities near threshold for
a thumb twitch; latencies appear not to change
appreciably in this intensity region ( H u m e and
Cant 1978; Small et al. 1980). However, dif-
ferences in electrode placement may affect latency.
There have been two approaches to recording
short-latency SEPs; one uses a non-cephalic refer-
ence site such as hand or knee (e.g., Cracco and
Cracco 1976; Kritchevsky and Wiederholt 1978;
T. ALLISON ET AL.
Desmedt and Cheron 1981a, b), while the other
uses a frontal scalp reference (e.g., Matthews et al.
1974; Jones 1977; H u m e and Cant 1978). Each
method has advantages and disadvantages. For
clinical purposes we chose the latter approach
because it provides better signal-to-noise ratios.
Since only 4 recording channels were available in
this study it was not possible to compare latencies
using both types of derivation.
In summary, these considerations suggest that
our AEP and SEP norms could be used elsewhere
if stimulating and recording conditions approxi-
mate those described here. Comparison of our
methods and results with those reported in the
literature suggests that modest differences in tech-
nique have no appreciable effect on latencies,
especially on interpeak latencies. However, VEP
P100 latency appears to depend on stimulus
parameters whose interactive effects are unclear.
Characterization of normative data
Sensitive detection of latency increases due to
neurological disorder requires that normative
standards take into account all other systematic
sources of variance in latency. Here we discuss
what appear to be the most important sources of
such variance in light of these and other normative
data.
Age. Most normative studies have not ana-
lyzed latency as a function of age. Our results
demonstrate significant age-related changes in most
peak and interpeak latencies. Aside from the in-
creased sensitivity obtained by taking age into
account when assessing a patient, these results are
also relevant to issues regarding maturational and
aging changes in sensory systems; they will be
discussed in a later paper.
Sex. Previous workers have described sex dif-
ferences for some AEP and VEP components and
suggested that they were due to ~ex differences in
brain size (e.g., Stockard et al. 1978, 1979). We
assessed this hypothesis as follows. The brain
volume of males is on average about 10% greater
than in females (Dekaban and Sadowski 1978). It
is reasonable to assume that this relatively small
variation in brain volume is isometric; that is,
differences in all dimensions are proportional (e.g.,
the difference in distance between the cochlear
N O R M A T I V E A U D I T O R Y , VISUAL A N D S O M A T O S E N S O R Y EPs
°I
,.o, .... ~ t F~ ~
102
i oo
P2 P3 P4 P5 P6 P7 P9 PI2 P2 P2 P2 P2 P2 P2 P2 P4
P3 P4 P5 P6 P7 P9 PI2 P(~
Fig. 8. Predicted and observed sex differences in latency. Thin
and thick lines are respectively the mean _ 3 S.D. values of the
predicted ratio of m a l e / f e m a l e linear distance within the brain,
Bar graphs show the observed ratios of m a l e / f e m a l e AEP peak
latencies (left), AEP interpeak latencies (center) and VEP P100
peak latency (right); all values based on entire subject sample,
nucleus and inferior colliculus is accompanied by
a proportional difference in the length of the optic
nerve). If so, any pathway in the brain will vary in
length as the cube root of brain volume (Schmidt-
Nielsen 1975). The extensive data on age- and
sex-related variation in brain weight (which is
proportional to volume) of Dekaban and Sadowski
(1978) were used to determine the mean and S.D.
of the cube root of the m a l e / f e m a l e ratio in brain
volume; these are 1.034 + 0.008. In Fig. 8 this
ratio is compared with the ratio of m a l e / f e m a l e
latency for each AEP and VEP component. All sex
differences in latency are within the predicted
limits, indicating that brain size is the relevant
variable underlying the sex differences. However,
since no estimate of individual brain size was
obtained, the effect of brain size on latency can be
accounted for best in the present data by the use
of separate male and female norms, as in Figs. 2
and 3.
