- Email: [email protected]
EEG findings in fetal alcohol syndrome and Down syndrome children
ELSEVIER
Q Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
EEG findings in fetal alcohol syndrome and Down syndrome children W.M. Kaneko a,*, E.L. Phillips a, E.P. Riley b, C.L. Ehlers
a
a Department ofNeuropharmacology, CVN-14, The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, CA 92037, USA b Department of Psychology, San Diego State University, San Diego, CA 92182-0350, USA Accepted for publication: 19 August 1995
Abstract
Results from previous studies evaluating the electroencephalograms (EEGs) of infants born to alcoholic mothers suggest that the neonatal EEG may be a sensitive measure of prenatal ethanol exposure. Few studies, however, have examined EEG records of adolescent children with fetal alcohol syndrome (FAS). The present study investigated the resting EEG recordings of 18 matched triads of FAS, Down syndrome, and normal control subjects. Significant reductions in mean power of the alpha frequencies (7.5-12 Hz) were seen for both clinical groups, however, each syndrome appeared to have distinct EEG spectral distributions. Down syndrome children overall had diffuse EEG slowing while the EEG records of the FAS children showed reduced power, particularly in the alpha frequencies in the absence of significant slow activity. In the Down syndrome children, significant decreases in alpha power was seen in posterior cortical regions, whereas FAS children were more affected in the left hemisphere. This study suggests that certain EEG variables may be helpful in characterizing the neurophysiology of FAS. Keywords: EEG; Fetal alcohol syndrome (FAS); Alpha activity; Slow waves
1. Introduction It has been suggested that, even in the absence of dysmorphology, neonatal EEG may be the most sensitive indicator of prenatal alcohol toxicity (Chernick et al., 1983; Ioffe et al., 1984). Abnormalities observed in sleep EEG of 3-day-old infants exposed to alcohol prenatally have been found to persist as long as 6 weeks after birth (Havlicek et al., 1977). This suggests that such EEG effects do not represent alcohol withdrawal and may reflect more permanent brain changes. More evidence that EEG recordings may be helpful for early diagnosis of alcohol effects on brain development is provided by Ioffe and Chernick (1990) who found that the degree of abnormality in the EEG power at birth correlated with subsequent motor and mental development. In that study, increases in power during REM sleep were associated with an adverse outcome in motor development, whereas increases in power during quiet sleep were related to an adverse outcome in mental development.
* Corresponding author. Tel.: (619) 554-7236; Fax: (619) 554-6494.
Few studies have evaluated EEGs in adolescents with fetal alcohol syndrome (FAS). In a preliminary report of two children with severe FAS (ages 13 and 14 years), we found moderate abnormalities of the awake EEG (Mattson et al., 1992). Theta was found to be the dominant EEG rhythm, with occasional delta activity predominantly in the posterior regions and missing or sparse 8 Hz alpha activity in the posterior head regions. No focal abnormalities or gross asymmetries were found in those subjects. To date, no studies have been conducted which performed spectral analyses of awake EEG in school-age children diagnosed as having FAS, especially as compared to other children with cognitive impairments, such as Down syndrome children. Down syndrome and FAS children have several common characteristics. These children are commonly cognitively impaired, growth retarded, hyperactive and have attention deficits and craniofacial dysmorphologies. While Down syndrome can be easily diagnosed with a chromosome assay, the diagnosis of FAS is more difficult. Thus, there is a need for more clinically discriminative indicators of FAS. The aim of the present study was to explore EEG variables in children with FAS. If the EEG of FAS chil-
0013-4694/96/$15.00 © 1996 Elsevier Science Ireland Ltd. All rights reserved SSDI 0 0 1 3 - 4 6 9 4 ( 9 5 ) 0 0 1 8 9 - 1
EEG 94168
W.M. Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
dren could be characterized distinctly from the EEG records of other patients with a congenital disorder which also included cognitive disabilities, craniofacial dysmorphology, and growth retardation, then the specific EEG variables may correctly categorize children into either group. Therefore, the purpose of this study was (1) to determine whether EEG abnormalities found in infants diagnosed with FAS are similar to those of school-age FAS children; (2) to characterize and evaluate the EEG recordings of children with FAS through spectral analysis; (3) to compare FAS subjects with children with Down syndrome; and (4) to compare these findings with normal controls matched for age, sex and ethnicity.
2. Methods 2.1. Subjects
Subjects were 18 triads of children who were matched on sex, age and, for the most part, ethnicity. The first group (n = 18) consisted of children diagnosed as having FAS by Dr. Kenneth Jones. Children in the"second group (n = 18), diagnosed as having Down syndrome, were referrals from the Down Syndrome Association of San Diego. Children in the third group (n = 18) were normal controls with no family history of neurological disorders and no history of exposure to alcohol or any other drug prenatally. A total of 54 children between the ages of 4 and 15 years (average age of 9.1 ___3.2 years) were evaluated. The 18 children in the FAS group consisted of 10 females and 8 males from various ethnic backgrounds: 9 Caucasians, 7 African Americans, 1 La.tino, and 1 of mixed Latino/ Caucasian parentage. The parent or legal guardian of each subject completed questionnaires regarding their child's demographics (e.g., ethnic background, age, gender, etc.) and medical history (e.g., hearing or visual complications, medications, handedness, etc.). Each parent was paid $35.00 for his/her child's participation, the children received small toys or stickers. 2.2. IQ assessments
Subjects were scheduled for an appointment at the Center for Behavioral Teratology at San Diego State University for intelligence testing. Full Scale IQ, Verbal IQ, and Performance IQ scores were obtained using the Wechsler Intelligence Scale for Children-Revised version (WlSC-R) or the Wechsler Preschool and Primary Scale of Intelligence (WPPSI), depending on age and ability of the individual. The subjects were given an additional $10.00 or a toy of their choice for participating. 2.3. Electrophysiological recordings
Electroencephalographic (EEG) recordings were obtained from all subjects between 8:00 a.m. and 4:00 p.m.
21
by a registered EEG technician at the General Clinical Research Center in the Green Hospital of the Scripps Clinic in La Jolla, CA. On the test date, procedures were fully explained to both parent/guardian and subject. Parents or guardians were required to sign a consent form before any procedures were performed. Individual goldplated electrodes or tin electrodes set inside a tightly fitting electrode cap were used for recordings. The placement of the electrodes was based on the international 10-20 system (Jasper, 1958). All electrode impedances were below 5 kO. Six channels of EEG data (F3-C3, C3-P3, P3-O1, F4-C4, C4-P4, and P4-O2) were collected. The subjects were instructed to lie comfortably, remain quiet and relaxed, and informed that it was important to be awake and have their eyes closed during the entire recording period. Subjects who could not close their eyes on instruction or had excessive eye movement were assisted in closing their eyes by gentle application of the hands or application of soft eye patches. EEG was recorded on paper for 5-10 min, using a Nihon-Kohden polygraph and on magnetic tape (Vetter Model D) for off-line analysis. The bandpass was 1-70 Hz and the sensitivity was 7 /zV/mm. Paper records were monitored during recording for signs of drowsiness and for any eye or movement artifacts, All paper records were visually analyzed, scored for abnormalities and rated for amplitude and wave form or morphological characteristics. Artifact-free EEG was then selected for computer analysis and cortical EEG variables of awake recordings were examined. 2.4. Data analysis
The EEG samples were digitized at a rate of 128 Hz (samples/sec) and the Fourier transforms of consecutive 4 sec epochs for each channel were calculated. The power spectrum was produced using a Macintosh computer and software developed by Ehlers and Havstad (1982). For statistical analyses, the transformed data were compressed into 8 frequency bands: 0.5-1 Hz; 1-2 Hz; 2-4 Hz; 4-7.5 Hz; 7.5-9 Hz; 9-12 Hz; 12-20 Hz; and 20-40 Hz. Mean spectral power density (/~V2/octave) over each band frequency was calculated by summing the raw power spectral values within the band, multiplying by a scale factor derived from the calibration signal to produce the total power in the band in microvolts squared, and dividing by the width of the band in octaves. This width is the logarithm of the ratio of the maximum and minimum frequencies in the band, divided by the log of the two. In addition to mean spectral power, peak frequency for each of the same frequency bands was also estimated. Peak frequency is the frequency at which power peaks within a certain frequency range. For each EEG frequency range at each recording site, 2-way between-subjects analyses of variance (ANOVAs), with 2 levels for sex (i.e., female and male) and 3 levels
W.M. Kaneko et al. / Electroencephalographyand clinical Neurophysiology 98 (1996) 20-28
22
for group (i.e., control, Down syndrome, FAS), were used to statistically evaluate each of the dependent measures (i.e., mean spectral power and peak frequency). Tukey HSD tests were utilized for post hoc analyses. Rather than use the traditional 0.05 level of significance, a more conservative 0.01 level was used because of the multiple comparisons defined and tested in this study.
