Restitution of Visual Function in Patients with Cerebral Blindness

J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

Journal ofNeurology, Neurosurgery, andPsychiatry, 1979, 42, 312-322

Restitution of visual function in patients with cerebral blindness

J. ZIHL AND D. VON CRAMON From the Max-Planck-Institut fur Psychiatrie, Miinchen, Germany

SUMMARY Patients with postchiasmatic visual field defects were trained at the border of their visual field. Using a psychophysical method, light-difference thresholds were determined repeatedly in this visual field area. Improvement in contrast sensitivity and increase in size of the visual field could be obtained by this training procedure. The improvement was confined to the trained visual field area and showed interocular transfer indicating its central nature. Although only contrast sensitivity was trained, the observed improvement was not limited to this visual function. Visual acuity, critical flicker fusion, and colour perception also showed an improvement suggesting an association of these functions. The improvement was restricted to the training period-no spontaneous recovery was observed between or after the periods of training. It is suggested that a lesion in the central visual system does not always result in a complete and permanent loss of function. The critical level of function that normally has to be reached for a Protected by copyright. sufficient neuronal sensitivity may be obtained by systematic visual stimulation in the area between the intact and blind parts of the visual field. This increase in neuronal sensitivity leads to an improvement in visual performance.

Lesions within the central visual pathways do not may be abrupt or gradual (Korner and Teuber, necessarily result in absolute and permanent visual 1973). Visual fields with sharp scotoma boundaries field defects (Teuber, 1975). Perimetric "blind" or show only a slight recovery of function, whereas impaired visual field areas may fluctuate in sensi- for visual fields with rather gradual borders of the tivity (Bender and Teuber, 1946). Recovery of defect some degree of recovery can be observed visual function has been observed in patients with (Teuber, 1974; Zihl et al., 1977b). postchiasmatic lesions although return of vision is The question arises whether the level of rarely complete (Bergmann, 1957; Symonds and neuronal sensitivity can be modulated by system- Mackenzie, 1957). Reappearance takes place in a atic external stimulation in order to improve re-

definite temporal order: sensation of light and covery of visual function. An increase in sensitivity http://jnnp.bmj.com/ motion, form perception, colour perception in impaired or perimetrically "blind" parts of the (Poppelreuter, 1917; Riddoch, 1917). visual field may result in the restitution of vision Bender and Teuber (1946) suggested that fluctua- in these areas and thus in an enlargement of the tions in the extent and size of a cortical scotoma visual field. Evidence for such an increase in sen- are correlated with the variation of neuronal sitivity in impaired or scotomatous visual field sensitivity in the cortical area which is affected by parts has come from experiments with monkeys the lesion. In the corresponding visual field parts- (Cowey and Weiskrantz, 1963; Cowey, 1967).

that is, at the border of the visual field-fluctua- Visual stimulation in the area between intact and on September 29, 2021 by guest. tion and extinction of visual function can be blind parts of the visual field leads to a decrease of observed. For these parts of the visual field a the scotoma. Observations in human subjects are diurnal variation of light sensitivity was found in agreement with the effect of stimulation in (Zihl et al., 1977a). The transition between the monkeys. Stimulation of impaired visual field areas intact parts of the visual field and the scotoma with cortically reduced visual functions increases sensitivity and visual field size, although fatigue Address for correspondence and reprint requests: J. Zihl, Max- Planck-Institut fur Psychiatrie, Kraepelinstrasse 10, 8000 Munchen 40 may take place rapidly and result in an overall Federal Republic of Germany. reduction of visual function (P6ppel et al., 1978). Accepted 6 October 1978 The purpose of this paper is to present evidence 312 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

Restitution of visual function in patients with cerebral blindness 313 that diminished visual function can be improved had a diameter of 116 min visual angle; target by systematic stimulation of the impaired areas of luminance was 32 cd/M2. The patient pushed a the visual field. Since light sensation reappears button when he could detect flickering of the first in blind areas, this visual function was trained target (for ascending series) or when flicker was no in order to improve visual performance. To demon- longer detectable (for descending series). For test- strate the central nature of the increase in con- ing visual acuity, the patient's task was to trast sensitivity obtained by the training procedure, indicate whether he saw a circular or a diamond- interocular transfer was tested. We also investi- shaped target. Luminance of the targets was gated whether improvement takes place only in a 32 cd/M2 and they were presented for one second. defined area of the visual field (according to the Order of presentation was randomised. Errorless central representation of the retina). In addition, performance was chosen as the criterion of acuity. periods without training were included to show Critical flicker frequency and visual acuity were that the increase in sensitivity depends on the determined with the Tubinger perimeter. For test- training session and is not the result of spon- ing binocular colour perception, the Farnsworth- taneous recovery. Munsell 100-hue test was used (Farnsworth, 1943).