In contrast to the auditory and visual pathways,
only a small portion of the total length of the
somatosensory afferent pathway is intracerebral;
the contribution of differences in brain size to the
observed sex differences in SEP peak latencies is
therefore minimal. Not surprisingly, several lines
of evidence indicate that these sex differences re-
sult from differences in body size and proportion:
(1) In children there was no significant sex dif-
ference in height, and correspondingly no dif-
ference in latency. (2) Latencies are highly corre-
lated with height or arm length (e.g., Matthews et
al. 1974; H u m e and Cant 1978; Kritchevsky and
631
I~
Wiederholt 1978), and we find similar high corre-
lations of latency with height in males ( r =
0.73-0.81, range across components) and in
females (r = 0.62-0.78). In adults the partial corre-
lations of latency with height (with age effects
removed) ranged from 0.69 to 0.76 and were
PIO0 significant for all components. (3) Adult males
were on average 8.4% taller than females (Fig. 7),
similar to the value of 7.6% calculated from the
data of Dekaban and Sadowski (1978) and similar
to the mean difference of 8.7% in peak latencies
averaged across components. Analysis of covari-
ance showed that most of the sex difference in
latency could be accounted for by this height
difference. (4) However, for all components the
sex difference remained significant even when the
effects of height were removed. That is, latencies
were on average significantly longer in males than
in females of the same height. These height-inde-
pendent sex differences are probably due to non-
isometric differences in body proportions. From
the data of Altman and Dittmer (1962) and Bayer
and Bayley (1959) it can be calculated that arm
length is 1.6%, and shoulder width 4.7%, longer in
males than in females of the same height. Taking
both differences into account it can be calculated
that wrist-C7 length is on average 2.1% longer in
males than in females of the same height. (5)
Because arm and shoulder dimensions were not
obtained, their covariation with height and latency
is unknown in this study. However, the above
considerations suggest that isometric and non-iso-
metric sex differences in median nerve length
plausibly account for the sex differences in latency.
In contrast to peak latencies, there was no
significant height-related change in any SEP inter-
peak latency. All sex differences remained when
the effects of height were removed, and were sig-
nificant for all interpeak latencies except N10-P26.
They probably reflect the fact noted above that,
independent of height, male shoulder width is on
average 4.7% greater than female shoulder width.
Assuming that the distance from the bachial plexus
to the spinal cord is isometric with shoulder width,
the sex differences in interpeak latencies are plau-
sibly explained as due to the relatively longer
length in males of that segment of the median
nerve. The U L N s of Figs. 4 - 6 provide the best
632
available prediction of interpeak latency in the
absence of a direct measure of shoulder width.
While our results do not rule out other possible
factors (e.g., body temperature) in accounting for
the observed sex differences in latencies, they show
that the differences are plausibly explained by
known sex differences in brain and body size.
They also suggest that predictors of somatosensory
pathway length such as arm length, shoulder width
and neck length, and predictors of brain size such
as head circumference or external cranial volume
(e.g., Jorgensen et al. 1961), might provide better
predictors of latency than height and sex.
Interpeak and left-right differences. In assess-
ing brain stem dysfunction the use of interpeak
latencies to remove the effects of latency variation
in the auditory nerve volley is well established, as
is the use of the P4-P6 interpeak latency (e.g.,
Stockard et al. 1978; Chiappa et al. 1979). SEP
interpeak analysis has been limited so far to the
use of N13-N20 as a measure of central conduc-
tion time (e.g., Hume and Cant 1978; Hume et al.
1982). Use of the other SEP interpeak latencies
should also provide more sensitive detection of
central lesions than use of peak latencies, since
latency variation due to variation in length of the
median nerve from wrist to brachial plexus is
eliminated. In addition, the present results show
that much of the age-related increase in AEP and
SEP central peak latencies is secondary to periph-
eral changes in AEP P2 and SEP N10 latencies;
thus the ULNs for most interpeak latencies are
independent of age. LR differences should provide
a sensitive measure of lateralized abnormality since
each patient serves as his own control, and may
reveal abnormality not demonstrated by analysis
of peak or interpeak latency (e.g., Asselman et al.
1975; Selters and Brackmann 1977; Halliday 1978).
