Evolution of daytime quiet sleep components in early treated phenylketonuric infants

Q Brain & Development 1996: 18: zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGF 201-206 ELSEVIER Original article Evolution of daytime quiet sleep components phenylketonuric infants in early treated Giulia Fanny De Giorgis a, Enrico Nonnis a, Fosca Crocioni a, Paola Gregori b, Maria Pia Rosini b, Vincenzo Leuzzi ‘, Albert0 Loizzo d3* zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA di Scienze Neurologiche e Psichiatriche dell’ Et2 Eoolutiua, lAGersit& La Sapienza, Roma. Ita!\ a Dipurtimento ’ Sercizio di Neuropsichiatria Infantile, Ospedale La Scarpetta, Roma, Italy ’ Istituto di M edicina Sperimentale, Uniuersitir La Sapienza, Roma, Italy ’ Istituto Superiore di Sanit& riale Regina Elena 299, 00161. Roma, Italy Received 21 July 1994; accepted 13 January 1996 The maturational patterns of ‘track alternant’ (TA) and sleep spindles obtained from 16 early detected phenylketonuric (PKU) children during their first months of life were compared with others that were evaluated in recordings taken from 42 controls of the same age group. The TA maturation evolved significantly later in the PKU group than in the control group during the 5th-8th week (the TA score for the PKU group was 64% vs. 10% in the control group, P < 0.001). Afterwards, during the 9th-12th week the score for the PKU group was 27% vs. 0% in the controls (P < 0.002). The sleep spindle evolution score also matured significantly later in the PKU than the control group: the score was 31% in PKU children vs. 85% in controls for the 5th-8th week of age (P < O.Ol), and it was 66% vs. 96% for the 9th-12th week (P < 0.02). After the 12th week, TA pattern could not be detected, and spindles reached complete maturation in the PKU children as well. Our results show a consistent delay in the maturation of TA and spindle scores in PKU children. This trend of delay is parallel to the plasma phenylalanine normalization, but not necessarily dependent only on it. In conclusion, we suggest that studies on the critical maturational periods of different sleep components (TA and sleep spindles) might provide a sensitive tool for early diagnosis of neurophysiological brain alterations during the first trimester of life in a population of children ‘at risk’. Keywords: Phenylketonuria; Quiet sleep; Track altemant; Sleep spindle; EEG development; Critical period 1. INTRODUCTION ‘TracC alternant’ (TA) is defined as a basic EEG pattern, characterized by bursts of large amplitude activity with marked attenuation between the bursts. It is most striking in prematurely born babies (‘track discontinu’), less so in babies born at term, and becomes increasingly difficult to distinguish until it can no longer be identified beyond 5-8 weeks of age. [l-4]. The appearance of sleep spindle bursts of relatively sinusoidal 12-14 Hz waves may be observed in full-term newborns at birth. Clearly distinguishable discrete bursts appear only between the ages of 3 and 9 weeks post term during quiet sleep. These reach adult-like matured patterns at lo-15 weeks of age [3,5,6]. * Corresponding author. Fax: (39) (6) 4440053. 0387-7604/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved PZI SO387-7604(96)00005-8 Serendipitous observations described in our previous paper [7] showed that children suffering with early treated phenylketonuria (PKU) had clear TA EEG patterns during quiet sleep at an age that was more mature than described as normal in the literature [3,8]. On the other hand, sleep spindles recorded in PKU children on later ages, up to the 12th month, showed alterations in morphology, frequency, synchrony and symmetry, even after the onset of dietary therapy. Since TA and sleep spindle EEG patterns undergo their most complete morphological definition during this period in normal full term children (critical periods, [9]), we decided to study the evolution of TA and sleep spindles in a group of PKU children. Our occasional and descriptive observations were previously reported [7]. The present investigation was designed to see whether using semi-quantitative methods, more precise information could be gained. Our working hypothesis was directed only to study the evolution of TA and sleep spindles; no evident G.F. De Giorgis et al. / Brain & DeLlelopment 1996; 18: 201- 206 202 alterations in the morphology of the single elements were noted within the limits of the age considered (apart from occasional observation of ‘spiky’ spindles). In all cases, one of the aims was to verify whether these data could be related to the evolution of biochemical data, i.e., plasma phenylalanine and tyrosine levels, and to verify whether biochemical injury during early postnatal period could induce permanent or long-lasting neurophysiological alterations. 