An Electrophysiological Study of School-aged Children with a History of Failure to Thrive during Infancy ROSCOEA. DYKMAN, PHILIP C . LOlZOU, PEGGY T. ACKERMAN, PATRICK H . CASEY, W . BRIAN MCPHERSON
Department of Pediatrics, University of Arkansas for Medical Sciences, Arkansas Children's Hospital Research Institute, and Arkansas Children's Nutrition Center
Abstract--Sixty-five subjects, ages 8 to 12, participated in a visual electrophysiological study. Twenty-two of the subjects had received a diagnosis of nonorganic failure-to-thrive (FTT) before the age of three. The remaining 43 subjects had no history of FFT and served as Controls. IQs were obtained with the abbreviated WISC-III, and the Controls were split into two groups, LO IQ and HI IQ, to provide a LO IQ Control group with an average IQ equivalent to the FTT group. Event-related brain potentials (ERPs) were recorded from five scalp locations during a cued continuous performance task (CPT). Subjects had to press a button every time they saw the letter "X" following the letter "A" (50 targets out of 400 stimuli). During the CPT, the FTT subjects made marginally more errors of omission to targets than the LO IQ Control group and significantly more errors of omission than the HI IQ Control subjects. The groups did not differ significantly on errors of commission (false alarms) or reaction times to targets. ERP averages revealed a group difference in amplitude in a late slow wave for the 50 non-X stimuli (false targets) that followed the letter A. This difference was greatest over frontal sites, where the FTT group had a more negative going slow wave than each control group. Late frontal negativity to No Go stimuli has been linked with post-decisional processing, notably in young children. Thus, the FFT subjects may have less efficient inhibitory processes, reflected by additional late frontal activation.
FAILURETO THRIVE(FTT) is a term used by pediatricians to describe infants and toddlers with abnormally low weight for age and gender and/or with an abnormally low weight gain over a period of time (Casey, 1992). Although there are no strict criteria for FTT, the FTT diagnosis usually implies the child weighs less than the 5th percentile on the National Center for Health Statistics (NCHS) curves (referenced against children of same gender and gestation adjusted age) and the child's weight to height ratio is less than the 25th percentile (on the NCHS curves). Additionally, genetic growth expectation is considered, i.e., the parents' height and weight. Thus, for a small child in a small family to be labeled FTT, the child would have to be low in weight for height or show low weight gain velocity, since weight for age would normally be low. Some children above the 5th percentile in weight are called FTT if they show abnormally slow weight gain. Abnormal weight gain can be defined as moving down across two major percentile lines on the
Address for correspondence: Roscoe A. Dykman, Ph.D., Dept. of Pediatrics/C.A.R.E., Arkansas Children's Hospital, 800 Marshall Street, Little Rock, AR 72202.
Integrative Physiological and Behavioral Science, October-December 2000, Vol. 35, No. 4,284-297.
284
ELECTROPHYSIOLOGICAL STUDY OF SCHOOL-AGED CHILDREN
285
NCHS curves (e.g., from above the 50th percentile to below the 25th percentile) over a period of a month or two. The causes of FTT are often divided into three categories: organic, non-organic, and mixed. Organic cases of FTT result from disease or genetic conditions, while environmental factors (e.g., inadequate parenting skills) are theorized to precipitate non-organic FTT. Mixed FTT has both organic and non-organic influences. Whatever the origin of FTT, it is generally associated with undernutrition, whether it is a disease that blocks or interferes with the absorption of nutrients or simply a lack of the proper food intake. It is long known that important developments of the brain and cognitive processes take place during the first three years of a child's life (McCall, Hogarty, & Hulbert, 1972). Infants and toddlers experience a rapid increase in neuronal connections crucial for language, executive functioning, and other cognitive tasks (Dawson & Fischer, 1994). Because of the accelerated cerebral growth during this period, any significant reduction in nutrition that results in FTT carries a high risk for negative effects for intellectual development later in life (Cravioto & Arrieta, 1983; Dobbing, 1976; Pollitt, 1995). The effects of FTT have been detected in both short-term and long-term outcomes. In a relatively short-term study, Skuse, Pickles, Wolke, and Reilly (1994) examined a cohort of 1558 full-term infants (singletons) containing 47 cases with serious growth delay in the first year of life. They found that 37 percent of the variance in cognitive and psychomotor functioning at the age of 15 months could be explained by developmental progress in the first few postnatal months. Studies of the long-term effects of poor growth and nutrition show similar results. Galler and associates (Galler, & Ramsey, 1989; Galler, Ramsey, Forde, Salt, & Archer, 1987) followed 129 malnourished Barbadian children from infancy into grade school and contrasted them with a carefully matched control group. They found mean IQ and achievement scores significantly lower in the previously malnourished children, and fully 60 percent of that group were considered to have attention deficit disorder (ADD). In a similar follow-up study of 48 preschool children earlier labeled FTT, Drotar and Sturm (1992) found deficits in behavioral organization, ego control, ego resiliency, and problem solving. In a small