Quantitative assessment of abnormality. False-
positive and false-negative error rates in patient
testing are strongly influenced by the criterion for
normality. Most workers use +2.5 or +3 S.D. as
the ULN. We prefer the +3 S.D. criterion on the
assumption that it is a reasonable compromise
between minimizing false-positive and false-nega-
tive errors, especially when all modalities are tested
and thus many measures obtained. However, other
S.D. criteria may be obtained by interpolation or
T. ALLISON ET AL.
extrapolation from the mean and +3 S.D. regres-
sion lines of Figs. 2-6. It was not feasible to
determine a false-negative rate. The average false-
positive rate across all components in Figs. 2-6 is
0.38%, similar to the predicted value of 0.15%.
Although not shown in Figs. 2-6, it is useful to
plot the -3 S.D. line (lower limit of normal) for
each component. When wave forms are abnormal
and identification of components difficult the
question may arise whether a peak is (for example)
a late N12 or an early N13. If the latency falls
below the -3 S.D. line for NI3 it can be surmised
that the peak is a late N12.
Although all components show significant
latency changes related to sex, age or height, the
effect of these changes on the ULN is sometimes
modest. The question arises to what extent all
statistically s i g n i f i c a n t - but perhaps clinically
unimportant - - predictors need to be taken into
account in assessing a patient. There is no single
answer to this question since it depends on the
distribution of latencies in patients with a particu-
lar disorder in relation to the distribution of nor-
mal latencies. Optic neuritis usually produces a
large increase in PI00 latency (e.g., Halliday 1978),
and the additional sensitivity provided by separate
male and female norms will change the false-posi-
tive and false-negative rates only slightly. How-
ever, in non-demyelinating lesions P100 latency
abnormalities are less striking (e.g., Halliday 1978),
and in our experience AEP and SEP latency
abnormalities are also usually subtle regardless of
the type of presumed lesion. In these situations the
use of all significant predictors is correspondingly
more important. In adults, ULNs for males are
often about 1 S.D. above the female values, as are
ULNs in the 60-70 age group compared with the
20-30 age group. Failure to take age and sex into
account will thus have substantial effects on
false-positive and false-negative rates. To take an
extreme case, Figs. 2-6 show that norms based on
a predominantly female sample of subjects in the
20-40 age range will lead to a substantial
false-positive rate when applied to older male pa-
tients (also see Stockard et al. 1978, 1979).
Component identification. The extent to which
components are undetectable or vary in wave form
in normal subjects is important in assessing clini-
NORMATIVE AUDITORY, VISUAL AND SOMATOSENSORY EPs
cal recordings. All AEP components except P12
were detectable in over 98% of subjects. Variations
in AEP wave form have been described (e.g.,
Stockard et al. 1978; Chiappa et al. 1979); in
particular, P5 and P6 are sometimes fused. We
found that both peaks could nearly always be
detected, although P5 was sometimes only an in-
flection on the rising phase of P6 (Fig. 1). How-
ever, P3, P7, P9 and P12 were sometimes small or
difficult to quantify, particularly in older subjects;
absence of these components in the presence of an
otherwise normal wave form should be interpreted
cautiously. With the recording conditions used
here the base-to-peak amplitude of P6 is always
larger than that of P2; the converse finding can be
regarded as abnormal.
VEP P100 was present in all subjects; P60, N75
and N145 peaks were present in most subjects but
have little current clinical utility and were not
analyzed. P100 wave form is variable in young and
old subjects due to several factors including a
broader potential, two or more peaks in the latency
range of P100, and greater interhemispheric dif-
ferences in wave form and latency than in the
20-59 age group. The decreases in visual acuity
and increases in ocular opacity commonly seen in
the elderly probably cannot account for the latency
increase (Halliday 1978), and in our 60-95 age
group there was no correlation between P100
latency and visual acuity (r = 0.04).
The identification of SEP components il-
lustrated in Fig. 1 has been discussed (Allison and
Hume 1981; Allison et al. 1982). Variations in
wave form can lead to an appreciable difference in
measured latency in 3 cases: (1) In about 15% of
cases it was not possible to differentiate N 12a and
b, or N13a and b; if the single peak could not be
assigned on morphologic or topographic grounds
it was considered N12a or N13a. (2) In over 90%
of cases N20 was the largest negative peak in the
15-25 msec latency range and followed the inflec-
tion labeled P18; the recording of Fig. 1 is repre-
sentative. However, occasionally the largest nega-
tive deflection occurred before N20, which then
appeared as a negative-going peak on the rising
phase of P22. This variation in wave form reflects
the likelihood that the parietal negativity in this
latency range consists of an earlier negative poten-
633
tial, probably of subcortical origin (Chiappa et al.