3. R e s u l t s
3.1. Group characteristics There were no significant differences in age between the 3 groups of children. The average age of both the F A S group and the Down syndrome group was 9.92 years, while the average age of the normal controls was 9.88 years. IQ scores for 12 of the 16 Down syndrome children tested could not be calculated due to poor cooperation a n d / o r severe mental impairment. Their Full Scale IQ scores were below 40. For the 4 testable Down syndrome children, the Full Scale IQ ranged from 43 to 47. IQ scores for 16 of the 18 F A S children ranged from 41 to 96, with a mean Full Scale IQ score of 74 + 16.0, which was significantly lower than the 16 matched normal controls who had IQs that ranged from 86 to 137, with a mean Full Scale IQ score of 109 -1- 13. Several children from the two clinical groups were taking medication at the time of testing. Three Down syndrome children were taking methylphenidate, 3 F A S
children were taking methylphenidate or cylert, and 4 other F A S subjects were on other drugs: two on anticonvulsants, one on phenothiazines, and one on a small dose of imipramine at bedtime (see Table 1).
3.2. Overall description of E E G records Although subjects were instructed to keep their eyes closed during the recording, 7 of the 18 Down syndrome children were unable to do so. Therefore, their recordings were obtained while their eyes remained open. Samples of actual EEG records of a representative triad (i.e., control, Down syndrome, F A S ) are presented in Fig. 1. The spectral plots for the same subjects can be found in Fig. 2. All controls were clinically rated as normal, whereas 12 out of 18 EEG records of the children with Down syndrome were interpreted as clinically abnormal, and 9 of the 18 EEG records of the F A S subjects had an overall rating o f borderline or abnormal (see Table 2). In general, they were found to be immature a n d / o r poorly developed. Low voltage or amplitude was commonly seen in the EEG records of F A S children, while slower activity (i.e., delta and theta waves) was dominant in the E E G records of Down syndrome children. Six F A S children and 10 children with Down syndrome had excessive generalized slowing. Other abnormalities noted in the cognitively impaired children were poor wave formation, focal spikes, and asymmetrical EEG development. Table 2 shows clinical ratings and descriptions of the E E G records for all children evaluated.
Table 1 Demographic information for all subjects Gender Age F M F F F M F F F F M F F M M
M M M
Control
(years)
IQ
4 5 6 6 7 7 8 9 9 9 10 11 11 12 13 14 14 15
111 115 112 100 94 109 117 135 116 137 128 91 109 n/a 122 86 103 101
Non-ethnicmatch
Caucasian
n/a, not available. * Mellaril, Phenobarbital, Dalmane.
FAS
Down syndrome Medication IQ < 40 < 40 45 < 40 < 40 < 40 < 40 < 40 43 < 40 < 40 < 40 n/a < 40
Non-ethnicmatch Caucasian Caucasian Caucasian Caucasian Samoan Caucasian Hispanic/Caucasian Caucasian
n/a
47 < 40 43
Medication IQ
Caucasian Caucasian
Ritalin Ritalin
Ritalin
78 90 73 74 56 96 69 74 94 < 40 78 91 52 79 51 41 74 87
Non-ethnicmatch Hispanic/Caucasian Caucasian Caucasian Afro-American Afro-American Caucasian Afro-American Caucasian Caucasian Caucasian Hispanic Afro-American Afro-American Caucasian Caucasian Afro-American Afro-American Caucasian
Medication
Mellaril
Imipramine Epitol Ritalin Ritalin * Cylert
W.M. Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
3.3. Delta and them activ#y
greater mean power in the delta frequencies (1-4 Hz) in the fronto-central (F3-C3, F4-C4) and centro-parietal (C3P3, C4-P4) derivations when compared to the control and FAS groups. Differences seen between the 3 study groups on mean EEG power for all frequency bands in the centro-parietal cortex can be seen in Fig. 3. Mean power for theta activity (4-7.5 Hz) in the F3-C3 and F4-C4 leads was also found to be significantly higher for children with Down syndrome as compared to children
Spectral analyses revealed significant differences between the 3 groups of children for mean EEG power and peak frequency in all cortical recording sites. The mean EEG power of low frequency delta (0.5-1 Hz) for Down syndrome children was :significantly greater than for matched normal controls in all 6 cortical recording sites. The Down syndrome group demonstrated significantly
'
•
,
i
.i~
I
|,
" :
I,
"
i
,
;
!
!
o Co.P3
0 E4 Z 0 L)
F
:
, i,
23
."
!
. .~
.
.
i
~
.
•
.
.
=;~
~
!
.
,
~ i
i
'
0
c~ C~ Z >4 EQ Z
0
•
I
~
i.
i,
i
i l
i
'i
i
i
I
'i
.+
i
r~
,,f ¢
I
I
I
i ,~
c4 ,V
¢ i
~
d ~,
i
i
I I
1 sec
I
50~V
Fig. 1. Samples of EEG record,; of three 9-year-old females, a normal control (top panel), Down syndrome (middle panel), and FAS (bottom panel).
4 5
6
6 7
7 8 9
9
9
10 11 11
12 13
14
14
15
F
F F
M F F
F
F
M F F
M M
M
M
M
Age (years)
F M
Gender
9 Hz alpha
11 Hz alpha
9-10 I-Iz alpha
9-10 Hz alpha 8-9 Hz alpha
8 Hz alpha 8 Hz alpha 8-9 Hz alpha
8-9 Hz alpha
9 Hz alpha
9 Hz alpha 9 Hz alpha 9-10 Hz alpha
8 Hz alpha 8 Hz alpha
9 Hz alpha
Normal
Normal
Normal
Normal Normal
Normal Normal Normal
Normal
Normal
Normal Normal Normal
Normal Normal
Normal
Normal Normal
5-6 I-Lztheta
5 - 6 I-Iz theta
8-9 Hz alpha
9 Hz alpha 8-9 Hz alpha
1 1 Hz alpha 9 Hz alpha 8-9 Hz alpha
2 Hz delta and 5 Hz theta 6-7 Hz theta
7-8 I-lz theta 4-5 Hz theta 4-5 Hz theta
5 - 6 Hz theta 4-5 Hz theta
6 - 7 Hz theta
6-7 Hz theta 5 Hz theta
Dominant rhythm
8 Hz alpha 8-9 Hz alpha
Down syndrome
Dominant rhythm Characteristic feature
Control
Table 2 Clinical ratings of EEG records
Normal Normal Essentially normal; poorly developed; left hemisphere poorly developed; frontal beta Normal Essentially normal; poorly developed; left hemisphere poorly developed Essentially normal; poorly developed; frontal beta Generalized slowing with excess movement artifact Generalized slowing with sparse alpha present
Generalized slowing; frontal beta Asymmetrical; left focal spike, left central phase reversal at C3 Poorly developed; generalized slowing; frontal beta Generalized slowing; frontal beta Poorly developed; generalized slowing Some slowing Generalized slowing Poorly developed; generalized slowing; frontal beta Poorly developed; generalized slowing; frontal beta Generalized slowing, frontal beta
Characteristic feature
8 Hz alpha
9 Hz alpha
4 - 6 Hz theta
10-11 Hz alpha 5 - 6 Hz theta
8 Hz alpha 8 - 9 Hz alpha 8 - 9 Hz alpha
8 141 alpha
8 Hz alpha
7-8 Hz theta 8-9 Hz alpha 6 - 7 l-Iz theta
5 - 6 Hz theta 4 - 6 Hz theta
5-6 Hz theta
7 - 8 H.z theta 7 - 8 Hz theta
Dominant rhythm
FAS
Poorly developed with mild transient theta
Poorly developed; generalize slowing Normal
Normal Generalized slowing; frontal beta
Essentially normal; poorly developed; frontal beta Normal Normal Normal
Generalized slowing with sparse alpha present Generalized slowing Generalized slowing with frontal beta Poorly developed; some slowing Normal Generalized slowing; sparse alpha; left hemisphere poorly developed Normal
Normal Some slowing
Characteristic feature
I t~ oo
t-o
e~
t~
t~
e,
W.h£ Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
/
-"N.¢IIO
/Ira
t, -'m.lmm
~ ,
_
,
~i
"
"mlsm
,' .