study of somewhat older children (average age 121/2 years), Oates and colleagues (Oates, Peacock, & Forrest, 1985) found poorer language development, less developed reading skills, less social maturity, and a higher prevalence of behavioral problems in children with an early history of FTT (N=14) compared to a socio-demographically matched control group. There is some psychophysiological evidence showing the impact of retarded infant growth. Barnet et al. (1978) first reported abnormal auditory evoked potentials (AEPs) in early infancy malnutrition (marasmus). Flinn, Barnet, Lydick, and Lackner (1993) later recorded ERPs to click and name stimuli in malnourished infants when they were admitted to the hospital and when they were discharged. Malnourished infants' average brain potentials differed from controls on admission (fewer well defined peaks and lower amplitude waves) but not on discharge. However, malnourished infants had smaller AEP amplitudes to names than to clicks on discharge, but normal infants did not. Electrophysiological evidence for long-term effects of early childhood FTT is lacking, but it is reasonable to expect that such effects exist, given the finding by Galler et al. (1984) of a high incidence of attention deficit disorder in children with early malnutrition. ADD children have been shown to exhibit a smaller, later P300 component than normal children (Holcomb, Ackerman, & Dykman, 1985) and aberrant slow wave activity to a warning signal (Newton, Oglesby, Ackerman, & Dykman, 1994). While it is unlikely that
286
DYKMAN ET AL.
children with a history of FTT will show the same brain wave patterns as those with ADD, given the additional problems associated with FTT, there may be some similarities between FTT and ADD children's brain responses. The current study recorded event-related brain potentials (ERPs) of subjects between 8 12 years of age with and without a history of FTT while the subjects engaged in a cued continuous performance task requiring both attention and aspects of executive processing such as memory updating and inhibition (see Overtoom, Verbaten, Kemner et al., 1998; Pennington, 1993). It was predicted that the children with a history of FTT would commit more errors of commission and omission than the control groups. It was also predicted that the FTT group would show abnormal ERPs. However, specific predictions were not possible given the lack of previous ERP work with formerly FTT children. METHODS
Subjects Clinical subjects (N=22) were recruited from a roster of former patients in the Growth and Development Clinic at Arkansas Children's Hospital. The project recruiter attempted to reach caregivers of all (82) former physician diagnosed non-organic FTT patients currently in the age range of 8 to 12 years old. Twenty-eight caregivers were reached. After the recruiter explained the nature of the project and offered $100 for participation, all 28 of the caregivers agreed to have their children participate. However, one subject was not admitted to the study because the child had developed cerebral palsy, and five subjects did not complete the entire protocol. Control subjects (N=43) were recruited via advertisements placed on the bulletin boards of the general pediatrics outpatient clinic at Arkansas Children's Hospital, at housing projects, at mobile home parks, at city colleges, and at laundromats. The majority of the control subjects were general pediatric clinic outpatients who were in for routine checkups, and none had history of FTT. The caregivers of the control subjects also were given $100 for their participation. All subjects and their caregivers signed informed consent forms approved by our local Institutional Review Board. In order to provide an IQ match for the FTT subjects, the control subjects were divided into two groups. The 18 subjects with a WISC-III abbreviated IQ less than 94 were placed in the LO IQ group and the 25 subjects with an IQ of 94 or greater formed a group called HI IQ. A median split on IQ (at 100) was not used because the lower half of the controls had an average IQ that was statistically greater than the FTT group. Basic demographic and IQ data for the three groups are shown in Table 1. The children in both groups were given, in addition to the abbreviated WISC-III, the Wechsler Individual Achievement Test and the Visual Motor Integration Test. Their caregivers completed the Child Behavior Checklist (Achenbach, 1991) and the Kaufman Brief Intelligence Test. Caregivers also answered questions that allowed estimation of current socioeconomic status (SES). The children were assessed for current height and weight. The findings from the complete assessment will be reported elsewhere (Dykman et al.; in press). In brief, these data showed the formerly FTT group to be still retarded in growth and to have poorer cognitive/achievement outcomes as well as more behavioral problems than a SES matched control group. Because this was not a planned follow-up study of the children earlier diagnosed with non-organic FTT, neither the family situation nor infant attachment during that period of
ELECTROPHYSIOLOGICAL STUDY OF SCHOOL-AGED CHILDREN
287
growth deficiency was systematically studied. Thus, we could not with any certainty ascertain the prevalence and degree of neglect in this sample. The best that can be said about this convenience sample is that each formerly FTT child demonstrated growth deficiency that was not found to stem from an organic cause. A planned follow-up study of non-organic FTT infants/toddlers is under way in our Nutrition Research Center and, after time, can provide more definitive background data.