2. MATERIALS AND METHODS 2.1. Patients The clinical experiments conformed to the Declaration of Helsinki. Sixteen PKU children were studied; all of them were detected from 1981 to 1990 through the Newborn Routine Screening Program by the ‘Servizio Speciale Malattie Genetiche e Metaboliche’ of the University ‘La Sapienza’ of Rome [lo]. All children were born at term with a mean gestational age of 39.8 weeks (two children with a GA of 38 weeks; one of 39 weeks; eleven of 40 weeks; one of 41 weeks and one of 42 weeks) following normal pregnancy in all but one (impending pre-term delivery). Nine of them were born after normal delivery, six after cesarean section, one with the aid of a vacuum device. No signs of perinatal hypoxic-ischemic syndrome were detected. Mean weight at birth was 3438 g k 109 g S.E.M. Gestational age was timed from onset of the last normal menstruation. As early as possible (two children at 3 weeks of age, four at 4 weeks; four at 5 weeks; four at 6 weeks; one at 9 weeks and one at 11 weeks), all the children were entrusted to our Institute for a program of follow-up examination (for detailed descriptions, cf. [7,11]). Neurological examination was performed at a mean age of 6 k 0.8 weeks of age: 14 children appeared normal and 2 showed the following transient neurological signs: brisk tendon reflexes, low threshold for the Moro response, low-frequency high-amplitude tremor of the limbs (hyperexcitable syndrome). Physical, neurological and developmental examinations were performed at the beginning of the dietary therapy, then at 3, 6, 9 and 12 months of age. placement was frontopolar, central, parasagittal, temporal, occipital and ear, according to the lo-20 International System; recordings were performed in the morning between 09.00 h and 12.00 h on eight channels of EEG (OTE-Biomedica, Firenze) with 0.1-35 Hz band pass filtering, on paper speed of 30 mm/s, in a sound-proof room. A technician coded infant’s behavior on the paper, according to the criteria suggested by Sterman et al [12], i.e., eyes closed/open; eyes movement; neck and legs muscular tone; movement of the head, body or limbs; frequency and regularity of breathing. Duration of quiet sleep in each EEG recording was at least 20 min, starting from the EEG observation of spindles or TA, but in some cases recording was prolonged as necessary. Following the data published by some investigators [ 13-151, we felt that the duration of quiet sleep episodes and sleep components obtained from naps are fulfilled by the temporal schedule used in our paper. No polygraphic recording was performed in these infants. 2.3. EEG evaluation methods The tracings were read independently by 2 investigators (G.F.D.G and A.L.), who scored the presence of TA and spindles. TA was defined as follows: bursts of symmetrical and synchronous large amplitude slow (0.5-3 Hz) waves. occasionally superimposed by rapid low voltage waves and sharp waves, lasting 2-4 s, separated by 4-8 s of attenuated activity of mixed frequencies [2]. When this pattern was recorded clearly at least two times during the whole EEG recording, it was scored 1 (Fig. 1A). Since TA becomes increasingly difficult to distinguish until it can no longer be identified beyond 6-7 weeks after birth, a score of 0.5 was given for less clearly distinguishable patterns (i.e., a TA pattern showing an amplitude ratio of 2:l or less, vs. the background pattern of attenuated activity), or patterns which showed asynchronous (i.e., showing at least a 4 s interhemispheric temporal gap) and asymmetric bursts (i.e., when amplitude was at least twice the amplitude of contralateral figures) (Fig. lB>. Absence of TA was scored as 0. The score was given to the whole tracing after detection of at least two TA patterns which could be attributed by the investigator to the higher score (0.5 or 1). Since the score was attributed to the ‘maturity’ and not to the length of the tracing, no data are reported on the absolute number or frequency of TA patterns over the course of time. For 2.2. EEG example: