Residual Visual Function after Loss of Both Cerebral Hemispheres in Infancy

Abstract

purpose. To investigate whether and what kind of visual function is still present in the absence of both cerebral hemispheres.

methods. Binocular visual function of five children who had suffered the loss of both cerebral hemispheres and the visual fields of 30 controls 5 to 12 months of age were examined according to a perimetric method based on forced-choice, preferential-looking methods.

results. Results show that after the destruction of both cerebral hemispheres, a stimulus presented binocularly beyond 5° eccentricity did not elicit a response. However, two children were still able to fixate steadily and to follow a stimulus presented binocularly within the central 5°, with eye and head movements despite the absence of both cerebral hemispheres. One child responded only to a moving face or a moving drum with black and white stripes presented binocularly within the central 5° but not to a moving spot of light. The binocular visual field of 30 controls 5 to 12 months of age almost reached the dimensions of the adult binocular visual field.

conclusions. Neural structures in the midbrain, including the superior colliculi and the pretectum, seem to be able to mediate visual function in the foveal and macular regions. These structures are, however, unable to mediate the presence of a functional visual field beyond 5° eccentricity.

It has been shown that patients who suffered from damage to the visual cortex of one cerebral hemisphere and even patients who underwent cerebral hemispherectomy were not completely blind in the visual hemifield contralateral to the damaged or removed cerebral hemisphere. Some of these patients were still able to detect, locate, and discriminate visual stimuli when asked to guess the presence, absence, or location of stimuli or when asked which of two stimuli was present. ^{1 2 3 4 5 6 7 8 9 10 11 12 13 14 15} Given that the patients described themselves as blind and unable to have seen any of the stimuli, these residual abilities were termed blindsight. Hemispherectomized patients were even able to detect the presence of light targets in the visual hemifield contralateral to the hemispherectomy with full awareness even if the influence of light scatter reaching the good visual field was excluded. ¹⁶  

It was assumed, that in hemispherectomized patients the visual functions of patients are mediated by the remaining cerebral hemisphere ¹⁷ or that subcortical structures, such as the superior colliculi, the pretectum, and the pulvinar, play a decisive role in mediating residual visual function and that cortical processing is not necessarily required. ^{18 19} The lateral geniculate nucleus may also be involved in mediating residual visual function after loss of the striate cortex. After destruction of the striate cortex, 99% of the neurons in a monkey’s dorsal lateral geniculate nucleus degenerate. ²⁰ The surviving neurons project to extrastriate visual areas. ^{21 22} When only the striate cortex was destroyed in patients, blindsight might also have been mediated by the surviving neurons in the lateral geniculate nucleus. 

Whether and to what extent the brain stem is able to mediate visual function in the absence of the cerebral hemispheres can only be answered by investigating residual visual abilities in humans in whom both cerebral hemispheres are missing. Although some of these anencephalic children are apparently completely blind, ^{23 24} in rare cases the responses of hydranencephalic infants have been reported to diffuse illumination of the eyes, ^{25 26 27 28 29} fixation, and visual tracking of objects. ^{16 30 31} Shewmon et al. ¹⁶ reported steady fixation and signs of visual discrimination of objects in a hydranencephalic child with frontal lobe remnants. 

These reports about the visual abilities of children lacking both cerebral hemispheres lead us to assume a visual field has developed. However, the visual fields of these children had never been assessed. It was, therefore, unclear whether in humans a visual field can develop in the absence of both cerebral hemispheres. In the present study, the visual fields of five children 9 to 87 months of age in whom both cerebral hemispheres were completely absent or in whom only remnants of the frontal, temporal, or occipital lobes were left were compared with the visual fields of healthy controls 5 to 12 months of age. Because these children were unable to understand and follow instructions, perimetric methods used for adult patients could not be applied. In these children, the extension of the visual field was examined using a perimetric method based on forced-choice, preferential-looking methods. ^{17 32 33 34}  

Results showed that the patients may still be able to follow a stimulus with eye and head movements if both cerebral hemispheres are missing and if the midbrain, including the superior colliculi and pretectum, is preserved. However, even if the children survived to the age of 7 years, no visual field developed. Control studies, which exclude an interpretation of the results of these experiments due to the influence of scattered light, have already been reported. ³³  

Patients and Methods

Patients

Five brain-damaged children 9 to 87 months of age (Table 1)were tested. In one child (SA), the occipital lobe of both cerebral hemispheres was missing because of prenatal cerebral damage, as verified by magnetic resonance imaging (MRI). In four children, both cerebral hemispheres were completely absent, or only remnants of the cerebral hemispheres were present. The visual fields of these patients could be assessed reliably because they were able to hold their heads upright and did not suffer from eye movement disturbances that interfered with arc perimetry. 

