MENTAL RETARDATION AND DEVELOPMENTAL DISABILITIES
RESEARCH REVIEWS 11: 317 324 (2005)
NEUROPHYSIOLOGIC EVALUATION OF EARLY
COGNITIVE DEVELOPMENT IN HIGH-RISK
INFANTS AND TODDLERS
Raye-Ann deRegnier*
Department of Pediatrics, Northwestern University Feinberg School of Medicine, Prentice Women s Hospital, Chicago, Illinois
New knowledge of the perceptual, discriminative, and memory ca-
young infants are rudimentary, there is evidence of memory and
pabilities of very young infants has opened the door to further evaluation of
early language processing in healthy infants from the time of
these abilities in infants who have risk factors for cognitive impairments. A
birth [DeCasper and Fifer, 1980; DeCasper and Spence, 1986;
neurophysiologic technique that has been very useful in this regard is the
Rose et al., 1988; Fagan, 1990; Geva et al., 1999; deRegnier et
recording of event-related potentials (ERPs). The event-related potential
(ERP) technique is widely used by cognitive neuroscientists to study cogni- al., 2000]. Application of these techniques to the study of
tive abilities such as discrimination, attention, and memory. This method
high-risk infants has revealed infant origins of differences in
has many attractive attributes for use in infants and children as it is relatively
cognitive processing that were formerly thought to surface in
inexpensive, does not require sedation, has excellent temporal resolution,
childhood. This article will review the use of one technique,
and can be used to evaluate early cognitive development in preverbal in-
fants with limited behavioral repertories. In healthy infants and children, event-related potentials (ERPs), in the study of early cognitive
ERPs have been used to gain a further understanding of early cognitive
development in high-risk infants and toddlers.
development and the effect of experience on brain function. Recently, ERPs
have been used to elucidate atypical memory development in infants of
TECHNICAL ASPECTS OF ERPS
diabetic mothers, difficulties with perception and discrimination of speech
ERPs have been used for many years in adults to evaluate
sounds in infants at risk for dyslexia, and multiple areas of cognitive differ-
ences in extremely premature infants. Atypical findings seen in high-risk cognitive processes such as attention, memory, and language
infants have correlated with later cognitive outcomes, but the sensitivity
[Coles and Rugg, 1995; Naatanen, 1995]. The technique nat-
and specificity of the technique has not been studied, and thus evaluation
urally lends itself to the study of developing cognitive processes
of individual infants is not possible at this time. With further research, this
in infants as it is noninvasive, relatively inexpensive, and can be
technique may be very useful in identifying children with cognitive deficits
during infancy. Because even young infants can be examined with ERPs, this performed in unsedated, preverbal infants with limited behav-
technique is likely to be helpful in the development of focused early inter-
ioral responses. Although it is not yet possible to determine the
vention programs used to improve cognitive function in high-risk infants
precise neural structures that generate each of the ERP compo-
and toddlers. © 2005 Wiley-Liss, Inc.
nents, the temporal resolution of ERPs is superior to all other
MRDD Research Reviews 2005;11:317 324.
currently available neuroimaging techniques [Nelson and
Monk, 2001; DeBoer et al., 2005a]. Functional information
Key Words: neurophysiologic assessment; cognition; infants; high risk
gained through use of the ERP can be complementary to
anatomical information gained through current neuroimaging
techniques such as magnetic resonance imaging (MRI). Because
of these characteristics, ERPs have been used in a growing
any infants and children are at increased risk for cog-
number of studies of high-risk and healthy infants and toddlers.
nitive impairments due to genetic or hereditary con-
In an ERP, brain activity is recorded from the scalp while
Mditions, perinatal illnesses, brain injuries, and brain
the participant is repeatedly presented with stimuli that vary
malformations. Typically, impaired cognitive function is diag-
along perceptual or cognitive dimensions, such as memory.
nosed at preschool or school age, long after the presumed
Brain activity recorded in response to the stimuli is thought to
pathogenic process has occurred. Earlier diagnoses of specific
arise from postsynaptic potentials of pools of cortical neurons
cognitive abnormalities could enable the development of fo-
that are physically oriented to yield dipolar fields that are volume
cused intervention programs to improve outcomes. This future
conducted to the surface of the scalp. [Nelson, 1994; Coles and
goal requires a solid understanding of early cognitive develop-
ment in infancy and early childhood.
Early theories of infant development focused primarily on
*Correspondence to: Raye-Ann deRegnier, MD, Department of Pediatrics,
sensorimotor development [Ginsberg and Opper, 1979], and
Northwestern University Feinberg School of Medicine, Prentice Women s Hos-
infants were not considered to show evidence of true cognitive
pital, Room 404B, 333 East Superior Street, Chicago, IL 60611. E-mail:
function until the emergence of object permanence and lan- [e-mail]
Received 19 August 2005; Accepted 29 August 2005
guage abilities during the toddler years. In more recent times,
Published online in Wiley InterScience (www.interscience.wiley.com).
sophisticated neuroimaging, behavioral, and neurophysiologic
DOI: 10.1002/mrdd.20085
studies have revealed that, although cognitive processes in very
© 2005 Wiley-Liss, Inc.
effects that will be discussed below. Un-
fortunately, there are still relatively few
true developmental ERP studies in
which development of attention, mem-
ory, or language discrimination is studied
at several different ages and the range of
normal has not been well defined for
scalp distribution, amplitudes, and laten-
cies of specific ERP components. These
studies are needed to understand normal
developmental processes and individual
differences. Understanding individual
differences in normally developing chil-
dren is vitally important in interpreting
differences that are seen in studies of low-
and high-risk groups of infants.
Due to the ease of recording and
low rate of artifacts, many ERP studies
have evaluated very young infants during
Fig. 1. Individual newborn infant s ERP demonstrating positive peak (circled) and negative slow
wave. Arrows indicate peak amplitude and latency measurements on respective axes. sleep. However, differences in state, at-
tention, or arousal are known to affect
cognition in infants and must be con-
trolled in ERP studies as much as possi-
Rugg, 1995]. The raw recordings consist scalp s surface. Significant brain activity ble. For very young infants, sleep state
of the total brain activity including activ- may occur in response to the testing but has been shown to have both qualitative
ity that is directly related to the cognitive may not be recorded at the surface of the and quantitative effects upon ERPs. For
process under investigation plus simulta- scalp. This means that failure to find ev- example, auditory cortical potentials re-
neous unrelated brain activity. Brain ac- idence of cognitive differences between corded during quiet sleep in neonates
tivity that is not related to the cognitive stimuli in an ERP study is not decisive tend to be of larger amplitude than po-
process being studied is assumed to be evidence of lack of function. tentials recorded during active sleep or
random and, when the electroencepha- General techniques and methods wakefulness (states that are similar to each
logram is averaged over many trials, this for using ERPs in adults and children other) [Novak et al., 1989; Duclaux et
random activity will average to zero and have been reviewed extensively by sev- al., 1991]. Duclaux et al. [1991] also ob-
the activity related to the cognitive pro- eral authors [Regan, 1989; Coles and served that components recorded during
cess being studied will emerge as a series Rugg, 1995; Nelson and Monk, 2001; active sleep appeared to have shorter la-
of components or slow wave activity DeBoer et al., 2005a, 2005b] and readers tencies than those recorded during quiet
[Regan, 1989], Figure 1. Components are referred to these sources for addi- sleep. Hemispheric asymmetry appears to
are typically referred to as peaks that are tional background information. There be affected by sleep state in adults and
measured in terms of the peak amplitudes are however, several specific consider- may also be affected in infants as well,
(the maximal voltage measurement ations relevant to the study of infants. with more hemispheric asymmetry noted
above the baseline) and latencies (the It should be noted that infant ERPs during quiet sleep [Duclaux et al., 1991].