if the EEG had three TA patterns scored 1, and 15 TA patterns scored 0.5, the final score would be 1. The TA score in In the present study only PKU children were included which each tracing was evaluated as the mean score of the two investihad EEGs that were recorded according to a longitudinal experigators. The total score in each age group was calculated as the mental design used in our laboratories since 1980. The first EEG sum of the mean scores of the tracings, divided by the number of was recorded as early as possible at the first medical examination tracings in the given age group, multiplied by 100 and expressed before dietary therapy; the second one within 4 weeks after as a percentage. dietary therapy, the third one during the 3rd (sometimes 4th) This score was plotted as the ordinate value on a graph, month of age. A total of 48 EEGs were recorded during daytime against the mean age of children of the same age group (exphysiological sleep in 16 children, but three EEGs were excluded pressed in weeks, f 1 S.E.M.). EEGs recorded in selected age from evaluation since depth or duration of sleep did not fulfill the groups, corresponding approximately to the first 4 months of life schedule requirements and recordings could not be repeated. after birth, were arbitrarily considered together, i.e., up to the 4th Therefore, in the present paper 45 EEGs are included: of these, week; 5-8 weeks; 9-12 weeks; and above 12 weeks. 16 were recorded in the children from a few hours up to 24 h An analogous method was adopted for the scoring of spindles. before dietary therapy (mean recording age 5 + 0.6 weeks, range Three kinds of spindles were identified (cf. [3,5,6,16]): (1) rudi3-l 1 weeks); 16 were recorded about 3 weeks after dietary mentary spindles, called ‘prespindles’, which have very low therapy (mean age 8 f 0.5 weeks, range 5-12 weeks); 13 were amplitude, variable frequency (about 14-16 Hz) and are not recorded at a mean age of 14 f 0.4 weeks, range 1 l-zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 18 weeks. formed in organized bursts; (2) grade I spindles, which are well Anamnestic data on the children were taken from the files regulated at 12- 14 Hz. occur at the vertex only in brief bursts from the hospitals where they were born. The EEG electrode G.F. De Giorgis et al/Brain & Development 203 1996; 18: 201-206 2.4. Control group A total of 42 EEGs were obtained from 42 neurologically normal children aged 2- 12 weeks in the same conditions as PKU children, after having obtained informed consent from their parents. All children had histories of normal pregnancy and delivery, except five: two of them were born after cesarean section and three with the aid of a vacuum device. Birth was recorded at a mean gestational age of 39.3 weeks (range 38-42). Mean weight was 3312 i 1 I5 g. Neurological examination was performed at a mean age of 7 i 0.7 weeks. Also in the case of control children data were gathered from hospital records. 2.5. Chemical analysis of plasma aminoacid levels Phenylalanine (Phe) and tyrosine (Tyrl levels were evaluated through column chromatography, according to previously described methods [7,11]. Plasma determination was performed every 2 days for 20 days after the onset of the diet; then. two times a week up to the 3rd month of age. followed by once a week up to the 6th month of age. At our Centre, the therapeutic target aimed for during the first year of life was to maintain plasma Phe values between 20 and 50 pmol/dl. 2.6. Statistical O2c4 J‘2 02c3 ; Fig. 1. A: sample of typical score I TA. The figure is taken by a PKU infant (tracing N. 55927), age 37 days. No qualitative differences could be identified between TA patterns in PKU or control infants, therefore A sample is referred to both populations. B: sample of a typical score 0.5 TA. The figure is taken by a PKU infant (tracing N. 583411, age 40 days. Also in this case, this pattern is typical in both PKU or control infants, depending on their maturational stage. evaluation of data The reliability of scores given independently to the same tracing by the two investigators was 89% for TA and 92% for sleep spindles (the same score for TA was given to 38 out of 45 tracings of PKU children, and to 39/42 of controls: the same score for spindles was given to 40/45 PKU and to 40/42 controls). EEG data were expressed in the graph as a percentage of the full-maturity score for each age group. The