Patient SA was a 38-month-old girl born in the 35th week suffering from spastic tetraparesis. MRI showed a cavity filled with cerebrospinal fluid (CSF) occupying the occipital lobes, the parietal lobes, and most of the frontal and temporal lobes of both cerebral hemispheres (Fig. 1) . Occipital lobes and parietal lobes of both cerebral hemispheres were completely absent. Dorsal aspects of the temporal and frontal lobes were also missing. Only the cortex and white matter of the ventral temporal lobe and of the anterior and ventral frontal lobes were preserved. 

Patient PL was a 28-month-old boy born in the 27th week suffering from hydranencephaly (Fig. 2) . The frontal, parietal, and occipital walls of the CSF-filled space consisted of remnants of brain tissue. No gyri could be identified in the occipital, parietal, and temporal lobes. Optic radiation was completely absent. The optic nerves, the cerebellum, the pons, and the midbrain, including the superior and inferior colliculi, were preserved. 

AG was a 7-year-old boy born at term suffering from hydranencephaly and spastic diparesis. The cortex and white matter of the cerebral hemispheres were completely absent and had been replaced by CSF (Fig. 3) . The optic nerves and the brain stem, including the midbrain, were well preserved. 

Patient HE was a 12-month-old girl born at term. The patient suffered from hydranencephaly. MRI showed that the cerebral hemispheres had been replaced by CSF-filled spaces (Fig. 4) . Only remnants of cerebral tissue were observed in the anterior part of the frontal lobe. The optic nerves, the cerebellum, the pons, and the midbrain, including the superior and inferior colliculi, were preserved. 

Patient ER was a 28-month-old girl born at term suffering from hydranencephaly and spastic tetraparesis. MRI revealed that the cerebral hemispheres, which had been replaced by a CSF-filled space, contained septumlike remnants of glial tissue (Fig. 5) . The optic nerves, the pons, and the midbrain, including the superior and inferior colliculi, were all intact. 

Controls

In the control study, the visual fields of 30 healthy children, 18 boys and 12 girls 5 to 12 months of age (mean ± SD, 9.24 ± 1.96 months), were assessed using the arc perimeter. 

Visual Perimetry

Sizes of the visual fields and functional luminance-difference thresholds were assessed using a specially designed, noncommercial arc perimeter consisting of a semicircular, semitranslucent white screen of 41-cm radius. During perimetric testing of the visual field, the child was held on an assistant’s or a parent’s lap. The child’s head was positioned at the center of the perimeter facing the semicircular screen at a distance of 41 cm. The child’s head was supported and stabilized by hand. On the screen, the fixation point (diameter, 1.5°) and the target (2.5°) appeared. Target luminance could be varied between 0 cd/m² and 26,000 cd/m². Background luminance was 0.03 cd/m². Head position, fixation, and eye movements were controlled by an infrared sensitive videocamera displayed on a high-resolution monitor. During perimetric testing, an assistant controlled fixation of the central fixation point and eye and head movements on the video monitor. An investigator and an assistant independently judged whether an eye movement was directed toward a target. The investigator did not know where the targets would appear. To exclude the presence of hemispatial neglect, spontaneous eye movements were recorded when no visual stimulus was present in the perimetric arc. 

The visual field was examined in steps of 5° or 10° along the horizontal and the vertical meridians and along the 45°, 135°, 225°, and 315° meridians up to 90° eccentricity. Borders of blind areas were assessed in steps of 2° with stationary targets along the meridians. Luminance of the target was 5 cd/m². If the target was not detected, the measurement was repeated with a target luminance of 40 cd/m² because targets with a luminance below 50 cd/m² had no light scatter. ¹⁰  

At the beginning of each trial, the fixation point flickered with a frequency of 4 Hz. When the child directed his or her gaze to the fixation point, the flicker frequency of the fixation point was reduced. To prevent anticipatory eye movements, each child’s gaze was drawn, before each trial, to the location where the fixation point had been presented. After the fixation point had flickered at a rate of 4 Hz, the target appeared at least three times after 2.5 seconds at the location of the fixation point (fixation trials). An experimental trial or a blank trial followed only if the child maintained fixation for at least 2.5 seconds in at least three fixation trials. Children who did not maintain fixation for at least 2.5 seconds were excluded from the test. 