"m~m c
25
,
"n~m
~
'
iNl131
"mllm "I21m
r' " m m m ' 2
sm I,
CONTROL
DOWN
SYNDROME
FAS
Fig. 2. EEG spectral plots from the centro-parietal lead of three 9-year-old girls, a normal control (left panel), Down syndrome (middle panel), and FAS
(right panel)• The x-axis representsfrequency(Hz), the y-axisreflectstime (min), and the z-axis representsEEG power(/xV2/octave).
with FAS. The frequency at which power peaked in the theta frequency range was significantly lower for the Down syndrome children than for the controls. This decrease in
--~
LEFT
CONTROL
~
~ •
S'n~ROME
•
FAS p~.O06
•"
t~<.oos
(DO~
v~.
peak frequency was present in all cortical regions. Compared to the FAS group, significantly lower peak frequency of theta activity was also found in the Down syndrome group, however, this was only present in the F4-C4 lead. Differences between the 3 groups on peak frequency of theta activity in the centro-parietal region are demonstrated in Fig. 4.
FAS)
3.4. Alpha and beta activity o~o
~"
>
50
.5-i
I-2
2-4
4-7,5
7.5-9
9-12
12-20
20-40
RIGHT 150
! o~ 100
0
.s-1
1-2
2-4
,-;.~
FREQUENCY
7.;-9 BAND
9-12
12-20 20-40
(Hz)
Fig. 3. M e a n + S.E.M. E E G spe~ral power in the centro-parietalcortical
lead from 18 normal control, 18 Down syndrome and 18 FAS children across all frequency bands.
FAS children showed significantly lower mean power in the alpha (9-12 Hz) frequency range for the left frontocentral (F3-C3) and left parieto-occipital (P3-O1) leads when compared to normal controls• The Down syndrome group displayed significantly lower mean power in the 7.5-9 Hz alpha frequency range in the left posterior lead (P3-O1) when compared to the normal control group and in the 9-12 Hz alpha range for the right posterior lead (P4-O2). For Down syndrome children, the frequency at which power peaked in the slow alpha frequency range (7.5-9 Hz) in all cortical regions was significantly lower than that of normal controls. FAS children also had significantly lower peak frequencies in the slow alpha range compared to controls; however, this decrease was present only in the right central lead (C4-P4).: 'Differences between the 3 groups on peak frequency Of alpha activity in the centroparietal region are demonstrated in Fig. 4. Mean power in the beta frequency (12-20 Hz) range for the centro-parietal regions of the Down syndrome children were significantly higher when compared to FAS children.
W.M. Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
26 G oo 7
THETA ACTIVITY (4-7.5 Hz)
D co~o~ [] D O W N s~n~D~OME • FAS
I
iiiiiiiiiiiiiiiiiiiii
~!ii!iii!!i~!ii
!i!iiii!!ii!
4. Discussion
ALPHA ACTIVITY (7.5-9 HZ)
8.25 "
8. O0
-
-
~
-
-
LEFT
RIGHT CENTRAL-PARIETAL
AREA
Fig. 4. Mean + S.E.M. peak frequency of the theta (top graph) and alpha (lower graph) ranges from the centro-parietal lead in 18 normal control, 18 Down syndrome and 18 FAS children.
3.5. Discriminant function analysis Eighteen matched triads, for a total of 54 subjects, were used for the DFA. Each of the 3 variables, mean power of alpha of the left parieto-occipital lead, mean power of theta
Table 3 Values and results from the discriminant function analysis (DFA) using EEG variables (18 control/Down syndrome/FAS matched triads) Factor 1 2
Canonical correlation 0.705 0.313
Proportion of variation 0.49 0.10
Wilks' lambda = 0.454 Group
Control Down syndrome FAS Predictor variable Left parieto-occipital mean power of alpha Right fronto-central mean power of theta Rigt fronto-central theta peak frequency Wilks' lambda
Classification and number of cases (%) Correct
Incorrect
10 (56) 13 (72) 10 (56)
8 (44) 5 (28) 8 (44)
df
of the right fronto-central lead, and theta peak frequency of the right fronto-central lead, significantly differentiated the 3 groups ( F (6, 98) = 7.91, P < 0.0005). DFA values are presented in Table 3. Fifty-six percent (10/18) of the normal controls, 72% (13/18) of the Down syndrome children and 56% (10/18) of the children with FAS were correctly classified by the DFA. These percentages are greater than the percentage of correct predictions expected on the basis of chance.
F
P
2, 51
6.806
0.002
2, 51
6.224
0.004
2, 51
13.581
0.0005
6, 98
7.914
0.0005
This study characterized EEG variables in FAS children and compared them to Down syndrome children. A putative EEG spectral pattern of children with FAS was revealed, which overall consisted of reduced alpha activity in the absence of significant EEG slowing. Mean power for theta (4-7.5 Hz) activity in FAS children did not differ from matched normal controls, however, the Down syndrome group was found to have increased mean theta power compared to FAS children in the fronto-central cortical regions. Increases in mean power in theta frequencies in the right fronto-central cortical area were also noted in Down syndrome children as compared to normal controls. Children with Down syndrome were also found to have slower theta activity compared to normal controls, as revealed by the lower peak frequency values. These results are consistent with previous findings stating that the most common EEG abnormality found in infants and children with Down syndrome is excessive theta and slower activity (Clausen et al., 1977; Tangye, 1979; Schmid et al., 1985; Epstein, 1986; Devinsky et al., 1990). Increases in slow EEG activity have been related to slow maturation of the brain (Ellingson et al., 1973). In Down syndrome patients, slowing is more common in young children, however, it is rare during infancy. The frequency of abnormal EEGs found in Down syndrome children appears to increase with age, when full maturity of the brain is achieved (Sepp~il~iinen and Kivalo, 1967; Ellingson et al., 1970, 1973; Ellingson and Peters, 1980; Schmid et al., 1985). The EEGs of the Down syndrome subjects may have a different developmental course than those of controls. Developmental aspects of the EEG remain unknown in FAS. The present study demonstrates that alpha activity (7.512 Hz) was significantly reduced in FAS children in the absence of EEG slowing. This effect was significant in left cortical regions, F3-C3 and P3-O1. Mean power for the alpha frequency ranges was also significantly decreased in Down syndrome subjects as compared to the controls. However, this reduction in alpha was present only in posterior areas, P3-O1 and P4-O2. These data suggest that the left hemisphere may be vulnerable to prenatal alcohol exposure while the posterior regions may be particularly
W.M. Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
affected in Down syndrome children. Slower alpha frequencies for all recording sites were also seen in children with Down syndrome as compared to normal controls. FAS children, however, showed slower alpha activity compared to normal controls only in the right parieto-occipital region. This suggests that the peak frequency of the alpha activity in certain cortical regions may also differentiate the two clinical groups in this study. Case studies on two of the patients in the present study using MRI techniques have been conducted (Mattson et al., 1992). These two chi]ldren were found to have decreased overall cranial and cerebellar volumes and a decrease in the proportional volume of the basal ganglia. No differences in cortical volume or subcortical fluid volumes, cortical grey matter, limbic and non-limbic cortex, or diencephalic structures were found between the two children prenatally exposed to alcohol and the controls. Other studies have reported neuropathological conditions in infants exposed to alcohol prenatally. Errors in migration of neuronal and glial elements may be one source of the brain malformations seen (Clarren et al., 1978). There is great variability in the nature and severity of the brain abnormalities in children with FAS. Our finding that the left hemisphere is more affected by alcohol than the right hemisphere is interesting since asymmetrical effects are present in the photo of the brain :from an FAS infant shown by Clarren et al. (1978). However, further studies of larger subject populations will be necessary to make appropriate correlations. A negative correlation between mental maturation and alpha frequency has been suggested in some population studies (Keezer, 1939; Fillingson and Lathrop, 1973; Schmid et al., 1985). Although there was a significant IQ difference between control,; and FAS children, no significant correlations between Full Scale IQ and mean power of alpha activity or between Full Scale IQ and peak frequency of alpha activity were observed at the P > 0.01 level. The lack of correlation may be related to the wide age span of the subject sample used in this study. Seven of the 18 Down syndrome children in the present study had their eyes open during the majority of the recording session. During "eyes open" conditions, it is the dominant rhythm that is attenuated. An evaluation of the raw data in relation to eye status and dominant rhythm revealed that all but one of the "eyes open" Down syndrome subjects had theta activity as the dominant rhythm. In addition, in the children whose eyes remained mainly open during the EEG recording, alpha frequencies were not necessarily present when they closed their eyes. This suggests that the significant decreases in alpha activity observed in Down syndrome subjects may not be explained entirely by the "eyes open" status differences among the groups. Reductions in alpha may also be caused by increased coherence between two electrodes in a bipolar montage. Further studies will require other montages to investigate this issue.