Procedure During the continuous performance task (CPT), subjects watched a monitor as a sequence of single letters appeared, one every three seconds. A total of 400 letters were presented, with each letter staying on the screen for one second. The letter "A" appeared 100 times. The letter "X" also appeared 100 times, 50 of those times were after the letter " A ' . Subjects were instructed to press a key, located on the arm of the chair in which they sat, every time they saw a letter X following the letter A (a total of 50 or 12.5 percent of the time). A goodwill reward of $2 was given at the end of the procedure. No feedback was given during the test. There were two equally probable categories of letters that were of primary interest: 1) the letter "X" appearing after the letter "A" or "targets"; and 2) any other letter than "X" that appeared after the letter "A", called "false targets". The two types of stimuli are sometimes labeled "Go" and "No Go" (Overtoom et al., 1998). The elapsed time between the onset of a letter and any key press that occurred within two seconds of that onset was recorded and stored on magnetic tape. This task is our own modification of the standard AX (cued) version of the CPT. It differs primarily in the inclusion of equal numbers of targets and false targets. The false targets or No Go stimuli are theorized to require greater inhibition than background letters not preceded by an A. ERP Recording. Subjects were fitted with an electrode cap (Electrode-Cap International) to hold the five active electrodes to the scalp. The locations for these electrodes included one midline lead, Cz, and two pairs of bilateral leads, F3, F4, P3, and P4. All leads were referenced to left mastoid. Blink activity of the right eye was recorded from two electrodes taped to the skin about 15 mm apart. The inferior lead was 9 to 10 mm below the lower lid margin, and the lateral electrode was 5 to 9 mm medial to the outer canthus. Impedance on all EEG channels was reduced to below 5000 ohms. EEG was recorded for one-second epochs starting 100 ms before each letter presentation. The EEG signal was amplified by a Grass Model 12 Neurodata Acquisition system using a bandpass filter with-3dB cutoffs of 0.3 Hz and 100 Hz. The amplified EEG was then sent to a Compupro 68000 microcomputer equipped with an analog to digital converter. The computer digitized the EEG at a rate of 256 Hz and stored the data on magnetic tape, as well as displaying the EEG for each trial on a computer monitor after each trial. On-line checking for eye and EEG artifact provided feedback to the experimenters on the quality of the data, thus alerting them to urge subjects not to blink when artifacts were excessive. The computer was programmed to reject any trial during which activity measured 100 microvolts or more in any channel. But, additional trials were not run to make up for rejected trials.
ERP Data Analyses In order to reduce the large number of possible comparisons, we opted to focus on our
288
D Y K M A N ET AL.