time from the onset of the stimulus to the tend to show more variable responses On the other hand, studies of auditory
point of maximal deflection). Slow waves than older children and adults [Kushne- discrimination in the newborn infant
are slow drifts in the voltage recordings renko et al., 2002a, 2002b; Leppanen et have shown that infants are able to detect
of the ERP that are seen with sustained al., 2004]. The reasons for this variability changes in speech sounds in both quiet
mental processing. Typically, there is no are not completely understood but likely and active sleep [Hirasawa et al., 2002;
discrete peak to measure and slow waves are due to a combination of individual Martynova et al., 2003]. An additional
are assessed by integrating the area under differences in brain maturation, state, or issue surrounding sleep state in young
the voltage time curve or by computing attentional differences and to the fact that infants is that behavioral and electroen-
the average amplitude across specified it is not possible to directly instruct in- cephalographic assessments of sleep state
time units [Nelson, 1994]. fants on the task to be performed in the may be discordant. Although behavioral
The most common use of ERPs in study. assessments are considered to be the gold
infants and toddlers is to evaluate the Individual differences in matura- standard for evaluation of sleep [Tho-
development of specific cognitive func- tion are to be expected in infants and man, 1990], there have been no studies
tions, such as auditory discrimination young children due to ongoing develop- that attempt to describe differences in
[Cheour et al., 1998, 2002] or memory ment of the brain. General developmen- ERPs using different methods of assess-
[Nelson and Collins, 1991; de Haan and tal changes seen in ERPs are the emer- ment of sleep states. It should be noted
Nelson, 1997; deRegnier et al., 2000]. gence of components, increases in that exhaustive studies of the effects of
This does not require localization of the amplitudes, and shortening of latencies sleep state have not been performed and
neural generators of the ERP activity. [Weitzman and Graziani, 1968; Novak et most studies have controlled for state or
Furthermore, it should be recognized al., 1989; Kushnerenko et al., 2002a]. excluded infants based upon their sleep
that neural activity recorded at the scalp These changes correlate with synapto- states.
surface represents only that portion of the genesis and myelination of the brain In behavioral studies, differences in
total brain activity that has a dipolar ori- [Eggermont, 1988], and a number of arousal or attention during wakefulness
entation and can be recorded at the ERP studies have shown developmental may also affect the assessed cognitive

318 MRDD RESEARCH REVIEWS COGNITIVE EVALUATION USING ERPS DEREGNIER
abilities. Geva et al. [1999] have demon- latency auditory responses from ex- guage and later speech. It would be ex-
strated that differences in arousal level in tremely premature infants, these wave- pected that infants who have difficulties
awake infants after feedings predict vari- forms undergo a prolonged maturational with this task may show difficulties with
ability of behavioral responses in tasks of period and the adult waveforms are not language and reading and indeed these
infant visual memory in the newborn and fully realized until 20 years of age [Pon- early auditory perceptual abilities have
1-month-old infant, whereas, in older ton et al., 2000; Kushnerenko et al., been exploited in ERP studies of high-
infants, responses are less variable and do 2002a]. During the first year of life, dra- risk infants. Some of these studies will be
not differ with feeding-related arousal matic changes are seen in the morphol- described below.
states. Richards and Casey [1991] have ogy of the responses, coinciding with The use of ERPs also has been
extensively studied visual attention in in- rapid development of the cerebral cortex beneficial in evaluating the development
fants and toddlers have described three at this age. In the extremely preterm in- of the word learning. Several investiga-
phases of attention that correlate with fant, the predominant ERP response to tors have evaluated toddler s brain re-
heart rate and predict distractibility (ori- an auditory stimulus is a simple negative sponses to known and unknown words
enting, sustained attention, and inatten- wave with a very long peak latency [Molfese, 1989, 1990; Mills et al., 1997,
tiveness). Richards has recently demon- [Weitzman and Graziani, 1968]. By term 2004]. In these studies, parents provided
strated that ERPs obtained during these gestation, a positive peak is seen over the information regarding words that were
three phases of attention differ, with the midline electrode sites in the majority of thought to be understood by the toddler
largest attentional component [negative infants [Novak et al., 1989; deRegnier et and these words were presented in ERP
component (NC)] described during the al., 2002] but, by the end of the first year, tasks along with words that were not
sustained attentional phase [Richards, a more stable series of positive and neg- known. In all studies, known words elic-
2003]. It is therefore apparent that ERPs ative peaks emerges [Kushnerenko et al., ited different ERP patterns from un-
are affected by sleep state in sleeping in- 2002a]. With further development over known words, with differences being
fants and attentional and arousal states in the first year of life, the amplitudes of noted in the amplitude and scalp distri-
awake infants. these components increase in amplitude bution of the ERP activity. With con-
Thus, even in low-risk infants and their latencies decrease. These tinued development, Mills et al. [1997,
studied in a single laboratory setting, in- changes are thought to be indicative of 2004] described a change in brain activity
dividual differences in ERP findings can synaptogenesis, increased synaptic effi- from being generalized over both hemi-
be due to both individual differences in ciency and myelination of the brain spheres (at 13 17 months) to becoming a
brain maturation as well as state differ- [Eggermont, 1988]. more localized pattern over the temporal
ences between infants that may be diffi- and parietal areas of the left hemisphere
cult to control. When subtly different Auditory Perception (at 20 months). The more localized pat-
paradigms are used by different investi- Although immature in morphol- tern was associated with a larger vocab-
gators, it can be very difficult to directly ogy compared with adults and older chil- ulary rather than age per se.
