In the past two decades, new techniques have been developed that allow a greater appreciation of the level of vision in the preverbal child. However, these techniques have limitations. Preferential looking, because of its ambulatory nature, is easy to use, particularly in the treatment of deprivation amblyopia, but it must be performed correctly to be reliable. Pattern visual evoked potential (VEP) response is theoretically a good method of visual assessment, but because of the high number of uninterpretable results in the young child, we have stopped using it for this indication. For most pediatrie ophthalmologists, clinical judgment, involving simple observation of visual behavior and a few basic maneuvers, remains the method of choice for visual assessment in the younger child.
However, in some instances a more precise idea of the degree of visual disability in one or both eyes is desired. Brain electrical activity mapping (BEAM), which allows the study of electrical visual reactivity on a computerized electroencephalogram (EEG), has interested us from the time of the earliest reports of its use in older children.1 The purpose of this study was to test the usefulness and reliability of BEAM in younger children as well as its ability to provide information additional to that of other methods, such as clinical testing, preferential looking, and VEP. We worked in conjunction with a department of neurological science, which was accustomed to perform pediatrie recording and was equipped with a mapping apparatus.2 We report here our results and the problems we encountered during more than 4 years of use of this technique in young children.
MATERIALS AND METHODS
BEAM, or quantified EEG recording, provides a numerical and cartographic representation of the EEG by means of computerized global analysis of cerebral activity on the basis of the rapid transformation principle of Fourier.3·4 Cerebral activity is studied at different points on the scalp.
The mapping apparatus consists of classical EEG recording equipment connected to a computer (Fig 1). The EEG recording is carried out at regular intervals by means of 12 electrodes attached to the child's scalp. Periods free from artifacts (which may be caused, for example, by movements of the child) are chosen for analysis. The EEG is analyzed by means of a computer. For every sequence and every electrode chosen, a range of power from O to 30 hertz is used. The result of this analysis of cerebral energy in microvolts squared is given electrode by electrode, either cartographically, with the use of colors or shades of gray, or numerically, for each band of frequency of brain activity.
FIGURE 1: Brain electrical activity mapping (BEAM) apparatus and procedure.
For BEAM analysis to be carried out, the visual stop reaction of the EEG is analyzed and quantified.5'7 When the subject's eyes are closed, there is a high electrical activity in the occipital zone. With the subject's eyes open this activity is reduced, and this reduction of activity can be quantified with an index of visual reactivity (VR), ranging from O to 1. VR compares absolute values of cerebral activity in microvolts squared at the level of 4 occipital electrodes in the alpha band of the EEG, with references from the eyes open (EO), or each eye separately, and then the eyes closed (EC). The formula is derived as follows: VR-(EC- EO) / EC. The closer the VR value is to 1, the better it is; the VR may be O in the absence of reduced activity on eye opening and may even be negative.
An examination is considered normal (Fig 2) when a VR exists on opening of both eyes or when the differential VR (DVR), or difference in the visual reactivity between the right and left eye, is in the region of zero, thus demonstrating an interocular symmetry of perception.
For this study, 120 children, aged 1 month to 16 years (average age, 24 months), underwent at least one BEAM examination. A total of 156 examinations were performed. All children underwent the same type of examination, which was repeated during follow up. They all had clinical testing, including assessment of visual development, fixation, pursuit, cover test, prism response, and binocular vision testing, if the subject's age allowed it. Visual acuity was measured with Teller acuity cards for preverbal subjects or with age related charts for verbal subjects. Thirty-four children, both verbal and preverbal, underwent VEP with the use of patterns to test visual acuity, and flash VEPs were obtained if necessary to assess optic nerve function.
The BEAM was performed with a commercially available apparatus (Cartovar 2000, Alvar Electronic, Paris, France). The technique is noninvasive, as the 12 electrodes are attached to the scalp. Testing lasts approximately 1 hour. Awakening of the children was controlled during the testing. Four measurements were carried out under photopic conditions in a neutral room during an artifact-free period of at least 30 seconds, with the subject's eyes open, eyes closed, right eye open, and left eye open. The subject's attention was monitored throughout the examination.
The analysis of the recordings was carried out in masked fashion by a neurologist (M.T.), who studied the presence, organization, and intensity of the different cerebral rhythms and then studied the VR in the alpha and other frequency bands. Each mapping was systematically compared with that of the EEG, which allowed us to eliminate, for example, movement-related artifacts and any associated neurologic disease. Following this, VR was calculated under three of the four experimental conditions: right eye open, left eye open, and both eyes open.