At the beginning of each experimental trial, the fixation point also flickered with a frequency of 4 Hz. When the child directed his or her gaze to the fixation point, the flicker frequency of the fixation point was reduced. When the fixation point had disappeared and the child was still directing his or her gaze to the center of the perimetric screen where the fixation point had been shown, the target was presented for 4 seconds (experimental trials). If eye movement did not occur within 4 seconds after the onset of the target, the target was removed, and the trial was terminated. 

The visual field of each patient was assessed in two ways. In one measurement session, the target was shown at first between 15° and 50° eccentricity in the left or in the right half of the visual field. In the following trial, the target appeared at a different location in the same or in the opposite visual hemifield or the following trial was a blank trial. The target position and visual hemifield in which the target was shown varied in random order. The visual field was also assessed by showing the target at least six times at the same eccentricity in either the left or the right visual hemifield. The sequence with which the target was shown in the left or in the right visual hemifield varied randomly. 

Half the trials were blank trials identical to the experimental trials, except that no target appeared after the offset of the fixation point. Experimental trials and blank trials were of equal length and followed in random order. Intertrial intervals varied randomly between 5 and 8 seconds. A target was regarded as detected only if (1) it elicited a valid saccadic eye movement (i.e., one saccade reaching the target or one saccade directed toward the target followed by a small correction saccade reaching the target) followed by fixation of the target in at least three successive trials, (2) if the eye movement occurred within a time interval of 4 seconds after the onset of the target, and (3) if, in at least three successive blank trials, there was no eye movement or, in successive blank trials, there were only searching eye movements consisting in a succession of saccades. Searching eye movements were invalid. 

Trials were regarded as invalid if a patient only looked, in one or two of three successive trials, at a target or if the patient made an eye movement in the direction of the target undershooting or overshooting the target more than 5°. Invalid trials did not occur. An area of the visual field was regarded as blind if the patient did not respond to a target in this area. Details of the control of scattering light have been reported. ³³  

When children did not direct their eyes to the fixation point, the position of the target in the visual field was estimated by the distance of the target from the middle of the cornea. As soon as the eye was at rest, a spot of light flickering with a frequency of 8 Hz and a luminance of 1000 cd/m² and 25,000 cd/m² was presented at various distances between 5° and 90° on eight meridians of the visual field. In addition, the same targets were moved at a velocity of approximately 1°/s from the periphery toward the center of the visual field. When the presentation of the target was interrupted by an eye movement, the target was presented again at the point on the cornea at which the presentation had been interrupted in the preceding trial. The target was again moved toward the center of the visual field. The target was moved along the 0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315° meridians of each eye. 

Additional Ophthalmologic Investigations

Optokinetic nystagmus was tested by showing the children a rotating drum with black and white stripes of a frequency of 0.7 cyc/deg (length of the stripes, 100 cm; radius of the drum, 20 cm; luminance of black stripes, 8 cd/m²; luminance of white stripes, 90 cd/m²; velocity of stripe movement, 2°/s and 5°/s), which was shown at a distance of 30 cm from the patient’s eyes. Eye movements were assessed using a light spot (diameter, 4 mm; luminance, 45 cd/m² or 300 cd/m²) or a 14 × 14-cm² black (luminance, 8 cd/m²) or white (luminance, 90 cd/m²) object, which moved at a distance of 30 cm horizontally or vertically at a velocity of approximately 3°/s. Optical alignment and pupillary reflexes were tested with routine methods. 

Responses to Acoustic Stimuli

If the patient did not turn his or her eyes and head toward a visual stimulus, we tested whether this was because of a general paucity of eye and head movements or if the patient directs his or her eyes and head toward an acoustic stimulus. Responses to acoustic stimuli were measured presenting an acoustic stimulus 20 times for 5 seconds to each ear. The sound level of the stimuli was 84 dB, measured with an impulse precision sound-level meter and an artificial ear (Brüel & Kjaer, Naerum, Denmark). The distance of the sound generator from the ear was 20 cm. The sequence of presentations to the left or to the right ear was random. The control experiment consisted of 20 blank trials in which head movements to the left and 20 blank trials in which head movements to the right were examined. The blank trials were identical to the test trials, except that the sound was not turned on. A stimulus was regarded as detected when the head was turned significantly more frequently toward the stimulus within 5 seconds than away from the stimulus and when the head was turned significantly more frequently toward the stimulus than the head was turned toward this side in blank trials. Head movements were recorded with a videocamera and were displayed on a high-resolution monitor. 

