Electroencephalography and clinical Neurophysiology, 84 (1992) 32-43 © 1992 Elsevier Scientific Publishers Ireland, Ltd. 0168-5597/92/$05.00
32
E V O P O T 91030
The late somatosensory evoked potential in premature and term infants. I. Principal component topography Walt Karniski Department of Pediatrics, MDC Box I5, University of South Florida College of Medicine, Tampa, FL 33612 (U.S.A.) (Accepted for publication: 3 June 1991)
Summary Very little is known about the topographic distribution of the cortical somatosensory evoked potential in premature infants. Principal component analysis (PCA) was applied to the wave forms generated from right median nerve stimuli over a relatively long sweep (483 msec post stimulus) at 16 electrodes in 53 infants with postconceptual ages from 31 to 40 weeks, subdivided into 5 groups by 2 week increments. Factor scores were averaged across subjects, within groups and displayed as topographical maps. Four factors accounted for 7 1 - 7 6 % of the variance in each of the 5 groups and the factors extracted from the P C A performed independently in each group were markedly consistent. The first factor ( N 1 / P 1 ) had a left posterior m i n i m u m and a left frontal-central maximum and probably represents a tangential dipole located in the post-central gyrus. The second factor (N2) was characterized by a consistent left central m i n i m u m with a systematic developmental change in the maximum that seemed to imply that its neural generator was changing in orientation as the infants matured. A third factor (N3) accounted for the most variance and appeared to represent the first evidence of activity in the ipsilateral cortex. Finally, a very late fourth factor appeared only in the more mature groups, with uncertain localization. The topographic maps of the factor scores for these 4 factors appear to account for independent generators in the SEP of the premature and term infant. Key words: SEP; Topography; Premature; Infants
Multiple peaks occur in the first 20 msec following a somatosensory stimulus in normal adults and each of these peaks appears to correspond to the activation of neural structures in the peripheral somatosensory pathway (see Desmedt 1988 for a review). For example, the central maximum negative peak occurring at approximately 20 msec (N20) represents the arrival of the peripheral afferent volley at the primary somatosensory cortex in the post-central gyrus (Broughton 1969; Allison et al. 1980; Wood et al. 1985; Allison et al. 1989a). Most studies of the SEP in full-term and young infants have focused on the N20, usually referred to as the N1, and the developmental changes of both latency and amplitude have been well described (Desmedt and Manil 1970; Desmedt and Debecker 1972; Cullity et al. 1976; Desmedt et al. 1976; Laget 1982; Hashimoto et al. 1983; Willis et al. 1984; Lauffer and Wenzel 1986; Tomita et al. 1986; Bartel et al. 1987; Zhu et al. 1987; Laureau et al. 1988; Taylor and Fagan 1988; George and Taylor 1991; Karniski et al. 1992). However, the later components of the SEP, presumably reflecting
Correspondence to: Walt Karniski, M.D., D e p a r t m e n t of Pediatrics, M D C Box 15, University of South Florida College of Medicine, 12901 Bruce B. Downs Blvd., Tampa, FL 33612 (U.S.A.).
cortical activity (Allison et al. 1989b), have generally received much less scrutiny. The SEP in premature infants has received infrequent attention (Hrbek et al. 1973; Gallai et al. 1986; Klimach and Cooke 1988; Bongers-Schokking et al. 1990) and these studies have also focused on the N1 or preceding peaks. Furthermore, because only a few electrodes have been used in most of these studies, little is known about the topographic distribution of the SEP. Thus, this study is an examination of the topographic distribution of the SEP in premature infants, beginning with the N1 and continuing into the time period in which cortical activation predominates.
Methods
Subjects Fifty-three infants were recruited from the Neonatal Intensive Care Unit and Newborn Nursery at Tampa General Hospital. In order to insure that they were relatively free of neurologic disorders, infants were included only if they met all of the following criteria: (1) 1 and 5 min Apgar scores of at least 4 and 7 respectively, (2) length, weight, and head circumference appropriate for gestational age, (3) absence of major congenital malformations, (4) absence of known
SEP PC TOPOGRAPHY IN PREMATURE INFANTS
33
neurologic abnormalities or evidence of seizure disorder, (5) normal cranial ultrasound in infants less than 34 weeks ~estation at birth or less than 1500 g birth weight, (6) absence of asymmetric limb lengths, and (7) stable clinical condition without assisted ventilation or oxygen requirements. Parental consent was obtained for all infants. Because of the importance of an accurate assessment of post-menstrual age (PMA) for this study, an infant was included only if the post-menstrual age as assessed by examination (Dubowitz et al. 1970) was within 2 weeks as estimated by history (date of mother's last menstrual period). The PMA at birth ranged from 28 to 40 weeks. Infants were grouped by PMA in 2 week increments (31-32 weeks, 33-34 weeks, 35-36 weeks, 37-38 weeks, 39-40 weeks). Post-menstrual age is defined as the gestational age at birth plus chronological age at the time of testing and thus represents the time that has elapsed since the mother's last menstrual period. Because the time of the last menstrual period usually antedates conception by 2 weeks, the PMA is approximately 2 weeks longer than the time from conception. Thus, the PMA represents a maturational age that is more relevant than chronological age. At the time of testing, the infants' chronological ages ranged from 1 to 65 days, with a mean of 13 days (Table I). There was no difference in chronological age across PMA groups ( F (4, 48) = 1.33, P = 0.21). Black and white infants were approximately equally distributed across all groups. The male-to-female ratio was approximately equal for the entire group of infants (23:30), but the 31-32 week and the 39-40 week groups had more males, whereas the middle 3 groups were predominantly female (Fisher's exact test, P = 0.02). Although there was a tendency for Apgar scores of the less mature infants to be somewhat lower, the Apgar scores were generally not significantly different across groups at 1 min ( F (4, 48) = 2.43, P = 0.06) or 5 min of age ( F (4, 48) = 2.05, P = 0.10).
Procedures All infants were tested during sleep. In the very premature infants, sleep stage was difficult to deter-
mine so that sleep stage was not formally evaluated. Environmental temperature and body temperature were maintained constant throughout the test period. Stimuli consisted of median nerve stimulation at the right wrist, with a stimulus duration of 100 msec and a variable frequency of 0.9-1.7/sec. Stimulus intensity was set below sensory threshold, then gradually increased until a slight thumb twitch was observed. The intensity was then reduced until the twitch was barely visible. Two blocks of 300 stimuli each were presented to each infant, 15 rain apart, then the two blocks were averaged together. Single trials with voltage in excess of a pre-set level for each infant were automatically excluded from subsequent averaging. Sixteen pediatric gold electrodes (5 mm in diameter) were used with a linked ears reference at the following sites: Fpl, Fp2, F3, F4, F7, F8, C3, C4, T3, T4, T5, T6, P3, P4, O1 and 02. Many other SEP studies have used electrodes posterior to the central electrodes in an attempt to locate the electrode directly over the somatosensory cortex (see Picton 1978). While this might be appropriate for N1 which is most likely generated there, there is no agreement about the source of the 3 components occurring after N1 in the pre-term infants described in this study. Thus the more standard 10-20 system electrode positions were used. A Cadwell 8400 was used for signal generation and data acquisition. Impedance was maintained below 5000 I2. The sweep extended from 50 msec prior to the stimulus to 483 msec after the stimulus. Voltage was sampled at a frequency of 960 Hz. An analog high-pass filter of 1 Hz was used. Desmedt et al. (1974) have recommended a low-pass filter of 3000 Hz in studies of the SEP in adults, in order not to miss the fast transients in the N1. Although this may be valid for adults, the N1 and especially the later components of interest in this study are of a much slower frequency. Thus, we used a low-pass filter of 100 Hz as recommended by BongersSchokking et al. (1989). In another study of these same infants (Karniski et al. 1992), this filter setting allowed for easy identification of an N1 in most infants and made it easier to identify the later peaks which were of greater interest.
TABLE I Group characteristics. Post-menstrual age group (weeks)
No.
Sex (% male)
Race (% Black)
Apgar (1 min)
Apgar (5 min)
Mean chron, age at testing (days)
31-32 33-34 35-36 37-38 39-40
13 14 9 9 8
76.9 28.6 22.2 22.2 62.5
46.2 50.0 44.4 44.4 12.5
7.6 6.5 7.7 7.5 8.5
8.3 8.2 8.5 8.5 9.1
13 19 13 10 8
Total (or avg.)
53
43.4
41.5
13
34
Data analysis Any analysis of evoked potentials must begin with the assumption that they are produced by multiple neural generators with independent locations, orientations, onsets and durations. Principal component analysis (PCA) is often successful in identifying a relatively small number of independent factors that account for a large majority of the variance of the evoked potential across a group of subjects. These independent factors have been shown to reflect orthogonal generators within the brain (Donchin and Heffley 1978; Skrandies and Lehmann 1982; Lehmann 1987). Each factor is represented by a factor loading at each variable (time point). By plotting the factor loadings across time for each factor, it can be determined that a given factor is "active" during a time segment when the factor loadings are maximum. A factor score is generated for each factor, for each subject and for each electrode and these factor scores are measures of how much that condition (electrode) is weighted on its corresponding factor. By averaging the factor scores across subjects within a group for each electrode, a topographical map of the factor scores is generated which represents the scalp topography of each factor. PCA with varimax rotation was performed with 512 time points as the variables over the 533 msec sweep. Five separate PCAs were performed, once for each of the 5 PMA groups. Factor loadings for the first 6 factors were graphed as wave forms. Factor scores averaged for each electrode across subjects within a group were displayed as topographic maps and then matched with their appropriate factor loading wave forms. Factor loading wave form and factor score map pairs were then sorted so that times of peak loading, shape of the loading wave form and factor score maps were all similar within a sorted PMA group. If a factor did not produce a factor loading wave form with a peak that approached or exceeded 0.5 or - 0 . 5 , it was discarded.