number of patients and controls in each group is specified in Fig. 3 (in parentheses). Note that control children had reached full TA and spindles maturity within the 12th week. In the abscissa the mean age was represented for each group, plus or minus the S.E.M. For statistical evaluation. the data were confronted using the Mann-Whitney U-test, and the Wilcoxon signed rank test. A) FPl T3 which last 0.4-I s, have low amplitude (5-15 p,V>, and do not show fusiform modulation (Fig. 2A); (3) grade II spindles, which have 12- 14 Hz in frequency, are vertex dominant but may occur simultaneously in anterior temporal regions, show a greater tendency to symmetry and synchrony, have a duration of l-3 or more s, have greater amplitude (20-70 uV>, and fusiform modulation (Fig. 2B). In this paper, prespindles were not considered (score 0). Recordings showing only grade I spindles were attributed score 0.5; those showing grade II spindles were scored 1. Also in this case, when at least two positive grade-l spindles, or at least two grade-2 spindles could be identified in the same EEG, the higher score was attributed to the whole tracing. Final evaluation was computed using the same criteria as for TA and an analogous developmental trend was plotted on a graph. 8) FP~ T3 T 5OP” 1s Fig. 2. A: sample of typical score 0.5 spindle. This pattern was taken by a PKU infant EEG (tracing N. 57763). age 32 days, and was considered typical in both PKU and control infants. B: sample of typical score-l spindle. This pattern was taken by the same infant as in A (tracing N. 581091, age 64 days, and was considered typical in both PKU and control infants. After this age. a few PKU infants may show some slight morphological spindle alterations i.e.. the so-called “spiky” spindles [7]. 204 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA G.F. De Giorgis et al/Brain & Decdopment 1996; 18: 201- 206 The biochemical data were expressed for each week as the mean for Phe and Tyr plasma levels, referred to the age of children after birth. 0 PHE, JIM / dL 0 TYR, PM / dL 100 60 3. RESULTS 60 The decreasing score of TA pattern is plotted in the upper part of Fig. 3. In control children the TA score, which was 91% up to the 4th week of age, dropped down rapidly to 10% during the 5th-8th week, and to 0% thereafter. In PKU children, TA was fully present in 100% of our cases up to the 4th week of age (not consistently different from controls), and dropped less sharply to 64% during the 5th-8th week (difference significant vs. controls at P < 0.001); afterwards, during the 9th-12nd week of age, it went down to 27% (difference significant vs. controls at P < 0.0021, thus showing a clear-cut shift towards a later age vs. controls. Evolution of spindles followed an analogous trend (lower part of Fig. 3): i.e., 11% in controls vs. 8% in PKUs during the weeks up to the 4th (difference not consistent); 85% vs. 31% during the 5th-8th week (significant at P < 0.01); and 96% vs. 66% during the 9th-12th week (significant at P < 0.02). Biochemical data paralleled neurophysiological findings: the highest mean plasma Phe level was recorded during the 3rd week after birth (at that time, the mean level was 181 i 37 pmol/dl), then it dropped down sharply up until the 8th week of life (mean value 38 + 6 pmol/dl), when it stabilized in the range of 20-50 b T T AGE (weeks) Fig. 4. Time course of plasma phenylalanine and tyrosine levels, expressed as weekly mean in all PKU children. Abscissa: age in weeks from birth. Ordinate: mean plasma aminoacid level in pmol/dl, plus 1 S.E.M. (log scale). p,mol/dl. The other parameter considered in our group, namely the plasma Tyr mean level, was always in the 5-10 pmol/dl range, and it did not show any correlation to the Phe level or the neurophysiological parameters (Fig. 4). 