major interest (selective attention to target stimuli and inhibition to false target stimuli). The CZ data are not included in the statistical analyses because we wished to evaluate hemispheric differences. Group comparisons are reported separately for the frontal and parietal regions inasmuch as the various components are more prominent in one region than the other (see Figure 1). The P200 peak is most pronounced frontally, as is the N300 peak. Indeed, there is no N300 for targets at parietal sites. The P300 peak is most pronounced in the parietal region. The group contrasts for the late slow wave cannot legitimately compare targets versus false targets since the targets required a motor response, which could affect this late component. The F values for condition and site are evaluated at 1 and 62 degrees of freedom. F values involving the three groups are evaluated at 2 and 62 df. All analyses were performed with the SPSS statistical package. Figure 1 shows the grand mean ERPs (FTT plus the two control groups) to the target and false target stimuli. Inspection of these waveforms revealed several components of interest. These included the P200 peak, the N300 peak, the P300 peak, and the late slow wave. The P200 component, a large positive peak at frontal sites and a smaller peak at parietal sites, appeared at approximately 200 ms post-stimulus onset at parietal sites and at about 250 ms at frontal sites. It appears to be increased for targets relative to false targets. It was measured as the amplitude of the largest positive peak between 150 and 300 ms. The N300 frontal component appears to be larger for false targets than targets and is not visible to targets at parietal sites. It was measured as the amplitude of the largest negative peak between 325 and 425 ms. The P300 elicited by the targets appears to be the classic P300 (Johnson & Donchin, 1978). It is measured here as the amplitude of the largest positivity that occurred between 300 and 430 ms and is distinct from the P200 peak that occurs earlier. The P300 is largest over the parietal sites and appears slightly later over frontal sites. The last component of interest is the slow wave activity that occurred between 650 and 900 ms after the presentation of stimuli. This feature was measured as the average amplitude value from 650 to 900 ms. All four components were referenced to the average values from 100 ms prior to stimulus onset. Figure 2 compares the ERPs of the three groups to the target stimuli. At the F4 (right frontal) site, the HI IQ group appears to have a larger P200 peak and a smaller N300 peak than the FTT group. At parietal sites, the LO IQ group has a less broad P300 component, and the HI IQ group seems to have a deeper P300 at the left parietal site. Figure 3 compares the ERPs of the three groups to the false target stimuli. The HI IQ group seems to have a larger P200 peak at the right frontal site, and the FTT group seems to have the steepest N300 component. The slow wave at frontal sites is positive for the two control groups and negative going for the FTT group. At parietal sites, the LO IQ group has a more aberrant waveform between 200 and 600 msec. Results Table 1 presents demographic data for the formerly FTT children and the LO IQ and HI IQ control groups. The FTT and LO IQ groups are well matched. The HI IQ group is somewhat younger and has fewer African American children. Table 2 gives behavioral data results from the Continuous Performance Task (CPT) for the three groups. ANOVAs showed a significant group difference for mean number of correct hits to the 50 target stimuli (F2,62 = 8.57, p=.001) but no group difference for mean number of false alarms (presses to non-targets) or mean reaction time to target
ELECTROPHYSIOLOGICAL
STUDY OF SCHOOL-AGED
289
CFtlLDREN
,g:
~t
a.O "r" II
"!,
g.
f '.'2""'""
9
,:g ,r_.~
1~ .
E
I
0
i
I
9"
O
o
o
-
7
g
9
o Art
r
i
c~
.-
.e .~ - - [...,
'k
G.
,e-.
~D "O'~
/
g
O O r
~D
Otj .'a
tad
8 t,..
0
O
i
"7
9
o ^~t
d
,-
"7
o.
o
d
.'-
~Z rl
290
DYKMAN ET AL.
1
L~
I|
/
ssSS
/
8C~
i
9
t-
I,,
II 0
I
9
i
A~
O~
8
8~T
E E
8 O O i
i
L~
|
'7
9
Ad
i
i
!
o
d
!
--
ELECTROPHYSIOLOGICAL
STUDY
OF SCHOOL-AGED
291
CHILDREN
B
II
~
E
8
g 9 O
C~
,...I r e-,
i
i
~!.
I
i
u!.
~._.;
'7
A~t
A~
r_je
8~
/)•i'
8
8
8
8 /
oO
I
8 O,I
O
O
O
o
re rrl i
i
'7
9
!
o
Ad
o
"7
,
i
c?
o Ad
o f.t.
292
DYKMANET AL.
Table 1 Means and standard deviations for age and IQ, plus race and gender distribution for the three groups of subjects. The LO IQ and FTT groups did not differ on any of these measures.
Mean (sd) age in months Mean (sd) IQ % AfricanAmerican % Female
FTT N=22
LO IQ N=18
HI IQ N=25
121 (14) 76 (14) 59 50
121 (18) 82 (10) 67 28
109 (11) 111 (12) 20 52
stimuli. Follow-up t-tests on correct hits showed a marginal difference between the FTT and LO IQ groups (t=1.80, p=.08, 38dr) and a marked difference between the FTT and HI IQ groups (t=4.16, p<.001,45 df). The HI and LO 1Q groups also differed on correct hits (t=2.04, p=.048). Also analyzed were the number of target and false target trials saved for ERP averaging. There were more false target than target trials saved (42.3 vs 38.9, F=13.48, p<.001), and the groups differed overall (F=10.30, p<.001). The FTT group had an across condition average of 35.5 trials saved as compared with 43.6 for the HI IQ group and 42.5 for the LO IQ group. Thus, the decision was made to use trials saved as a covariate in the event of a significant group difference on any ERP measure. P200 Peaks
The ANOVA for the P200 peak at frontal sites yielded a large condition effect (F=56.56, p<.001) attributable to the much larger peak to target than false target stimuli. There was a marginal site effect (F=3.72, p=.058, F4>F3). There was no main effect for group, but there was a site x group interaction (F=5.24, p=.008). This arose because both control groups had somewhat larger peaks at F4 (right hemisphere) whereas the FTT group had somewhat larger peaks at F3 (left hemisphere). However, the groups did not differ at either F3 or F4 when the sites were analyzed separately. The P200 peak at parietal sites also yielded a significant condition effect (F=39.34, p<.001) because of the much larger peak to the targets. No other effects emerged.