Methods TRAINING PROCEDURE Visual training was performed with the Tubinger VISUAL FIELD EXAMINATION perimeter. The method of limits was used to im- The visual fields were tested by dynamic and static prove sensitivity to light stimuli. Monocular and perimetry using the Tubinger perimeter (Sloan, binocular contrast thresholds were determined 1971; Aulhorn and Harms, 1972). The patients several times at a constant eccentricity in the fixated a red spot of light (10 or 30 min visual area between visual field and scotoma. The inter- angle in diameter) at a distance of 330 mm. Back- stimulus interval between presentations was about Protected by copyright. ground luminance was 3.2 cd/M2. Head position three seconds; time between measurements was in was held stable by a chin rest; eye position was the range of 10-30 seconds in order to prevent controlled through a telescope. fatigue (Singer et al., 1977). As a control of the For dynamic perimetry, the target (diameter patient's detection strategy, blanks were inter- 116 min visual angle; luminance 32 cd/M2) was spersed among targets. Usually, 15-20 contrast moved with a constant velocity of about 20/s from thresholds were determined in one session with a the periphery towards the centre of the visual field. break of about five minutes after five measure- The patient had to push a button whenever he ments. detected the target. The border of the visual field The patient was instructed to fixate a red spot was determined for 16 meridians, and the sequence of light in the centre of the homogeneously illumi- of measurements along the different meridians was nated sphere (background luminance 3.2 cd/M2), randomised. The visual fields were tested under and to direct his attention to the area where train- monocular and binocular conditions. ing was performed. Within training sessions, the Static perimetry was performed by measuring patients were reinforced for their fixation and binocular increment threshold using the method of their attention. They were not informed about an http://jnnp.bmj.com/ limits. Fixation point and background luminance increase or decrease in sensitivity until they experi- were the same as for dynamic perimetry. Target enced a progress in vision in their everyday life diameter was 69 min visual angle; presentation outside the laboratory. Training sessions were per- time 500 ms. Target luminance was chosen well formed daily at the same time for each patient; above or below threshold level and was then de- usually one session took an hour. creased or increased in steps of 0.1 log units. When the patient detected three targets consecutively, CONTROL OF TRAINING EFFECTS the luminance of the first of these was defined as Visual fields were mapped before and after training on September 29, 2021 by guest. the threshold for the ascending method. For the sessions using dynamic perimetry. In addition light- descending series of threshold measurements, the difference threshold was measured along the luminance of the last seen target was defined as meridian selected for training before and after the threshold-that is, when after this target was training; the difference was taken as change in seen, three consecutive targets were not detected sensitivity. For determination of the retinotopic by the patient. dependence of improvement, increment threshold In addition critical flicker frequency (CFF), was determined at various points surrounding the visual acuity, and colour perception were tested position of training. To test interocular transfer, for both eyes. For measurement of CFF the target training was performed monocularly. Transfer was J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