Data Analysis and Statistical Evaluation

The frequency of eye and head movements to the left half and to the right half of space was determined for each patient in the presence and in the absence of the target. Differences between the frequency of eye and head movements to the left and to the right were calculated for each experiment. For each patient, these differences were tested on significance using the McNemar χ² test for dependent samples. Means of the latencies of saccadic eye movements to the left and to the right were compared using the t-test for dependent samples. 

This study was approved by the ethics committee of the University of Munich and was performed in accordance with the tenets of the Declaration of Helsinki. Parents of all children gave their informed consent before their children’s inclusion in the study. 

Results

Controls

Binocular visual fields of children 5 to 12 months of age in the control group had reached almost normal size (Table 2) . On the horizontal meridian, the mean extension of the right half of the binocular visual field was 85.2° (SD, ±3.8°), and the mean extent of the left half of the binocular visual field was 86.0° (SD, ±3.6°). On the vertical meridian, the mean extent of the upper half of the binocular visual field was 55.3° (SD, ±5.1°), and the mean extent of the lower binocular visual field was 63.8° (SD, ±7.2°). In seven children aged 5 to 7 months, both binocular visual hemifields had already reached an extent between 80° and 90°. In these children, the upper binocular visual field had reached an extent of 50° to 62° on the vertical meridian. On the vertical meridian, the lower binocular visual field had reached an extent of 60° to 70°. 

Visual Function after Bilateral Loss of the Occipital and the Parietal Lobes

Patient SA with hydranencephaly followed a drum with black and white stripes of a spatial frequency of 0.7 cyc/deg (luminance of black stripes, 8 cd/m²; luminance of white stripes, 90 cd/m²), which was shown at a distance of 30 cm from the child’s eyes, with eye and head movements in 18 of 20 trials when the whole drum was moved horizontally in both directions at a velocity of 3°/s or 10°/s in the absence of acoustic stimuli. There was, however, no response to the moving black and white stripes of a stationary rotating drum. 

When an assistant moved his face (luminance of the skin, 22.9–35.5 cd/m²) horizontally at a velocity between 1°/s and 3°/s, the child fixated the face steadily and followed the face with a succession of small saccades and head movements in all 20 trials. When the drum or the face moved from the periphery of the visual field toward the center, the child did not respond. The child fixated the drum or the face only when the center of the visual field was reached. When a 14 × 14-cm² black object (luminance, 8 cd/m²) or white object (luminance, 90 cd/m²) was moved 20 times at the same distance with the same velocity in front of the child’s face, the child did not respond. These results were irrespective of the sequence in which the child’s visual abilities were tested with the drum, the face, the black object, or the spot of light. The child closed his eyelids when the eyes were exposed to diffused light with a luminance of 60,000 cd/m². Direct and indirect pupil reactions were observed in both eyes. 

The child did not respond to light targets up to a luminance of 25,000 cd/m² when the stimuli were presented between 5° and 80° eccentricity in the left or the right binocular visual hemifield (Fig. 3) . The patient did not follow a target (diameter, 4 mm) up to a luminance of 25,000 cd/m² when the stimulus moved at a distance of 30 cm horizontally or vertically with a velocity of 3°/s or 10°/s. 

When an 84-dB sound was presented at a distance of 20 cm from the left or the right ear, the child turned his head toward the acoustic stimulus. The child responded to all 20 presentations to the left ear and to 18 of 20 presentations of the stimulus to the right ear. In blank trials, when the loudspeaker was near the left ear 20 times, the child spontaneously turned his head to the left four times. When the loudspeaker was near the right ear 20 times but no sound was produced, the child spontaneously turned his head to the right twice. The difference between head turns toward the loudspeaker in test trials and in blank trials was highly significant (McNemar χ² for dependent samples; P < 0.001). 