compare data between studies and derive dren, auditory cortical responses from in-
normative data for the finer aspects of the fants and young children nonetheless Auditory Discrimination
recordings such as peak latencies. All of reflect the acoustic properties of the Beyond auditory perception and
this means that ERP differences between evoking stimuli [Molfese and Molfese, word learning, auditory discrimination is
groups of high- and low-risk infants may 1997]. Studies evaluating differences in another important ability that can be
be difficult to interpret unless state is ERPs evoked by speech sounds with studied in young children using ERPs.
routinely assessed and the study is driven phonetical differences provide an impor- Here, the perception of individual
by a hypothesis that is based upon known tant method of evaluating the develop- sounds or words themselves is not the
neurobiologic differences between the ment of speech perception. For example, focus of the study, but rather the brain s
studied groups. It is also very important Molfese et al. have evaluated the devel- responses to a change in a stream of
to study adequate numbers of participants opment of cortical responses to speech sounds. In a typical auditory change par-
so that individual differences that are seen and/or nonspeech stimuli in several stud- adigm, a frequent sound is occasionally
within groups of high- or low-risk in- ies [Molfese, 1973; Molfese and Molfese, interrupted by a novel, deviant sound. In
fants will not skew the group findings, 1997]. In one of the earliest infant ERP the theory put forth by Naatanen [1995],
potentially masking or falsely amplifying studies, Molfese evaluated children be- when a sound is repeated frequently, a
differences between groups. tween 1 week and 11 years of age com- short-lived, sensory memory trace is es-
pared with young adults, using consonant tablished. The neural representation of
TYPES OF COGNITIVE syllables, monosyllabic words, a piano the novel sound is compared with the
FUNCTIONS STUDIED USING chord, and white noise. At all ages, left memory trace, and a mismatch response
ERPS hemisphere amplitudes were larger than is generated. In adults, this mismatch re-
Although the preceding discussion right hemisphere amplitudes for the sponse is a negative wave (called a mis-
has focused on methodological caveats speech sounds, with a reversed pattern match negativity) that is recorded over
and sources of variability, ERPs have for the nonspeech sounds [Molfese, the frontal and central areas of the scalp
been very useful in the evaluation of in- 1973]. Further study of auditory percep- [Naatanen, 1995]. Studies of infants and
fants auditory perception, auditory dis- tion has also shown differences in ERPs children have shown evidence of similar
crimination, and auditory and visual that are elicited by individual speech waveform morphologies [Alho et al.,
memory. In the adult, auditory cortical sounds, such as /ba/, /da/, and /ga/ in 1990; Cheour-Luhtanen et al., 1996;
responses (indexing auditory perception) newborn infants [Molfese and Molfese, Cheour et al., 1998; Hirasawa et al.,
show a well-ordered pattern of positive 1997; Guttorm et al., 2001]. Differential 2002], even from preterm infants; ini-
and negative components that are labeled responding to acoustic differences in tially it was suspected that this type of
as P1, N1, P2, and N2 [Regan, 1989]. speech sounds is clearly important for the discriminative response might be the on-
Although it is possible to record long further development of receptive lan- togenetically earliest discriminative re-

MRDD RESEARCH REVIEWS COGNITIVE EVALUATION USING ERPS DEREGNIER 319
sponse of the human brain [Cheour-Lu- slow wave (NSW), and the positive slow fants, to further understand cognitive
htanen et al., 1996]. However, with wave (PSW). The NC is a component processes in high-risk children. Some
wider application of the auditory change that is thought to index obligatory or studies have performed statistical correla-
paradigm, it has become apparent that automatic attention to a stimulus [Nel- tions between early ERP activity and
the mismatch response in very young son, 1994; Nelson and Monk, 2001; later development, but no studies have
infants ( 3 months of age) is more vari- Richards, 2003]. The NSW is a wave evaluated the reliability, sensitivity, or
able than previously recognized, with that emerges when a familiar stimulus is specificity of specific ERP patterns, and,
some authors finding no discriminative paired with an entirely novel stimulus. at this time, the usefulness of ERPs in
responses and other authors finding pos- The wave is thought to represent the prediction of developmental disorders in
itive rather than negative responses [Lep- detection of novelty, meaning that the individual children is untested. Large
panen et al., 1997; Therien et al., 2004]. infant has detected that a stimulus is new longitudinal studies of healthy and high-
Some variability in research findings can or different, but the stimulus itself has not risk infants and children are needed be-
be attributed to differences in the meth- been encoded. As encoding abilities im- fore this technique can be regarded as
odology of the studies, but Leppanen et prove and infants are able to begin to clinically useful. Despite this caveat, ERP
al. [2004] have demonstrated that some encode a novel stimulus during the studies of high-risk children studied to
of this variability is due to different levels course of the ERP study, a PSW is de- date have offered new information on
of maturity, particularly in newborn in- tected in response to a repeated novel the neural underpinnings of later cogni-
fants. With this greater appreciation of stimulus paired with a familiar stimulus. tive impairments. The most intensively
the variability in auditory change re- It is presumed that repeated presentations studied groups include infants of diabetic
sponses, it is no longer clear whether of a novel stimulus result in encoding of mothers, children with familial risk for
findings generated in infants are analo- the novel stimulus so that during the dyslexia, and premature infants, and these
gous to those found in adults. course of the study, the novel stimulus groups will be discussed in turn.
However, this qualification has not itself becomes  partially familiar. It is
limited the usefulness of the auditory thought the PSW represents updating of Infants of Diabetic Mothers (IDMs)
change paradigm. This paradigm has memory for the  partially familiar stim- Infants of poorly controlled dia-
been used to study the development of ulus [Nelson and Collins, 1991; Nelson betic mothers are known to have diffi-
auditory discrimination in infants. If and deRegnier, 1992; Nelson and Monk, culties with cognitive development that
acoustic differences are too small to be 2001]. originate in the fetal period [Stehbens et
perceived, no mismatch response will be Although the number of infant and al., 1977; Rizzo et al., 1997]. Poorly
noted. This offers the ability to evaluate toddler ERPs is still relatively small, it is controlled diabetes may lead to fetal met-
infant development of perception of apparent that this technique can be used abolic derangements that can be delete-
acoustic differences between sounds, and to describe the early perceptual, discrim- rious to fetal brain development, includ-
to evaluate the role of experience in inative, and memory abilities of very ing hypoglycemia, ketonemia, acidosis,
learning language. For example, it has young, preverbal infants with limited be- chronic hypoxia, and brain iron deple-
been shown that 2.5 5 h of auditory havioral responses. Though a great deal tion [Georgieff and Rao, 2001]. In ani-
training during sleep enabled newborn of normative data is still required, knowl- mal models, both brain iron depletion
infants to discriminate acoustically similar edge of cognitive function in these very and hypoxia result in aberrant develop-
Finnish vowel sounds that were not dis- young infants opens up the possibility of ment of neural structures important for
criminated prior to the training [Cheour early assessment of cognitive function in explicit or declarative memory [deUngria
et al., 2002]. The auditory change para- the high-risk infant with congenital, fa- et al., 2000]. Thus it would be predicted
digm has also been used in a number of milial, or perinatal risk factors. that children of poorly controlled dia-
studies to evaluate the discriminative betic mothers may have difficulties with
abilities of high-risk children compared STUDIES OF HIGH-RISK explicit memory and this hypothesis has
with control children, as will be discussed INFANTS recently been tested using ERPs.