Results
Wave forms and voltage maps The temporal and topographic characteristics of the SEP in premature infants are best understood by si, multaneously examining the wave forms (Fig. 1) and voltage topographic maps (Fig. 2). The first 15 msec of the sweep in each P M A group consists of a stimulus artifact and far-field components that were not relevant to this study and will not be discussed further. The first negative deflection after 15 msec appears biphasic in all P M A groups, primarily in the central and parietal electrodes. The first of these two peaks is maximal near 25 msec in the 31-32 week group and decreases
W. KARNISKI
to approximately 20 msec in the 39-40 week group. It appears to have the same amplitude across all electrodes, producing a relatively flat voltage map for each group, indicating that this peak probably represents a far-field potential. Near 25 msec in all groups, the left and right central and parietal electrode wave forms begin to diverge, and this divergence indicates the beginning of the N1. In the topographical maps, the N1 occurs as a negative minimum in the left central electrodes starting at 51 msec in the 31-32 week group and in the left posterior temporal and parietal electrodes starting at 51 msec, 46 msec, 36 msec and 31 msec in the last 4 groups respectively. A small positive peak follows the N1, becoming larger and earlier with maturation. However, the topographical maps reveal that the maximum actually begins much earlier than it appears in the wave forms, very near to the onset of the N1. In the 31-32 week group, it initially appears ipsilateral to the stimulus on the right at 51 msec, while for the other 4 groups, it begins left frontal (61, 46, 36, and 31 msec). The maximum then changes differently for each group, but these changes progress systematically across groups. For the first 2 groups the maximum becomes bifrontal at 120 msec and 110 msec, for the third group the maximum becomes bifrontal and more central near 90 msec, and for the 2 older groups the maximum remains unilateral, but becomes more central (80 and 61 msec). For the first 2 groups, this topographic change occurs coincident with the onset on the N2, while in the latter 3 groups it occurs earlier than the N2 onset. However, in each of these 3 more mature groups, the maximum changes again with N2 onset. These topographic changes in the maximum would suggest at least 2 subcomponents for the P1, with the first subcomponent (referred to here as the Pla) temporally associated with the N1. The most prominent feature of the wave forms is the large second negative component (N2) between 100 and 300 msec. It is maximal at C3 for all groups and with maturation it decreases in amplitude, latency and duration. The early part of the N2 has an associated frontal maximum in all 5 groups, which is topographically distinct from the P l a described above. However, while the minimum appears stable in location at C3, the maximum "moves" from a bilateral frontal location to a right temporal focus for the 31-32 and 35-36 week groups, to a right temporal-parietal focus for the 33-34 week group, and to a right central focus for the 37-38 and 39-40 week groups. Eventually, in all of the groups, the maximum is seen on the right side, ipsilateral to the stimulus. Although this maximum remains long after the N2 has apparently disappeared in the first 2 groups, the right central maximum seems to fade or change concurrently with the N2 in the latter 3 groups.
SEP PC T O P O G R A P H Y IN P R E M A T U R E I N F A N T S
35
RIGHT MEDIAN NERVE SEP CENTRAL
FRONTAL
PARIETAL
PMA 0N'KS)
.
31-32
...................
aa-a4
---,..........,...._,.-
35-36
37-38 N1
N2
39-40 ELECTRODES LEFT--
I
.
RGHT. . . . . . .
0
100
.
. 200
.
l"vl 300
400
0
P1 . 100
.
. 200
P2 .
300
400
I
.
0
100
.
. 200
. 300
400
Fig. l. G r a n d average wave forms for left and right frontal, central and parietal electrodes to right median nerve stimuli for each P M A group. T h e developmental evolution of 4 peaks can be seen from 31 weeks P M A to 40 weeks. N1, P1, N2 and P2 all show decreasing latency with maturation. A third low amplitude negative wave near the end of the sweep is suggested. (Linked ear reference, negative up.)
The wave forms indicate that the N2 simply trails off in all the groups, but the topographical maps reveal a third negative, low amplitude component. For all of the groups, the amplitude is more negative on the right and on the maps it appears as a right sided minimum emerging after the N2 (405, 390, 360, 300 and 300 msec in the 5 groups respectively). This N3 appears more frontal for the more immature groups and more central for the 37-38 and 39-40 week groups. It is still present at the end of the sweep in all groups.
Principal component analysis Three important, general features emerged from the PCA. First, the maps of the factor scores have a close correspondence to certain persistent or prominent features of the voltage topographical maps. Secondly, this feature extraction results in a significant reduction of data. In each application of the PCA, the first 4 factors account for 71-76% of the total variance. Third, although there are differences in the PCAs performed across the groups, there is a very strong consistency in
Fig. 2. Voltage topography for right median nerve SEP in p r e m a t u r e and term infants. T h e temporal evolution of the right median nerve SEP topography can be seen. Maturational effects are seen as earlier onset and shorter duration for certain topographic features as infants mature. Systematic topographic changes with maturation can be seen as well. (All maps were created using the voltage from 16 electrodes referenced to linked ears. Midline electrode values were interpolated. Maps are shown at 5 msec increments until 70 msec, at 10 msec increments to 150 msec and then at 15 msec subsequently.)
36
W. KARNISKI
PMA (wks)
31-32
33-34
35-36
37-38
39-40 LATENCY (ms) 21
165
180
195
26
31
36
41
46
51
56
61
210
225
240
255
270
285
300
315
SEP PC T O P O G R A P H Y IN P R E M A T U R E INFANTS
37
1.6
luv
.8 66
70
80
90
100
110
120
130
140
150 0
-.8
-1.6
330
345
360
375
390
405
420
435
450
465
38
W. KARNISKI
TABLE II Factor number and percent of variance for each factor in Fig. 3. Factor Range (msec) Peak (msec) 31-32 weeks 33-34 weeks 35-36 weeks 37-38 weeks 39-40 weeks
N1/P1 30-66 47
N2 N3 ? 139-181 155-483 345-? 160 287 419
3 6.4% 4 6.6% 3 12.2% 3 14.5% 6 5.2%
2 21.3% 2 20.4% 4 6.4% 2 16.1% 3 11.0%
1 43.3% 1 33.5% 1 30.3% 1 34.4% 1 37.3%
2 24.4% 2 13.8%
the results obtained across each group. This consistency lends credence to the hypothesis that the resultant factor score maps represent spatially and temporally distinct E R P components (Donchin and Heffiey 1978). The factors in Fig. 3 have been ordered chronologically from left to right rather than by the order in which the factors emerged. The order in which they emerged, the time range of peak activity and the amount of variance for which they account are shown in Table II. The time range of maximal activity was determined by averaging the factor loading wave forms across all groups, then identifying the time period in which the average loadings were above 0.5. This generates an artificial time range that is only approximate for all the groups, because the less mature infants had factors with somewhat longer latencies, while the latencies of the factors in the older infants were less. The first factor after the stimulus was active between 30 and 66 msec. It was present in all 5 groups and accounted for 5 - 1 5 % of the total variance from each PCA. Factor score topography was very consistent across groups, with a left posterior minimum and a left frontal-central maximum for all groups except for the 31-32 week group. In these very premature infants, the minimum was more central with the maximum more frontal and slightly more prominent on the right. The second factor had factor loadings that were active between 139 and 181 msec, accounting for 6 - 2 1 % of the total variance. The factor score topography showed a left central minimum and a right central maximum for the first 3 groups, while the last 2 groups
each had 2 different maximums in the right frontal and bilateral occipital areas. The third factor emerging from each group accounted for the greatest amount of variance in each separate P C A (30-43%) and covered a broad range from 155 to 480 msec. The maximum is in the left central area in the 31-32 week group and becomes more lateral as the infants mature. More importantly, if one scans down this third column, a systematic change across groups is clearly visible. The more prominent right frontal minimum and the less prominent right occipital minimum both move centrally, meeting as a single topographic minimum at C4 by the time the infants reach a full-term equivalent. A fourth factor emerged in the very late part of the SEP wave form only for the 35-36 and 39-40 week groups beginning near 345 msec and this factor was still active at the end of the sweep at 483 msec. It accounted for a relatively large amount of variance (14-24%). Although both factors have a left central minimum, the 35-36 week group has a maximum in the right occipital area. On the other hand, the older group has a maximum which is well circumscribed to the right central area, with a second smaller maximum in the left occipital area.
Discussion A traditional analysis of the SEP wave forms in this study indicates that 4 major peaks occur between 30 and 450 msec, after the primary somatosensory cortex is first activated (N1, P1, N2 and P2). These peaks consistently decrease in latency with maturation, as shown in other studies of the first negative peak (N1). A more systematic study of the peak latency and amplitude changes with maturation in premature infants can be found in a separate study of these same children (Karniski et al. 1992). However, when peaks overlap temporally, it is difficult to distinguish whether the overlap is due to multiple generators that are simultaneously active or whether the two "peaks" represent opposite ends of a single dipole. It was not the purpose of this study to localize these apparent dipoles nor the generators associated with them. However, the factor score maps generated by the P C A (Fig. 3) do seem to extract the most prominent features of the serial topographic maps (Fig.
Fig. 3. Principal components for right median nerve SEP in premature and term infants. Principle component analysis was performed separately for each PMA group with 512 points along a 533 msec sweep for 16 electrodes as variables. Factor loadings are displayed as wave forms in the bottom row, with a topographic map of factor scores displayed above. (Midline electrode values were extrapolated.) Although the factor loading wave forms appear very different across groups, the peaks tend to occur in the same approximate time periods, with a tendency for shorter latencies for the more mature infants. The first 4 factors accounted for 71-76% of the total variance in each PCA performed for each PMA group. The factor loading wave form was matched with its appropriate factor score map, then factor loading wave forms were sorted chronologicallyby peak loading times. A factor was identified if factor loading wave forms and factor score maps were similar across groups.