Visual Function in the Absence of Both Cerebral Hemispheres

Patient PL with hydranencephaly (Fig. 5)followed a light target (diameter, 4 mm; luminance, 100 cd/m²) with a sequence of small saccades when the target was moved 10 times horizontally in both directions at a velocity of approximately 3°/s at a distance of 30 cm from the eyes. However, the child showed no response to light targets in the left or in the right half of the binocular visual field if the target had a luminance up to 25,000 cd/m². There was no optokinetic nystagmus to a drum with black and white stripes with a spatial frequency of 0.7 cyc/deg (luminance of black stripes, 8 cd/m²; luminance of white stripes, 90 cd/m²), which was shown at a distance of 30 cm from the boy’s eyes moving horizontally or vertically, and there was no visual thread response. If a 14 × 14-cm black or white object (luminance, 8 cd/m² or 90 cd/m²) moved at a distance of 30 cm in front of the boy, he did not follow the object with eye or head movements. The child neither fixated nor followed the drum or an assistant’s face when it moved 20 times horizontally at a velocity between 1°/s and 3°/s horizontally at a distance of 30 cm from the child’s eyes. The child did not respond if the drum or the assistant’s face moved 20 times from the periphery of the visual field toward the visual field center. The child closed his eyes when they were diffusely illuminated by a light with a luminance of 60,000 cd/m². Direct and indirect pupil reactions were observed in both eyes. 

The child turned his head toward an acoustic stimulus of 84 dB, which was 20 cm from his left or right ear. The child responded in 16 of 20 presentations of stimulus to the left ear and in all 20 presentations of stimulus to the right ear. In the blank trials, where the sound source was near the left ear 20 times, the child turned his head spontaneously to the left twice. When the loudspeaker was near the right ear 20 times but no sound was produced, the child turned his head spontaneously to the right three times. The difference between head turns toward the sound source in test trials and in blank trials was highly significant (McNemar χ² for dependent samples; P < 0.001). 

AG with hydranencephaly closed his eyes when they were diffusely illuminated by a light with a luminance of 60,000 cd/m². Direct and indirect pupil reactions were observed in both eyes. 

The child did not respond to light targets with a luminance up to 25,000 cd/m² presented between 10° and 90° eccentricity in the binocular visual field. In addition, he did not follow a light spot up to a luminance of 25,000 cd/m² when the stimulus moved 20 times at a distance of 30 cm horizontally or vertically at a velocity of 3°/s or 10°/s. There was also no response to a rotating drum with black and white stripes of a spatial frequency of 0.7 cyc/deg (luminance of black stripes, 8 cd/m²; luminance of white stripes, 90 cd/m²), which was shown 20 times at a distance of 30 cm from the child’s eyes. When the whole drum was moved 20 times horizontally in both directions at a velocity of 3°/s or 10°/s in the absence of acoustic stimuli, the child did not follow the drum with eye or head movements. There was also no response to a moving 14 × 14-cm² black (luminance, 8 cd/m²) or white (luminance, 90 cd/m²) object shown 20 times at a distance of 30 cm before the child’s eyes. The child neither fixated nor followed a face that moved 20 times in a horizontal direction at a velocity between 1°/s and 3°/s. 

The child turned his head toward an acoustic stimulus of 84 db, which was 20 cm from his left or right ear. The child responded to all 20 presentations to the left ear and to all 20 presentations of the stimulus to the right ear. In the blank trials in which the loudspeaker was near the left ear 20 times and near the right ear 20 times but no sound was produced, there was no response in 40 trials. 

Patients ER and HE closed their eyes when they were diffusely illuminated by light at a luminance of 60,000 cd/m². Direct and indirect pupil reactions were observed in both eyes. 

However, the children did not respond to visual stimuli up to a luminance of 25,000 cd/m² in the right half or the left half of the binocular visual field. No eye or head movement was directed toward the stimulus in 40 trials. The children showed no optokinetic nystagmus to black and white stripes of a spatial frequency of 0.7 cyc/deg moving horizontally or vertically, and there was no visual thread response in 20 trials. If a 14 × 14-cm black or white object (luminance, 8 cd/m² or 90 cd/m²) moved 20 times at a distance of 30 cm in front of them, they did not follow the object with eye or head movements. The children neither fixated nor followed the drum or an assistant’s face in 20 trials when it moved horizontally at a velocity between 1°/s and 3°/s horizontally and a distance of 30 cm from the children’s eyes. The children did not respond if the drum or the assistant’s face moved 20 times from the periphery of the visual field toward the visual field center. The children did not respond to an acoustic stimulus of 84 dB that was presented 40 times 20 cm from the left or the right ear. 