Three groups of subjects could be isolated from the analysis. Group 1 consisted of 17 normal children. Because we did not have any reference for the normal results of mapping in children younger than 18 months, we carried out 17 examinations in children who were free from all neurologic or ophthalmologic disease, aged between 1 month and 3 years. These controls allowed us to study the normal characteristics of brain mapping and VR in very young children, in whom brain electrogenesis is immature. Group 2 consisted of 55 children, 44 of whom had clinical evidence of amblyopia (strabismic, deprivation, or anisometropic). In the other 11 cases, amblyopia was suspected but clinical and other methods left the initial diagnosis in doubt, despite presence of an evident symmetric (n = 4 cases) or asymmetric (n = 7 cases) ocular disease. Group 3 consisted of 48 children who had a pathologic visual development without unilateral amblyopia. Twenty-nine had strabismus, and 19 had various types of visual maturation delay.
Of the 156 examinations performed, 20 (12.8%) were uninterpretable. Fifteen of these were under the age of 3 months and had a cerebral electrogenesis too immature to give a VR; the 5 others were older children in whom an artifact-free recording could not be obtained. In all 20 cases BEAM could be obtained in another attempt.
FIGURE 2: Normal results of brain electrical activity mapping (BEAM) in a 3-yearold child with normal vision. Four maps are analyzed in the occipital area (electrodes L, C, G, and J). Top right, with both eyes closed there is intense occipital activity (dark gray). Top left, with both eyes open there is extinction of occipital activity (light gray), with a visual reactivity (VR) of 0.80 which represents a good reactivity. Bottom left, with the left eye (LE) open the VR is good, 0. 75. Bottom right, with the right eye (RE) open, the VR is also good, 0.80. The differential VR (DVR) between the two eyes is less than 0.1.
Group 1: Control Study
The results of the control study were previously published in detail.8 Under the age of 8 months there is no interpretable alpha activity in terms of VR in the posterior zones. The alpha rhythm is not fully mature until the age of 12 months, but a measurable VR, begins to appear in this rhythm from the age of 8 months on.9·10 We therefore decided to look at the aspect of mapping and VR on slower rhythms, which appear earlier in the brain electrogenesis sequence. We studied the theta rhythm (3.5 to 7 hertz), which appears toward the age of 3 months and which also has a stop reaction in the occipital zones on eye opening and is thus measurable in terms of VR. It is on the theta rhythm that it is best to study VR for subjects younger than 1 year of age.
After the age of 8 months alpha activity with VR gradually appears and VR in the theta rhythm gradually disappears. This rhythm persists but at a far lesser intensity.
In this study we found substantial individual variation between subjects regarding the precise timing of appearance of VR in the different frequencies. However, in all children observed, the theta rhythm with reactivity was present at the age of 6 months, and the alpha rhythm with reactivity was able to be analyzed toward the age of 10 months. Between 6 and 10 months, we used the rhythm that had the best VR, varying from one child to another. BEAM did not appear to be a dependable examination in children under the age of 3 months.
Group 2: Amblyopia
The results of studies of BEAM in children with amblyopia were published previously in part.11·12
Fifty-five children aged 2 months to 16 years (average age, 30 months) had known or suspected amblyopia, as demonstrated on other examinations.
The important factor in the study of amblyopia is the DVR: DVR - VRl - VR2, where VRl is the visual reactivity of the normal eye and VR2 is the visual reactivity of the amblyopic eye. This index was greater when profound amblyopia existed (Fig 3).
Alteration of DVR, The DVR was greater when an organic amblyopia was clinically profound, and it measured less in strabismic amblyopia. In anisometropic amblyopia, on the other hand, DVR was almost normal despite the presence of a clinically profound amblyopia (Fig 4): Twenty-two children, with an average age of 30 months, had an organic amblyopia. The average DVR was 0.36, with a range from 0. 12 to 0.74. Twelve children, with an average age of 48 months, had strabismic amblyopia. The average DVR was 0.17, with a range from 0.05 to 0.37. Eleven children, with an average age of 21 months, had anisometropic amblyopia. The average DVR was 0.12, with a range from 0.02 to 0.40.
FIGURE 3: Results of brain electrical activity mapping (BEAM) indicating probability of amblyopia in a 4-year-old child who had a traumatic cataract of the left eye (LE) 6 months earlier. The visual reactivity (VR) is good for the maps with eyes open (EO) and the right eye (RE) open (0. 73 and 0.68, respectively). The VR of the left eye (LE) is only 0.04, with a differential VR (DVR) of 0.64. Profound amblyopia was found after surgery, and there was a good improvement in visual acuity after patching of t he RE.
FIGURE 4: Differential visual reactivity (DVR) as a function of the type of amblyopia.