1980; Desmedt and Cheron 1981b; Allison 1982;
Kimura and Yamada 1982), superimposed on the
later N20 deflection. In any case, in this study N20
was measured as the negative peak following P18.
(3) The wave form of P22 and P26 was variable.
Subjects may have a single peak, two clear peaks,
or a major peak with the other peak appearing as
an inflection. All SEP components except N12b,
N13b and P26 were detectable in over 96% of
subjects. However, some components (e.g., NI4
and P15) appear only as minor peaks or inflec-
tions. As a general rule only the absence or latency
abnormality of N10, N12 or N13 (whether or not
differentiable into subpeaks) or N20 provide strong
evidence of abnormality.
Comparison with Committee recommendations.
The methods used in this study are similar to those
being formulated by the Committee on Evoked
Potentials of the American EEG Society (AES;
Chatrian et al. 1982), and by a similar committee
of the International Federation (IF; Halliday et al.
1983), with the following exceptions: (1) AES rec-
ommends AEP testing at several stimulus intensi-
ties, and some workers advocate testing at higher
stimulus frequencies, when results using standard
parameters are equivocal. If these procedures are
used, norms must be established for each condi-
tion since stimulus intensity and frequency affect
latency. (2) AES and IF suggest U L N criteria
ranging from +2 to +3 S.D. For reasons noted
above we favor the +3 S.D. criterion. However, we
agree with AES that 'different laboratories may
legitimately adopt different normal limits because
of different clinical situations.' (3) SEP montages
recommended by AES and IF are slightly different
than those used here, but will likely not result in
different latencies except that Erb's point is 2-3
cm proximal to the mid-clavicular placement used
here, and N10 latency is therefore longer (0.24 +
0.06 msec). If Erb's point is used, the ULNs for
N I0 and associated interpeak latencies should be
changed accordingly. (4) P100 latency measured at
O~ and 02 is not appreciably different at the
medial 3 locations of the montage recommended
by AES and IF. The latter montage is preferable
when VEP topography rather than latency is rele-
vant, for example in the study of visual field
defects.
634
Summa~
To determine standards of normality for audi-
tory, somatosensory and visual evoked potentials
commonly used in the assessment of neurological
disease, 8 AEP, 1 VEP and 12 SEP components
were recorded to stimulation of left and right ears,
eyes, and median nerves in 286 normal subjects
ranging in age from 4 to 95 years. Peak and
interpeak latencies, and left-right differences in
latency, were analyzed as a function of age, sex,
and estimates of brain and body size. Major fea-
tures of the results were: (1) Peak latencies of all
components showed statistically significant in-
creases in latency with age except that VEP P100
latency decreased significantly between 4 and 19
years and did not change between 20 and 59 years.
(2) In adults the peak latencies of all components
were significantly later in males than in females.
For AEPs and VEPs these differences were ex-
plained by sex differences in brain size, and for
adult SEPs were explained by sex differences in
arm and shoulder dimensions. No significant sex
differences in VEP and SEP latencies were seen in
children. (3) Most interpeak latencies showed sig-
nificant differences in relation to age or sex. (4)
Age and sex are useful predictors of latency for
nearly all peak and interpeak latencies; in addi-
tion, height is a useful predictor of SEP peak
latencies. (5) Left-right latency differences showed
little age-related, and no sex-related, change.
The interlaboratory use of these or other
normative data was discussed. It was concluded
that these AEP and SEP norms can probably be
used in other laboratories if stimulating and re-
cording conditions are similar. However, VEP re-
suits are difficult to transfer due to the poorly
understood effects of variation in stimulus condi-
tions. Some issues regarding the optimal char-
acterization of norms were also discussed.