SEP PC TOPOGRAPHY IN PREMATURE INFANTS
39
MAPS OF FACTOR SCORES FROM PRINCIPAL COMPONENT ANALYSIS PMA (wks)
N1/P1
N2
N3
31-32 1.0
~v
33-34
,5
35-36
0
37-38
-.5
-1.0
39-40
05
FACTOR LOADINGS
0
-05
31-32 33-34 35-36 37-38 39-40
........ -1.0
...... ...... - - -
0
100
2OO
30-66
3OO
40O
0
100
2OO
3o0
40O
0
139-181
IO0
2oo
3OO
155-483
4OO
0
IO0
2OO
3OO
4OO
345-?
Peaks(ms) ZWART
40
2). Furthermore, these factor maps appear to have topographic analogs in studies of the SEP in adults. The first factor emerging from the PCA has a left posterior minimum and a left frontal-central maximum for all groups except the 31-32 week infants. By the temporal distribution of the factor loadings and the topography of the factor scores, this factor appears to represent the N 1 / P l a described above in the analysis of the voltage maps. In topographic studies of the SEP in adults, an N20 is localized to the contralateral posterior temporal scalp area (Desmedt and Cheron 1980; Desmedt and Bourguet 1985; Tsuji and Murai 1986; Desmedt et al. 1987; Tsuji et al. 1988a). At the same time, these studies disclose a P20 with a bilateral frontal distribution that is slightly more prominent on the side contralateral to the stimulus. Since the first factor extracted by PCA in this study (N1/Pla) embodies the same topography, it is probably analogous to the adult N20/P20. It is now fairly well established that the N20/P20 represents a single, tangentially oriented dipole located in area 3b of the post-central gyrus (Broughton 1969; Allison 1980; Maugui~re et al. 1983; Slimp et al. 1986; Allison et al. 1989a). The topographic studies of the SEP in adults also reveal a P22 that follows and often overlaps the N20/P20. A radially oriented dipole in area 4 of the pre-central gyrus has been proposed as the generator for this component (Desmedt and Bourguet 1985; Desmedt et al. 1987). The wave forms in this study indicate that a distinct positive peak follows the N1. However, in the maps, the early part of the maximum (Pla) is clearly associated with the N1, while the later part of this maximum (Plb) develops parallel with the N2. The maximum for the 35-38 week infants shifts more central before the onset of the N2 and this may represent the "true" P1. However, PCA does not reveal a separate factor that is either active in the time period between the N1 and N2 or that would account for the voltage topography in this time period. Although a "true" P1 that is not part of N1 or N2 dipoles may exist in this time period in premature infants, this study does not provide convincing evidence for it. However, in a separate study of the effects of maturation on latency, amplitude and topography in the same infants in this study (Karniski et al. 1992), the P1 was found to have a different voltage topography from the N1 and the P1 amplitude and latency were found to mature at different rates than for the N1, providing some evidence that N1 and P1 have different neural generators. An important caveat is necessary whenever PCA is utilized to draw conclusions about independent generators of evoked potential peaks. Independent factors emerge from PCA only when an experimental manipulation produces a systematic change in voltage amplitude (Donchin and Heffiey 1978). In this study, differ-
W. KARNISKI
ent maturational levels provide the experimental manipulation. The fact that the N1 and Pla appear on factor score maps simultaneously provides evidence, but not proof, for their lack of independence. If some experimental manipulation influences only one of the peaks and not the other, only then will they emerge as separate factors. The weakness of PCA is that one can never test for all possible experimental variables and therefore this lack of independence can never be proven through PCA. Desmedt and Manil (1970) described a second positive peak appearing between the N1 and N2, and after the P1, only during slow wave sleep in full-term infants. This pattern was seen in a only a few infants and is not visible in the grand average. This is a potential but unlikely explanation for the change in the topography of the P1, because separate components did not emerge from the PCA. The second factor has a left central minimum and a maximum that varies systematically with maturation. This factor most likely represents the N2 and this is supported by at least two findings. In the few studies of premature infants that examined the N2 (Desmedt and Manil 1970; Hrbek et al. 1973; Gallai et al. 1986) the latency of the N2 was similar to that found in this study (139-181 msec). In addition, the latency and topography (left central minimum) of this second factor are similar to the latency and topography of the N2 as seen in the voltage maps from this study. However, the topography of the second factor varied with maturation. In each group, the minimum was centrally located, but the maximum shifted. In the 31-32 week group, the maximum is right temporal. In the 33-36 week infants it becomes right central and becomes bi-peaked in the 37-40 week infants. If a single generator is assumed for each factor, the very premature infants' maps could be explained by a tangential dipole, while the older infants' maps would more likely suggest a radial dipole. It is possible that as the brain of the premature infant matures, the orientation of some generators would shift or rotate in space. The cortex of the 25 week fetus is lissencephalic, or almost completely smooth with little evidence of fissures or gyri. The sylvian fissure is evident in the third month post conception, but the central sulcus is barely visible until 6 months (Bailey and Miller 1927; Patten 1968; Crelin 1974). Over the next 15 weeks, the fetal brain progresses through extensive structural, microscopic and functional changes. The central sulcus deepens and the pre-central gyrus and post-central gyrus (containing the motor and somatosensory cortices respectively) become visible and fold in. As the association areas of the frontal lobe increase in size, the central sulcus moves more posterior. Myelin begins to develop around neuronal axons, neuronal cells differentiate, migrate to the surface of the cortex and begin to establish multiple
SEP PC TOPOGRAPHY IN PREMATURE INFANTS cell-to-cell connections (Herschkowitz 1988; Mrzljak et al. 1988), Thus it is not surprising that these structural changes would be reflected in a change in orientation of some of the dipoles that generate SEP components and that appears to be what is happening with this factor. The analog of the N2 in this study to the SEP in adults is difficult to determine. A large negative peak following the adult N 2 0 / P 2 0 and P22 has been found as early as 24 msec by Tsuji et al. (1988a), at 26 msec by Tsuji et al. (1988b) and at 30 msec by Desmedt and colleagues (Desmedt and Bourguet 1985; Desmedt et al. 1987). This peak is much larger in amplitude than the peaks preceding it and, regardless of latency, each of these peaks has a pre-central and frontal topography. Desmedt and Bourguet (1985) proposed a source in the supplementary motor area for this peak. Allison et al. (1989a) studied the SEP in normal adults with a small patch of 64 electrodes over the sensorimotor cortex and with transcortical electrodes. They describe a P25-N35 wave that has a maximum over the postcentral gyrus, medial to the locus of the earlier N 2 0 / P 2 0 and they propose a radial dipole located in area 1 of the post-central gyrus. In a second study Allison et al. (1989b) examined peaks occurring in the time period from 40 to 250 msec. They described a P45-N80-P180 with a probable generator in area 3b of the post-central gyrus and a P50-N90-P190 with a probable generator in area 1 of the somatosensory cortex. The N2 factor in the current study appears analogous to the N 2 4 / N 2 6 / N 3 0 peaks described above. Its topography is consistent with a radially oriented dipole in the post-central gyrus, similar to that described by Allison et al. (1989b), for the 37-40 week infants only. It is postulated that both functional and structural differences in the less mature infants produce the difference in topography. A longitudinal or cross-sectional study, which follows the development of this factor as the SEP matures into its adult form, would help answer this question. The third factor in Fig. 3 was actually the first factor to emerge in each of the PCAs and thus accounted for the largest proportion of variance in each analysis. It is active in the same time period as the very slow, third negative deflection that is seen in the very late part of the wave forms and thus will be referred to as the N3. It has not been previously described, because most studies have not used post-stimulus sweeps as long as those employed in this study (483 msec). Although the origin of this component cannot be determined from this study alone, it does appear to have a minimum ipsilateral to the side of stimulation, especially in the 39-40 week group. In the wave forms, the ipsilateral electrodes seem to indicate this peak near the end of the sweep, while the contralateral electrodes produce a wave form that appears relatively flat in the same time
41 period. Thus this component seems to represent activity in the opposite hemisphere from that of the earlier components and thus ipsilateral to the stimulus. If so, it represents the first evidence of activation of the ipsilateral hemisphere in the premature SEP. Finally, an additional factor emerged in the 35-36 and 39-40 week groups. Both groups show a left-sided minimum at C3 and a right-sided maximum, but the topographies of these two factors are actually rather different. Because this factor represents such a large amount of variance, it is difficult to dismiss, and yet there is no analog in the voltage maps or in the wave forms that would help identify this factor. All previous studies of the SEP in premature infants have focused on either the early, peripheral far-field components or on the N1, which from studies in adults is presumed to represent the afferent volley arriving at the cortex in the post-central gyrus. However, the central nervous system activity that occurs after this point has generally been ignored in premature infants. In this study, the voltage peaks and topography of the SEP were examined over a much longer period of time after stimulus (483 msec) than in other studies of premature infants. Principal component analysis extracted 3 major factors. The first factor had a contralateral parietal-temporal minimum and a contralateral frontal-central maximum analogous to the adult N20/P20. The second factor had a contralateral central minimum and a variable maximum across groups. It appeared to be analogous to the N30 described by Desmedt and colleagues and the N 2 4 / N 2 6 described by Tsuji and colleagues. A third factor with an ipsilateral minimum may represent previously undescribed activity of the ipsilateral cortex. These factors represented the predominant features of the voltage topographic maps and they were consistently replicated in each of the 5 PMA groups. Systematic topographic changes in these factors were seen with maturation, which appear to reflect functional and structural changes as the infants mature. The author wishes to thank Loretta Lease, R.EEGT, for her attention to detail in data acquisition, Richard Weibley, M.D., and Lance Wyble, M.D., for their identification of infants qualifying for this study, Rodney Vanderploeg, Ph.D., for his review of an earlier version of this paper, R. Clifford Blair, Ph.D., for assistance in data analysis and Roger Daniels for assistance in the Neonatal ICU. A special thanks is extended to Dietrich Lehmann, M.D., who supported the author's sabbatical and stimulated thinking that lead to the final form of this paper. This research was funded in part by the State of Florida, Department of Education, Florida Diagnostic and Learning Resource System and Children's Medical Services, State of Florida.