4. DISCUSSION -L/, , , , , , , , , , , , , 2 4 6 6 10 12 14 Mean age (in weeks f 4.e.m.) Fig. 3. Upper part: time course of TA disappearance during the first weeks of life in PKU and control children. Lower part: time course of spindle maturation in PKU and control children. The difference between the curves is significant at least at the 1% level (Wilcoxon signed rank test, and Mann-Whitney (i-test). Abscissa: mean age in weeks, f 1 S.E.M. Ordinate: score computed for the same-age groups. In parentheses: the number of EEG tracings for each group in the upper as well as lower part of the figure. Asterisks indicate significant differences between groups of age: * P < 0.02; * * P < 0.01. Data drawn from our control population confirm previous findings in the literature. Ellingson and Peters [3,17] performed semiquantitative monitoring of maturational patterns of TA and sleep spindles in different physiological or pathological conditions. These authors observed complete TA disappearance (i.e., mature pattern) in 100% of their full-term (estimated gestation period 40 f 1 weeks) control group at the 7th week of age. In accordance with them, TA disappearance was observed at the 7th week of age in our controls. Moreover, they observed rolandic sleep spindle mature pattern in 100% of controls at the 9th week of age, vs. the 10th week in our control children. In a series of investigations, the same authors [3,17,18] found that both the TA and spindle pattern maturation were consistently accelerated (as in the premature children) or, vice-versa delayed (as in the trisomy 21 children) vs. the control group, following their peculiar physio-pathological condition. In our experimental conditions, the maturational pattern of both phenomena is retarded, and their evolution seems to follow the evolution of Phe plasma levels (Fig. 41, therefore we can suggest that plasma Phe normalization accelerates TA disappearance and spindle maturation in PKU infants. No evident alteration was observed for Tyr plasma levels during the same time, neither were there any correlations between Phe and Tyr levels. We must point out some possible biases contained in our procedures. (1) In the present paper we attributed a digital value to the EEG parameters. Such an approach may be a source of mistakes, since EEG patterns are considered mainly to be descriptive. G.F. De Giorgis et al. / Brain & Decelopment However, in our preliminary studies, several parameters were considered and analyzed before obtaining reproducible results. In our experimental conditions no parameters, even the ones derived from spectral analysis of EEG [19,20], turned out to be more reliable than the visual semi-quantitative EEG analysis performed independently by two experienced clinicians in controlled conditions and with the aid of behavioral serial observations, although others (e.g., [21] suggested more efficient algorithms for the automatic analysis of sleep pattern in newborns). (21 EEG recordings in our PKU children were performed according to an experimental design (longitudinal) which was different from the one adopted for controls (cross-sectional). However, the developmental trend of TA and spindles in our control children is almost identical to the one described by other investigators [3], who adopted a longitudinal design in their experiments for control children, and this would indirectly confirm our data. In any case, some pathophysiological suggestions may be inferred by observations from other laboratories and our studies. Since a delayed maturation of TA is accompanied by a delayed maturation of spindles and, vice-versa, an anticipated maturation of TA is accompanied by an anticipated maturation of spindles, we may suggest that both neurophysiological phenomena depend on the maturation of two biochemical backgrounds which are ‘programmed’ to proceed parallel. As an alternative hypothesis, we may suggest that both phenomena depend on a single biochemical background, whose evolutive steps may influence the maturation delay observed in both sleep patterns. These findings strongly suggest that the critical period in the development of nervous structure(s) which ‘produce’ TA and spindle patterns is substantially complete before birth. In fact, during intrauterine life, the fetus brain is protected by the mother’s enzymatic function from a hazard of an exceedingly high Phe plasma level. When a fetus is exposed to high Phe levels during intra uterine life (as it happens in the case of babies born from PKU mothers) the PKU child’s brain maturation is often severely affected [22]. In a previous paper [ 1 I] we found significant alterations of visual (F-VEPS) and short latency auditory potentials (BAEPs) in early treated PKU children during the first year of life. We showed that different components of evoked responses exhibited different maturational delays. according to the rate of their postnatal evolution. We suggested that the neurophysiological components still in their sensitive periods of development [9] received the most severe damage. On the