Table 2 Behavioral responses of the three groups recorded during the Continuous Performance Task. Shown are the number of subjects with less than a 90% hit rate and the number who made two or more false alarms as well as mean number of hits and false alarms and mean reaction times. FTT N=22
Less than 90% hits Mean correct hits Two or more false alarms Mean false alarms Mean reaction times (ms)
16(73%) 38.4+8.0 13(59%) 4.9+6.3 666+81
LO IO N=18
HI 10 N=25
7(39%) 42.8+7.2 10(56%) 4.1+5.2 650+64
7(28%) 46.5+4.2 8(32%) 2.0+3.5 639_+76
293
ELECTROPHYSIOLOGICALSTUDY OF SCHOOL-AGED CHILDREN
Table 3 Hemispheric Effects Site
Site x Condition
P200
F4>F3
N300
F3>F4
F3>F4 larger to targets
P300 P300 Slow wave
F4>F3 P3>P4
F4>F3 larger to targets P3>P4 larger to targets
Site x Group
F4>F3 F4>F3 F3>F4 F3>F4 F3=F4 F3=F4 F4>F3
HI IQ LO IQ FrT HI IQ LO IQ FIT HI IQ greater
P4>P3
HI IQ
N 3 0 0 Peak The N300 peaks at frontal sites produced a highly significant condition effect (F=137.12, p<.001) because the peak to the false targets was much more negative going than the peak to the targets. There was also a site effect (F=4.73, p=.033) and a condition x site interaction (F=4.49, p=.038). The sites had nearly identical peak amplitudes to the false targets, but the F4 (right frontal) site was not as negative going as the F3 site to the targets. Additionally, there was a site x group interaction (F=5.28, p=.008). This effect reflects the fact that the FTT and LO IQ groups had nearly identical amplitudes at F3 and F4 whereas the HI IQ group had more negative going peaks at F3 than F4. Since there was no visible N300 to targets in the parietal region, the groups were contrasted only for the false target condition. There was no group or site difference. P300 Peak The P300 component at frontal sites yielded significant main effects for condition (F---64.97, p<.001) and site (F1,68 = 11.36, p=.001), as well as a condition x site interaction (F=5.73, p=.020). These effects arose because target values were larger than false target values and because F4 (right hemisphere) values were larger than F3 values. The interaction reflects the fact that this hemispheric difference was greater for targets than false targets. There was also a significant site x group interaction (F=4.02, p=.023). The hemispheric difference (right>left) was greater in the HI IQ control group than in the FTT and LO IQ groups. The P300 peak at parietal sites, where it is most pronounced, produced no group effects. Main effects emerged for condition (F=188.96, p<.001) and site (F=8.13, p=.006), modified by a condition x site interaction (F=4.50, p=.038). Targets elicited much larger peaks than false targets and the left (P3) values were greater than the right (P4) hemisphere values, but the left>right difference was more marked to targets than false targets. Slow Wave The slow wave was evaluated separately for the false target and target conditions, as noted earlier. The ANOVAs for the slow wave averages to the target stimuli showed no group differences either at the frontal or parietal sites. There were also no site differences. The analysis of false targets at frontal sites yielded a significant group effect (F=3.30, p=.044). The FTT group had more negative going values than both the HI and LO IQ groups.
294
DYKMAN ET AL.