References
Allison, T., Goff, W.R., Williamson, P.D. and Van Gilder, J.C. On the neural origin of early components of the somatosensory
42 evoked potential. In: J.E. Desmedt (Ed.), Clinical Uses of Cerebral, Brainstem and Spinal Somatosensory Evoked Potentials. Prog. Clin. Neurophysiol., Vol. 7. Karger, Basel, 1980: 51-68. Allison, T., McCarthy, G., Wood, C.C., Darcey, T.M., Spencer, D.D. and Williamson, P.D. Human cortical potentials evoked by stimulation of the median nerve. I, Cytoarchitectonic areas generating short-latency activity. J, Neurophysiol., 1989a, 62: 694-710. Allison, T., McCarthy, G., Wood, C.C., Williamson, P.D. and Spencer, D.D. Human cortical potentials evoked by stimulation of the median nerve. II. Cytoarchitectonic areas generating Iongla'~ency activity. J. Neurophysiol., 1989b, 62: 711-722. Bailey, F.R and Miller, A.M. Text-book of Embryology. William Wood and Co., New York, 1927. Bartel, P., Conradie, J., Robinson, E., Prinsloo, J. and Becker, P. The relationship between median nerve somatosensory evoked potential latencies and age and growth parameters in young children. Electroenceph. clin. Neurophysiol., 1987, 68: 1811-186. Bongers-Schokking, C.J., Colon, E.J,, Hoogland, R.A., Van den Brande, J.L. and De Groot, C.J. The somatosensory evoked potentials of normal infants: influence of filter bandpass, arousal state and number of stimuli. Brain Dev., 1989, 1l: 33-39. Bongers-Schokking, J.J., Colon, E.J., Hoogland, R.A., Van den Brande, J.L. and De Groot, C.J. Somatosensory evoked potentials in term and preterm infants in relation to postconeeptual age and birth weight. Neuropediatrics, 1990, 21: 32-36. Broughton, R.J. In: E. Donchin and D.B. Lindsley (Eds.), Average Evoked Potentials. U.S. Government Printing Office, Washington, DC. NASA SP-191, 1969: 79-84. Cretin, E.S. Development of the nervous system: a logical approach to neuroanatomy. Clin. Symp., 1974, 26: 1-32. Cullity, P., Franks, C.I., Duckworth, T. and Brown, B.H. Somatosensory evoked cortical responses: detection in normal infants. Dev. Med. Child Neurol., 1976, 18:11-18. Desmedt, J.E. Somatosensory evoked potentials. In: T.W. Picton (Ed.), Human Event-Related Potentials. Elsevier, Amsterdam, 1988. Desmedt, J.E. and Bourguet, M. Color imaging of parietal and frontal somatosensory potential fields evoked by stimulation of median or posterior tibial nerve in man. Electroenceph. clin. Neurophysiol., 1985, 62: 1-17. Desmedt, J.E. and Cheron, G. Somatosensory evoked potentials to finger stimulation in healthy octogenarians and in young adults: wave forms, scalp topography and transit times of parietal and frontal components. Electroenceph. clin. Neurophysiol., 1980, 50: 404-425. Desmedt, J.E. and Debecker, J. The somatosensory cerebral evoked potentials of the sleeping human newborn. In: C.D. Clemente, D.P. Purpura and F.E. Mayer (Eds.), Sleeping and the Maturing Nervous System. Academic Press, New York, 1972: 229-239. Desmedt, J.E. and Manil, J. Somatosensory evoked potentials of the normal human neonate in REM sleep, in slow wave sleep and in waking. Electroenceph. clin~ Neurophysiol., 1970, 29: 113-126. Desmedt, J.E., Brunko, E., Debecker, J. and Carmeliet, J. The system bandpass required to avoid distortion of early components when averaging somatosensory evoked potentials, Electroenceph. clin. Neurophysiol., 1974, 37: 407-410. Desmedt, J.E., Brunko, E. and Debecker, J. Maturation of the somatosensory evoked potentials in normal infants and children, with special reference to the early NI component, Electroenceph. clin. Neurophysiol., 1976, 40: 43-58. Desmedt, J.E., Nguyen, T.H. and Bourguet, M. Color imaging of human evoked potentials with reference to the N20, P22, P27 and N30 somatosensory response. Electroenceph. clin. Neurophysiol., 1987, 68: 1-22. Donchin, E. and Heffley, E. Multivariate analysis of event-related potential data: a tutorial review. In: D.A. Otto (Ed.), Multidisciplinary Perspectives in Event-Related Brain Potential Research,
W. KARNISK1 EPA-600/9-77-043, U.S, Environmental Protection Agency, Washington, DC, 1978: 555-572. Dubowitz, L.M.S., Dubowitz, V. and Golbert, C. Clinical assessment of gestational age in the newborn infant. J. Pedlar., 1970, 77: 1-10. Gallai, V., Mazzotta, G., Cagini, L., Del Gatto, F. and, Angelotti, F. Maturation of SEPs in pre-terrn and full-term neonates. In: V. Gallai (Ed.), Maturation of the CNS and Evoked Potentials. Elsevier, Amsterdam, 1986: 95-1/16. George, S.R. and Taylor, M.J. Somatosensory evoked potentials in neonates and infants: developmental and normative data. Electroenceph, clin. Neurophysiol., 1991.80: 94-1[12. Hashimoto, T., Masanobu, T., Hiura, K., Endo, S., Fukuda, K., Tamura, Y., Mori, A. and Miyao, M. Short latency somatosensory evoked potential in children. Brain Dev., 1983, 5: 390-396. Herschkowitz, N. Brain development in the fetus, neonate and infant. Biol. Neonate, 1988, 54: t-19. Hrbek, A., Karlberg, P. and Olsson, T. Development of visual and somatosensory evoked responses in pre-term newborn infants. Electroenceph. c[in. Neurophysiol., 1973, 34: 225-232. Karniski, W., Wyble, L., Lease, L. and Blair, R.C. The late somatosensory evoked potential in premature and term infants, l[. Topography and latency development. Electroenceph. clin. Neurophysiol., 1992, 84: 44-54. Klimach, V.J. and Cooke, R.W.I. Maturation of the neonatal somatosensory evoked response in pre-term infants. Dev. Med. Child Neurol., 1988, 30: 2118-214. Laget, P. Maturation of somesthetic evoked potentials in normal children. In: G.A. Chiarenza and D. Papakostopoulos (Eds.), Clinical Application of Cerebral Evoked Potentials in Pediatric Medicine. Excerpta Medica, Amsterdam, 1982: 185-206. Lauffer, H. and Wenzel, D~ Maturation of central somatosensory conduction time in infancy and childhood. Neuropediatrics, 1986, 17: 72-74. Laureau, E., Majnemer, A., Rosenblatt, B. and Riley, P. A longitudinal study of short latency somatosensory evoked responses in healthy newborns and infants. Electroenceph. olin. Neurophysiol., 1988, 71: 100-1118. Lehmann, D. Principles of spatial analysis. In: A.S. Gevins and A. R&mond (Eds.), Handbook of Electroencephalography and Clinical Neurophysiology, Revised Series, Vol. 1. Elsevier, Amsterdam, 1987: 309-354. Maugui~re, F., Desmedt, J.E. and Courjon, J. Astereognosis and dissociated loss of frontal or parietal components of somatosensory evoked potentials in hemispheric lesions: detailed correlations with clinical signs and computerized tomography scanning. Brain, 1983, 106:271-311. Mrzljak, L., Uylings, H.B., Kostovic, 1. and Van Eden, C.G. Prenatal development of neurons in the human prefrontal cortex. L A qualitative Golgi study. J. Comp. Neurol., 1988, 271: 355-386. Patten, B.M, Human Embryology, 3rd edn. McGraw-Hill, New York, 1968. Picton, T.W., Woods, D. and Campbell, K. Methodology and meaning of human evoked potentials scalp distribution studies. In: D.A. Otto (Ed.), Multidisciplinary Perspectives in Event-Related Brain Potential Research, 600/0-77-043. U.S. Environmental Protection Agency, Washington, DC, 1978: 515-522. Skrandies, W. and Lehmann, D. Spatial principal components of multichannel maps evoked by lateral visual half-field stimuli. Electroenceph. clin. Neurophysiol., 1982, 54: 662-667. Slimp, J.C., Tamas, L.B., Stolov, W.C. and Wyler, A.R. Somatosensory evoked potentials after removal of somatosensory cortex in man. Electroenceph. clin. Neurophysiol., 1986, 65:111-I 17. Taylor, M.J. and Fagan, E.R. SEPs to median nerve stimulation: normative data for paediatrics. Electroenceph. clin. Neurophysiol., 1988, 71: 323-330. Tomita, Y., Nishimura, S. and Tanaka, T. Short latency SEPs in
SEP PC TOPOGRAPHY IN PREMATURE INFANTS infants and children: developmental changes and maturational index of SEPs. Electroenceph. clin. Neurophysiol., 1986, 65: 335-343. Tsuji, S. and Mural, Y. Scalp topography and distribution of cortical somatosensory evoked potentials to median nerve stimulation. EIectroenceph. clin. Neurophysiol., 1986, 65: 429-439. Tsuji, S., Mural, Y. and Kadoya, C. Topography of somatosensory evoked potentials to median nerve stimulation in patients with cerebral lesions. Electroenceph. clin. Neurophysiol., 1988a, 71: 28{)-288. Tsuji, S., Murai, Y. and Hashimoto, M. Frontal distribution of early cortical somatosensory evoked potentials to median nerve stimulation. Eleclroenceph. clin. Neurophysiol., 1988b, 71: 273-279.
43 Willis, J., Seales, D. and Frazier, E. Short latency somatosensory evoked potentials in infants. Electroenceph. clin. Neurophysiol., 1984, 59: 366-373. Wood, C.C., Cohen, D., Cuffin, B.N., Yarita, M. and Allison, T. Electrical sources in human somatosensory cortex: identification by combined magnetic and potential recordings. Science, 1985, 227: 1051-1053. Zhu, Y., Georgesco, M. and Cadilhac, J. Normal latency values of early cortical somatosensory evoked potentials in children. E[ectroenceph, c[in. Neurophysiol., 1987, 68: 471-474.
