Neurobiology of Aging 26 (2005) 1297–1299
Age-sensitivity of P3 in high-functioning adults?
Anders M. Fjell∗, Kristine B. Walhovd
Institute of Psychology, University of Oslo, P.B. 1094 Blindern, 0317 Oslo, Norway
Received 8 February 2005; accepted 10 February 2005
In their interesting paper, Daffner et al. [Daffner KR, Ryan KK, Williams DM, Budson AE, Rentz DM, Scinto LFM, et al. Age-related
differences in novelty and target processing among cognitively healthy high performing adults. Neurobiol Aging, 2005;26:1283–95] argue
that previous studies have found changes in ERP components in response to novel and target stimuli due to two methodological factors: (1)
lack of control for differences in level of cognitive status between age groups, and (2) not controlling for a non-specific age-related processing
difference for all stimulus types (standards, targets, and novel). The questions raised by Daffner et al. are interesting, but based on existing
literature, their conclusion seems premature. In the following, we will present examples from empirical literature as well as re-analyses of
some of our own work to illustrate problematic aspects of Daffner et al.’s position.
© 2005 Published by Elsevier Inc.
Keywords: ERP; P3a; P3b; Aging; Cognition; Neuropsychology
In our paper “Life-span changes in P3a” published in
Psychophysiology , we reported correlations between
visual P3 to distractors and age of .52 (latency) and −.53
(amplitude), both at Cz. The sample was large (n=103) and
the population mean (IQ=115, 84th percentile), and IQ did
not correlate with age. Thus, the results were obtained in a
did not differ in terms of normative cognitive function across
age groups. The age-effects can, therefore, not be explained
by differences in normative level of cognitive status.
The same conclusion was reached in a previous study by
administered to a large adult life-span sample with limited
overlap with the above described study (n=71, age 22–95).
The mean IQ in this sample was also 115, and did not cor-
relate with age. Correlations between P3 to target and age
?Commentary to Daffner et al. Age-related differences in novelty and tar-
∗Corresponding author. Tel.: +47 84 51 29; fax: +47 22 84 51 29.
E-mail address: firstname.lastname@example.org (A.M. Fjell).
were demonstrated (.52 and −.52, for latency and amplitude,
respectively, both at Pz).
Yet another example: In a study of auditory novelty-P3,
Friedman et al.  demonstrated age-decreases in P3 ampli-
tude to distractors. In this study, the screening procedures
were excellent: older participants were determined to be free
of depression, dementia, and limitations in the activities of
daily living as assessed by the Short CARE, normal on a
complete medical and neurological examination, and a neu-
ropsychological test battery was administered. Participants
in the different age groups had similar scores on different
reported measures, including verbal and performance IQ and
112.8 to 118.9 (verbal) and 103.5 to 110.1 (performance).
acknowledge, a large number of ERP-studies focusing on
the P3-complex demonstrate changes in latency, amplitude,
and topography with increasing age (starting with Goodin et
al. [7,11–13]. For instance, a meta-analysis of 32 studies of
age-effects of P3 latency  was published nearly 10 years
P3 latency exists. The lowest P3-age correlation reported by
Polich  was .33. Even though the correlations for visual
0197-4580/$ – see front matter © 2005 Published by Elsevier Inc.
A.M. Fjell, K.B. Walhovd / Neurobiology of Aging 26 (2005) 1297–1299
stimuli were lower than for auditory or somatosensory stim-
uli, the median correlation was .54. However, we disagree
with Daffner et al.’s view on the reason for these findings. In
our opinion, there is no reason to assume that the observed
age-effects on the P3-complex in such a large number of
studies are solely due to normative differences in cognitive
function between age groups. Actually, a larger number of
ences in actual cognitive function. It is true that the criteria
for inclusion in different studies of age-effects on the P3-
complex have varied, but even so, some studies arguably
meets the standard set by Daffner et al. At least, this is the
case for the three studies reviewed above, where the mean
normative cognitive function was in the superior range and
not related to age. The main point, however, is that most of
the studies of age-effects on P3 have come to approximately
similar results. Seven studies using visual tasks are included
in the meta-analysis by Polich . Adding our own , we
have eight studies. The correlation between P3 and age in
these eight studies varies from .33 to .63, with samples from
24 to 172 participants. Even with different screening proce-
dures, correlations within the same range are demonstrated,
dures, differences in paradigms used, and strategies for data
Daffner et al. have gone to great length to ensure superior
re-analyze the data from [2,3]. Here, it is vital to distinguish
between normative function and actual function. While e.g.
performance abilities are highly sensitive to the effect of age,
groups of different age may have the same normative level of
is taken to mean normative ability. In this study, a 3-stimuli
visual oddball task was used, comparable to the novelty task
used by Daffner et al. (for studies of the relationship between
novelty-P3 and P3 to distractor, see, e.g., Simons et al. ).
Some additional data have been gathered since these reports
[2,3] were published, so the sample pool consists of 129 par-
IQ≥85, no reported psychiatric or neurological diseases or
injuries, including Parkinson’s disease, multiple scleroses,
sample of 69 participants rather evenly distributed across the
adult life-span (mean age=54, S.D.=21, range 21–89), with
a mean IQ of 122, about the same level as the groups’ esti-
mated IQ in Daffner et al’s study. In our sample of superiorly
functioning participants, there was a non-significant positive
tractor at Cz correlated .53 (latency) and −.47 (amplitude)
with age (p<.0001). T-tests of Fisher z-transformed cor-
relations yielded non-significant differences between these
two correlations and the correlations reported in . Thus,
restricting the analyses to participants with superior general
mental abilities did not change the relationship between P3
Next, to ensure that all participants had high neuropsy-
chological function in addition to superior general abilities,
we used a criterion inspired by Daffner et al. To be further
included in our super sample, all participants must score at
digit symbol substitution (WAIS-R; ), Controlled Word
Association Test (COWAT; [9,16]), Trail Making Test A ,
and Stroop color/word . Because many of the partici-
tests, we were able to pick six tests with some resemblance
to the six used by Dafnner et al. Age was non-correlated
now consisted of 45 participants. In this neuropsychologi-
cally superior sample, P3 distractor latency and amplitude
correlated .62 (p<.0001) and −.34 (p<.02), respectively.