314 J. Zihl and D. von Cramon tested by comparing contrast threshold in the tion. Some patients, however, exhibited residual corresponding visual field area of the covered eye vision (Poppel et al., 1973). Sometimes they "felt" before and after training. In order to control objects appearing within their blind half-field. "spontaneous" recovery visual field size and visual Furthermore, all patients with the boundary of sensitivity between the end of the foregoing the scotoma near the fovea (along the horizontal session (or period) and the beginning of the next meridian) had problems in reading because words session (or period) were compared. In addition, appeared "cut off" on the right or left side. All visual acuity, critical flicker fusion or colour per- patients were highly motivated although they were ception were tested before and after training. informed that this visual training was only an attempt to improve their visual performance. Subjects Results Training was performed in 12 patients who had suffered visual field defects because of postchi- In the first training session, the repeated measure- asmatic lesions. The clinical details are summarised ment of light-difference threshold resulted in a in Table 1. The lesions had been verified by com- decrease in sensitivity. In most of the patients the puter tomography or cerebral angiography or both increase in sensitivity was obtained after three to (Zihl et al., 1978). All patients except cases 6 and five sessions (Fig. IA). In one patient (case 5), 10 suffered from cerebrovascular diseases. The improvement in contrast sensitivity was first ob- time between the first evidence of the visual field served during the eighth session. This decrease in defect and start of training varied from one month sensitivity in the first sessions may be the result (case 6) to nearly six years (case 8). No remark- of fatigue, even though longlasting breaks were able spontaneous recovery of visual function was introduced between measurements. After this first Protected by copyright. observed in the majority of cases. In case 6 light phase of training a rather remarkable improve- and movement perception returned in the per- ment was obtained within one session (Fig. 1B). iphery of the right visual half-field soon after the The increase in contrast sensitivity was always operation. In case 10 perception of light and greater for the periphery of the visual field (up movement and some kind of colour and form to 0.5 log units) than for the region near the perception recovered in the periphery of the visual fovea (0.2 log units; see Fig. 1B and Table 2). The field, whereas the central scotoma remained improvement in contrast sensitivity showed almost unchanged. total interocular transfer (Fig. 2). Increase in All patients showed good performance in fix- light sensitivity was not restricted to the trained ation. No obvious neurological defects (that is, for visual field position. As can be seen in Fig. 2, oculomotor, motor, and speech behaviour or atten- improvement extended over an area of 8° when tion and memory) were found when training was training was performed at 140 eccentricity along started. The main problem for these patients was the 450 meridian. The area within which improve- the failure to avoid obstacles at the side of their ment was also found increased with increasing scotoma and the restriction of their visual orienta- eccentricity. Figure 3 shows this effect. When http://jnnp.bmj.com/ Table 1 List of patients with clinical details Case Sex Age Diagnosis Visualfield defect Time since (yr) lesion* 1 M 42 Occlusion of the left post. cerebral artery with infarction Right hemianopia 0/11 2 M 64 Occlusion of the left calcarine artery with infarction Anopia of right upper quadrant 0/6 3 F 33 Occlusion of the right post. cerebral artery with infarction Left hemianopia 0/5 4 F 35 Partial occlusion of the left middle cerebral post. cerebral Right hemianopia 3/1 arteries with infarction

5 M 57 Bilateral occipital haemorrhage Right hemianopia. Anopia of 4/9 on September 29, 2021 by guest. left lower quadrant 6 F 61 Tumour in the left occipital lobe (operated) Paracentral right sided scotoma 0/1 7 M 19 Right temporobasal epidural haematoma (operated) Left hemianopia 0/2 8 M 52 Embolic occlusion of the right middle cerebral artery Left hemianopia 5/11 9 M 31 Papilloma of the choroidal plexus of the left posterior Right hemianopia 0/7 horn (operated) 10 F 33 Cerebral hypoxia Central scotoma 0/6 1 1 F 45 Intracerebral haemorrhage of the putamen-claustrum type Anopia of left lower quadrant 0/1 12 M 48 Bilateral partial occlusion of the middle cerebral artery Incomplete right hemianopia 0/2 and partial occlusion of the left post. cerebral artery with infarction Time (in years/months) from evidence of visual field defect to beginning of training. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

Restitution of visual function in patients with cerebral blindness 315 Table 2 Maximal increa;e in contrast sensitivity for A various eccentricities (mean of five patients) C4 E 16 Eccentricity Maximal increase Range SESSION 3 (degrees) (log units) (log units) - u 1 0.10 2 0.14 0.10-0.20 25 3 0.18 0.10-0.30 4 0.24 0.20-0.30 5 0.26 0.20-0.30 la-J 8 0.32 0.20-0.40 10 0.32 0.20-0.40 1- 40 - 1 5 0.40 0.30-0.50 1 8 0.40 0.30-0.50 w 20 0.41 0.30-0.50 24 0.44 0.30-0.60 SESSION 1 I- 64 - strated in Fig. 4. This patient (case 10) showed a central scotoma (Fig. 4A, hatched area). Training 1 5 10 15 was performed in the visual field of the left eye, TRIAL NUMBER while the right eye was covered by an-occluder. With one eye closed the patient was able to keep fixation stable using a black circle as a "fixation" point for the blind spot. During the first period of B training, the threshold was measured along the in right half-field. (N horizontal meridian the visual Protected by copyright. E 10 As can be seen in Fig. 4B, the training effect was 12°12ECCENTRICITYECCEN T RI CI TY A r not limited to the trained area but was obtained IN A rwithin the whole "blind" central region (Fig. 4B, 29.10.76). After this overall effect of visual stimu- -D 16 - lation any further increase in sensitivity was limited to the area actually trained. After approxi- -J mately two months of training, no more visual field defect could be detected and normal contrast I- 25 w sensitivity was obtained within the previous central 0 scotoma (Fig. 4A). Increase in contrast sensi- 30 ECCENTRICITY tivity showed complete transfer to the occluded I-- 40 right eye. Static perimetry demonstrated identical sensitivity profiles for the visual field of both eyes r , (Fig. 4B). 1 5 10 15 The increase in visual field size obtained by http://jnnp.bmj.com/ T RI A L NUM B ER training of contrast sensitivity varied from patient to patient (see Table 3). In some of the patients Fig. Course of I threshold during (cases 1, 5, 6, 7, and 10) visual function improved night-diference after one training sessiotInconsisting of 15 measurements in the area between visual field and scotoma (case 9). Increase iin contrast sensitivity was obtained a In first time in the thi'rd session (A, 30 eccentricity/00 few trials (12-40 thresholds measurements). merid-an). The impirovement is greater for the other patients, up to 260 trials were necessary to periphery of the visual field (0.4 log units at 12°) than increase visual field size by one degree (cases 4, 1 and In the latter the slope of for the foveal regicrn (B, 0.2 log units at 3°). 8, 1, 12). patients, on September 29, 2021 by guest. the light-difference threshold between visual field training was performed at 8° eccentricity along and scotoma was steep, whereas in the other cases the horizontal mi eridian, the area of improvement the slope was rather gradual. extended over approximately 10 degrees (Fig. 3A), Increase in visual field size and contrast sen- whereas for a training position near the fovea sitivity for four patients is shown in Figs. 5 and 6. (10) an increase in sensitivity was obtained for In the two patients shown in Fig. 5 training had this position onlI y (Fig. 3B). Interocular transfer a dramatic effect on restitution of visual field. The and retinotopic dlependence of the training effect visual field of the first patient (Fig. 5A, case 5) was in the course olf a training period are demon- extremely reduced and behaviourally he was blind. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