27
In the present study, 10 of the 36 children from the clinical groups were not drug free, and several children were taking similar drugs. Specific effects on on-going EEG have been associated with certain medication (Fink, 1968; Shagass, 1968; Kiloh et al., 1981). Methylphenidate may produce some decreases in alpha, whereas neuroleptics and antidepressants, such as Mellaril and imipramine, may increase the amount of alpha activity, while sedatives or tranquilizers may lead to an increase in fast activity (20-30 Hz). Greater amounts of beta activity were found in the Down syndrome children when compared to FAS children. This is interesting since only three of these children were on methylphenidate and none were on antiepileptics which can also increase beta. However, no clear correlations were found between medication status and EEG beta as quantitated with spectral analyses. Therefore, this increase in beta activity appears to be related to diagnosis and consistent with previous findings (Schmid et al., 1985; Devinsky et al., 1990). The present study offers new data on resting EEGs in school-age children diagnosed with FAS. The EEGs of FAS children were characterized and differentiated from another cognitively disabled group. In summary, Down syndrome children displayed excessive amounts of delta, theta and beta activity as compared to FAS children and controls, whereas both clinical groups had less alpha and lower alpha frequencies. Our findings of slower alpha activity in both clinical groups support previous hypotheses of cerebral immaturity in Down syndrome children and suggest cerebral immaturity in those prenatally exposed to alcohol. However, unlike the children with Down syndrome, the reduction in alpha activity seen in the FAS children is in the absence of overall EEG slowing. Significant reductions in alpha mean power were observed in both clinical groups; however, this abnormality was found to be distributed in posterior cortical regions of children with Down syndrome and generally found in the left hemisphere of FAS children. Thus, differences in EEG abnormality type and the cortical distribution of the abnormalities may differentiate these two syndromes. Future studies utilizing a larger electrode array may further contribute to the understanding of the neurophysiology of these two syndromes.
Acknowledgements This study was supported by NIAAA 00098 to Cindy L. Ehlers, Alcohol Research Center Grant 06420 to Floyd E. Bloom and General Clinical Research Center (GCRC) Grant 00833. The authors would like to thank Dr. Sarah Mattson for aiding in the estimation of IQs, Dr. Kenneth Lyons Jones for providing patients, and Dr. James Havstad for writing the computer programs.
28
W.M. Kaneko et aL / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
References Chernick, V., Childiaeva, R. and Ioffe, S. Effects of maternal alcohol intake and smoking on neonatal electroencephalogram and anthropometric measurements. Am. J. Obstet. Gynecol., 1983, 146: 41-47. Clarren, S.K., Alvord, E.C., Sumi, S.M., Streissguth, A.P. and Smith, D.W. Brain malformations related to prenatal exposure to ethanol. J. Pediat., 1978, 92: 64-67. Ctausen, J., Sersen, E.A. and Lidsky, A. Sleep patterns in mental retardation: Down's syndrome. Electroenceph. clin. Neurophysiol., 1977, 43: 183-191. Devinsky, O., Sato, S., Conwit, R.A. and Schapiro, M.B. Relation of EEG alpha background to cognitive function, brain atrophy, and cerebral metabolism in Down's syndrome: age-specific changes. Arch. Neurol., 1990, 47: 58-62. Ehlers, C.L. and Havstad, J.W. Characterization of drug effects on the EEG by power spectral band time series analysis. Psychopharmacol. Bull., 1982, 18: 43-47. Ellingson, R.J. and Lathrop, G.H. Intelligence and frequency of the alpha rhythm. Am. J. Ment. Defic., 1973, 78: 334-338. Ellingson, R.J. and Peters, J.F. Development of EEG and daytime sleep patterns in Trisomy-21 infants during the first year of life: longitudinal observations. Electroenceph. clin. Neurophysiol., 1980, 50: 457466. Ellingson, R.J., Menolascino, F.J. and Eisen, J.D. Clinical-EEG relationships in mongoloids confirmed by karyotype. Am. J. Ment. Defic., 1970, 74: 645-650. Ellingson, R.J., Eisen, J.D. and Ottersberg, G. Clinical electroencephalographic observations on institutionalized mongoloids confirmed by karyotype. Electroenceph. clin. Neurophysiol., 1973, 34: 193-196. Epstein, C.J. The Neurobiology of Down Syndrome. Raven Press, New York, 1986. Fink, M. EEG classification of psychoactive compounds in man: review and theory of behavioral associations. In: D.H. Efron (Ed.), Psy-
chopharmacology: a Review of Progress. Raven Press, New York, 1968: 497-508. Havlicek, V., Childiaeva, R. and Chernick, V. EEG frequency spectrum characteristics of sleep states in infants of alcoholic mothers. Neuropediatrics, 1977, 8: 360-373. Ioffe, S. and Chernick, V. Prediction of subsequent motor and mental retardation in newborn infants exposed to alcohol in utero by computerized EEG analysis. Neuropediatrics, 1990, 21: 11-17. Ioffe, S., Childiaeva, M.D. and Chernick, M.D. Prolonged effects of maternal alcohol ingestion on the neonatal electroencephalogram. Pediatrics, 1984, 74: 330-335. Jasper, H.H. The 10-20 electrode system of the International Federation. Electroenceph. clin. Neurophysiol., 1958, 10: 371-375. Keezer, G. Intelligence level and occipital alpha rhythm in the Mongolian type of mental retardation. Am. J. Psychol., 1939, 52: 503-532. Kiloh, L.G., McComas, A.J., Osselton, J.W. and Upton, A.R.M. Clinical Electroencephalography (4th Edn.). Butterworth, Stoneham, MA, 1981. Mattson, S.N., Riley, E.P., Jernigan, T.L., Ehlers, C.L., Delis, D.C., Jones, K.L., Stern, C., Johnson, K.A., Hesselink, J.R. and Bellugi, U. Fetal alcohol syndrome: a case report of neuropsychological, MRI, and EEG assessment of two children. Alcohol. Clin. Exp. Res., 1992, 16: 1001-1003. Schmid, R.G. Sadowsky, K., Weinmann, H.M., Tirsch, W.S. and P~Sppi, S.J. Z-transformed EEG power spectra of children with Down syndrome vs. a control group. Neuropediatrics, 1985, 16: 218-224. Sepp~il~iinen, A.M. and Kivalo, E. EEG findings and epilepsy in Down's syndrome. J. Ment. Defic. Res., 1967, 11: 116-125. Shagass, C. Pharmacology of evoked potentials in man. In: D.H. Efron (Ed.), Psychopharmacology: a Review of Progress. Raven Press, New York, 1968: 483-492. Tangye, S.R. The EEG and incidence of epilepsy in Down's syndrome. J. Ment Defic. Res., 1979, 23: 17-24.