32
E V O P O T 91030
The late somatosensory evoked potential in premature and term infants. I. Principal component topography Walt Karniski Department of Pediatrics, MDC Box I5, University of South Florida College of Medicine, Tampa, FL 33612 (U.S.A.) (Accepted for publication: 3 June 1991)
Summary Very little is known about the topographic distribution of the cortical somatosensory evoked potential in premature infants. Principal component analysis (PCA) was applied to the wave forms generated from right median nerve stimuli over a relatively long sweep (483 msec post stimulus) at 16 electrodes in 53 infants with postconceptual ages from 31 to 40 weeks, subdivided into 5 groups by 2 week increments. Factor scores were averaged across subjects, within groups and displayed as topographical maps. Four factors accounted for 7 1 - 7 6 % of the variance in each of the 5 groups and the factors extracted from the P C A performed independently in each group were markedly consistent. The first factor ( N 1 / P 1 ) had a left posterior m i n i m u m and a left frontal-central maximum and probably represents a tangential dipole located in the post-central gyrus. The second factor (N2) was characterized by a consistent left central m i n i m u m with a systematic developmental change in the maximum that seemed to imply that its neural generator was changing in orientation as the infants matured. A third factor (N3) accounted for the most variance and appeared to represent the first evidence of activity in the ipsilateral cortex. Finally, a very late fourth factor appeared only in the more mature groups, with uncertain localization. The topographic maps of the factor scores for these 4 factors appear to account for independent generators in the SEP of the premature and term infant. Key words: SEP; Topography; Premature; Infants
Multiple peaks occur in the first 20 msec following a somatosensory stimulus in normal adults and each of these peaks appears to correspond to the activation of neural structures in the peripheral somatosensory pathway (see Desmedt 1988 for a review). For example, the central maximum negative peak occurring at approximately 20 msec (N20) represents the arrival of the peripheral afferent volley at the primary somatosensory cortex in the post-central gyrus (Broughton 1969; Allison et al. 1980; Wood et al. 1985; Allison et al. 1989a). Most studies of the SEP in full-term and young infants have focused on the N20, usually referred to as the N1, and the developmental changes of both latency and amplitude have been well described (Desmedt and Manil 1970; Desmedt and Debecker 1972; Cullity et al. 1976; Desmedt et al. 1976; Laget 1982; Hashimoto et al. 1983; Willis et al. 1984; Lauffer and Wenzel 1986; Tomita et al. 1986; Bartel et al. 1987; Zhu et al. 1987; Laureau et al. 1988; Taylor and Fagan 1988; George and Taylor 1991; Karniski et al. 1992). However, the later components of the SEP, presumably reflecting
Correspondence to: Walt Karniski, M.D., D e p a r t m e n t of Pediatrics, M D C Box 15, University of South Florida College of Medicine, 12901 Bruce B. Downs Blvd., Tampa, FL 33612 (U.S.A.).
cortical activity (Allison et al. 1989b), have generally received much less scrutiny. The SEP in premature infants has received infrequent attention (Hrbek et al. 1973; Gallai et al. 1986; Klimach and Cooke 1988; Bongers-Schokking et al. 1990) and these studies have also focused on the N1 or preceding peaks. Furthermore, because only a few electrodes have been used in most of these studies, little is known about the topographic distribution of the SEP. Thus, this study is an examination of the topographic distribution of the SEP in premature infants, beginning with the N1 and continuing into the time period in which cortical activation predominates.
Methods
Subjects Fifty-three infants were recruited from the Neonatal Intensive Care Unit and Newborn Nursery at Tampa General Hospital. In order to insure that they were relatively free of neurologic disorders, infants were included only if they met all of the following criteria: (1) 1 and 5 min Apgar scores of at least 4 and 7 respectively, (2) length, weight, and head circumference appropriate for gestational age, (3) absence of major congenital malformations, (4) absence of known
SEP PC TOPOGRAPHY IN PREMATURE INFANTS
33
neurologic abnormalities or evidence of seizure disorder, (5) normal cranial ultrasound in infants less than 34 weeks ~estation at birth or less than 1500 g birth weight, (6) absence of asymmetric limb lengths, and (7) stable clinical condition without assisted ventilation or oxygen requirements. Parental consent was obtained for all infants. Because of the importance of an accurate assessment of post-menstrual age (PMA) for this study, an infant was included only if the post-menstrual age as assessed by examination (Dubowitz et al. 1970) was within 2 weeks as estimated by history (date of mother's last menstrual period). The PMA at birth ranged from 28 to 40 weeks. Infants were grouped by PMA in 2 week increments (31-32 weeks, 33-34 weeks, 35-36 weeks, 37-38 weeks, 39-40 weeks). Post-menstrual age is defined as the gestational age at birth plus chronological age at the time of testing and thus represents the time that has elapsed since the mother's last menstrual period. Because the time of the last menstrual period usually antedates conception by 2 weeks, the PMA is approximately 2 weeks longer than the time from conception. Thus, the PMA represents a maturational age that is more relevant than chronological age. At the time of testing, the infants' chronological ages ranged from 1 to 65 days, with a mean of 13 days (Table I). There was no difference in chronological age across PMA groups ( F (4, 48) = 1.33, P = 0.21). Black and white infants were approximately equally distributed across all groups. The male-to-female ratio was approximately equal for the entire group of infants (23:30), but the 31-32 week and the 39-40 week groups had more males, whereas the middle 3 groups were predominantly female (Fisher's exact test, P = 0.02). Although there was a tendency for Apgar scores of the less mature infants to be somewhat lower, the Apgar scores were generally not significantly different across groups at 1 min ( F (4, 48) = 2.43, P = 0.06) or 5 min of age ( F (4, 48) = 2.05, P = 0.10).
Procedures All infants were tested during sleep. In the very premature infants, sleep stage was difficult to deter-
mine so that sleep stage was not formally evaluated. Environmental temperature and body temperature were maintained constant throughout the test period. Stimuli consisted of median nerve stimulation at the right wrist, with a stimulus duration of 100 msec and a variable frequency of 0.9-1.7/sec. Stimulus intensity was set below sensory threshold, then gradually increased until a slight thumb twitch was observed. The intensity was then reduced until the twitch was barely visible. Two blocks of 300 stimuli each were presented to each infant, 15 rain apart, then the two blocks were averaged together. Single trials with voltage in excess of a pre-set level for each infant were automatically excluded from subsequent averaging. Sixteen pediatric gold electrodes (5 mm in diameter) were used with a linked ears reference at the following sites: Fpl, Fp2, F3, F4, F7, F8, C3, C4, T3, T4, T5, T6, P3, P4, O1 and 02. Many other SEP studies have used electrodes posterior to the central electrodes in an attempt to locate the electrode directly over the somatosensory cortex (see Picton 1978). While this might be appropriate for N1 which is most likely generated there, there is no agreement about the source of the 3 components occurring after N1 in the pre-term infants described in this study. Thus the more standard 10-20 system electrode positions were used. A Cadwell 8400 was used for signal generation and data acquisition. Impedance was maintained below 5000 I2. The sweep extended from 50 msec prior to the stimulus to 483 msec after the stimulus. Voltage was sampled at a frequency of 960 Hz. An analog high-pass filter of 1 Hz was used. Desmedt et al. (1974) have recommended a low-pass filter of 3000 Hz in studies of the SEP in adults, in order not to miss the fast transients in the N1. Although this may be valid for adults, the N1 and especially the later components of interest in this study are of a much slower frequency. Thus, we used a low-pass filter of 100 Hz as recommended by BongersSchokking et al. (1989). In another study of these same infants (Karniski et al. 1992), this filter setting allowed for easy identification of an N1 in most infants and made it easier to identify the later peaks which were of greater interest.
TABLE I Group characteristics. Post-menstrual age group (weeks)
No.
Sex (% male)
Race (% Black)
Apgar (1 min)
Apgar (5 min)
Mean chron, age at testing (days)
31-32 33-34 35-36 37-38 39-40
13 14 9 9 8
76.9 28.6 22.2 22.2 62.5
46.2 50.0 44.4 44.4 12.5
7.6 6.5 7.7 7.5 8.5
8.3 8.2 8.5 8.5 9.1
13 19 13 10 8
Total (or avg.)
53
43.4
41.5
13
34
Data analysis Any analysis of evoked potentials must begin with the assumption that they are produced by multiple neural generators with independent locations, orientations, onsets and durations. Principal component analysis (PCA) is often successful in identifying a relatively small number of independent factors that account for a large majority of the variance of the evoked potential across a group of subjects. These independent factors have been shown to reflect orthogonal generators within the brain (Donchin and Heffley 1978; Skrandies and Lehmann 1982; Lehmann 1987). Each factor is represented by a factor loading at each variable (time point). By plotting the factor loadings across time for each factor, it can be determined that a given factor is "active" during a time segment when the factor loadings are maximum. A factor score is generated for each factor, for each subject and for each electrode and these factor scores are measures of how much that condition (electrode) is weighted on its corresponding factor. By averaging the factor scores across subjects within a group for each electrode, a topographical map of the factor scores is generated which represents the scalp topography of each factor. PCA with varimax rotation was performed with 512 time points as the variables over the 533 msec sweep. Five separate PCAs were performed, once for each of the 5 PMA groups. Factor loadings for the first 6 factors were graphed as wave forms. Factor scores averaged for each electrode across subjects within a group were displayed as topographic maps and then matched with their appropriate factor loading wave forms. Factor loading wave form and factor score map pairs were then sorted so that times of peak loading, shape of the loading wave form and factor score maps were all similar within a sorted PMA group. If a factor did not produce a factor loading wave form with a peak that approached or exceeded 0.5 or - 0 . 5 , it was discarded.