In follow-up analyses, we contrasted the FTT group with the LO IQ group alone and the HI IQ group alone and used number of trials saved as a covariate. The FTT vs LO IQ contrast yielded an F1,37 of 7.19, p=.011, and the FTT vs HI IQ contrast yielded an F 1,44 of 5.28, p=.026. Combining the LO and HI IQ groups and contrasting the resultant control group with the FTT group gave an F of 7.99, p=.006. At parietal sites, there was a site x group interaction (F=3.43, p=.038), because the HI IQ group had somewhat more positive values at P4 than P3 whereas the other groups showed similar values at the two sites. Discussion
The cued (AX) version of the Continuous Performance Task is somewhat more challenging than the non-cued (respond to all Xs) version. Even so, children of the age range here studied generally make relatively few errors of commission (sometimes called false alarms) or omission (failure to respond to targets). Thus, it is not surprising that the formerly FTT children made no more errors, on average, than the LO IQ control group and that the mean reaction times of the FTT and LO IQ groups did not differ. The FTT children did perform more poorly than the HI IQ control group, particularly in correct detections, or inversely, errors of omission. Indeed, as Table 2 shows, 73 percent of the FTT children were less than 90 percent accurate in detecting target stimuli. The two stimulus conditions (targets and false targets) differed markedly on all three peaks (P200, P300, and N300), but there were no condition x group interactions. All three groups displayed larger P200 and P300 peaks to targets than false targets, and all three groups had larger N300 peaks to false targets than targets (frontal only). As explained previously, the slow wave could not be legitimately analyzed for condition effects since the children made a motor response to targets only. We had expected the P300 wave at parietal sites to show a major condition effect because this is the most frequently replicated finding in tasks that have an odd-ball (infrequent vs frequent) feature. In the CPT task used here, the target stimuli occurred on 12.5 percent of trials. It had not been expected that the P200 peak would also reflect a major condition difference. This difference, we believe, arises from the fact that the P200 peak is more shallow to false targets. We suggest that this more shallow peak occurs because of the larger subsequent N300 to false targets. This larger N300 to false targets than targets probably reflects priming, or cueing, of the target Xs by the preceding As. That is, the N300 appears to be in the N400 family. A large literature demonstrates that priming or cueing of a target stimulus by a preceding stimulus results in a smaller N400 (see Kutas & Iragui, 1998). Kutas and Hillyard (1980) first demonstrated this effect in sentences with incongruous ending words (The man spread his toast with mud.). The effect is also apparent when the preceding stimulus is semantically or phonetically related to the second (or target) stimulus (e.g., nurse, doctor, or hat, cat). The target stimulus is presumably processed more easily because it is expected. In the present task, when the A is not followed by the expected X, we propose that the non-X (false target) is not as easily integrated, and this extra processing effort is reflected in the larger N300. Whatever the dynamics, the recognition of targets and false targets clearly begins before the peak of the large parietal P300 wave near 375 msec. Indeed, in adults the frontal N2 peak for NoGo (false target) trials is larger than for Go (target) trials (Eimer, 1993; Jodo & Kayama, 1992), and this peak therefore has been interpreted to reflect inhibition. Overtoom and associates (1998) were thus led to study the frontal N2
ELECTROPHYSIOLOGICAL STUDY OF SCHOOL-AGED CHILDREN
295
peak in children with Attention Deficit Hyperactivity Disorder (ADHD) performing the AX version of the CPT. However, they found no difference in frontal N2 amplitudes to NoGo (false target) stimuli between the ADHD group and normal controls. On the other hand, a subgroup of the ADHD sample having comorbid Oppositional Defiant Disorder did have smaller frontal N2 peaks to the NoGo stimuli. Overtoom et al. did not analyze the late slow wave. In the present study, hemispheric differences were not nearly as robust as condition differences, and most involved interactions with group or condition. These effects are summarized in Table 3. The most consistent findings are that the target stimuli elicited greater hemispheric effects than the false targets and that the HI IQ group showed more evidence of hemispheric lateralization than the LO IQ and FTT groups. Functional brain imaging results suggest that in young adults there is a lateralized right frontal activation on the CPT, especially during NoGo trials (Haeger, Volz, Gaser et al., 1998; Strik, Fallgatter, Brondeis, & Pascual-Marqui, 1998). Fallgatter, Wiesbeck, Weigers, and Strik (1998) also found a right frontal activation on NoGo trials using ERP field potentials. Of particular pertinence, Casey, Castellanos, Giedd and associates (1997), found evidence of right prefrontal abnormality in an MRI study of ADHD children performing three response inhibition tasks. The only clear-cut difference between our FTT and control groups was in the slow wave to false targets at frontal sites. The LO and HI IQ groups had similarly less negative going components than the FTT group. Although this late slow wave has been little studied, the available evidence suggests it may index post-response evaluation. Wijker (1991) found a negative slow wave in 5-year old children that decreased in amplitude between 7 and 9 years of age. He suggested that since older children can process information more efficiently than younger children, fewer resources are needed by the older ones in the postresponse evaluation processing period, resulting in a lower amplitude negative slow wave. Lavoie, Robaey, Stauder, and associates (1998) analyzed the negative slow wave in 5year-old children born prematurely. These investigators used a visual odd ball task and required a motor response both to the frequent and infrequent (25 percent) stimuli. The preterm boys had a larger frontal slow wave than the preterm girls and a smaller P300. These investigators suggested that the premature boys had less efficient pre-response processing and thus required more resources in the post-response evaluation, reflected by a larger negative going frontal slow wave. Evidence suggests that premature boys are less able than premature girls to recover from early cerebral insult (Raz, Lauterbach, Hopkins et al., 1995). Our cued CPT task and the Lavoie et al. (1998) odd ball task have the related frequent/ infrequent feature but obviously differ considerably otherwise. Also, we did not find a group difference for the P300 component, nor did the FTT and LO IQ groups differ in reaction times or number of errors. Thus, we cannot make the same argument as Lavoie and associates that the negative slow wave was more pronounced in the clinical group because their pre-response processing was less efficient than in the LO IQ control group. The more pronounced and quicker developing negative going frontal slow wave to false targets in the FTT group may reflect simply a faster return to baseline. Or the FTT group could be less secure in the decision to inhibit a response and thus have needed to engage in further evaluation. Behavioral ratings on the FTT children (reported in Dykman et al., in press) revealed the group to have abnormally high scores on the externalizing factor of the Child Behavior Checklist, which assesses problems with attention, aggression, and delinquency. Other investigators have also reported attention and behavioral conformity prob-
296
DYKMAN ET AL.