later. Timely identification of infants Memory development in a cohort
with congenital, hereditary, or perina- of IDMs from Minnesota has been stud-
Infant Memory tally acquired cognitive difficulties may ied from birth using ERPs and behavioral
The development of auditory and improve our understanding of cognitive testing (deRegnier et al., 2000; Nelson et
visual recognition memory has also been development in high-risk children and al., 2000, 2003; Sidappa et al., 2004; De-
studied using ERPs. Recognition mem- facilitate the development of focused Boer et al., 2005b]. It is known that
ory is a type of explicit memory that is early intervention programs for specific newborn infants are capable of recogniz-
dependent on the hippocampus and disabilities, similar to the situation that ing the maternal voice [DeCasper and
other medial temporal lobe structures now exists for early diagnosis of congen- Fifer, 1980]. For the first test, ERPs were
[Broadbent et al., 2002]. It is the earliest ital hearing loss. A specific approach is recorded at term gestation (40 weeks;
type of memory to develop and many likely to be most accurate, e.g., early mean postnatal age of 10 days) while
behavioral studies have described the rec- testing of memory deficits would be most infants listened to the maternal voice
ognition abilities of infants [Rose et al., useful in prediction of later memory abil- compared with a stranger s voice [de-
1988, 2001; Fagan, 1990]. It is therefore ities. Regnier et al., 2000]. The word  baby
possible to create ERP tasks that are Toward this end, ERPs have been was chosen as a stimulus. The stranger s
based on behavioral information about used to evaluate neural correlates of cog- voice was the voice of the previous
the constraints of early memory develop- nitive function in a number of groups of mother and therefore varied for each in-
ment. high-risk infants and children. It should fant. In the control group s ERPs, the
ERP waveforms that have been be noted that studies to this date have maternal voice elicited a positive wave
observed in memory studies include the focused on evaluating differences in brain peaking at approximately 290 ms. This
negative component (NC), the negative activity between high-and low-risk in- same peak was noted in response to the

320 MRDD RESEARCH REVIEWS COGNITIVE EVALUATION USING ERPS DEREGNIER
stranger s voice, but the ERP elicited by fants memory networks in a similar behavioral manifestations of memory
the stranger s voice also elicited a nega- manner as in animal models. deficits. A recent study [DeBoer et al.,
tive slow wave (consistent with novelty Not only have infants with diabetic 2005b] evaluated IDMs at 12 months of
detection). In the IDM group, these mothers shown neurophysiologic evi- age for behavioral evidence of memory
findings were attenuated and no signifi- dence of differences in auditory recogni- using an elicited imitation paradigm. In
cant differences were noted between the tion memory, these findings have per- an elicited imitation paradigm, sequences
ERPs elicited by the maternal and strang- sisted in visual recognition memory at 6 of events leading to the creation of a toy
er s voice. Follow-up through 1 year of months [Nelson et al., 2000]. Healthy are modeled for the infant. Infants are
age showed that the presence of the neg- 6-month-old infants show behavioral then provided with the supplies needed
ative slow wave for the stranger s voice and ERP evidence of recognition of the to recreate the toy, either immediately or
was associated with higher scores on the maternal face [de Haan and Nelson, after a delay. In this study, infants of
Bayley Scales of Infant Development in 1997; Nelson et al., 2000], and a fol- diabetic mothers were found to have dif-
both the control and IDM groups [de- low-up study evaluated this ability in ficulties with elicited imitation after a
Regnier et al., 2000]. Since no children 6-month-old IDMs compared with con- delay, providing behavioral substantia-
in this study had delayed development, trol infants tested at the same age [Nelson tion of the previous ERP findings.
the negative slow wave was interpreted as et al., 2000]. ERPs were recorded while Though long-term follow up of this co-
a marker of better developmental each infant was shown digital photo- hort continues, this series of ERP and
progress. graphs of the maternal face and a strang- behavioral studies was useful in translat-
The authors speculated that the er s face. As expected from previous ing information gleaned from animal
differences in neonatal memory in the studies, the control infants showed a models of iron deficiency into informa-
IDM group may have been due to peri- larger NC to the maternal face that was tion about memory development in hu-
natal brain iron deficiency. IDMs are followed in the recording by a positive man infants. These findings have led the
known to be at risk for perinatal iron slow wave for the stranger s face. In con- Minnesota IDM research group to initi-
deficiency, and, in animal models, this trast, the IDMs showed no significant ate a trial of early iron therapy for infants
causes iron depletion in the explicit differences between the maternal and of diabetic mothers with suspected brain
memory network that is important for stranger s ERPs. Consequently, this iron deficiency to try to improve mem-
recognition memory [deUngria et al., study showed that the presumed deficits ory development in these infants.