Discussion

Development of the Visual Field in Healthy Infants

Although the morphologic development of the human fovea has not yet been completed at the age of 13 months, ³⁵ children are able to detect and locate visual targets and to fixate them steadily. Children 5 months of age are already able to direct their eyes spontaneously to a flickering fixation point in the middle of a perimetric arc, and they shift their gaze immediately to a target when the fixation point disappears and the target appears. The present study shows that in humans, the binocular visual field nearly reaches adult size by 5 to 7 months of age, when the binocular visual field is assessed with a perimetric technique preventing the influence of extraocular and retinal light scatter. ^{17 32 33 34} The results are in agreement with the results of Mayer et al. ³⁶ and Lewis and Maurer, ³⁷ who also observed the extent of the temporal visual field in 6-month-old infants resembling that of adults. Mohn and Van-Hof-van-Duin ³⁸ found visual fields of almost adult extent by 1 year of age. Delaney et al. ³⁹ reported that even 30-month-old children had not yet attained visual field extents like those of adults. This divergence may be attributed to different methods used to assess visual fields in infants. However, it cannot be excluded that the course of development of the visual field is not equal in all children and that in some children the visual field expands faster than in others. Irrespective of these divergent results, at the age of 9 months, the visual field of healthy children is nearly that of adults, whereas no visual field greater than 5° eccentricity developed in the hydranencephalic children reported in the present study. When cerebral hemispheres are missing, it appears that only foveal or perifoveal visual fields can develop. 

Can Midbrain Structures Mediate Vision in Humans?

In earlier studies ^{1 5 6 9 40 41} in which residual visual functions were present in the visual hemifield contralateral to a corticoectomized or removed cerebral hemisphere, the superior colliculi were considered the “critical structure” for the residual visual functions. In these studies, patients were able to detect and locate visual targets and to discriminate between visual patterns presented in the visual hemifield contralateral to the corticoectomized or removed cerebral hemisphere by guessing. Werth ¹⁷ reported that after hemispherectomy, a patient was able to detect, with full awareness, the presence of light targets in a normally extended visual field. 

After damage to the occipital lobe, including the striate visual cortex, many children suffer from permanent homonymous hemianopia. ^{42 43} In a review of 67 children with visual impairment resulting from cerebral cortex lesions or periventricular damage to the white matter, Hoyt ⁴³ found spontaneous improvement of visual function in 43 children. These children were not completely blind; rather, visual function was preserved. After systematic visual field training, visual function developed within 3 months in 15 of 22 ³² and in 11 of 19 cerebrally blind children ³³ whose conditions had not improved spontaneously for at least 1 year. To date, reports about infants in whom both cerebral hemispheres are absent also suggest that their visual functions are not always completely lost. Although in many children missing both cerebral hemispheres no visual fixation and/or visual tracking could be elicited, ^{23 25 27 44} some hydranencephalic infants responded to diffuse illumination of the eyes ^{25 26 27 28 29} or steadily fixated objects and visually tracked moving objects. ^{16 30 31} Shewmon et al. ¹⁶ assumed from the clinical observation of a hydranencephalic child with frontal lobe remnants that even discrimination of persons and objects may be possible. These different findings may be attributed to differences in the extent and nature of the brain damage and also to differences in the plasticity of the remaining functional brain tissue. 

Patients SA (Fig. 1)and PL (Fig. 2)also showed residual visual function. The children followed a moving target with eye and head movements and closed their eyes to diffuse illumination despite the nearly total absence of both cerebral hemispheres. Patient SA steadily fixated a face or a drum with vertical stripes and tracked both objects when they moved in front of her. However, she neither fixated nor tracked a black object or a spot of light, even if the light spot had a much higher luminance than the drum or the face. Thus, the child’s residual visual function was not limited to seeing that an object was present. The patients did not, however, respond when a light target, a face, or a drum with vertical stripes was shown beyond 5° eccentricity. The patients never oriented toward a visual stimulus outside the foveal region or reached for an object. Precise perimetric measurement of the area between the fovea and 5° eccentricity was not possible in these children. Because light stimuli that were presented beyond 5° eccentricity did not elicit eye movement, it appeared that the children were unable to detect stimuli outside the fovea or outside the macular region. 