other hand, in the cases described herein, we can suggest that the nervous structures which determine TA and spindle patterns are no longer in their critical period of maturation at birth: yet, their final evolution is sensitive to the toxic injury induced by high Phe levels in plasma and brain, and dietary therapy facilitates normal neurophysiological evolution. In another previous paper [7], we found a full TA in the EEG of one infant aged 11 weeks, recorded before dietary therapy. The EEG was again recorded 2 weeks later. i.e., 2 weeks after dietary therapy, and TA was no longer detected. This previously described case as well suggests that normalization of plasma Phe level after dietary therapy may facilitate favourable TA evolution. Moreover, in the same paper we showed a significant correlation between the latency (in days) to dietary therapy and the score of EEG epileptiform alterations during the first year of life in the same children. Therefore, a hypothesis suggesting a corre- 205 1996: 18: 201-206 lation also between sleep pattern maturation (as it occurs with other neurophysiological parameters, [7,11]) and plasma chemical alteration from birth to dietary therapy is worth consideration. This hypothesis could be further studied in the future by monitoring the sleep of children born from PKU mothers since evidently, during pregnancy PKU mothers cannot exert efficient Phe metabolism, due to their own enzymatic alteration, therefore their children’s brains do not receive sufficient environmental protection (i.e., they live in a high Phe level medium, [23]) as compared to the brains of children born to normal mothers. Other hypotheses and mechanisms cannot be discarded, for example those related to the unbalance of brain neuromediators following high Phe levels in plasma: a decreased synthesis rate of serotonin and dopamine has been shown in adolescent cerebrospinal fluid (CSF) after discontinuation of dietary treatment, together with enhancement of plasma and CSF Phe level [24]. Moreover, a dietary-induced reduction of endogenous brain serotonin level induces progressive disappearance of EEG sleep spindles in the experimental animal [25]. In experimental models of hyperphenylalaninemia serotonin (5-HTz) receptors decrease. together with dopamine (D2) receptors [26]. Although we cannot attribute the genesis of the spindle phenomenon to the modulation of one neuromediator activity alone (e.g., serotonin and/or dopaminel it is reasonable to suggest that unbalance of some neuromediators following high Phe levels in the brain may block the biochemical substrate of spindle formation in PKU newborns, till the unbalance is corrected through dietary therapy. Analogous hypothesis may be forwarded also for TA, although no papers have been published in this field, as it is zyxwvutsrqponmlkjihgfedcb in our knowledge. In conclusion, we suggest that studies on the critical maturational periods of different sleep components and of evoked potentials might provide a sensitive tool for early indication of delay of the development of some neurophysiological parameters during the first trimester of life in a population of children ‘at risk’. Our methods however cannot give useful indications for diagnostic purposes in the single child. On the other hand, studies in inborn errors of metabolism which induce for defining early sensitive physiological brain injury components may provide further indications for different during perinatal life. periods of development neuro- ACKNOWLEDGEMENTS Thanks are due to Dr. G.A. Zapponi for statistical help, to Ms. Paola Capriani, Ms. Manuela Luzi and Mr. Adrian0 Scuderani for their expert technical contribution, and Ms. Susan Holt for English revision. REFERENCES Dreyfus-Brisac C. The electroencephalogram of the premature infant and full-term newborn: normal and abnormal development of waking and sleeping patterns. In: Kellaway P. Petersen I. eds. Neurological and electroencephalographic correlatice .studies in i@ncy. New York: Grune and Stratton, 1964: 186-207. Parmelee AH Jr, Schulte FJ, Akiyama Y. Wenner WH, Schultz MA. Stem E. Maturation of EEG activity during sleep in premature infants. Electroencephalogr Clin Neurophysiol 1968; 24: 3 19-29. Ellingson RJ, Peters JF. Development of EEG and daytime sleep patterns in normal full-term infants during the first 3 months of life: Electroencephalogr C/in Neurophysiol longitudinal observation. 1980: 49: 112-24. 