2000]. To evaluate whether the neonatal in auditory recognition memory seen in
findings in the IDM group could be re- the full-term newborn were again found Infants at Risk for Dyslexia
lated to perinatal brain iron depletion, a in the visual modality at 6 months of age. The relationship between early
second group of IDMs were tested and Cross-modal recognition memory speech perception and later reading skills
grouped according to their iron status at refers to the ability to recognize objects or dyslexia has been the source of a num-
birth [Sidappa et al., 2004]. Ferritin is an in different sensory modalities than the ber of studies correlating early ERPs with
iron storage protein in the body, and very original presentation. In infant studies, later cognitive function. Dyslexia, or
low ferritin values at birth ( 35 g/dL) infants are typically tested for visual rec- reading disorder, is a learning disability
have been associated with depletion of ognition of objects that previously were that is known to have a familial basis and
brain iron (see Sidappa et al., 2004 for a palpated but not seen. Behavioral studies therefore may be amenable to early diag-
discussion). The next newborn study have shown when 8-month-old infants nosis [Lyytinen et al., 2004]. The under-
stratified IDMs into two groups, one palpate a simple but unique shape in the lying neuropsychological deficit that un-
with suspected brain iron deficiency (fer- dark, they show visual recognition mem- derlies dyslexia is not known, but
ritin 35 g/dL) and another presumed ory for the palpated object [Rose et al., difficulties with speech perception have
to have brain iron sufficiency (ferritin 1983]. Neurophysiologic studies of been suspected. Children with dyslexia
35 g/dL). The infants were tested at 40 8-month-old infants have shown that, may have very early deficits with speech
weeks postmenstrual age and a mean similar to the maternal face, the familiar perception that may be useful in the early
postnatal age of 2 weeks. The IDMs palpated object elicits a larger NC com- diagnosis of dyslexia that is formally di-
thought to be brain iron sufficient ponent than does the novel object and, agnosed much later. Given that ERPs
showed findings similar to control infants similar to the situation with the stranger s have been used to evaluate early neural
from the previous study (a negative slow face, there is a positive slow wave for the underpinnings of speech perceptions,
wave for the stranger s voice), whereas ERP from the novel object [Nelson et several authors have used ERPs to test
the IDMs suspected to be brain iron de- al., 1993]. However, in a study of infants infants at risk for dyslexia and have de-
ficient showed markedly attenuated find- of diabetic mothers, the slow wave was tected early differences in auditory corti-
ings, with no differences seen between not seen, indicating that alterations in cal responses to speech sounds. In the
the ERPs derived from the maternal and memory development seen in the new- ongoing Jyvaskyla Longitudinal Study of
stranger s voices. In this group of infants, born infant were persistent through 8 Dyslexia [Lyytinen et al., 2004], children
there was a correlation between the se- months of age [Nelson et al., 2003] and with a strong family history for dyslexia
verity of iron deficiency (measured by noted in multiple modalities. have been followed since birth with a
the ferritin concentration) and the ERP With this series of ERP studies, combination of ERP and behavioral
patterns, with infants having higher fer- deficits of auditory, visual, and cross- tests. ERPs were conducted in the new-
ritin values showing ERP patterns typical modal recognition memory were de- born period in response to consonant-
of healthy control infants. Although it is scribed in infants of diabetic mothers. vowel speech sounds (/ba/, /da/, and
necessary to more fully evaluate the role Furthermore, specific ERP patterns re- /ga/). Those newborns with a family his-
of other factors such as chronic hypoxia corded from the neonate were associated tory of dyslexia showed a difference in
in these results, the authors have specu- with mental developmental test scores at the hemispheric patterns of activation for
lated that perinatal brain iron deficiency 1 year of age. However, these findings the speech sound /ga/, with larger and
affects the development of human in- would not be of concern if there were no more positive responses recorded in the

MRDD RESEARCH REVIEWS COGNITIVE EVALUATION USING ERPS DEREGNIER 321
high-risk group over the right hemi- skills at 3 years [Molfese and Molfese, dren [Aylward, 2002]. In contrast to
sphere compared with children at low 1985]. IDMs or infants at risk for dyslexia, no
risk for dyslexia, possibly indicating poor Molfese et al. [2001] also evaluated longitudinal studies have followed pre-
the predictive value of newborn ERPs term children from birth to school age
hemispheric specialization for language
in the high-risk children. Five-year out- coupled with 3-year evaluations of the using ERPs and behavioral or standard-
comes of these children have been re- home environment, language develop- ized testing. However, some shorter fol-
ment, and early intelligence testing in low-up studies and cross-sectional studies
ported. The ERP findings that were
prediction of 8-year reading scores, using have evaluated auditory perception and
characteristic of the risk group showed a
the WRAT-R. Multiple regression discrimination in premature infants.
significant association with poorer recep-
models were created to evaluate these Kurtzberg et al. [1984] tested pre-
tive language at 2.5 years of age. Further-
early factors in prediction of the 8-year mature, very low birth weight infants
more, poor verbal memory skills at age 5
reading scores. Newborn ERP peak la- responses to speech sounds (/da/ and
years were predicted by neonatal ERP
tency measurements over the right tem- /ta/) and a 800-Hz tone at term (40
findings [Guttorm et al., 2001].
poral region and the left frontal area in weeks postmenstrual age) compared with
These same authors also studied in-
response to the speech sound /gi/ con- full-term infants. Monthly ERPs were
fants at risk for dyslexia at 6 months of
tributed significantly to the regression conducted until 3 months corrected age.
age in comparison with low-risk infants.
model and were correlated with the The premature infants showed immature
For this study, they utilized an auditory
8-year reading scores. The authors hy- ERP patterns for the speech sounds at 40
change paradigm in which a long vowel
pothesized that the results of their mul- weeks with progressive normalization of
served as the frequent stimulus and a
tiple studies indicate the children who the ERPs by 3 months corrected age.
short vowel served as the deviant stimu-
process language information more Similar differences were seen in the pre-
lus. They demonstrated that larger
quickly from birth have advantages in term infants responses to the 800-Hz
change detection responses over the left
learning language and, later, in learning tone, but these findings did not achieve
hemisphere predicted better receptive
to read. statistical significance. The authors theo-
language skills at 2.5 years and better
Overall these studies suggest that rized that relative immaturity of auditory
verbal memory skills at 3.5 years. Larger
very early brain responses to speech sounds cortical processing of speech might con-
change detection response over the right
are important in language acquisition. The tribute to subsequent difficulties with
hemisphere was associated with poorer
speed of these responses and hemispheric speech and language development.
verbal memory at 5 years [Lyytinen et al.,
lateralization both appear to be important, An auditory change paradigm was
2004]
though a consensus has not yet been used in a more recent study of preterm
Molfese et al. also reported data
reached regarding the basic underlying al- infants, including a group of infants with
consistent with these findings in a group
terations in the neural pathways leading to intrauterine growth restriction [Fellman
of children studied from birth through
dyslexia and poor reading skills. et al., 2004]. In this study, ERPs were
age 8 using ERPs and behavioral testing
recorded from preterm infants and con-
[Molfese and Molfese, 1985, 1997;
Premature Infants trol infants while infants were tested for
Molfese, 2000; Molfese et al., 2001].