Q Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
EEG findings in fetal alcohol syndrome and Down syndrome children W.M. Kaneko a,*, E.L. Phillips a, E.P. Riley b, C.L. Ehlers
a
a Department ofNeuropharmacology, CVN-14, The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, CA 92037, USA b Department of Psychology, San Diego State University, San Diego, CA 92182-0350, USA Accepted for publication: 19 August 1995
Abstract
Results from previous studies evaluating the electroencephalograms (EEGs) of infants born to alcoholic mothers suggest that the neonatal EEG may be a sensitive measure of prenatal ethanol exposure. Few studies, however, have examined EEG records of adolescent children with fetal alcohol syndrome (FAS). The present study investigated the resting EEG recordings of 18 matched triads of FAS, Down syndrome, and normal control subjects. Significant reductions in mean power of the alpha frequencies (7.5-12 Hz) were seen for both clinical groups, however, each syndrome appeared to have distinct EEG spectral distributions. Down syndrome children overall had diffuse EEG slowing while the EEG records of the FAS children showed reduced power, particularly in the alpha frequencies in the absence of significant slow activity. In the Down syndrome children, significant decreases in alpha power was seen in posterior cortical regions, whereas FAS children were more affected in the left hemisphere. This study suggests that certain EEG variables may be helpful in characterizing the neurophysiology of FAS. Keywords: EEG; Fetal alcohol syndrome (FAS); Alpha activity; Slow waves
1. Introduction It has been suggested that, even in the absence of dysmorphology, neonatal EEG may be the most sensitive indicator of prenatal alcohol toxicity (Chernick et al., 1983; Ioffe et al., 1984). Abnormalities observed in sleep EEG of 3-day-old infants exposed to alcohol prenatally have been found to persist as long as 6 weeks after birth (Havlicek et al., 1977). This suggests that such EEG effects do not represent alcohol withdrawal and may reflect more permanent brain changes. More evidence that EEG recordings may be helpful for early diagnosis of alcohol effects on brain development is provided by Ioffe and Chernick (1990) who found that the degree of abnormality in the EEG power at birth correlated with subsequent motor and mental development. In that study, increases in power during REM sleep were associated with an adverse outcome in motor development, whereas increases in power during quiet sleep were related to an adverse outcome in mental development.
* Corresponding author. Tel.: (619) 554-7236; Fax: (619) 554-6494.
Few studies have evaluated EEGs in adolescents with fetal alcohol syndrome (FAS). In a preliminary report of two children with severe FAS (ages 13 and 14 years), we found moderate abnormalities of the awake EEG (Mattson et al., 1992). Theta was found to be the dominant EEG rhythm, with occasional delta activity predominantly in the posterior regions and missing or sparse 8 Hz alpha activity in the posterior head regions. No focal abnormalities or gross asymmetries were found in those subjects. To date, no studies have been conducted which performed spectral analyses of awake EEG in school-age children diagnosed as having FAS, especially as compared to other children with cognitive impairments, such as Down syndrome children. Down syndrome and FAS children have several common characteristics. These children are commonly cognitively impaired, growth retarded, hyperactive and have attention deficits and craniofacial dysmorphologies. While Down syndrome can be easily diagnosed with a chromosome assay, the diagnosis of FAS is more difficult. Thus, there is a need for more clinically discriminative indicators of FAS. The aim of the present study was to explore EEG variables in children with FAS. If the EEG of FAS chil-
0013-4694/96/$15.00 © 1996 Elsevier Science Ireland Ltd. All rights reserved SSDI 0 0 1 3 - 4 6 9 4 ( 9 5 ) 0 0 1 8 9 - 1
EEG 94168
W.M. Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
dren could be characterized distinctly from the EEG records of other patients with a congenital disorder which also included cognitive disabilities, craniofacial dysmorphology, and growth retardation, then the specific EEG variables may correctly categorize children into either group. Therefore, the purpose of this study was (1) to determine whether EEG abnormalities found in infants diagnosed with FAS are similar to those of school-age FAS children; (2) to characterize and evaluate the EEG recordings of children with FAS through spectral analysis; (3) to compare FAS subjects with children with Down syndrome; and (4) to compare these findings with normal controls matched for age, sex and ethnicity.
2. Methods 2.1. Subjects
Subjects were 18 triads of children who were matched on sex, age and, for the most part, ethnicity. The first group (n = 18) consisted of children diagnosed as having FAS by Dr. Kenneth Jones. Children in the"second group (n = 18), diagnosed as having Down syndrome, were referrals from the Down Syndrome Association of San Diego. Children in the third group (n = 18) were normal controls with no family history of neurological disorders and no history of exposure to alcohol or any other drug prenatally. A total of 54 children between the ages of 4 and 15 years (average age of 9.1 ___3.2 years) were evaluated. The 18 children in the FAS group consisted of 10 females and 8 males from various ethnic backgrounds: 9 Caucasians, 7 African Americans, 1 La.tino, and 1 of mixed Latino/ Caucasian parentage. The parent or legal guardian of each subject completed questionnaires regarding their child's demographics (e.g., ethnic background, age, gender, etc.) and medical history (e.g., hearing or visual complications, medications, handedness, etc.). Each parent was paid $35.00 for his/her child's participation, the children received small toys or stickers. 2.2. IQ assessments
Subjects were scheduled for an appointment at the Center for Behavioral Teratology at San Diego State University for intelligence testing. Full Scale IQ, Verbal IQ, and Performance IQ scores were obtained using the Wechsler Intelligence Scale for Children-Revised version (WlSC-R) or the Wechsler Preschool and Primary Scale of Intelligence (WPPSI), depending on age and ability of the individual. The subjects were given an additional $10.00 or a toy of their choice for participating. 2.3. Electrophysiological recordings
Electroencephalographic (EEG) recordings were obtained from all subjects between 8:00 a.m. and 4:00 p.m.
21
by a registered EEG technician at the General Clinical Research Center in the Green Hospital of the Scripps Clinic in La Jolla, CA. On the test date, procedures were fully explained to both parent/guardian and subject. Parents or guardians were required to sign a consent form before any procedures were performed. Individual goldplated electrodes or tin electrodes set inside a tightly fitting electrode cap were used for recordings. The placement of the electrodes was based on the international 10-20 system (Jasper, 1958). All electrode impedances were below 5 kO. Six channels of EEG data (F3-C3, C3-P3, P3-O1, F4-C4, C4-P4, and P4-O2) were collected. The subjects were instructed to lie comfortably, remain quiet and relaxed, and informed that it was important to be awake and have their eyes closed during the entire recording period. Subjects who could not close their eyes on instruction or had excessive eye movement were assisted in closing their eyes by gentle application of the hands or application of soft eye patches. EEG was recorded on paper for 5-10 min, using a Nihon-Kohden polygraph and on magnetic tape (Vetter Model D) for off-line analysis. The bandpass was 1-70 Hz and the sensitivity was 7 /zV/mm. Paper records were monitored during recording for signs of drowsiness and for any eye or movement artifacts, All paper records were visually analyzed, scored for abnormalities and rated for amplitude and wave form or morphological characteristics. Artifact-free EEG was then selected for computer analysis and cortical EEG variables of awake recordings were examined. 2.4. Data analysis
The EEG samples were digitized at a rate of 128 Hz (samples/sec) and the Fourier transforms of consecutive 4 sec epochs for each channel were calculated. The power spectrum was produced using a Macintosh computer and software developed by Ehlers and Havstad (1982). For statistical analyses, the transformed data were compressed into 8 frequency bands: 0.5-1 Hz; 1-2 Hz; 2-4 Hz; 4-7.5 Hz; 7.5-9 Hz; 9-12 Hz; 12-20 Hz; and 20-40 Hz. Mean spectral power density (/~V2/octave) over each band frequency was calculated by summing the raw power spectral values within the band, multiplying by a scale factor derived from the calibration signal to produce the total power in the band in microvolts squared, and dividing by the width of the band in octaves. This width is the logarithm of the ratio of the maximum and minimum frequencies in the band, divided by the log of the two. In addition to mean spectral power, peak frequency for each of the same frequency bands was also estimated. Peak frequency is the frequency at which power peaks within a certain frequency range. For each EEG frequency range at each recording site, 2-way between-subjects analyses of variance (ANOVAs), with 2 levels for sex (i.e., female and male) and 3 levels
W.M. Kaneko et al. / Electroencephalographyand clinical Neurophysiology 98 (1996) 20-28
22
for group (i.e., control, Down syndrome, FAS), were used to statistically evaluate each of the dependent measures (i.e., mean spectral power and peak frequency). Tukey HSD tests were utilized for post hoc analyses. Rather than use the traditional 0.05 level of significance, a more conservative 0.01 level was used because of the multiple comparisons defined and tested in this study.