Results
Wave forms and voltage maps The temporal and topographic characteristics of the SEP in premature infants are best understood by si, multaneously examining the wave forms (Fig. 1) and voltage topographic maps (Fig. 2). The first 15 msec of the sweep in each P M A group consists of a stimulus artifact and far-field components that were not relevant to this study and will not be discussed further. The first negative deflection after 15 msec appears biphasic in all P M A groups, primarily in the central and parietal electrodes. The first of these two peaks is maximal near 25 msec in the 31-32 week group and decreases
W. KARNISKI
to approximately 20 msec in the 39-40 week group. It appears to have the same amplitude across all electrodes, producing a relatively flat voltage map for each group, indicating that this peak probably represents a far-field potential. Near 25 msec in all groups, the left and right central and parietal electrode wave forms begin to diverge, and this divergence indicates the beginning of the N1. In the topographical maps, the N1 occurs as a negative minimum in the left central electrodes starting at 51 msec in the 31-32 week group and in the left posterior temporal and parietal electrodes starting at 51 msec, 46 msec, 36 msec and 31 msec in the last 4 groups respectively. A small positive peak follows the N1, becoming larger and earlier with maturation. However, the topographical maps reveal that the maximum actually begins much earlier than it appears in the wave forms, very near to the onset of the N1. In the 31-32 week group, it initially appears ipsilateral to the stimulus on the right at 51 msec, while for the other 4 groups, it begins left frontal (61, 46, 36, and 31 msec). The maximum then changes differently for each group, but these changes progress systematically across groups. For the first 2 groups the maximum becomes bifrontal at 120 msec and 110 msec, for the third group the maximum becomes bifrontal and more central near 90 msec, and for the 2 older groups the maximum remains unilateral, but becomes more central (80 and 61 msec). For the first 2 groups, this topographic change occurs coincident with the onset on the N2, while in the latter 3 groups it occurs earlier than the N2 onset. However, in each of these 3 more mature groups, the maximum changes again with N2 onset. These topographic changes in the maximum would suggest at least 2 subcomponents for the P1, with the first subcomponent (referred to here as the Pla) temporally associated with the N1. The most prominent feature of the wave forms is the large second negative component (N2) between 100 and 300 msec. It is maximal at C3 for all groups and with maturation it decreases in amplitude, latency and duration. The early part of the N2 has an associated frontal maximum in all 5 groups, which is topographically distinct from the P l a described above. However, while the minimum appears stable in location at C3, the maximum "moves" from a bilateral frontal location to a right temporal focus for the 31-32 and 35-36 week groups, to a right temporal-parietal focus for the 33-34 week group, and to a right central focus for the 37-38 and 39-40 week groups. Eventually, in all of the groups, the maximum is seen on the right side, ipsilateral to the stimulus. Although this maximum remains long after the N2 has apparently disappeared in the first 2 groups, the right central maximum seems to fade or change concurrently with the N2 in the latter 3 groups.
SEP PC T O P O G R A P H Y IN P R E M A T U R E I N F A N T S
35
RIGHT MEDIAN NERVE SEP CENTRAL
FRONTAL
PARIETAL
PMA 0N'KS)
.
31-32
...................
aa-a4
---,..........,...._,.-
35-36
37-38 N1
N2
39-40 ELECTRODES LEFT--
I
.
RGHT. . . . . . .
0
100
.
. 200
.
l"vl 300
400
0
P1 . 100
.
. 200
P2 .
300
400
I
.
0
100
.
. 200
. 300
400
Fig. l. G r a n d average wave forms for left and right frontal, central and parietal electrodes to right median nerve stimuli for each P M A group. T h e developmental evolution of 4 peaks can be seen from 31 weeks P M A to 40 weeks. N1, P1, N2 and P2 all show decreasing latency with maturation. A third low amplitude negative wave near the end of the sweep is suggested. (Linked ear reference, negative up.)
The wave forms indicate that the N2 simply trails off in all the groups, but the topographical maps reveal a third negative, low amplitude component. For all of the groups, the amplitude is more negative on the right and on the maps it appears as a right sided minimum emerging after the N2 (405, 390, 360, 300 and 300 msec in the 5 groups respectively). This N3 appears more frontal for the more immature groups and more central for the 37-38 and 39-40 week groups. It is still present at the end of the sweep in all groups.
Principal component analysis Three important, general features emerged from the PCA. First, the maps of the factor scores have a close correspondence to certain persistent or prominent features of the voltage topographical maps. Secondly, this feature extraction results in a significant reduction of data. In each application of the PCA, the first 4 factors account for 71-76% of the total variance. Third, although there are differences in the PCAs performed across the groups, there is a very strong consistency in
Fig. 2. Voltage topography for right median nerve SEP in p r e m a t u r e and term infants. T h e temporal evolution of the right median nerve SEP topography can be seen. Maturational effects are seen as earlier onset and shorter duration for certain topographic features as infants mature. Systematic topographic changes with maturation can be seen as well. (All maps were created using the voltage from 16 electrodes referenced to linked ears. Midline electrode values were interpolated. Maps are shown at 5 msec increments until 70 msec, at 10 msec increments to 150 msec and then at 15 msec subsequently.)
36
W. KARNISKI
PMA (wks)
31-32
33-34
35-36
37-38
39-40 LATENCY (ms) 21
165
180
195
26
31
36
41
46
51
56
61
210
225
240
255
270
285
300
315
SEP PC T O P O G R A P H Y IN P R E M A T U R E INFANTS
37
1.6
luv
.8 66
70
80
90
100
110
120
130
140
150 0
-.8
-1.6
330
345
360
375
390
405
420
435
450
465
38
W. KARNISKI
TABLE II Factor number and percent of variance for each factor in Fig. 3. Factor Range (msec) Peak (msec) 31-32 weeks 33-34 weeks 35-36 weeks 37-38 weeks 39-40 weeks
N1/P1 30-66 47
N2 N3 ? 139-181 155-483 345-? 160 287 419
3 6.4% 4 6.6% 3 12.2% 3 14.5% 6 5.2%
2 21.3% 2 20.4% 4 6.4% 2 16.1% 3 11.0%
1 43.3% 1 33.5% 1 30.3% 1 34.4% 1 37.3%
2 24.4% 2 13.8%
the results obtained across each group. This consistency lends credence to the hypothesis that the resultant factor score maps represent spatially and temporally distinct E R P components (Donchin and Heffiey 1978). The factors in Fig. 3 have been ordered chronologically from left to right rather than by the order in which the factors emerged. The order in which they emerged, the time range of peak activity and the amount of variance for which they account are shown in Table II. The time range of maximal activity was determined by averaging the factor loading wave forms across all groups, then identifying the time period in which the average loadings were above 0.5. This generates an artificial time range that is only approximate for all the groups, because the less mature infants had factors with somewhat longer latencies, while the latencies of the factors in the older infants were less. The first factor after the stimulus was active between 30 and 66 msec. It was present in all 5 groups and accounted for 5 - 1 5 % of the total variance from each PCA. Factor score topography was very consistent across groups, with a left posterior minimum and a left frontal-central maximum for all groups except for the 31-32 week group. In these very premature infants, the minimum was more central with the maximum more frontal and slightly more prominent on the right. The second factor had factor loadings that were active between 139 and 181 msec, accounting for 6 - 2 1 % of the total variance. The factor score topography showed a left central minimum and a right central maximum for the first 3 groups, while the last 2 groups
each had 2 different maximums in the right frontal and bilateral occipital areas. The third factor emerging from each group accounted for the greatest amount of variance in each separate P C A (30-43%) and covered a broad range from 155 to 480 msec. The maximum is in the left central area in the 31-32 week group and becomes more lateral as the infants mature. More importantly, if one scans down this third column, a systematic change across groups is clearly visible. The more prominent right frontal minimum and the less prominent right occipital minimum both move centrally, meeting as a single topographic minimum at C4 by the time the infants reach a full-term equivalent. A fourth factor emerged in the very late part of the SEP wave form only for the 35-36 and 39-40 week groups beginning near 345 msec and this factor was still active at the end of the sweep at 483 msec. It accounted for a relatively large amount of variance (14-24%). Although both factors have a left central minimum, the 35-36 week group has a maximum in the right occipital area. On the other hand, the older group has a maximum which is well circumscribed to the right central area, with a second smaller maximum in the left occipital area.
Discussion A traditional analysis of the SEP wave forms in this study indicates that 4 major peaks occur between 30 and 450 msec, after the primary somatosensory cortex is first activated (N1, P1, N2 and P2). These peaks consistently decrease in latency with maturation, as shown in other studies of the first negative peak (N1). A more systematic study of the peak latency and amplitude changes with maturation in premature infants can be found in a separate study of these same children (Karniski et al. 1992). However, when peaks overlap temporally, it is difficult to distinguish whether the overlap is due to multiple generators that are simultaneously active or whether the two "peaks" represent opposite ends of a single dipole. It was not the purpose of this study to localize these apparent dipoles nor the generators associated with them. However, the factor score maps generated by the P C A (Fig. 3) do seem to extract the most prominent features of the serial topographic maps (Fig.
Fig. 3. Principal components for right median nerve SEP in premature and term infants. Principle component analysis was performed separately for each PMA group with 512 points along a 533 msec sweep for 16 electrodes as variables. Factor loadings are displayed as wave forms in the bottom row, with a topographic map of factor scores displayed above. (Midline electrode values were extrapolated.) Although the factor loading wave forms appear very different across groups, the peaks tend to occur in the same approximate time periods, with a tendency for shorter latencies for the more mature infants. The first 4 factors accounted for 71-76% of the total variance in each PCA performed for each PMA group. The factor loading wave form was matched with its appropriate factor score map, then factor loading wave forms were sorted chronologicallyby peak loading times. A factor was identified if factor loading wave forms and factor score maps were similar across groups.
SEP PC TOPOGRAPHY IN PREMATURE INFANTS
39
MAPS OF FACTOR SCORES FROM PRINCIPAL COMPONENT ANALYSIS PMA (wks)
N1/P1
N2
N3
31-32 1.0
~v
33-34
,5
35-36
0
37-38
-.5
-1.0
39-40
05
FACTOR LOADINGS
0
-05
31-32 33-34 35-36 37-38 39-40
........ -1.0
...... ...... - - -
0
100
2OO
30-66
3OO
40O
0
100
2OO
3o0
40O
0
139-181
IO0
2oo
3OO
155-483
4OO
0
IO0
2OO
3OO
4OO
345-?