Premature infants, particularly those discrimination of a change in harmonic
Neonatal ERPs from children diagnosed
born at less than 32 weeks gestation, are at tone frequency. Infants were tested at
with dyslexia at 8 years differed from
high risk for long-term cognitive problems term and 6 and 12 months of age, cor-
ERPs obtained from neonates who were
including lowered intelligence, memory rected for prematurity. In the infants
later found to be normal readers. Several
impairments, attention deficit disorders, tested at term, the control group showed
differences were found between normal
learning disabilities, and difficulties with evidence of auditory change detection at
readers, poor readers, and dyslexic chil-
verbal processing [Rose et al., 1988, 2001; several time points in the ERP, whereas
dren in their neonatal ERP responses to
Aylward, 2002]. Premature infants are a the responses were atypical or absent in
speech sounds over both right and left
heterogeneous group of infants who may the preterm infants with and without in-
hemispheres. The investigators used a
experience multiple risk factors, including trauterine growth restriction. Some of
discriminant function analysis to identify
intracranial hemorrhage and white matter the ERP findings at each age (term, 6,
areas of difference between the control
damage, perinatal nutritional deficits, an al- and 12 months corrected age) were
and dyslexic children s neonatal ERPs.
tered environment, and physiological insta- found to correlate with 2-year tests of
They found that measures of peak ampli-
bility with recurrent episodes of hypoxia, mental development (the Bayley Scales
tudes and latencies in the neonatal period
hypotension, or acidosis [Mattia and de- of Infant Development).
correctly classified 13 of 17 dyslexic chil-
Regnier, 1998; Maalouf et al., 1999; Hack Older preterm children have also
dren, a finding well above chance levels
and Taylor, 2000]. Because some of the been the subject of some ERP studies.
of prediction [Molfese, 2000].
factors associated with premature infants Jansson-Verkasalo et al. [2004] in Finland
This same group also used ERPs
development may be modifiable, it is very evaluated 4-year-old former preterm
longitudinally, every 6 months from
important to develop techniques to under- children using ERPs and the auditory
birth through 3 years of age in 16 infants
stand early brain function in these infants, change paradigm using syllables and fol-
in response to synthetic speech stimuli
so that interventions can be tested without lowed these children until 6 years of age,
that had been previously found to pro- the long time lag associated with traditional at which time an object naming test and
duce hemispheric difference effects as
assessment of cognitive function. For this additional ERP was performed. The pre-
well as discrimination effects. The best
group of infants, ERPs are likely to be a term infants showed reduced amplitudes
predictor of later development was found
useful adjunct to anatomic imaging proce- of the auditory change response com-
to be lateralized speech sound discrimi- dures that are currently very popular. pared with full-term control infants. The
nation ability. Children with left hemi- Early language acquisition may be amplitude of the 4-year auditory change
sphere generated discrimination of con- delayed and later complex verbal process- response correlated with the 6-year re-
sonant sounds showed better language ing may be problematic in preterm chil- sponse, and the absence of the auditory

322 MRDD RESEARCH REVIEWS COGNITIVE EVALUATION USING ERPS DEREGNIER
change response at 4 years predicted infants both showed evidence of auditory perceptual and cognitive function are seen
naming difficulties at 6 years. This study change detection, but the ERP patterns between low-risk infants and infants with a
showed stability of the ERP findings were opposite from each other. For the variety of risk factors, including prematu-
over 2 years and revealed that difficulties control group, the auditory change re- rity, uncontrolled maternal diabetes, and
with auditory change detection were as- sponse was a positive wave over the fron- familial disorders such as dyslexia. In a small
sociated with language difficulties. tal scalp, whereas the preterm infants number of studies thus far, these differences
These studies show evidence of ERPs showed a negative auditory change in very early brain function have been pre-
persistent alteration of the neural path- response over frontal and lateral scalp dictive of later difficulties. It should be
ways for perception and discrimination sites. There was no correlation between noted that the sensitivity, specificity, and
of speech sounds in preterm infants from the results of the auditory change para- predictive value of individual infants ERPs
term gestation through the later pre- digm and the maternal voice recognition have not yet been tested. Furthermore, no
school years. Alterations in perception paradigm. The findings were consistent systematic study of individual differences
and discrimination of speech sounds may with altered development of multiple has been performed in low-risk infants.
be indicative of later difficulties with lan- neural pathways, including those in- This will be important because there is a
guage and cognitive development. The volved with auditory recognition mem- great deal of variability seen in infant ERPs
sensitivity and specificity of the ERP ory and auditory discrimination. that may or may not be indicative of cog-
findings has not been tested to determine These studies indicate that cogni- nitive impairments. A great deal of further
whether ERPs could be used as predic- tive impairments in preterm infants al- research is needed before ERPs can be used
tors of language development and verbal most certainly have their origins in the to evaluate individual infants, but the tech-
processing for individual infants. perinatal time period. These data and re- nique shows promise in early diagnosis and
In addition to difficulties with lan- cent MRI studies [Maalouf et al., 1999] assessing new therapies for infants at risk for
guage and verbal development, preterm suggest that many preterm infants expe- later cognitive impairments. f
infants also have been shown to have rience brain injury or disruption of the
deficits in visual recognition memory normal developmental trajectory. For the REFERENCES
Alho K, Sainio K, Sajaniemi N, et al. 1990. Event-
that persist over the first year of life and children who ultimately have normal
related brain potential of human newborns to
correlate with later intelligence [Rose et outcomes, it seems possible that families
pitch change of an acoustic stimulus. Electro-
al., 1988, 2001]. Additionally, a syn- are providing a nurturing environment
encephal Clin Neurophysiol 77:151 155.
drome of developmental amnesia as well that facilitates positive changes in brain
Aylward GP. 2002. Cognitive and neuropsycho-
as a growing spectrum of clinical mem- development and function. Several of the logical outcomes: More than intelligence
quotient (IQ) scores. Ment Retard Dev Dis-
ory deficits have been described in ado- studies described here have shown effects
abil Res Rev 8:234 240.
lescents or young adults who were born of the environment or experience on
Broadbent NJ, Clark RE, Zola S, et al. 2002. The
prematurely [Isaacs et al., 2000, 2003; ERP manifestations of auditory percep-
medial temporal lobe and memory. In: Squire
Vargha-Khadem et al., 2001]. MRI stud- tion or recognition memory [Cheour et
LR, Schacter DL, editor. Neuropsychology
ies have shown that these findings are al., 1998,;2002; Molfese et al., 2001; de- of memory. 3rd edition. New York: Guilford
Press. p 3 23.
associated with decreased hippocampal Regnier et al., 2002]. The ERP tech-
Cheour-Luhtanen M, Alho K, Sainio K, et al.
volumes [Isaacs et al., 2003]. These be- nique has the potential to evaluate
1996. The ontogenetically earliest discrimina-
havioral and MRI studies indicate that changes in brain development and func-
tive response of the human brain. Psycho-
preterm infants are at increased risk for tion that are associated with interventions physiology 33:478  481.
Cheour M, Ceponiene R, Lehtokoski A, et al.
hippocampal injury that can be associated to improve developmental outcomes,
1998. Development of language-specific pho-
with behavioral deficits if severe. even in the preverbal infant. Although
neme representations in the infant brain. Nat
ERPs have recently been used to underutilized for this purpose currently,
Neurosci 1:351 353.
evaluate memory development at term this may be the most promising use of
Cheour M, Martynova O, Naatanen R, et al. 2002.
(40 weeks) postmenstrual age in ex- this technique in our efforts to help im- Speech sounds learned by sleeping newborns.