The assumption that subcortical structures mediate residual visual function is in agreement with findings of animal studies. If monkeys with bilateral occipital lesions are forced to use their residual visual functions, vision may recover to such an extent that the animals show almost normal visually guided behavior. These visual functions may be mediated by the extrageniculostriate (second) visual pathway, including the superior colliculi and pretectum. ^{45 46 47 48} However, not all studies about residual visual function in monkeys that appeared before the development of MRI included serial sectioning of the whole occipital lobe. Detailed anatomic studies have shown that the superior colliculi and pretectum may be able to mediate residual visual functions. After destruction of the striate cortex in monkeys, only P beta neurons of the retina degenerate, whereas P alpha and P gamma neurons are preserved. ^{49 50} Azzopardi et al. ⁵¹ have shown that after hemispherectomy in humans, retinal ganglion cells that could respond to visual stimuli survive. After hemispherectomy in monkeys, approximately 70% of the neurons in the superior colliculi, which receive afferents from retinal P gamma neurons, ^{52 53} survive. ⁵⁴ It is, therefore, possible that after hemispherectomy, the brain stem, including the superior colliculi and the pretectum, is able to process visual stimuli shown in the visual hemifield contralateral to the removed cerebral hemisphere. However, both patients, who still fixated and visually tracked objects, had some remnants of their cerebral hemispheres. Patient SA still had remnants of the cortex and white matter of the ventral temporal lobe and the anterior and ventral frontal lobes. In patient PL the frontal, parietal, and occipital walls of the CSF-filled space that had replaced the cerebral hemispheres consisted of remnants of brain tissue. In earlier studies ^{16 30 31} in which fixation and visual tracking of objects was reported in hydranencephalic children, remnants of frontal, temporal, parietal, or occipital tissue were observed. One cannot completely exclude that these remnants of brain tissue at least partially mediated the residual visual functions. MRI used in the present study was able to detect brain tissue of 1 mm³. Therefore, we cannot assume that visual functions were mediated by undetected brain tissue. 

Results of the present study demonstrate that these tissue remnants are not sufficient for the development of a visual field beyond the foveal and parafoveal areas. Although it cannot be ruled out that the residual visual functions of patients SA and PL were partly mediated by remnants of brain tissue, it may well be that visual functions, such as fixation and visual tracking, are possible if only the brain stem is left. Patient AG demonstrated that auditory stimuli can still be detected and that the patient is still able to perform eye and head movements to auditory stimuli if both cerebral hemispheres are completely absent (Fig. 3) . 

Are the Results Caused by Disturbances in Eye Movement or by Nonvisual Cues?

In a retrospective case series of 50 children with cortical injury and 50 children with subcortical perinatal injury, Brodsky et al. ⁵⁵ reported eye movement disturbances that were associated with cortical and subcortical perinatal injury. The children reported in the present study had no eye movement disturbances, such as gaze deviation, exotropia, esotropia, or tonic downgaze. Children with such disturbances were excluded from the present study because reliable perimetric testing is not possible. 

In patients PL, SA, and AG, it was observed that the nonresponse of patients to targets outside the macular region could not be explained by a general paucity of eye and head movements toward stimuli. Paucity of eye movements might have been attributed to gaze palsy, oculomotor apraxia, or hemispatial neglect. Patients with hemispatial neglect demonstrate a paucity of spontaneous eye and head movements directed in the neglected half of space. ^{56 57} Patients PL, SA, and AG, who did not respond to visual stimuli outside the macular region, still turned their eyes and heads toward auditory stimuli presented in the left or right half of space. They also exerted spontaneous eye movements in the left and right half of space, excluding the presence of gaze palsy, oculomotor apraxia, or hemispatial neglect. When the children did not respond to targets with a luminance below 40 cd/m², where no scattering light was present, the target luminance was increased up to 25,000 cd/m² and 60,000 cd/m². These targets may warm up the illuminated retinal area, perhaps proving a cue that could elicit lid closure or eye movement directed toward the stimulus, or it might provoke pursuit eye movements. In all investigations, the light used was emitted from light-emitting diodes (LEDs), which produce negligible amounts of heat. Werth and Seelos ³³ presented stimuli of 25,000 cd/m² to cerebrally blind children who did not respond to 50-cd/m² or 1000-cd/m² targets. These children also did not respond to stimuli of 25,000 cd/m². Even a target luminance of 60,000 cd/m² did not elicit a response in blind children. 

A reliable perimetric measurement was possible in all children. The position of the children’s heads could be controlled well. Healthy controls and hydranencephalic children with preserved foveal vision were able to steadily fixate the fixation point. It was shown earlier ^{17 32 33 34} that the extent and even the luminance difference thresholds can be measured reliably in healthy and in brain-damaged children with the perimetric method used in the present study. However, the hydranencephalic children did not respond to the targets beyond 5° eccentricity. They did not even exert hypometric or hypermetric eye movement in the direction of the target. 