lems in formerly FTT children (Galler & Ramsey, 1989; Oates et al., 1985). Recall that NoGo (false target) stimuli have been shown in brain imaging studies to activate right frontal sources (Haeger et al., 1998; Strik et al., 1998). These imaging studies concur with brain lesion studies in linking the right frontal cortex to response inhibition. Note that as early as the P200 peak, the two control groups exhibited right frontal lateralization whereas the FTT group did not (see Table 3). However, this right lateralization for P200 was as strong to targets as false targets; and for N300, we found F3>F4 (left lateralization) in the HI IQ group, more so for targets than false targets. Thus, the present results are not in complete agreement with prior related studies. The most conservative interpretation of the frontal slow wave difference between our FTT and control groups is that the more pronounced late negativity in the FTT group is either indicative of brain immaturity as theorized by Wijker (1991), or additional postresponse processing due to greater uncertainty regarding the false targets (Lavoie, 1998). It is tempting to link this finding to impaired inhibition in the FTT group, given their adverse scores on the externalizing scale of the Child Behavior Checklist. Acknowledgement This research was supported by a grant from the United States Department of Agriculture
References Achenbach, T.M. (1991). Manual for the Child Behavior Checklist~4-18 and 1991 Profile. Burlington, VT: University of Vermont, Dept. of Psychiatry. Barnet, A.B., Weiss, I.P., Sotillo, M.V., Ohlrich, E.S., Shkurovich, M., & Cravioto, J. (1978). Abnormal auditory evoked potentials in early infancy malnutrition. Science, 201,450-452. Casey, B.J., Castellanos, F.X., Giedd, J.M., Marsh, W.L., Hamburger, S.D., Schubert, A.B., Vauss, Y.C., Vaituzis, A.C., Dickstein, D.P., Sarfatti, S.E., & Rapoport, J.L. (1997). Implication of right frontostriatal circuitry in response inhibition and attention-deficit/hyperactivity disorder. Journal of American Academy of Child and Adolescent Psychiatry, 36,374-383. Casey, P.H. (1992). Failure to thrive. In M. D. Levine, W. B. Carey, & A. C. Crocker (Eds.), Developmental Behavioral Pediatrics. (pp. 379-383). Philadelphia, PA: W.B. Saunders Co. Cravioto, J., & Arrieta, R. (1983). Malnutrition in childhood. In M. Rutter (Ed.), Developmental Neuropsychiatry. (pp. 32-51). New York, NY: Guilford. Dawson, G., & Fischer, K.W. (1994). Human behavior and the developing brain. New York, NY: Guilford. Dobbing, J (1976). The biology of human fetal growth. London: Taylor & Francis. Drotar, D., & Sturm, L. (1988). Prediction of intellectual development in young children with early histories of non-organic failure-to-thrive. Journal of Pediatric Psychology, 13, 281-296. Dykman, R.A., Casey, P.H., Ackerman, P.T., & McPherson, W.B. Behavioral and cognitive status in schoolaged children with a history offailure to thrive during early childhood. (in press). Eimer, M. (1993). Effects of attention and stimulus probability on ERPs in a Go/NoGo task. Biological Psychology, 35, 123-138. Fallgatter, A.J., Wiesbeck, G.A., Weijers, H.G., Boening, J., & Strik, W.K. (1998). Event-related correlates of response suppression as indicators of novelty seeking in alcoholics. Alcohol & Alcoholism, 33,475-481. Flinn, J.M., Barnet, A.B., Lydick, S., & Lackner, J. (1993). Infant malnutrition affects cortical auditory evoked potentials. Perceptual and Motor Skills, 76, 1359-1362. Galler, J.R., Ramsey, F.C., Forde, V., Salt, P., & Archer, E. (1987). Long term effects of early kwashiorkor compared with marasmus. Journal of Pediatric Gastroenterology and Nutrition, 6, 847-854. Galler, J.R., & Ramsey, F. (1989). A follow-up study of the influence of early malnutrition on development: Behavior at home and at school. Journal of the American Academy of Child and Adolescent Psychiatry, 28, 254-261. Haeger, F., Volz, H.-P., Gaser, C., Mentzel, H.-J., Kaiser, W.A., & Sauer, H. (1998). Challenging the anterior attentional system with a continuous performance task: A functional magnetic resonance imaging approach.