Mapping seemed to be a more dependable method for evaluation of the degree of deprivation or strabismic amblyopia. These results appeared to confirm that amblyopia is more serious, from a neurologic point of view, when it occurs with the phenomenon of suppression of vision.
Reliability of Beam in Amblyopia. When anisometropic amblyopia was excluded from analysis, 34 children presented with definite strabismic or organic amblyopia, as diagnosed from clinical examination, from asymmetry of the VEP, or on preferential looking. In all cases, we observed an asymmetry in VR. The DVR had an average value of 0.25 and was always greater than 0.10 in the alpha or theta rhythms, depending on the age of the patient and on the activity observed in each of these rhythms. In cases of anisometropia, when an asymmetry in VR was seen, this always reflected the clinical impression, but variations were less substantial and even nonexistent in some cases.
In seven cases we could not have any opinion on the presence or absence of amblyopia despite potentially amblyogenic findings (2 cases of proptosis, 2 cases of paracentral corneal scars, 1 case of unilateral recurrent uveitis, and 2 cases of Rieger"s anomaly with asymmetric abnormal pupils). In these cases BEAM was found to be asymmetric and the subsequent clinical course confirmed the presence of amblyopia.
FIGURE 5: Electrical signs of amblyopia, which appeared on brain electrical activity mapping (BEAM) before clinical signs in a child with normal visual development and no strabismus or amblyopia but with bilateral symmetrical congenital cataracts. At 12 months of age, there is asymmetry of visual reactivity (VR)1 and the differential VR (DVR) is 0.51, indicating a probability of amblyopia of the left eye (LE). At 15 months, strabismus becomes clinically evident, with the right eye (RE) fixing and LE amblyopic. At 18 months the cataract in the LE has become more dense on clinical examination.
Finally, in four cases (one case of proptosis and three cases of bilateral and symmetrical congenital cataracts) we had definite clinical evidence that amblyopia did not exist; in all four cases, we were surprised to find VR values that were completely asymmetric. In these cases signs of amblyopia appeared a few months later (Fig 5). Analysis of these last four cases indicated that electrical signs of amblyopia might exist before clinical signs become evident.
Evolution of VR After Treatment of Amblyopia. Eleven children aged between 5 months and 6 years (average age, 30 months) underwent at least two BEAM examinations. In all cases we observed an improvement in VR that took place just before clinical improvement in amblyopia became evident. Before treatment, the visual acuity varied between absence of fixation and 20/200, with an average DVR of 0.31 (range 0.04 to 0.70). After treatment the visual acuity varied between alternating fixation and 20/50, with an average DVR of 0.13 (range 0.03 to 0.40).
Prognostic Implications of BEAM. To determine if the initial results of the cerebral mapping could allow us to predict the eventual outcome of treated amblyopia, we carried out BEAM in 48 amblyopic children, after treatment of the deprivation cause when such existed but before treatment of the amblyopia. In nine of these children the amblyopia was unresponsive to treatment, and in the other 39 children the amblyopia was treated successfully (Fig 6).
Analysis showed that a very moderate difference in the average DVR existed between these two groups. Analysis of each individual case did not allow us to predict the ultimate outcome of the amblyopia. Indeed, there were children who had an initial DVR of 0.30 and who have never gain vision, and on the other hand, we encountered children with greater initial DVR of 0.70 whose vision returned to excellent.
Group 3: Pathological Visual Development
Alternating Strabismus. We studied a group of 29 children, aged 3 months to 7 years (average age 28 months), consisting entirely of cases of alternating strabismus without clinically evident amblyopia, 17 of which were early-onset esotropías. We compared the aspects of BEAM found in this group (presence, organization, intensity, and VR of the cerebral rhythms) to the aspects found in the control group.
The only remarkable observation was a predominance of crossed VR (for example, right cerebral cortex reactivity with the left eye open) as distinct from direct VR (left cerebral cortex reactivity with the left eye open) in all children under the age of 1 year. This finding persisted at all ages in children with early-onset esotropia or nystagmus.
FIGURE 6: Initial dii Thrent Eat visual Teattivity (DVR) and outcome of amblyopia after treatment. The initial results of brain electrical activity mapping (BEAM) do not allow prediction of the chances of success of treatment. There tins no obvious dijfrrence between the two groups of children, and in each treat ment group DVR varied considerably between cases despite identical outcomes.
Delayed Visual Maturation. We encountered 19 children, with an average age of 10 months (range 3 to 18 months), who had an obvious visual disability, such as an absence of pursuit or of fixation after the age of 6 months, wandering fixation, or eventually unreactive pupils.2·8 The clinical progression allowed us to divide these children into four subgroups on the basis of etiologic and pathologic factors, with each subgroup having its own particular type of mapping results.