316 J. Zihl and D. von Cramon

LEFT EYE RIGHT EYE 1.0- 1.0- Fig. 2 Interocular transfer of the training effect. Training was E4 3.2 - E4 a 2 - in E performed the nasal visual la- field of the left eye (450 uJ 10 10- meridian; 14° eccentricity, z0 - w - indicated by arrow), while the TER z - ' AFTER BEFORE 4 TRAINING BEFORE right eye was occluded (case 1). Z 4 \,;TRAINING 32- TRAJNNG x 3 2- TRAINING \ The increase in contrast uJ II sensitivity was also observed \Af in o after training the 100- I' o 100- t\ " corresponding region of the i"% temporal visual field of the 320 - 320 right eye (arrow) and shows the I. central nature of the 0 5 lb 15' 201. 0 5 10 15. improvement by training. NVF: 45 MERIDIAN TVF: L5 M'ERDIAN

A B 0.3 - - 0.3 - *

- 1.0- AFTER 1.0 - C-E TRAINING -- 21 C) / BEFORE TRAINING u 3.2- / - 3.2 _ Fig. 3 Light-difference Protected by copyright. f RE threshold before and after one w C session x training (case 10). The z 10 - AFTER TRAINING -10 z increase in visual sensitivity z I and the area within which z , 32- 32 m improvement was also found, were greater for 8° eccentricity w .n (0° meridian, A) than for 1° (' 100- 100 a 3 eccentricity (B). 320- 320

0 5 10 15 20 0 5 10 15 20 ECCENTRICITY (DEGREES) ECCENTRICITY (DEGREES)

He used acoustic and tactile cues to find his way the horizontal meridian (where training was per- around. Training was performed along the whole formed) was only 20. This rather slight enlarge- visual field border. At the end of the training ment, however, was sufficient to allow normal http://jnnp.bmj.com/ period the left upper quadrant was restituted. The reading. In the last patient shown in Fig. 6B patient is now able to walk using visual cues. He (case 9), training produced an increase of 15°, can perceive colour and form, and is able to read although the boundary of the scotoma before a newspaper. The second patient (Fig. 5B, case 6) training was at 40 eccentricity (horizontal mer- suffered from a right paracentral scotoma. As idian) for both patients. The slope, however, was with the first patient, training was performed rather gradual for case 9 (cf. static perimetry for along the whole visual field boundary. After 30 both cases at the right side of Fig. 6). Training training sessions, the right visual half-field showed was performed along the horizontal meridian and on September 29, 2021 by guest. normal contrast sensitivity except for a relative restitution of the visual field was obtained mainly scotoma at 80 eccentricity along the horizontal along this meridian. The enlargement of the visual meridian. Subjectively this patient exhibited nor- field in the upper and lower quadrant (about 20) mal visual perception. In another patient (Fig. 6A, is due to the overall effect in the first sessions as case 8), training did not produce such a marked stated above. increase in visual field size. Although as many From the data reported it is not yet clear trials were conducted as for the two patients re- whether the improvement in contrast sensitivity is ported above, the increase in visual field size along spontaneous or the result of the systematic exter- J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