It appears that functional cerebral hemispheres are a necessary condition for the presence of a visual field beyond the macular region. Basic foveal or macular visual functions are, however, present in the absence of both cerebral hemispheres if the patient has a functional midbrain, including the superior colliculi. 

Conclusions

Based on the results of the present study, it can be concluded that at the age of 5 to 12 months, the visual field has reached almost adult extent. If both cerebral hemispheres are missing and if only remnants of frontal or temporal tissue are present, foveal and possibly even perifoveal vision may still be possible. The infant may close his or her eyes to diffuse illumination and may still be able to fixate and track a moving object with eye and head movements and may visually discriminate objects to a certain extent. Beyond the fovea and possibly beyond an extent of 5° from the fovea, no visual field is present. Therefore, one can gather that either the brain stem alone or the brain stem together with tissue remnants in the frontal or temporal lobe may be able to mediate foveal or even macular vision. However, a visual field extending beyond the fovea or possibly beyond the macula presupposes the presence of functional cerebral hemispheres. In humans there is no evidence that residual visual functions such as blindsight, which may still be present beyond 5° eccentricity after damage to the occipital lobe or after hemispherectomy, are mediated by the midbrain. 

 Submitted for publication September 22, 2006; revised January 16, 2007; accepted April 27, 2007.

 Disclosure: R. Werth, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Brain-Damaged Children Investigated in the Present Study

Patient/Sex	Age (mo)	Lesion	Visual Field	Pupil	OKN	Tracking	Blink	Sound
SA/F	25	Hydranencephaly	Tunnel vision	+	−	+	+	+
PL/M	29	Hydranencephaly	Tunnel vision	+	−	+	+	+
AG/M	87	Hydranencephaly	Blind	+	−	−	+	+
HE/F	9	Hydranencephaly	Blind	+	−	−	+	−
ER/F	28	Hydranencephaly	Blind	+	−	−	+	−

 Lesion, cerebral lesion at the time of testing; age, age of the patient at the time of first testing; pupil, pupillary reflex; OKN, optokinetic nystagmus; tracking, visual tracking of objects or faces; blink, eye closure on diffuse illumination; sound, response to acoustic stimuli; +, present; −, absent.

Figure 1.

 

(A–C) MRI of patient SA shows a CSF-filled space occupying the occipital lobes, the parietal lobes, and most of the frontal and temporal lobes of both cerebral hemispheres. The occipital lobes and the parietal lobes of both cerebral hemispheres are completely absent. The dorsal aspects of the temporal and frontal lobes are also missing. Only the cortex and white matter of the ventral temporal lobe and the anterior and ventral frontal lobe are preserved.

Figure 2.

 

(A–C) MRI of patient PL, who has hydranencephaly. The cerebral hemispheres have been replaced by a CSF-filled space. Frontal and parietal walls of the CSF-filled space consist of remnants of brain tissue. No gyri can be identified in the occipital, parietal, or temporal lobe. Optic radiation is completely absent. The optic nerves, the cerebellum, the pons, and the midbrain, including the superior and inferior colliculi, are preserved.

Figure 3.

 

(A–C) MRI of patient AG reveals hydranencephaly. The cortex and white matter of both cerebral hemispheres are completely absent. The optic nerves and the brain stem, including the midbrain, are well preserved.

Figure 4.

 

(A–C) MRI of patient HE, a 12-month-old girl with hydranencephaly. Both cerebral hemispheres were replaced by a CSF-filled space. Only remnants of cerebral tissue were observed in the anterior part of the frontal lobe. The optic nerves, the cerebellum, the pons, and the midbrain, including the superior and inferior colliculi, were preserved.

Figure 5.

 

(A–C) MRI of patient ER, a 28-month-old girl with hydranencephaly. MRI revealed that the cerebral hemispheres had been replaced by CSF-filled space containing septum-like remnants of glial tissue. The optic nerves, the pons, and the midbrain, including the superior and inferior colliculi, were preserved.

Table 2.

 

Means of the Visual Field Extent of 30 Healthy Controls, 5 to 12 Months of Age, on Eight Meridians of the Binocular Visual Field

Meridian (°)	Visual Field Extent (°)
225	x = 84.2; SD = 4.7
180	x = 86.0; SD = 3.6
135	x = 66.9; SD = 7.3
90	x = 55.3; SD = 5.1
45	x = 67.2; SD = 7.2
0	x = 85.2; SD = 3.8
315	x = 84.2; SD = 4.9
270	x = 63.8; SD = 7.2