Nature 415:599 600.
tremely premature infants by Therien et prove the outcomes of high-risk infants.
Coles MGH, Rugg MD. 1995. Event-related brain
al. [2004]. Preterm infants with normal
potentials: an introduction. In: Rugg MD,
cranial ultrasound tests at 1 week and 1 CONCLUSION
Coles MGH, editors. Electrophysiology of
month of age were compared with full- The results of the ERP studies pre- mind: event-related potentials and cognition.
Volume 25. Oxford: Oxford University
term control infants for ERP responses to sented here indicate that memory abilities
Press. p 1 26.
the maternal voice compared with a and speech perception and discrimination
de Haan M, Nelson CA. 1997. Recognition of the
stranger s voice. For the preterm infant are present from birth and are important in
mother s face by 6-month-old infants: a neu-
group, no significant differences were the development of both specific and gen- robehavioral study. Child Dev 73:187 210.
DeBoer T, Scott LS, Nelson CA. 2005. ERPs in
found between the maternal and strang- eral cognitive abilities. These studies have
developmental populations. In: Handy TC,
er s voice ERPs. Lack of difference be- shown that ERPs can be used to evaluate
editor. Event-related potentials: a methods
tween the maternal and stranger s voice cognitive function and neural plasticity in a
handbook. Cambridge, MA: The MIT Press.
ERPs persisted even when an additional longitudinal fashion beginning prior to the
p 264  297.
familiarization period was provided to onset of language and sophisticated behav- DeBoer T, Wewerka S, Bauer PJ. 2005. Explicit
memory performance in infants of diabetic
increase the infants experience with the ior. It is clear that early perceptual abilities
mothers at 1 year of age. Dev Med Child Neu-
maternal voice (though 80% of the in- are tied to later cognitive skills. Further-
rol 47:525 531.
fants had been discharged for at least 6 more, although cognitive function is very
DeCasper AJ, Fifer WP. 1980. Of human bonding:
days before the ERP test was conducted). immature in infants and is not likely to be
newborns prefer their mothers voices. Sci-
ence 208:1174  1176.
An auditory change paradigm was also immutably fixed from birth, the specific
DeCasper AJ, Spence MJ. 1986. Prenatal maternal
used in this study to determine whether cognitive abilities seen in even the young-
speech influences newborns perception of
patterns of auditory discrimination cor- est infants do appear to germinate into
speech sounds. Infant Behav Dev 6:19 25.
related with memory abilities. Interest- more mature abilities in both healthy and
deRegnier R, Nelson CA, Thomas K, et al. 2000.
ingly, the control infants and preterm high-risk children. Very early differences in Neurophysiologic evaluation of auditory rec-

MRDD RESEARCH REVIEWS COGNITIVE EVALUATION USING ERPS DEREGNIER 323
ognition memory in healthy newborn infants Leppanen PHT, Eklund KM, Lyytinen H. 1997. Nelson CA, Henschel M, Collins PF. 1993. Neural
and infants of diabetic mothers. Disabil Re- Event-related brain potentials to change in correlates of cross-modal recognition mem-
habil 137:777 784. rapidly presented acoustic stimuli in new- ory by 8-month-old infants. Brain Cogn 29:
deRegnier R, Wewerka S, Georgieff MK, et al. borns. Dev Neuropsychol 13:175 204.
411 420.
2002. Influences of post-conceptional age and Leppanen, PHT, Guttorm TK, Pihko E, et al.
Nelson CA, Monk CS. 2001. The use of event-
postnatal experience on the development of
2004. Maturational effects on newborn ERPs
related potentials in the study of cognitive
auditory recognition memory in the newborn
measured in the mismatch negativity para-
development. In: Nelson CA, Luciana M.
infant. Dev Psychobiol 41:216 225.
digm. Exp Neurol 190:S91 S101.
editors. Handbook of developmental cogni-
deUngria M, Rao R, Wobken JD, et al. 2000.
Lyytinen H, Ahonen T, Eklund K, et al. 2004.
tive neuroscience. Cambridge, MA: MIT
Perinatal iron deficiency decreases cyto- Early development of children at familial risk
Press. p 125 136.
chrome c oxidase activity in selective regions
for dyslexia-follow-up from birth to school
Nelson CA, Wewerka S, Borscheid AJ, et al. 2003.
of the brain. Pediatr Res 48:169 176.
age. Dyslexia 10:146 178.
Electrophysiologic evidence of impaired
Duclaux R, Challamel MJ, Collet L, et al. 1991.
Maalouf EF, Duggan PJ, Rutherford MA, et al.
cross-modal recognition memory in
Hemispheric asymmetry of late auditory
1999. Magnetic resonance imaging of the
8-month-old infants of diabetic mothers. J
evoked response induced by pitch changes in
brain in a cohort of extremely preterm infants.
Pediatrics 142:575 582.
infants: influence of sleep stages. Brain Res
Disabil Rehabil 135:351 357.
Nelson CA, Wewerka S, Thomas KM, et al. 2000.
566:152 158.
Martynova O, Kirjavainen J, Cheour M. 2003.
Neurocognitive sequelae of infants of diabetic
Eggermont JJ. 1988. On the rate of maturation of
Mismatch negativity and late discriminative
mothers. Behav Neurosci 114:950 956.
sensory evoked potentials. Electroencephal
negativity in sleeping human newborns. Neu-
Novak GP, Kurtzberg D, Kreuzer JA, et al. 1989.
Clin Neurophysiol 70:293 305.
rosci Lett 340:75 78.
Cortical responses to speech sounds and their
Fagan JFI. 1990. The paired-comparison paradigm
Mattia FR, deRegnier RAO. Chronic physiologic
formants in normal infants: maturational se-
and infant intelligence. Ann NY Acad Sci
instability is associated with neurodevelop-
quence and spatiotemporal analysis. Electro-
608:337 364.
mental morbidity at one and two years in
encephal Clin Neurophysiol 73:295 305.
Fellman V, Kushnerenko E, Mikkola K, et al. 2004.
extremely premature infants. Pediatrics elec-
Ponton CW, Eggermont JJ, Kwong B, et al. 2000.