Peaks(ms) ZWART
40
2). Furthermore, these factor maps appear to have topographic analogs in studies of the SEP in adults. The first factor emerging from the PCA has a left posterior minimum and a left frontal-central maximum for all groups except the 31-32 week infants. By the temporal distribution of the factor loadings and the topography of the factor scores, this factor appears to represent the N 1 / P l a described above in the analysis of the voltage maps. In topographic studies of the SEP in adults, an N20 is localized to the contralateral posterior temporal scalp area (Desmedt and Cheron 1980; Desmedt and Bourguet 1985; Tsuji and Murai 1986; Desmedt et al. 1987; Tsuji et al. 1988a). At the same time, these studies disclose a P20 with a bilateral frontal distribution that is slightly more prominent on the side contralateral to the stimulus. Since the first factor extracted by PCA in this study (N1/Pla) embodies the same topography, it is probably analogous to the adult N20/P20. It is now fairly well established that the N20/P20 represents a single, tangentially oriented dipole located in area 3b of the post-central gyrus (Broughton 1969; Allison 1980; Maugui~re et al. 1983; Slimp et al. 1986; Allison et al. 1989a). The topographic studies of the SEP in adults also reveal a P22 that follows and often overlaps the N20/P20. A radially oriented dipole in area 4 of the pre-central gyrus has been proposed as the generator for this component (Desmedt and Bourguet 1985; Desmedt et al. 1987). The wave forms in this study indicate that a distinct positive peak follows the N1. However, in the maps, the early part of the maximum (Pla) is clearly associated with the N1, while the later part of this maximum (Plb) develops parallel with the N2. The maximum for the 35-38 week infants shifts more central before the onset of the N2 and this may represent the "true" P1. However, PCA does not reveal a separate factor that is either active in the time period between the N1 and N2 or that would account for the voltage topography in this time period. Although a "true" P1 that is not part of N1 or N2 dipoles may exist in this time period in premature infants, this study does not provide convincing evidence for it. However, in a separate study of the effects of maturation on latency, amplitude and topography in the same infants in this study (Karniski et al. 1992), the P1 was found to have a different voltage topography from the N1 and the P1 amplitude and latency were found to mature at different rates than for the N1, providing some evidence that N1 and P1 have different neural generators. An important caveat is necessary whenever PCA is utilized to draw conclusions about independent generators of evoked potential peaks. Independent factors emerge from PCA only when an experimental manipulation produces a systematic change in voltage amplitude (Donchin and Heffiey 1978). In this study, differ-
W. KARNISKI
ent maturational levels provide the experimental manipulation. The fact that the N1 and Pla appear on factor score maps simultaneously provides evidence, but not proof, for their lack of independence. If some experimental manipulation influences only one of the peaks and not the other, only then will they emerge as separate factors. The weakness of PCA is that one can never test for all possible experimental variables and therefore this lack of independence can never be proven through PCA. Desmedt and Manil (1970) described a second positive peak appearing between the N1 and N2, and after the P1, only during slow wave sleep in full-term infants. This pattern was seen in a only a few infants and is not visible in the grand average. This is a potential but unlikely explanation for the change in the topography of the P1, because separate components did not emerge from the PCA. The second factor has a left central minimum and a maximum that varies systematically with maturation. This factor most likely represents the N2 and this is supported by at least two findings. In the few studies of premature infants that examined the N2 (Desmedt and Manil 1970; Hrbek et al. 1973; Gallai et al. 1986) the latency of the N2 was similar to that found in this study (139-181 msec). In addition, the latency and topography (left central minimum) of this second factor are similar to the latency and topography of the N2 as seen in the voltage maps from this study. However, the topography of the second factor varied with maturation. In each group, the minimum was centrally located, but the maximum shifted. In the 31-32 week group, the maximum is right temporal. In the 33-36 week infants it becomes right central and becomes bi-peaked in the 37-40 week infants. If a single generator is assumed for each factor, the very premature infants' maps could be explained by a tangential dipole, while the older infants' maps would more likely suggest a radial dipole. It is possible that as the brain of the premature infant matures, the orientation of some generators would shift or rotate in space. The cortex of the 25 week fetus is lissencephalic, or almost completely smooth with little evidence of fissures or gyri. The sylvian fissure is evident in the third month post conception, but the central sulcus is barely visible until 6 months (Bailey and Miller 1927; Patten 1968; Crelin 1974). Over the next 15 weeks, the fetal brain progresses through extensive structural, microscopic and functional changes. The central sulcus deepens and the pre-central gyrus and post-central gyrus (containing the motor and somatosensory cortices respectively) become visible and fold in. As the association areas of the frontal lobe increase in size, the central sulcus moves more posterior. Myelin begins to develop around neuronal axons, neuronal cells differentiate, migrate to the surface of the cortex and begin to establish multiple
SEP PC TOPOGRAPHY IN PREMATURE INFANTS cell-to-cell connections (Herschkowitz 1988; Mrzljak et al. 1988), Thus it is not surprising that these structural changes would be reflected in a change in orientation of some of the dipoles that generate SEP components and that appears to be what is happening with this factor. The analog of the N2 in this study to the SEP in adults is difficult to determine. A large negative peak following the adult N 2 0 / P 2 0 and P22 has been found as early as 24 msec by Tsuji et al. (1988a), at 26 msec by Tsuji et al. (1988b) and at 30 msec by Desmedt and colleagues (Desmedt and Bourguet 1985; Desmedt et al. 1987). This peak is much larger in amplitude than the peaks preceding it and, regardless of latency, each of these peaks has a pre-central and frontal topography. Desmedt and Bourguet (1985) proposed a source in the supplementary motor area for this peak. Allison et al. (1989a) studied the SEP in normal adults with a small patch of 64 electrodes over the sensorimotor cortex and with transcortical electrodes. They describe a P25-N35 wave that has a maximum over the postcentral gyrus, medial to the locus of the earlier N 2 0 / P 2 0 and they propose a radial dipole located in area 1 of the post-central gyrus. In a second study Allison et al. (1989b) examined peaks occurring in the time period from 40 to 250 msec. They described a P45-N80-P180 with a probable generator in area 3b of the post-central gyrus and a P50-N90-P190 with a probable generator in area 1 of the somatosensory cortex. The N2 factor in the current study appears analogous to the N 2 4 / N 2 6 / N 3 0 peaks described above. Its topography is consistent with a radially oriented dipole in the post-central gyrus, similar to that described by Allison et al. (1989b), for the 37-40 week infants only. It is postulated that both functional and structural differences in the less mature infants produce the difference in topography. A longitudinal or cross-sectional study, which follows the development of this factor as the SEP matures into its adult form, would help answer this question. The third factor in Fig. 3 was actually the first factor to emerge in each of the PCAs and thus accounted for the largest proportion of variance in each analysis. It is active in the same time period as the very slow, third negative deflection that is seen in the very late part of the wave forms and thus will be referred to as the N3. It has not been previously described, because most studies have not used post-stimulus sweeps as long as those employed in this study (483 msec). Although the origin of this component cannot be determined from this study alone, it does appear to have a minimum ipsilateral to the side of stimulation, especially in the 39-40 week group. In the wave forms, the ipsilateral electrodes seem to indicate this peak near the end of the sweep, while the contralateral electrodes produce a wave form that appears relatively flat in the same time
41 period. Thus this component seems to represent activity in the opposite hemisphere from that of the earlier components and thus ipsilateral to the stimulus. If so, it represents the first evidence of activation of the ipsilateral hemisphere in the premature SEP. Finally, an additional factor emerged in the 35-36 and 39-40 week groups. Both groups show a left-sided minimum at C3 and a right-sided maximum, but the topographies of these two factors are actually rather different. Because this factor represents such a large amount of variance, it is difficult to dismiss, and yet there is no analog in the voltage maps or in the wave forms that would help identify this factor. All previous studies of the SEP in premature infants have focused on either the early, peripheral far-field components or on the N1, which from studies in adults is presumed to represent the afferent volley arriving at the cortex in the post-central gyrus. However, the central nervous system activity that occurs after this point has generally been ignored in premature infants. In this study, the voltage peaks and topography of the SEP were examined over a much longer period of time after stimulus (483 msec) than in other studies of premature infants. Principal component analysis extracted 3 major factors. The first factor had a contralateral parietal-temporal minimum and a contralateral frontal-central maximum analogous to the adult N20/P20. The second factor had a contralateral central minimum and a variable maximum across groups. It appeared to be analogous to the N30 described by Desmedt and colleagues and the N 2 4 / N 2 6 described by Tsuji and colleagues. A third factor with an ipsilateral minimum may represent previously undescribed activity of the ipsilateral cortex. These factors represented the predominant features of the voltage topographic maps and they were consistently replicated in each of the 5 PMA groups. Systematic topographic changes in these factors were seen with maturation, which appear to reflect functional and structural changes as the infants mature. The author wishes to thank Loretta Lease, R.EEGT, for her attention to detail in data acquisition, Richard Weibley, M.D., and Lance Wyble, M.D., for their identification of infants qualifying for this study, Rodney Vanderploeg, Ph.D., for his review of an earlier version of this paper, R. Clifford Blair, Ph.D., for assistance in data analysis and Roger Daniels for assistance in the Neonatal ICU. A special thanks is extended to Dietrich Lehmann, M.D., who supported the author's sabbatical and stimulated thinking that lead to the final form of this paper. This research was funded in part by the State of Florida, Department of Education, Florida Diagnostic and Learning Resource System and Children's Medical Services, State of Florida.