In the first subgroup, seven children recovered normal visual development between the ages of 12 and 18 months. In these cases we observed a delay in electrogenesis, which gradually reverted to normal. The theta rhythm appeared in all cases by the age of 6 months, and the alpha rhythm appeared in all cases by the age of 12 months.
We observed seven children in a second subgroup who had suffered serious neonatal injury. In all cases, cerebral palsy occurred. In every case, the electroretinogram was normal and the VEP was abnormal. Mapping was not able to be interpreted in these children because the EEG was always abnormal (hypsarythmia) and the VR was very weak.
In the third subgroup we encountered three children, aged between 7 and 9 months, who had presented wandering fixation. The results of ophthalmologic examination were absolutely normal, and a good defense reflex was present. In these three cases, mapping was normal for the age of the child. In all of these children we observed a gradual progression toward profound behavioral disorders, which eventually necessitated psychotherapy in a framework of serious family problems. Autism was suspected initially and is now confirmed in two of the three children.
In the last subgroup were two children who had clinical blindness and a flat electroretinogram, indicative of Leber congenital amaurosis. In both cases there was no visual reactivity on mapping, but the underlying EEG was normal.
BEAM was used in the early 1970s but had disappointing results when neurologists tried to use it for localization of pathologic cerebral lesions. The technique was of greater interest in the study of cerebral function.13·14 In the field of ophthalmology, Japanese researchers in the 1980s demonstrated a variation in the alpha occipital rhythm seen with modification of light stimuli that was called the visual stop reaction.5 It is found on the EEG of eye opening and on occipital electrodes in the alpha rhythm.6
Orssaud and colleagues7 used this visual stop reaction in the alpha band of the EEG, employing a cartographic method. The study was carried out in older children and adults. They showed that there was a decrease in cerebral activity on the alpha band of the EEG on the opening of either or both of nonamblyopic eyes but that this was not the case when one amblyopic eye was opened. They believed that the absence of VR in the amblyopic eye indicated cortical neutralization with a decrease in neuronal activity.
In our present study we showed that this noninvasive technique may be carried out easily and painlessly in very young children. As BEAM had never been used in this population, we first showed that it may be performed and interpreted at any age older than 3 months. The brain rhythm used for analysis must be chosen on the basis of the level of electrogenesis of the child. The method is, however, limited by the need for sophisticated equipment and for close cooperation between neurologist and ophthalmologist, who will produce interpretations that are as rigorous and as careful as possible.2 Any brain disease or even movement related artifacts may impair the results so that they are unable to be analyzed. Twenty of our 156 BEAM examinations (12.6%) were !ininterpretable, and this represents a fairly moderate proportion compared with that found with preferential looking or VEP.
BEAM was found to be very useful as the only examination able to "show* amblyopia. The method was routinely used to explain clearly to the children's parents that an amblyopia was present and what this condition was, and it eventually showed the effect of the treatment if a second examination was performed. Parents appreciated this explanation, especially during the treatment of severe deprivation amblyopia in preverbal children with unilateral cataracts, in which they found it more concrete than Teller visual acuity. It is well known that a good level of parental comprehension is an important factor in the successful treatment of amblyopia.
In our mappings of amblyopia, we found no false negatives and BEAM consistently confirmed the results of clinical and paraclinical assessment in cases of known amblyopia. When the DVR is greater than 0.1, this asymmetry may be regarded as proof of the existence of amblyopia, but this information did not allow us to predict the likelihood of cure of the amblyopia. BEAM cannot yet tell us, for example, in a case of unilateral cataract discovered at 6 months, if it is worth treating the amblyopia when the DVR is almost normal or, on the contrary, greatly altered.
The variation in VR seen in different types of amblyopia and, even more importantly, the lack of correlation with clinical signs, indicate that cerebral mapping seems to demonstrate amblyopia at a cortical level. Perhaps electrical activity, which reflects neuronal activity, shows alterations that appear and disappear before clinical signs of amblyopia. This could be of great interest in the management of diseases with a high risk for the development of amblyopia, for example, proptosis and bilateral developmental cataracts. We were amazed to find that in cases of bilateral and perfectly symmetrical developmental cataracts, which showed no clinical signs of amblyopia or binocular visual disability, results of BEAM were asymmetric and 2 or 3 months later clinical signs of amblyopia or strabismus were evident.