3. R e s u l t s
3.1. Group characteristics There were no significant differences in age between the 3 groups of children. The average age of both the F A S group and the Down syndrome group was 9.92 years, while the average age of the normal controls was 9.88 years. IQ scores for 12 of the 16 Down syndrome children tested could not be calculated due to poor cooperation a n d / o r severe mental impairment. Their Full Scale IQ scores were below 40. For the 4 testable Down syndrome children, the Full Scale IQ ranged from 43 to 47. IQ scores for 16 of the 18 F A S children ranged from 41 to 96, with a mean Full Scale IQ score of 74 + 16.0, which was significantly lower than the 16 matched normal controls who had IQs that ranged from 86 to 137, with a mean Full Scale IQ score of 109 -1- 13. Several children from the two clinical groups were taking medication at the time of testing. Three Down syndrome children were taking methylphenidate, 3 F A S
children were taking methylphenidate or cylert, and 4 other F A S subjects were on other drugs: two on anticonvulsants, one on phenothiazines, and one on a small dose of imipramine at bedtime (see Table 1).
3.2. Overall description of E E G records Although subjects were instructed to keep their eyes closed during the recording, 7 of the 18 Down syndrome children were unable to do so. Therefore, their recordings were obtained while their eyes remained open. Samples of actual EEG records of a representative triad (i.e., control, Down syndrome, F A S ) are presented in Fig. 1. The spectral plots for the same subjects can be found in Fig. 2. All controls were clinically rated as normal, whereas 12 out of 18 EEG records of the children with Down syndrome were interpreted as clinically abnormal, and 9 of the 18 EEG records of the F A S subjects had an overall rating o f borderline or abnormal (see Table 2). In general, they were found to be immature a n d / o r poorly developed. Low voltage or amplitude was commonly seen in the EEG records of F A S children, while slower activity (i.e., delta and theta waves) was dominant in the E E G records of Down syndrome children. Six F A S children and 10 children with Down syndrome had excessive generalized slowing. Other abnormalities noted in the cognitively impaired children were poor wave formation, focal spikes, and asymmetrical EEG development. Table 2 shows clinical ratings and descriptions of the E E G records for all children evaluated.
Table 1 Demographic information for all subjects Gender Age F M F F F M F F F F M F F M M
M M M
Control
(years)
IQ
4 5 6 6 7 7 8 9 9 9 10 11 11 12 13 14 14 15
111 115 112 100 94 109 117 135 116 137 128 91 109 n/a 122 86 103 101
Non-ethnicmatch
Caucasian
n/a, not available. * Mellaril, Phenobarbital, Dalmane.
FAS
Down syndrome Medication IQ < 40 < 40 45 < 40 < 40 < 40 < 40 < 40 43 < 40 < 40 < 40 n/a < 40
Non-ethnicmatch Caucasian Caucasian Caucasian Caucasian Samoan Caucasian Hispanic/Caucasian Caucasian
n/a
47 < 40 43
Medication IQ
Caucasian Caucasian
Ritalin Ritalin
Ritalin
78 90 73 74 56 96 69 74 94 < 40 78 91 52 79 51 41 74 87
Non-ethnicmatch Hispanic/Caucasian Caucasian Caucasian Afro-American Afro-American Caucasian Afro-American Caucasian Caucasian Caucasian Hispanic Afro-American Afro-American Caucasian Caucasian Afro-American Afro-American Caucasian
Medication
Mellaril
Imipramine Epitol Ritalin Ritalin * Cylert
W.M. Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
3.3. Delta and them activ#y
greater mean power in the delta frequencies (1-4 Hz) in the fronto-central (F3-C3, F4-C4) and centro-parietal (C3P3, C4-P4) derivations when compared to the control and FAS groups. Differences seen between the 3 study groups on mean EEG power for all frequency bands in the centro-parietal cortex can be seen in Fig. 3. Mean power for theta activity (4-7.5 Hz) in the F3-C3 and F4-C4 leads was also found to be significantly higher for children with Down syndrome as compared to children
Spectral analyses revealed significant differences between the 3 groups of children for mean EEG power and peak frequency in all cortical recording sites. The mean EEG power of low frequency delta (0.5-1 Hz) for Down syndrome children was :significantly greater than for matched normal controls in all 6 cortical recording sites. The Down syndrome group demonstrated significantly
'
•
,
i
.i~
I
|,
" :
I,
"
i
,
;
!
!
o Co.P3
0 E4 Z 0 L)
F
:
, i,
23
."
!
. .~
.
.
i
~
.
•
.
.
=;~
~
!
.
,
~ i
i
'
0
c~ C~ Z >4 EQ Z
0
•
I
~
i.
i,
i
i l
i
'i
i
i
I
'i
.+
i
r~
,,f ¢
I
I
I
i ,~
c4 ,V
¢ i
~
d ~,
i
i
I I
1 sec
I
50~V
Fig. 1. Samples of EEG record,; of three 9-year-old females, a normal control (top panel), Down syndrome (middle panel), and FAS (bottom panel).
4 5
6
6 7
7 8 9
9
9
10 11 11
12 13
14
14
15
F
F F
M F F
F
F
M F F
M M
M
M
M
Age (years)
F M
Gender
9 Hz alpha
11 Hz alpha
9-10 I-Iz alpha
9-10 Hz alpha 8-9 Hz alpha
8 Hz alpha 8 Hz alpha 8-9 Hz alpha
8-9 Hz alpha
9 Hz alpha
9 Hz alpha 9 Hz alpha 9-10 Hz alpha
8 Hz alpha 8 Hz alpha
9 Hz alpha
Normal
Normal
Normal
Normal Normal
Normal Normal Normal
Normal
Normal
Normal Normal Normal
Normal Normal
Normal
Normal Normal
5-6 I-Lztheta
5 - 6 I-Iz theta
8-9 Hz alpha
9 Hz alpha 8-9 Hz alpha
1 1 Hz alpha 9 Hz alpha 8-9 Hz alpha
2 Hz delta and 5 Hz theta 6-7 Hz theta
7-8 I-lz theta 4-5 Hz theta 4-5 Hz theta
5 - 6 Hz theta 4-5 Hz theta
6 - 7 Hz theta
6-7 Hz theta 5 Hz theta
Dominant rhythm
8 Hz alpha 8-9 Hz alpha
Down syndrome
Dominant rhythm Characteristic feature
Control
Table 2 Clinical ratings of EEG records
Normal Normal Essentially normal; poorly developed; left hemisphere poorly developed; frontal beta Normal Essentially normal; poorly developed; left hemisphere poorly developed Essentially normal; poorly developed; frontal beta Generalized slowing with excess movement artifact Generalized slowing with sparse alpha present
Generalized slowing; frontal beta Asymmetrical; left focal spike, left central phase reversal at C3 Poorly developed; generalized slowing; frontal beta Generalized slowing; frontal beta Poorly developed; generalized slowing Some slowing Generalized slowing Poorly developed; generalized slowing; frontal beta Poorly developed; generalized slowing; frontal beta Generalized slowing, frontal beta
Characteristic feature
8 Hz alpha
9 Hz alpha
4 - 6 Hz theta
10-11 Hz alpha 5 - 6 Hz theta
8 Hz alpha 8 - 9 Hz alpha 8 - 9 Hz alpha
8 141 alpha
8 Hz alpha
7-8 Hz theta 8-9 Hz alpha 6 - 7 l-Iz theta
5 - 6 Hz theta 4 - 6 Hz theta
5-6 Hz theta
7 - 8 H.z theta 7 - 8 Hz theta
Dominant rhythm
FAS
Poorly developed with mild transient theta
Poorly developed; generalize slowing Normal
Normal Generalized slowing; frontal beta
Essentially normal; poorly developed; frontal beta Normal Normal Normal
Generalized slowing with sparse alpha present Generalized slowing Generalized slowing with frontal beta Poorly developed; some slowing Normal Generalized slowing; sparse alpha; left hemisphere poorly developed Normal
Normal Some slowing
Characteristic feature
I t~ oo
t-o
e~
t~
t~
e,
W.h£ Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
/
-"N.¢IIO
/Ira
t, -'m.lmm
~ ,
_
,
~i
"
"mlsm
,' .