These coefficients are not significantly different, by means
of t-tests of the Fisher z-transformed correlations, from the
Thus, re-analyses of one of our own P3-data sets showed
that (1) age-effects on P3 amplitude and latency were indeed
included, (2) these effects did not change significantly from
cognitive function is non-related to age. Such a broad set of
neuropsychological data was not collected for the sample
used in Walhovd and Fjell , so these analyses cannot be
repeated with that sample. However, the sampling strategy,
screening-procedures, and IQ-level are almost identical in
these two studies, so there is no reason to assume the results
would be different in that sample. Since these results fit well
often-found age-effects on P3 amplitude and latency can be
explained by lack of careful control of differences in level of
age-relationship for the P3 to target in Daffner et al. may
be caused by the use of a 3-stimuli instead of a 2-stimuli
paradigm. This mimic what we have discovered ourselves in
the latency and amplitude of peaks in the P3-complex (P3a,
P3b) are sensitive to neurophysiological changes that come
with increasing age, even in very high-functioning adults.
latency and reduced amplitude of the P3-complex.
specific age-related processing difference, this phenomenon
seems to occur as an exception rather than a rule. In ,
no processing positivity in the P3 window is observed for
the standard stimuli for any of the age groups (see Fig. 1).
In another publication from the same study , age-effects
A.M. Fjell, K.B. Walhovd / Neurobiology of Aging 26 (2005) 1297–1299
Fig. 1. Grand average ERPs for a young and an old group of participants. The task was an auditory oddball paradigm (the study is presented in detail in ).
The two age groups had the same IQ (young: 114, old: 115). The activity to the target in the old group shows less amplitude and a slower peak. The standard
is not related to any positive activity in the traditional P3-window.
on P3 amplitude is clearly seen in the difference waves
(target–standard activation). In the Friedman et al. study,
standards did not yield positive activation in the P3 window
groups cannot explain the observed effects in these studies.
In sum, while acknowledging methodological strengths
in Daffner et al.’s study, we believe their interpretation of the
claim that previous studies have lacked control of cognitive
status between participants of different age is not justified.
more data on groups with more average cognitive function,
we look forward to see these results in the future. A fruitful
between high and lower functioning participants at different
focusing on whether the changes in ERPs with age are best
seen as “compensation” or “inefficiency.” A mean through
which to obtain such a goal will be to include independent
measures of individual differences.
 Daffner KR, Ryan KK, Williams DM, Budson AE, Rentz DM, Scinto
LFM, et al. Age-related differences in novelty and target processing
among cognitively healthy high performing adults. Neurobiol Aging
 Fjell AM, Walhovd KB. Life-span changes in P3a. Psychophysiology
 Fjell AM, Walhovd KB. On the topography of P3a and P3b—A
factor-analytic study with orthogonal Procrustes rotation to maximize
congruence. Brain Topogr 2003;15:153–64.
 Fjell AM, Walhovd KB. P300 and neuropsychological tests as mea-
sures of aging: scalp topography and cognitive changes. Brain Topogr
 Friedman D. Cognition and aging: a highly selective overview
of event-related potential (ERP) data. J Clin Exp Neuropsychol
 Friedman D, Kazmerski VA, Cyciwicz YM. Effects of aging on the
novelty P3 during attend and ignore oddball tasks. Psychophysiology
 Goodin DS, Squires KC, Henderson B, Starr A. Age-related variation
in evoked potentials to auditory stimuli in normal human subjects.
Electroenceph Clin Neurophysiol 1978;44:447–58.
 Heaton RK, Grant I, Matthews CB. Comprehensive norms for an
expanded Halstead-Reitan battery: demographic corrections, research
findings and clinical applications. Odessa, FL: Psychological Assess-
ment Resources; 1991.
 Lezak MD. Neuropsychological assessment, 3rd ed. Oxford: Oxford
University Press; 1995.
 MacLeod C. Half a century of research on the Stroop effect: an
integrative review. Psychol Rev 1991;109:163–203.
 Pfefferbaum A, Ford JM, Roth WT, Kopell BS. Age-related changes
in auditory event-related potentials. Electroencephalogr Clin Neuro-
 Pfefferbaum A, Ford JM, Wenegrat BG, Roth WT, Kopell BS.
Clinical application of the P3 component of event-related poten-
tials. I. Normal aging. Electroencephalogr Clin Neurophysiol
 Picton TW, Stuss DT, Champagne SC, Nelson RF. The effects
of age on human event-related potentials. Psychophysiology
 Polich J. Meta-analysis of P300 normative aging studies. Psy-
 Simons RF, Graham FK, Miles MA, Chen X. On the relationship of
P3a and the Novelty-P3. Biol Psychol 2001;56:207–18.
 Spreen O, Strauss E. A compendium of neuropsychological tests:
Administration, norms, and commentary. New York: Oxford Univer-
sity Press; 1991.
 Trahan DE, Larrabee GJ. Continuous visual memory test. Odessa,
FL: Psychological Assessment Resources; 1988.
 Walhovd KB, Fjell AM. The relationship between P3 and neu-
ropsychological function in an adult life-span sample. Biol Psychol
 Wechsler D. Wechsler adult intelligence scale—revised. San Antonio,
TX: The Psychological Corporation; 1981.