206 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA G.F. De Giorgis et al. /Brain & Development 1996; 18: 201-206 Grass A, Grass E. Glossary of terms used in electroencephalography. Quincy, MA: Grass Instrument Company, 1983. 5. Metcalf DR. EEG sleep spindle ontogenesis. Neuropaediatrie 1970; 1: 428-33. 6. Shibagaki M, Kiyono S, Watanabe K. Spindle evolution in normal and mentally retarded children: a review. Sleep 1982; 5: 47-57. 7. De Giorgis GF, Antonozzi 1, Del Caste110 PG, Rosano M, Loizzo A. EEG as a possible prognostic tool in phenylketonuria. Electroen4. cephalogr 8. 10. 11. 12. 13. 14. 15. 1980; and clinical Clin Neurophysiol 1990, 19. Giaquinto statistical S, Marciano F, Monod equivalence to newborn cephalogr Clin Neurophysiol neurophysiology, 76: 325-30. Louis J, Zhang JX, Revel M, Debilly G, Challamel MJ. Ontogenesis 165-71. Ellingson RJ, Peters JF. Development of EEG and daytime sleep patterns in trisomy-21 infants during the first year of life: longitudiElectroencephalogr Clin Neurophysiol 1980; 50: nal observations. In: Remond A, ed. Vol. 15B, Section 2. Amsterdam: Elsevier, 1972: 14-8. Lauder JM, Krebs H. Do neurotransmitters, neurohumors and hormones specify critical periods? In: Greenough WT, Juraska JM, eds. Deuelopmental neuropsychobiology. Orlando, FL: Academic Press, 1986: 119-74 Antonozzi I, Dominici R, Andreoli M, Monaco F. Neonatal screening in Italy for congenital hypothyroidism and metabolic disorders; hyperphenylalaninaemia, maple syrup urine disease and homocystinuria. J Endocr Invest 1980; 4: 357-63. Cardona F, Leuzzi V. Antonozzi I, Benedetti P, Loizro A. The development of auditory (BAEP) and visual (F-VEP) evoked potentials in early treated phenylketonuric children. Electroencephalogr 1991; 80: 8-15 Clin Neurophysiol Sterman MB, Harper RM, Havens B, Hoppenbrouwers T, McGinty DJ, Hodgman JE. Quantitative analysis of infant EEG development during quiet sleep. Electroencephalogr Clin Neurophysiol 1977; 43: 371-85 Coons S. Development of sleep and wakefulness during the first 6 months of life. In: Guilleminault C, ed. Sleep and its disorders in children. New York: Karger, 1987: 17-27. Hoppenbrouwers T. Sleep in infants. In: Guilleminault C, ed. Sleep and its disorders in children. New York: Karger, 1987: 1-15 Kohyama J, Iwakawa Y. Developmental changes in phasic sleep parameters as reflections of the brain-stem maturation: polysomnographical examinations of infants, including premature neonates. 50: 18. 55: 60-8. of electroencephalography Electroencephalogr 16. 1983; Metcalf DR. EEG in inborn errors of metabolism. Handbook 9. Clin Neurophysiol of nocturnal organization of sleep spindles: a longitudinal study Clin Neurophvsduring the first 6 months of life. Electroencephalogr iol 1992, 83: 289-96. 17. Ellingson RJ, Peters JF. Development of EEG and daytime sleep patterns in low risk premature infants during the first year of life: Electroencephalogr Clin Neurophysiol longitudinal observations. 457-66. 1977; N, Nolfe G. Applications of EEG recordings. Electroen42: 406-13. Gallozzi G, De Feo MR, Furio M, Loizzo A, Ricci GF, Spallina R. Criteria of univariate statistical analysis applied to computerized EEG for the classification of sleep-wakefulness cycle in prematures (in Italian). Ril: Ital EEG Neurofisiol Clin 1979; 2: 417-23. 21. Barlow JS. Computer charachterization of trace altemant and REM sleep patterns in the neonatal EEG by adaptive segmentation - an Clin Neurophysiol 1985; 60: exploratory study. Electroencephalogr 20. 163-73. Striver CS, Kaufman SK, Woo SLC. The hyperphenylalaninemias. In: Striver CS, Beaudet AL, Sly WS, Valle D eds. The metabolic basis of inherited diseases. New York: McGraw-Hill, 1989: 495-546. 23. Koch R. Levy HL, Matalon R et al. The international collaborative study of maternal phenylketonuria: status report 1994. Acta Paediatr 22. 1994; 24. 1I I-9. 407 Suppl: Lou H C, Glittler F, Lykkelund C, Bruhn P, Niederwieser A. Decreased vigilance and neurotransmitter synthesis after discontinuation of dietary treatment for phenylketonuria in adolescents. Eur J Pediatr 1985: 144: 17-20. Lanoir J, Ternaux JP, Pons C, Lagarde JM. Long-term effects of a tryptophan-free diet on serotonin metabolism and sleep-waking balance in rats. Exp Brain Res 1981; 41: 346-57. 26. Hommes FA. The effect of hyperphenylalaninaemia on the muscarinic acetylcholine receptor in the HPH-5 mouse brain. J Znher 25. Metab Dis 1993: 16: 962-74.