"m~m c
25
,
"n~m
~
'
iNl131
"mllm "I21m
r' " m m m ' 2
sm I,
CONTROL
DOWN
SYNDROME
FAS
Fig. 2. EEG spectral plots from the centro-parietal lead of three 9-year-old girls, a normal control (left panel), Down syndrome (middle panel), and FAS
(right panel)• The x-axis representsfrequency(Hz), the y-axisreflectstime (min), and the z-axis representsEEG power(/xV2/octave).
with FAS. The frequency at which power peaked in the theta frequency range was significantly lower for the Down syndrome children than for the controls. This decrease in
--~
LEFT
CONTROL
~
~ •
S'n~ROME
•
FAS p~.O06
•"
t~<.oos
(DO~
v~.
peak frequency was present in all cortical regions. Compared to the FAS group, significantly lower peak frequency of theta activity was also found in the Down syndrome group, however, this was only present in the F4-C4 lead. Differences between the 3 groups on peak frequency of theta activity in the centro-parietal region are demonstrated in Fig. 4.
FAS)
3.4. Alpha and beta activity o~o
~"
>
50
.5-i
I-2
2-4
4-7,5
7.5-9
9-12
12-20
20-40
RIGHT 150
! o~ 100
0
.s-1
1-2
2-4
,-;.~
FREQUENCY
7.;-9 BAND
9-12
12-20 20-40
(Hz)
Fig. 3. M e a n + S.E.M. E E G spe~ral power in the centro-parietalcortical
lead from 18 normal control, 18 Down syndrome and 18 FAS children across all frequency bands.
FAS children showed significantly lower mean power in the alpha (9-12 Hz) frequency range for the left frontocentral (F3-C3) and left parieto-occipital (P3-O1) leads when compared to normal controls• The Down syndrome group displayed significantly lower mean power in the 7.5-9 Hz alpha frequency range in the left posterior lead (P3-O1) when compared to the normal control group and in the 9-12 Hz alpha range for the right posterior lead (P4-O2). For Down syndrome children, the frequency at which power peaked in the slow alpha frequency range (7.5-9 Hz) in all cortical regions was significantly lower than that of normal controls. FAS children also had significantly lower peak frequencies in the slow alpha range compared to controls; however, this decrease was present only in the right central lead (C4-P4).: 'Differences between the 3 groups on peak frequency Of alpha activity in the centroparietal region are demonstrated in Fig. 4. Mean power in the beta frequency (12-20 Hz) range for the centro-parietal regions of the Down syndrome children were significantly higher when compared to FAS children.
W.M. Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
26 G oo 7
THETA ACTIVITY (4-7.5 Hz)
D co~o~ [] D O W N s~n~D~OME • FAS
I
iiiiiiiiiiiiiiiiiiiii
~!ii!iii!!i~!ii
!i!iiii!!ii!
4. Discussion
ALPHA ACTIVITY (7.5-9 HZ)
8.25 "
8. O0
-
-
~
-
-
LEFT
RIGHT CENTRAL-PARIETAL
AREA
Fig. 4. Mean + S.E.M. peak frequency of the theta (top graph) and alpha (lower graph) ranges from the centro-parietal lead in 18 normal control, 18 Down syndrome and 18 FAS children.
3.5. Discriminant function analysis Eighteen matched triads, for a total of 54 subjects, were used for the DFA. Each of the 3 variables, mean power of alpha of the left parieto-occipital lead, mean power of theta
Table 3 Values and results from the discriminant function analysis (DFA) using EEG variables (18 control/Down syndrome/FAS matched triads) Factor 1 2
Canonical correlation 0.705 0.313
Proportion of variation 0.49 0.10
Wilks' lambda = 0.454 Group
Control Down syndrome FAS Predictor variable Left parieto-occipital mean power of alpha Right fronto-central mean power of theta Rigt fronto-central theta peak frequency Wilks' lambda
Classification and number of cases (%) Correct
Incorrect
10 (56) 13 (72) 10 (56)
8 (44) 5 (28) 8 (44)
df
of the right fronto-central lead, and theta peak frequency of the right fronto-central lead, significantly differentiated the 3 groups ( F (6, 98) = 7.91, P < 0.0005). DFA values are presented in Table 3. Fifty-six percent (10/18) of the normal controls, 72% (13/18) of the Down syndrome children and 56% (10/18) of the children with FAS were correctly classified by the DFA. These percentages are greater than the percentage of correct predictions expected on the basis of chance.
F
P
2, 51
6.806
0.002
2, 51
6.224
0.004
2, 51
13.581
0.0005
6, 98
7.914
0.0005
This study characterized EEG variables in FAS children and compared them to Down syndrome children. A putative EEG spectral pattern of children with FAS was revealed, which overall consisted of reduced alpha activity in the absence of significant EEG slowing. Mean power for theta (4-7.5 Hz) activity in FAS children did not differ from matched normal controls, however, the Down syndrome group was found to have increased mean theta power compared to FAS children in the fronto-central cortical regions. Increases in mean power in theta frequencies in the right fronto-central cortical area were also noted in Down syndrome children as compared to normal controls. Children with Down syndrome were also found to have slower theta activity compared to normal controls, as revealed by the lower peak frequency values. These results are consistent with previous findings stating that the most common EEG abnormality found in infants and children with Down syndrome is excessive theta and slower activity (Clausen et al., 1977; Tangye, 1979; Schmid et al., 1985; Epstein, 1986; Devinsky et al., 1990). Increases in slow EEG activity have been related to slow maturation of the brain (Ellingson et al., 1973). In Down syndrome patients, slowing is more common in young children, however, it is rare during infancy. The frequency of abnormal EEGs found in Down syndrome children appears to increase with age, when full maturity of the brain is achieved (Sepp~il~iinen and Kivalo, 1967; Ellingson et al., 1970, 1973; Ellingson and Peters, 1980; Schmid et al., 1985). The EEGs of the Down syndrome subjects may have a different developmental course than those of controls. Developmental aspects of the EEG remain unknown in FAS. The present study demonstrates that alpha activity (7.512 Hz) was significantly reduced in FAS children in the absence of EEG slowing. This effect was significant in left cortical regions, F3-C3 and P3-O1. Mean power for the alpha frequency ranges was also significantly decreased in Down syndrome subjects as compared to the controls. However, this reduction in alpha was present only in posterior areas, P3-O1 and P4-O2. These data suggest that the left hemisphere may be vulnerable to prenatal alcohol exposure while the posterior regions may be particularly
W.M. Kaneko et al. / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
affected in Down syndrome children. Slower alpha frequencies for all recording sites were also seen in children with Down syndrome as compared to normal controls. FAS children, however, showed slower alpha activity compared to normal controls only in the right parieto-occipital region. This suggests that the peak frequency of the alpha activity in certain cortical regions may also differentiate the two clinical groups in this study. Case studies on two of the patients in the present study using MRI techniques have been conducted (Mattson et al., 1992). These two chi]ldren were found to have decreased overall cranial and cerebellar volumes and a decrease in the proportional volume of the basal ganglia. No differences in cortical volume or subcortical fluid volumes, cortical grey matter, limbic and non-limbic cortex, or diencephalic structures were found between the two children prenatally exposed to alcohol and the controls. Other studies have reported neuropathological conditions in infants exposed to alcohol prenatally. Errors in migration of neuronal and glial elements may be one source of the brain malformations seen (Clarren et al., 1978). There is great variability in the nature and severity of the brain abnormalities in children with FAS. Our finding that the left hemisphere is more affected by alcohol than the right hemisphere is interesting since asymmetrical effects are present in the photo of the brain :from an FAS infant shown by Clarren et al. (1978). However, further studies of larger subject populations will be necessary to make appropriate correlations. A negative correlation between mental maturation and alpha frequency has been suggested in some population studies (Keezer, 1939; Fillingson and Lathrop, 1973; Schmid et al., 1985). Although there was a significant IQ difference between control,; and FAS children, no significant correlations between Full Scale IQ and mean power of alpha activity or between Full Scale IQ and peak frequency of alpha activity were observed at the P > 0.01 level. The lack of correlation may be related to the wide age span of the subject sample used in this study. Seven of the 18 Down syndrome children in the present study had their eyes open during the majority of the recording session. During "eyes open" conditions, it is the dominant rhythm that is attenuated. An evaluation of the raw data in relation to eye status and dominant rhythm revealed that all but one of the "eyes open" Down syndrome subjects had theta activity as the dominant rhythm. In addition, in the children whose eyes remained mainly open during the EEG recording, alpha frequencies were not necessarily present when they closed their eyes. This suggests that the significant decreases in alpha activity observed in Down syndrome subjects may not be explained entirely by the "eyes open" status differences among the groups. Reductions in alpha may also be caused by increased coherence between two electrodes in a bipolar montage. Further studies will require other montages to investigate this issue.