References
Allison, T., Goff, W.R., Williamson, P.D. and Van Gilder, J.C. On the neural origin of early components of the somatosensory
42 evoked potential. In: J.E. Desmedt (Ed.), Clinical Uses of Cerebral, Brainstem and Spinal Somatosensory Evoked Potentials. Prog. Clin. Neurophysiol., Vol. 7. Karger, Basel, 1980: 51-68. Allison, T., McCarthy, G., Wood, C.C., Darcey, T.M., Spencer, D.D. and Williamson, P.D. Human cortical potentials evoked by stimulation of the median nerve. I, Cytoarchitectonic areas generating short-latency activity. J, Neurophysiol., 1989a, 62: 694-710. Allison, T., McCarthy, G., Wood, C.C., Williamson, P.D. and Spencer, D.D. Human cortical potentials evoked by stimulation of the median nerve. II. Cytoarchitectonic areas generating Iongla'~ency activity. J. Neurophysiol., 1989b, 62: 711-722. Bailey, F.R and Miller, A.M. Text-book of Embryology. William Wood and Co., New York, 1927. Bartel, P., Conradie, J., Robinson, E., Prinsloo, J. and Becker, P. The relationship between median nerve somatosensory evoked potential latencies and age and growth parameters in young children. Electroenceph. clin. Neurophysiol., 1987, 68: 1811-186. Bongers-Schokking, C.J., Colon, E.J,, Hoogland, R.A., Van den Brande, J.L. and De Groot, C.J. The somatosensory evoked potentials of normal infants: influence of filter bandpass, arousal state and number of stimuli. Brain Dev., 1989, 1l: 33-39. Bongers-Schokking, J.J., Colon, E.J., Hoogland, R.A., Van den Brande, J.L. and De Groot, C.J. Somatosensory evoked potentials in term and preterm infants in relation to postconeeptual age and birth weight. Neuropediatrics, 1990, 21: 32-36. Broughton, R.J. In: E. Donchin and D.B. Lindsley (Eds.), Average Evoked Potentials. U.S. Government Printing Office, Washington, DC. NASA SP-191, 1969: 79-84. Cretin, E.S. Development of the nervous system: a logical approach to neuroanatomy. Clin. Symp., 1974, 26: 1-32. Cullity, P., Franks, C.I., Duckworth, T. and Brown, B.H. Somatosensory evoked cortical responses: detection in normal infants. Dev. Med. Child Neurol., 1976, 18:11-18. Desmedt, J.E. Somatosensory evoked potentials. In: T.W. Picton (Ed.), Human Event-Related Potentials. Elsevier, Amsterdam, 1988. Desmedt, J.E. and Bourguet, M. Color imaging of parietal and frontal somatosensory potential fields evoked by stimulation of median or posterior tibial nerve in man. Electroenceph. clin. Neurophysiol., 1985, 62: 1-17. Desmedt, J.E. and Cheron, G. Somatosensory evoked potentials to finger stimulation in healthy octogenarians and in young adults: wave forms, scalp topography and transit times of parietal and frontal components. Electroenceph. clin. Neurophysiol., 1980, 50: 404-425. Desmedt, J.E. and Debecker, J. The somatosensory cerebral evoked potentials of the sleeping human newborn. In: C.D. Clemente, D.P. Purpura and F.E. Mayer (Eds.), Sleeping and the Maturing Nervous System. Academic Press, New York, 1972: 229-239. Desmedt, J.E. and Manil, J. Somatosensory evoked potentials of the normal human neonate in REM sleep, in slow wave sleep and in waking. Electroenceph. clin~ Neurophysiol., 1970, 29: 113-126. Desmedt, J.E., Brunko, E., Debecker, J. and Carmeliet, J. The system bandpass required to avoid distortion of early components when averaging somatosensory evoked potentials, Electroenceph. clin. Neurophysiol., 1974, 37: 407-410. Desmedt, J.E., Brunko, E. and Debecker, J. Maturation of the somatosensory evoked potentials in normal infants and children, with special reference to the early NI component, Electroenceph. clin. Neurophysiol., 1976, 40: 43-58. Desmedt, J.E., Nguyen, T.H. and Bourguet, M. Color imaging of human evoked potentials with reference to the N20, P22, P27 and N30 somatosensory response. Electroenceph. clin. Neurophysiol., 1987, 68: 1-22. Donchin, E. and Heffley, E. Multivariate analysis of event-related potential data: a tutorial review. In: D.A. Otto (Ed.), Multidisciplinary Perspectives in Event-Related Brain Potential Research,
W. KARNISK1 EPA-600/9-77-043, U.S, Environmental Protection Agency, Washington, DC, 1978: 555-572. Dubowitz, L.M.S., Dubowitz, V. and Golbert, C. Clinical assessment of gestational age in the newborn infant. J. Pedlar., 1970, 77: 1-10. Gallai, V., Mazzotta, G., Cagini, L., Del Gatto, F. and, Angelotti, F. Maturation of SEPs in pre-terrn and full-term neonates. In: V. Gallai (Ed.), Maturation of the CNS and Evoked Potentials. Elsevier, Amsterdam, 1986: 95-1/16. George, S.R. and Taylor, M.J. Somatosensory evoked potentials in neonates and infants: developmental and normative data. Electroenceph, clin. Neurophysiol., 1991.80: 94-1[12. Hashimoto, T., Masanobu, T., Hiura, K., Endo, S., Fukuda, K., Tamura, Y., Mori, A. and Miyao, M. Short latency somatosensory evoked potential in children. Brain Dev., 1983, 5: 390-396. Herschkowitz, N. Brain development in the fetus, neonate and infant. Biol. Neonate, 1988, 54: t-19. Hrbek, A., Karlberg, P. and Olsson, T. Development of visual and somatosensory evoked responses in pre-term newborn infants. Electroenceph. c[in. Neurophysiol., 1973, 34: 225-232. Karniski, W., Wyble, L., Lease, L. and Blair, R.C. The late somatosensory evoked potential in premature and term infants, l[. Topography and latency development. Electroenceph. clin. Neurophysiol., 1992, 84: 44-54. Klimach, V.J. and Cooke, R.W.I. Maturation of the neonatal somatosensory evoked response in pre-term infants. Dev. Med. Child Neurol., 1988, 30: 2118-214. Laget, P. Maturation of somesthetic evoked potentials in normal children. In: G.A. Chiarenza and D. Papakostopoulos (Eds.), Clinical Application of Cerebral Evoked Potentials in Pediatric Medicine. Excerpta Medica, Amsterdam, 1982: 185-206. Lauffer, H. and Wenzel, D~ Maturation of central somatosensory conduction time in infancy and childhood. Neuropediatrics, 1986, 17: 72-74. Laureau, E., Majnemer, A., Rosenblatt, B. and Riley, P. A longitudinal study of short latency somatosensory evoked responses in healthy newborns and infants. Electroenceph. olin. Neurophysiol., 1988, 71: 100-1118. Lehmann, D. Principles of spatial analysis. In: A.S. Gevins and A. R&mond (Eds.), Handbook of Electroencephalography and Clinical Neurophysiology, Revised Series, Vol. 1. Elsevier, Amsterdam, 1987: 309-354. Maugui~re, F., Desmedt, J.E. and Courjon, J. Astereognosis and dissociated loss of frontal or parietal components of somatosensory evoked potentials in hemispheric lesions: detailed correlations with clinical signs and computerized tomography scanning. Brain, 1983, 106:271-311. Mrzljak, L., Uylings, H.B., Kostovic, 1. and Van Eden, C.G. Prenatal development of neurons in the human prefrontal cortex. L A qualitative Golgi study. J. Comp. Neurol., 1988, 271: 355-386. Patten, B.M, Human Embryology, 3rd edn. McGraw-Hill, New York, 1968. Picton, T.W., Woods, D. and Campbell, K. Methodology and meaning of human evoked potentials scalp distribution studies. In: D.A. Otto (Ed.), Multidisciplinary Perspectives in Event-Related Brain Potential Research, 600/0-77-043. U.S. Environmental Protection Agency, Washington, DC, 1978: 515-522. Skrandies, W. and Lehmann, D. Spatial principal components of multichannel maps evoked by lateral visual half-field stimuli. Electroenceph. clin. Neurophysiol., 1982, 54: 662-667. Slimp, J.C., Tamas, L.B., Stolov, W.C. and Wyler, A.R. Somatosensory evoked potentials after removal of somatosensory cortex in man. Electroenceph. clin. Neurophysiol., 1986, 65:111-I 17. Taylor, M.J. and Fagan, E.R. SEPs to median nerve stimulation: normative data for paediatrics. Electroenceph. clin. Neurophysiol., 1988, 71: 323-330. Tomita, Y., Nishimura, S. and Tanaka, T. Short latency SEPs in
SEP PC TOPOGRAPHY IN PREMATURE INFANTS infants and children: developmental changes and maturational index of SEPs. Electroenceph. clin. Neurophysiol., 1986, 65: 335-343. Tsuji, S. and Mural, Y. Scalp topography and distribution of cortical somatosensory evoked potentials to median nerve stimulation. EIectroenceph. clin. Neurophysiol., 1986, 65: 429-439. Tsuji, S., Mural, Y. and Kadoya, C. Topography of somatosensory evoked potentials to median nerve stimulation in patients with cerebral lesions. Electroenceph. clin. Neurophysiol., 1988a, 71: 28{)-288. Tsuji, S., Murai, Y. and Hashimoto, M. Frontal distribution of early cortical somatosensory evoked potentials to median nerve stimulation. Eleclroenceph. clin. Neurophysiol., 1988b, 71: 273-279.
43 Willis, J., Seales, D. and Frazier, E. Short latency somatosensory evoked potentials in infants. Electroenceph. clin. Neurophysiol., 1984, 59: 366-373. Wood, C.C., Cohen, D., Cuffin, B.N., Yarita, M. and Allison, T. Electrical sources in human somatosensory cortex: identification by combined magnetic and potential recordings. Science, 1985, 227: 1051-1053. Zhu, Y., Georgesco, M. and Cadilhac, J. Normal latency values of early cortical somatosensory evoked potentials in children. E[ectroenceph, c[in. Neurophysiol., 1987, 68: 471-474.