Also of interest was the assessment of visual maturation with BEAM. In alternating strabismus we found no specific signs. An asymmetry of crossed and direct reactivity was found in the population of early-onset strabismus and in children younger than 6 months. It is too soon to conclude if BEAM can isolate early-onset strabismus (by findings of asymmetric brain reactivity) from secondary strabismus, as optokinetic nystagmus can. More studies must be carried out to confirm if BEAM may serve as a marker of visual pathways. The use of BEAM was more practical in evaluation of visual maturation delay, in which different results could be fitted with different outcomes. An electrical maturation delay was found in the classical aspect of pure visual maturation delay (Beauvieux disease), with a good visual outcome (a nystagmus was still present after 10 months). Neurologic impairment was isolated on the EEG, with extremely abnormal BEAM results, demonstrating the neurologic origin of the visual maturation delay. Our cases of tapetoretinal degeneration also showed no VR but a normal EEG. We were surprised to find among the cases of visual disinterest, which had normal BEAM results, a gradual evolution toward profound behavioral disorders. In all these cases, which had almost the same visual results, BEAM could separate different possible outcomes. We believe that more studies need to be done on this specific subject.
In conclusion, BEAM seems to be of use in the study of amblyopia and of normal and pathologic visual maturation in the child, as an adjunct to the other classical methods of visual assessment. This method may be used in the initial assessment of certain cases of organic or strabismic amblyopia and during treatment, as a way to confirm clinical evaluation. It cannot be used routinely because of the need for sophisticated technology and close cooperation with a neurologic staff If our observations are confirmed by other studies, BEAM may show specific electrical signs of vision, amblyopia, and visual pathways.
1. Oresaud C, Dulier JL. Objectiva t ion par cartographie cérébrale de Ja neutralisation corticale au cours de l'amblyopie fonctionnelle. Bulletin des Sociétés Ophtalmologie de France. 1989;3:375-378.
2. Thouvenin D, Tiberge M1 Arne JL, et al. Cerebral mapping: a new method for assessment of visual maturation and amblyopia in infants. Presented at the Seventh Internationa] Orthoptic Congress; June 2, 1991; Nuremberg, Germany.
3. Duffy FH, Barters P, Burchfield JL. Significance probability mapping: an aid in the topographic analysis of brain electrical activity. Electroencephalogr Clin Neurophysiol. 1981 i5:455-462.
4. Etevenon P, Peron-Magnan P, Pidoux P. Analyse de Fourier. Analyse statistique d'amplitude et quantitative de l'EEG temps réel. Revue d'Electroencephalographie et de Neurophysiologie Clinique. 1976;6: 246-254.
5. Inouye T, Sumitsuji N, Matsumoto K. EEG changes induced by light stimuli modulated with the subject alpha rhythm. Electroencepkalogr Clin Neurophysiol. 1980;49:135-142.
6. Srebro R. Reduction of EEG activity caused by a light flash. Electroencephalogr Clin Neurophysiol. 1982;54:707-710.
7. Orssaud C, Le Neindre P, Menard F, et al. Intérêt de la cartographie cérébrale dans l'étude de l'amblyopie. J Fr Ophtalmol. 1989;12: 111-118.
8. Thouvenin D, Tiberge M, Arne JL, et al. Intérêt de la cartographie cérébrale dans l'étude du développement visuel et de l'amblyopie chez l'enfant. Revue Otoneuro Ophthalmology. 1 993 ;2 3/24:49 -5 2.
9. Delay J1 Verdeaux G. Electroencephalogramme de l'enfant. In: Electroencéphalographie Clinique. Paris: Masson Publishers; 1979;234-288.
10. Monod N, Tharp B. Activité electroencephalographique normale du nouveau-né et du prématuré au cours de l'état de veille et du sommeil. Revue d'Electroencephalographie et de Neurophysiologie Clinique. 1977;7:302-315.
11. Thouvenin D, Arne JL, Tiberge M. Cartographie cérébrale et évaluation fonctionnelle des amblyopîes organiques chez l'enfant. Presented at the first International Symposium in Ophalmology; September 9, 1993; Bordeaux, France.
12. Thouvenin D. Cartographie cérébrale et etude du développement visuel normal et pathologique et de l'amblyopie du jeune enfant. Journal Français Orthoptie. 1993;25:37-46.
13. Duffy FH, Burchfield JL, Lombroso L. Brain electrical mapping (BEAM): a method for extending the clinical utility of EEG and evoked potential data. Ann Neurol. 1979;5:309-321.
14. Gachés J, DoI F. Electroencephalographie quantifiée (cartographie I et tomodensìtométrie crânienne. In: Encyclopédie Medico Chirurgicale, Neurologie. Paris: Masson Publishers; 1 986 :A50.