Atypical auditory event-related potentials in
tronic pages 102:e35, 1998.
preterm infants during the first year of life: a Maturation of human central auditory system
Mills DL, Coffey-Corina SA, Neville HJ. 1997.
possible sign of cognitive dysfunction? Pediatr activity: evidence from multi-channel evoked
Language comprehension and cerebral spe-
Res 56:291 297. potentials. Clin Neurophysiol 111:220  236.
cialization from 13 20 months. Dev Neuro-
Georgieff MK, Rao R. 2001. The role of nutrition
Regan D. 1989. Human brain electrophysiology:
psychol 13:397 446.
in cognitive development. In: Nelson CA,
evoked potentials and evoked magnetic fields
Mills DL, Prat C, Zangl R, et al. 2004. Language
Luciana M, editors. Handbook of develop-
in science and medicine. New York: Elsevier.
experience and the organization of brain ac-
mental cognitive neuroscience. Cambridge,
Richards JE. 2003. Attention affects the recognition
tivity to phonetically similar words: ERP ev-
MA: MIT Press. p 45 58.
of briefly presented visual stimuli in infants: an
idence from 14- and 20-month olds. J Cogn
Geva R, Gardner JM, Karmel BZ. 1999. Feeding-
ERP study. Dev Sci 6:312 328.
Neurosci 16:1452 1464.
based arousal effects on visual recognition
Richards JE, Casey BJ. 1991. Heart rate variability
Molfese DL. 1973. Cerebral asymmetry in infants,
memory in early infancy. Dev Psychol 35:
children and adults: auditory evoked re- during attention phases in young infants. Psy-
640 650.
chophysiology 28:43 53.
sponses to speech and noise stimuli. Diss Abstr
Ginsberg H, Opper S. 1979. Piaget s theory of
Int 34:1298. Rizzo TA, Metzger BE, Dooley SL, et al. 1997. Early
intellectual development. Englewood Cliffs,
Molfese DL. 1989. Electrophysiological correlates malnutrition and child neurobehavioral devel-
NJ: Prentice-Hall.
of word meanings in 14-month-old human opment: insights from the study of children of
Guttorm TK, Leppanen PHT, Richardson U, et al.
infants. Dev Neuropsychol 5:79  103. diabetic mothers. Child Dev 68:26 38.
2001. Event-related potentials and consonant
Molfese DL. 1990. Auditory evoked responses re- Rose SA, Feldman JF, Jankowski JJ. 2001. Atten-
differentiation in newborns with familial risk
corded from 16-month-old human infants to
tion and recognition memory in the 1st year
for dyslexia. J Learn Disabil 34:534 544.
words they did and did not know. Brain Lang
of life: a longitudinal study of preterm and
Hack M, Taylor HG. 2000. Perinatal brain injury
38:345 363.
full-term infants. Dev Psychol 37:135 151.
in preterm infants and later neurobehavioral
Molfese DL. 2000. Predicting dyslexia at 8 years of
Rose SA, Feldman JF, McCarton CM, et al. 1988.
function. JAMA 284:1973 1974.
age using neonatal brain responses. Brain
Information processing in seven-month-old
Hirasawa K, Kurihara M, Konishi Y. 2002. The
Lang 72:238 245.
infants as a function of risk status. Child Dev
relationship between mismatch negativity and
Molfese DL, Molfese VJ. 1985. Electrophysiological
59:589 603.
arousal level: can mismatch negativity be an
indices of auditory discrimination in newborn
index for evaluating the arousal level in in- Rose SA, Gottfried AW, Bridger WH. 1983. In-
infants: the basis for predicting later language
fants? Sleep Med Suppl 2:S45 48. fants cross-modal transfer from solid objects
performance? Infant Behav Dev 8:197 211.
Isaacs EB, Lucas A, Chong WK, et al. 2000. Hip- to their graphic representations. Child Dev
Molfese DL, Molfese VJ. 1997. Discrimination of
pocampal volumes and everyday memory in
54:686 694.
language skills at five years of age using event-
children of very low birthweight. Pediatr Res
Sidappa A, Georgieff MK, Wewerka S, et al. 2004.
related potentials recorded at birth. Dev Neu-
47:713 720.
Iron deficiency alters auditory recognition
ropsychol 13:135 156.
Isaacs EB, Vargha-Khadem F, Watkins KE, et al.
memory in newborn infants of diabetic moth-
Molfese VJ, Molfese DL, Modglin AA. 2001. New-
2003. Developmental amnesia and its rela-
ers. Pediatr Res 55:1034 1041.
born and preschool predictors of second-
tionship to degree of hippocampal atrophy.
Stehbens JA, Baker GL, Kitchell M. 1977. Out-
grade reading scores: an evaluation of cate-
Proc Acad Natl Sci USA 100:13060 13063.
come at ages 1, 3, and 5 years of children born
gorical and continuous scores. J Learn Disabil
Jansson-Verkasalo E, Korpilahti P, Jantti V, et al.
to diabetic women. Am J Obstet Gynecol
34:545 554.
2004. Neurophysiologic correlates of defi-
127:408 413.
Naatanen R. 1995. The mismatch negativity: a
cient phonological representations and object
Therien JM, Worwa CT, Mattia FR, et al. 2004.
powerful tool for cognitive neuroscience. Ear
naming in prematurely born children. Clin
Altered pathways for auditory discrimination
Hearing 16:6  18.
Neurophysiol 115:179  187.
and recognition memory in premature new-
Nelson CA. 1994. Neural correlates of recognition
Kurtzberg D, Hilpert PL, Kreuzer JA, et al. 1984.
borns. Dev Med Child Neurol 46:816  824.
memory in the first postnatal year of life. In:
Differential maturation of cortical auditory
Thoman EB. 1990. Sleeping and waking states in
Dawson G, Fischer K, editors. Human behav-
evoked potentials to speech sounds in normal
infants: a functional perspective. Neurosci
ior and the developing brain New York:
full term and very low-birthweight infants.
Biobehav Rev 14:93 107.
Guilford Press. p 269 313.
Dev Med Child Neurol 26:466 475.
Vargha-Khadem F, Gadian DG, Mishkin M. 2001.
Nelson CA, Collins PF. 1991. An event-related
Kushnerenko E, Ãeponienë R, Balan P, et al.
Dissociations in cognitive memory: the syn-
2002a Maturation of the auditory event-re- potential and looking time analysis of infants
responses to familiar and novel events: impli- drome of developmental amnesia. Philos Trans
lated potentials during the first year of life.
R Soc Lond Ser B Biol Sci 356:1435 1440.
NeuroReport 13:47 51. cations for visual recognition memory. Dev
Weitzman ED, Graziani LJ. 1968. Maturation and
Kushnerenko E, Ãeponienë R, Balan P, et al. Psychol 27:50 58.
2002b. Maturation of the auditory change Nelson CA, deRegnier R. 1992. Neural correlates topography of the auditory evoked response
detection response in infants: a longitudinal of attention and memory in the first year of of the prematurely born infant. Dev Psycho-
ERP study. NeuroReport 13:1843 1848. life. Dev Neuropsychol 8:119 134. biol 1:79  89.

324 MRDD RESEARCH REVIEWS COGNITIVE EVALUATION USING ERPS DEREGNIER