R6sum6
Potentiels dvoqu~s auditifs du tronc, potentiels
~voqu~s visuels au renversement de pattern et
potentiels ~voquOs somatosensoriels ~ courte latence:
relations entre latence, dlge, sexe et dimension du
cerveau et du corps
T. ALLISON ET AL.
Afin de d6terminer les conditions de normalit6
pour les potentiels 6voqu6s auditifs (PEA),
somatosensoriels (PES) et visuels (PEV) commun6-
ment utilisbs dans le diagnostic d'une affection
neurologique, on a enregistr6 8 composantes pour
le PEA, 1 pour le PEV et 12 pour le PES, pour des
stimulations des oreilles droite et gauche, des yeux
et des nerfs m6dians chez 286 sujets normaux ~g6s
de 4 ~t 95 ans. Les latences des pics et les latences
entre pics, ainsi que les diff6rences entre gauche et
droite ont 6t6 analys~s en fonction de l'~ge, du
sexe et d'estimations effectu6es pour la taille du
cerveau et du corps. Les principaux r6sultats ont
6t6: (1) Les latences des pics de toutes les com-
posantes croissaient significativement avec l'glge
sauf pour la P100 du PEV qui a d6cru significati-
vement entre 4 et 19 ans et n'a pas chang6 entre 20
et 59 ans. (2) Chez les adultes les latences des pics
de toutes les composantes 6talent significativement
sup6rieures chez les hommes que chez les femmes.
Pour les PEA et les PEV ces diff6rences s'ex-
pliquaient par des diff6rences de la taille du cerveau
et pour les PES d'adultes elles s'expliquait par des
diff6rence dans la dimension des bras et des 6-
paules entre sexe. Aucune diff6rence significative
n'est apparue entre sexe pour les PEV et PES des
enfants. (3) La plupart des latences entre pics
pr6sentaient des diffbrences significatives en rela-
tion avec l'~ge et le sexe. (4) L'~ge et le sexe
peuvent 6tre utiles pour pr6voir la latence de pres-
que tousles pics et celle des interpics, en outre la
taille permet de pr6voir les latences de pic du PES.
(5) Les diffbrences de latence entre droite et gauche
n'ont montr6 que peu de modifications en relation
avec l'~ge et aucune avec le sexe.
L'utilisation, entre laboratoires, de ces donnbes
ou d'autres donn6es normatives est discut6e. On
conclut que ces normes pour les PEA et les PES
peuvent probablement 6tre utilis6es dans d'autres
laboratoires si les conditions de stimulation et
d'enregistrement sont similaires. Toutefois les
r6sultats pour les PEV sont difficiles ~ transf6rer
car on connait mal les effets de variations des
conditions de stimulation. Quelques conclusions
concernant la caract6risation optimale des normes
sont 6galement envisag6es.
NORMATIVE AUDITORY, VISUAL AND SOMATOSENSORY EPs
We thank R. Bartozzi, J. Jasiorkowski, M. Reisenauer, and
D. Thompson for assistance, A. Hume and G. McCarthy for
collaboration and advice, and G.E. Chatrian and A.M. Halli-
day for helpful criticism of earlier versions of this paper.
References
Allison, T. Scalp and cortical recordings of initial somato-
sensory cortex activity to median nerve stimulation in man.
Ann. N.Y. Acad. Sci., 1982, 388: 671-678.
Allison, T. and Hume, A.L. A comparative analysis of short-
latency somatosensory evoked potentials in man, monkey,
cat, and rat. Exp. Neurol., 1981, 72:592-611.
Allison, T., Goff, W.R. and Wood, C.C. Auditory, somato-
sensory and visual evoked potentials in the diagnosis of
neuropathology: recording considerations and normative
data. In: D. Lehmann and E. Callaway (Eds.), Human
Evoked Potentials: Applications and Problems. Plenum,
New York, 1979: 1-16.
Allison, T., Wood, C.C., McCarthy, G., Hume, A.L. and Goff,
W.R. Short-latency somatosensory evoked potentials in man,
monkey, cat, and rat: comparative latency analysis. In: J.
Courjon, F. Maugui~re and M. Revol (Eds.), Clinical Appli-
cations of Evoked Potentials in Neurology. Raven Press,
New York, 1982: 303-311.