27
In the present study, 10 of the 36 children from the clinical groups were not drug free, and several children were taking similar drugs. Specific effects on on-going EEG have been associated with certain medication (Fink, 1968; Shagass, 1968; Kiloh et al., 1981). Methylphenidate may produce some decreases in alpha, whereas neuroleptics and antidepressants, such as Mellaril and imipramine, may increase the amount of alpha activity, while sedatives or tranquilizers may lead to an increase in fast activity (20-30 Hz). Greater amounts of beta activity were found in the Down syndrome children when compared to FAS children. This is interesting since only three of these children were on methylphenidate and none were on antiepileptics which can also increase beta. However, no clear correlations were found between medication status and EEG beta as quantitated with spectral analyses. Therefore, this increase in beta activity appears to be related to diagnosis and consistent with previous findings (Schmid et al., 1985; Devinsky et al., 1990). The present study offers new data on resting EEGs in school-age children diagnosed with FAS. The EEGs of FAS children were characterized and differentiated from another cognitively disabled group. In summary, Down syndrome children displayed excessive amounts of delta, theta and beta activity as compared to FAS children and controls, whereas both clinical groups had less alpha and lower alpha frequencies. Our findings of slower alpha activity in both clinical groups support previous hypotheses of cerebral immaturity in Down syndrome children and suggest cerebral immaturity in those prenatally exposed to alcohol. However, unlike the children with Down syndrome, the reduction in alpha activity seen in the FAS children is in the absence of overall EEG slowing. Significant reductions in alpha mean power were observed in both clinical groups; however, this abnormality was found to be distributed in posterior cortical regions of children with Down syndrome and generally found in the left hemisphere of FAS children. Thus, differences in EEG abnormality type and the cortical distribution of the abnormalities may differentiate these two syndromes. Future studies utilizing a larger electrode array may further contribute to the understanding of the neurophysiology of these two syndromes.
Acknowledgements This study was supported by NIAAA 00098 to Cindy L. Ehlers, Alcohol Research Center Grant 06420 to Floyd E. Bloom and General Clinical Research Center (GCRC) Grant 00833. The authors would like to thank Dr. Sarah Mattson for aiding in the estimation of IQs, Dr. Kenneth Lyons Jones for providing patients, and Dr. James Havstad for writing the computer programs.
28
W.M. Kaneko et aL / Electroencephalography and clinical Neurophysiology 98 (1996) 20-28
References Chernick, V., Childiaeva, R. and Ioffe, S. Effects of maternal alcohol intake and smoking on neonatal electroencephalogram and anthropometric measurements. Am. J. Obstet. Gynecol., 1983, 146: 41-47. Clarren, S.K., Alvord, E.C., Sumi, S.M., Streissguth, A.P. and Smith, D.W. Brain malformations related to prenatal exposure to ethanol. J. Pediat., 1978, 92: 64-67. Ctausen, J., Sersen, E.A. and Lidsky, A. Sleep patterns in mental retardation: Down's syndrome. Electroenceph. clin. Neurophysiol., 1977, 43: 183-191. Devinsky, O., Sato, S., Conwit, R.A. and Schapiro, M.B. Relation of EEG alpha background to cognitive function, brain atrophy, and cerebral metabolism in Down's syndrome: age-specific changes. Arch. Neurol., 1990, 47: 58-62. Ehlers, C.L. and Havstad, J.W. Characterization of drug effects on the EEG by power spectral band time series analysis. Psychopharmacol. Bull., 1982, 18: 43-47. Ellingson, R.J. and Lathrop, G.H. Intelligence and frequency of the alpha rhythm. Am. J. Ment. Defic., 1973, 78: 334-338. Ellingson, R.J. and Peters, J.F. Development of EEG and daytime sleep patterns in Trisomy-21 infants during the first year of life: longitudinal observations. Electroenceph. clin. Neurophysiol., 1980, 50: 457466. Ellingson, R.J., Menolascino, F.J. and Eisen, J.D. Clinical-EEG relationships in mongoloids confirmed by karyotype. Am. J. Ment. Defic., 1970, 74: 645-650. Ellingson, R.J., Eisen, J.D. and Ottersberg, G. Clinical electroencephalographic observations on institutionalized mongoloids confirmed by karyotype. Electroenceph. clin. Neurophysiol., 1973, 34: 193-196. Epstein, C.J. The Neurobiology of Down Syndrome. Raven Press, New York, 1986. Fink, M. EEG classification of psychoactive compounds in man: review and theory of behavioral associations. In: D.H. Efron (Ed.), Psy-
chopharmacology: a Review of Progress. Raven Press, New York, 1968: 497-508. Havlicek, V., Childiaeva, R. and Chernick, V. EEG frequency spectrum characteristics of sleep states in infants of alcoholic mothers. Neuropediatrics, 1977, 8: 360-373. Ioffe, S. and Chernick, V. Prediction of subsequent motor and mental retardation in newborn infants exposed to alcohol in utero by computerized EEG analysis. Neuropediatrics, 1990, 21: 11-17. Ioffe, S., Childiaeva, M.D. and Chernick, M.D. Prolonged effects of maternal alcohol ingestion on the neonatal electroencephalogram. Pediatrics, 1984, 74: 330-335. Jasper, H.H. The 10-20 electrode system of the International Federation. Electroenceph. clin. Neurophysiol., 1958, 10: 371-375. Keezer, G. Intelligence level and occipital alpha rhythm in the Mongolian type of mental retardation. Am. J. Psychol., 1939, 52: 503-532. Kiloh, L.G., McComas, A.J., Osselton, J.W. and Upton, A.R.M. Clinical Electroencephalography (4th Edn.). Butterworth, Stoneham, MA, 1981. Mattson, S.N., Riley, E.P., Jernigan, T.L., Ehlers, C.L., Delis, D.C., Jones, K.L., Stern, C., Johnson, K.A., Hesselink, J.R. and Bellugi, U. Fetal alcohol syndrome: a case report of neuropsychological, MRI, and EEG assessment of two children. Alcohol. Clin. Exp. Res., 1992, 16: 1001-1003. Schmid, R.G. Sadowsky, K., Weinmann, H.M., Tirsch, W.S. and P~Sppi, S.J. Z-transformed EEG power spectra of children with Down syndrome vs. a control group. Neuropediatrics, 1985, 16: 218-224. Sepp~il~iinen, A.M. and Kivalo, E. EEG findings and epilepsy in Down's syndrome. J. Ment. Defic. Res., 1967, 11: 116-125. Shagass, C. Pharmacology of evoked potentials in man. In: D.H. Efron (Ed.), Psychopharmacology: a Review of Progress. Raven Press, New York, 1968: 483-492. Tangye, S.R. The EEG and incidence of epilepsy in Down's syndrome. J. Ment Defic. Res., 1979, 23: 17-24.