ELECTROPHYSIOLOGICAL STUDY OF SCHOOL-AGED CHILDREN
297
European Archives of Psychiatry & Clinical Neuroscience, 248, 161-170. Holcomb, P.J., Ackerman, P.T., & Dykman, R.A. (1985). Cognitive event-related brain potentials in children with attention and reading deficits. Psychophysiology, 22,656-666. Jodo, E., & Kayama, Y. (1992). Relation of negative ERP component to response inhibition in a Go/No-go task. Electroencephalography & Clinical Neurophysiology, 82,477-482. Johnson, R.,& Donchin, E. (1978). On how P300 amplitude varies with the utility of the eliciting stimuli. Electroencephalography & Clinical Neurophysiology, 44,424-437. Kutas, M., & Hillyard, S.A. (1980). Reading senseless sentences: Brain potentials reflect semantic incongruity. Science, 207, 203-205. Kutas, M., & Iragui, V. (1998). The N400 in a semantic categorization task across 6 decades. Electroenc ephalography & Clinical Neurophysiology, 108, 456-471. Lavoie, M.E., Robaey, P., Stauder, J.E.A., Glorieux, J . , & Lefebvre, F. (1998). Extreme prematurity in healthy 5-year-old children: A re-analysis of sex effects on event-related brain activity. Psychophysiology, 35,679689. McCall, R.B., Hogarty, P.S., & Hulbert, N. (1972). Transitions in infant sensorimotor development and the prediction of childhood IQ. American Psychologist, 27, 728-748. Newton, J.E.O., Oglesby, M., Ackerman, P.T., & Dykman, R.A. (1994). Visual slow brain potentials in children with attention deficit disorder. Integrative Physiological & Behavioral Science, 29, 39-54. Oates, R.K., Peacock, A., & Forrest, D. (1985). Long-term effects of nonorganic failure to thrive. Pediatrics, 75, 36 Overtoom, C.C.E., Verbaten, M.N., Kemner, C., Kenemans, J.L., van Engeland, H., Buitelaar, J.K., Camfferman, G., & Koelega, H. (1998). Associations between event-related potentials and measures of attention and inhibition in the continuous performance task in children with ADHD and normal controls. Journal of American Academy of Child and Adolescent Psychiatry, 37, 977-985. Pennington, B.F. (1993). The working memory function of the prefrontal cortices: Implications for development and individual differences in cognition. In M.Haith et al. (Ed.), Future oriented processes in development. Chicago, IL: University of Chicago Press. Pollitt, E. (1995). The relationship between nutrition and behavioral development in children. Journal of Nutrition, 125, 2211 S-2284S. Raz, S., Lauterbach, M.D., Hopkins, T.L, Glogowski, B.K., Porter, C.L., Riggs, W.W., & Sander, C.J. (1995). A female advantage in cognitive recovery from early cerebral insult. Developmental Psychology, 31,958966. Skuse, D., Pickles, A. , Wolke, D., & Reilly, S. (1994). Postnatal growth and mental development: Evidence for a "sensitive period". Journal of Child Psychology and Psychiatry, 35, 521-545. Strik, W.K., Fallgatter, A.J., Brandeis, D., & Pascual-Marqui, R.D. (1998). Three-dimensional tomography of event-related potentials during response inhibition: Evidence for phasic frontal lobe activation.
Electroencephalography & Clinical Neurophysiology, 108,406--413. Wijker, W. (1991). ERP ontogenesis in childhood. University of Amsterdam: Unpublished doctoral dissertation.