Altman, P.L. and Dittmer, D.S. (Eds.) Growth. Fed. Amer.
Soc. Exp. Biol., Washington, D.C., 1962:608 pp.
Asselman, P., Chadwick, D.W. and Marsden, C.D. Visual
evoked responses in the diagnosis and management of
patients suspected of multiple sclerosis. Brain, 1975, 98:
261-282.
Bayer, L.M. and Bayley, N. Growth Diagnosis. University of
Chicago Press, Chicago, I11., 1959.
Celesia, G.G. Clinical applications of evoked potentials. In: E.
Niedermeyer and F. Lopes da Silva (Eds.), Electroen-
cephalography: Basic Principles, Clinical Applications and
Related Fields. Urban and Schwarzenberg, Baltimore, Md.,
1982: 665-684.
Celesia, G.G. and Daly, R.F. Effects of aging on visual evoked
responses. Arch. Neurol. (Chic.), 1977, 34: 403-407.
Chatrian, G.E., et al. Report of the Committee on Evoked
Potentials, American ElectroencephalographicSociety, 1982.
Chiappa, K.H., Gladstone, K.J. and Young, R.R. Brainstem
auditory evoked responses: studies of variations in 50 nor-
mal human subjects. Arch. Neurol. (Chic.), 1979, 36: 81-87.
Chiappa, K.H., Choi, S. and Young, R.R. Short latency
somatosensory evoked potentials following median nerve
stimulation in patients with neurological lesions. In: J.E.
Desmedt (Ed.), Clinical Uses of Cerebral, Brainstem and
Spinal Somatosensory Evoked Potentials. Prog. Clin. Neu-
rophysiol., Vol. 7. Karger, Basel, 1980: 264-281.
Courjon, J., Maugui~re, F. and Revol, M. (Eds.) Clinical Appli-
cations of Evoked Potentials in Neurology. Raven Press,
New York, 1982:577 pp.
Cracco, R.Q. and Cracco, J.B. Somatosensory evoked poten-
635
tials in man: far-field potentials. Electroenceph. clin. Neu-
rophysiol., 1976, 41: 460-466.
Dekaban, A.S. and Sadowsky, D. Changes in brain weights
during the span of human life: relation of brain weights to
body heights and body weights. Ann. Neurol., 1978, 4:
345-356.
Desmedt, J.E. and Cheron, G. Prevertebral (oesophageal) re-
cording of subcortical somatosensory evoked potentials in
man: the spinal P13 component and the dual nature of the
spinal generators. Electroenceph. clin. Neurophysiol., 1981a,
52: 257-275.
Desmedt, J.E. and Cheron, G. Non-cephalic reference re-
cording of early somatosensory potentials to finger stimula-
tion in adult or aging normal man: differentiation of
widespread N 18 and contralateral N20 from the prerolandic
P22 and N30 components. Electroenceph. clin. Neurophys-
iol., 1981b, 52: 553-570.
Dixon, W.J. (Ed.) BMDP Statistical Software. Univ. Calif.
Press, Berkeley, Calif., 1981:726 pp.
Donchin, E., Callaway, E., Cooper, R., Desmedt, J.E., Golf,
W.R., Hillyard, S.A. and Sutton, S. Publication criteria for
studies of evoked potentials in man. In: J.E. Desmedt (Ed.),
Progress in Clinical Neurophysiology, Vol. 1. Karger, Basel,
1977: 1-11.
HaUiday, A.M. Clinical applications of evoked potentials. In:
W.B. Matthews and G.H. Glaser (Eds.), Recent Advances
in Clinical Neurology. Churchill-Livingstone, Edinburgh,
1978: 47-73.
Halliday, A.M. et al. Standards of Clinical Practice for the
Recording of Evoked Potentials (EP). Appendix 1F of Rec-
ommendations for the Practice of Clinical Neurophysiol-
ogy, International Federation of Societies for Electroen-
cephalography and Clinical Neurophysiology. Elsevier,
Amsterdam, 1983: in press.
Hume, A.L. and Cant, B.R. Conduction time in central
somatosensory pathways in man. Electroenceph. clin. Neu-
rophysiol., 1978, 45: 361-375.
Hume, A.L., Cant, B.R., Shaw, N.A. and Cowan, J.C. Central
somatosensory conduction time from 10 to 79 years. Elec-
troenceph, clin. Neurophysiol., 1982, 54: 49-54.
Jewett, D.L. and Williston, J.S. Auditory-evoked far fields
averaged from the scalp of humans. Brain, 1971, 94:
681-696.
Jones, S.J. Short latency potentials recorded from the neck and
scalp following median nerve stimulation in man. Electro-
enceph, clin. Neurophysiol., 1977, 43: 835-863.
Jorgensen, J.B., Paridon, E. and Quaade, F. The correlation
between external cranial volume and brain volume. Amer. J.
phys. Anthropol., 1961, 19: 317-320.
Kimura, J. and Yamada, T. Short-latency somatosensory evoked
potentials following median nerve stimulation. Ann. N.Y.
Acad. Sci., 1982, 388: 689-694.
Kjaer, M. The value of brain stem auditory, visual and
somatosensory evoked potentials and blink reflexes in the
diagnosis of multiple sclerosis. Acta neurol, scand., 1980,
62: 220-236.
Kritchevsky, M. and Wiederholt, W.C. Short-latency somato-
636
sensory evoked potentials. Arch. Neurol. (Chic.), 1978, 35:
706-711.
Lieber, A. and Sohmer, H. Standard values of amplitude and
latency of cochlear audiometry (electro-cochleography) re-
sponses in different age groups. Arch. klin. exp. Ohr.-,
Nas.-, u. Kehlk.-Heilk., 1973, 203: 267-273.
Matthews, W.B., Beauchamp, M. and Small, D.G. Cervical
somato-sensory evoked responses in man. Nature (Lond.),
1974, 252: 230-232.
Schmidt-Nielsen, K. Scaling in biology: the consequences of
size. J. exp. Zool., 1975, 194: 287-307.
Selters, W.A. and Brackmann, D.E. Acoustic tumor detection
with brain stem electric response audiometry. Arch.
Otolaryngol., 1977, 103: 181-187.
Shaw, N.A. and Cant, B.R. Age-dependent changes in the
latency of the pattern visual evoked potential. Electroen-
ceph. clin. Neurophysiol., 1980, 48: 237-241.
Shearer, D.E. and Dustman, R.E. The pattern reversal evoked
potential: the need for laboratory norms. Amer. J. EEG
Technol., 1980, 20: 185-200.
Small, D.G., Beauchamp, M. and Matthews, W.B. Subcortical
somatosensory evoked potentials in normal man and in
patients with central nervous system lesions. In: J.E. De-
smedt (Ed.), Clinical Uses of Cerebral, Brainstem and Spi-
T. ALLISON ET AL.
nal Somatosensory Evoked Potentials, Prog. clin. Neuro-
physiol., Vol. 7. Karger, Basel, 1980: 190-204.
Sokol, S., Moskowitz, A. and Towle, V.L. Age-related changes
in the latency of the visual evoked potential: influence of
check size. Electroenceph. clin. Neurophysiol., 1981, 51:
559-562.
Starr, A. and Achor, L.J. Auditory brain stem responses in
neurological disease. Arch. Neurol. (Chic.), 1975, 32:
761-768.
Starr, A., Sohmer, H. and Celesia, G.G. Some applications of
evoked potentials to patients with neurological and sensory
impairment. In: E. Callaway, P. Tueting and S.H. Koslow
(Eds.), Event-Related Brain Potentials in Man. Academic
Press, New York, 1978: 155-221.
Stockard, J.J., Stockard, J.E. and Sharbrough, F.W. Nonpatho-
logic factors influencing brainstem auditory evoked poten-
tials. Amer. J. EEG Technol., 1978, 18: 177-209.
Stockard, J.J., Hughes, J.F. and Sharbrough, F.W. Visually
evoked potentials to electronic pattern reversal: latency
variations with gender, age, and technical factors. Amer. J.
EEG Technol., 1979, 19: 171-204.
Stohr, M., Dichgans, J., Diener, H.C. und Buettner, U.W.
Evozierte Potentiale. Springer, Berlin, 1982.