Restitution of visual function in patients with cerebral blindness 317

A 0.1 E 0.3 w 1.0 - 13.12.76 < 3.2 - \ 10C a - 32 Li 100-o 310.76 = 320 30 20 10 0 10 20 30 Fig. 4 Time course of ECCENTRICITY ( DEG) restitution of contra.st sensitivity in a case of central scotoma B (ca.se 10). The blind central part of the binocular visual E field (A, hiatchled area) show.s normal contrast sensitivity at I the end of thle training period (cf. contra.st thlresholds shown top righlt). Training was a:. J performed only in tile visual field of the left eye; the righlt eye was always occluded during tile training sessions. The 30t 0 30030 0 30 3t 0 30' scotomatous region (black area 29.10.76 12.11.76 29.11.76

in B) decreases in both visual Protected by copyright. fields in a similar way. A O01--~

O32

D 10- ,, 32- a: 100 320 l 30' 0. 30' 30' 0. 30' 30' 0O 30' 30' 0t 30O. 8.10.76 29.10.76 12.10.76 29.11.76

Table 3 Number of trials, increase in vi.sual field 6 and case 9). Improvement in contrast sensitivity size, and trials per degree restituted visual field was limited to the training session (Fig. 7A and B) and to the training period (Fig. 7C and D). Inter- Case Trial Increase in visutal Trials/zdegree numbnber field size vals without training did not show any influence http://jnnp.bmj.com/ (degrees) on contrast sensitivity in the area between intact 1 380 10-24 14 27 and blind parts of the visual field. No obvious 2 195 7-10 3 65 increase or decrease in visual field size was ob- 3 220 2-7 5 44 4 260 4-5 260* served, indicating that restitution of visual func- 5 315 25-50 25 12 tion depends strongly on systematic visual stimu- 6 535 1-28 27 20 along the boundary of the visual field. 7 180 5-12 7 26 lation 8 530 2-4 2 265* Furthermore, improvement obtained by visual

9 560 3-17 14 40 training did not disappear during periods without on September 29, 2021 by guest. 10 320 10-10 20 16 11 360 1-5 4 90* training or after training. Control measurements 12 210 8-9 1 210* six months after the end of training support the view of lasting restitution of visual function. *Patients with an abrupt transition betwseen intact and blind parts of their visual field. In all patients described, restitution of contrast sensitivity was followed by return of resolution of nal stimulation. To decide this question, contrast form, critical flicker fusion, and colour perception. thresholds between daily sessions and between The increase in various visual functions was not periods without visual training were compared. limited to the restored visual field area-these Results are shown in Fig. 7 for two patients (case visual functions also showed improvement in the J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

318 J. Zihl and D. von Cramon

A A E ° -

, 0.3 - I Li. 1.0 - z 3.2 - Z I10 - c 11w :- 32 - I 1, 100- '- 320 -

60 40 20 0 20 1 5 C 5 10 ECCENTRICITY DEG E7'.ENTRICI TY EG

B B

E 017 -E 0.1 - 0.3 - a- sV 28 4 78 _H L. 1,0 - z3.21~0 --~ LI X 3 2- 2 3'.78 032 -

L 100 - >\ 4 8,77 cr 32 - -j 320 -

0 10 2 30 40 1 ,22 ECCEN2 T1 CI TY E , --+ >E 1i'RI' '.Tf L't

Fig. 5 Binocular visual field (left) and profile of Fig. 6 Binzocular visual field (left) and contrast Protected by copyright. contrast sensitivity (right) before and after the sensitivity profile (right, horizontal meridians) in twvo period in a patient with a bilateral occipital training patients witli cerebral infarction (A, case 8; B, case 9) lesion (A, case 5) and in a patient with a left before anid after tile training period. occipital lesion (B, case 6). Hatched area inidicates enlargemnent of the visual field obtained by training. Contrast sensitivity is shown for the horizontal axis patients can be increased by systematic external of the visual field. The broken line in B indicates stimulation. Lowering light-difference threshold contrast sensitivity for the 0° meridian left visual in the area of diminished sensitivity between the half-field. intact and the blind parts of the visual field leads to an increase in contrast sensitivity and to an unimpaired visual field regions, for example, in the enlargement of the visual field in the trained foveal region (Table 4). Only in the patient with region. The increase in visual field size had a the central scotoma (case 10) did spatial resolution rather dramatic effect on the visual behaviour in remain unchanged in spite of the intact visual field most of the patients. Since other visual functions for contrast sensitivity after the training. --for example, colour and form perception-also

Colour perception was preserved in most of the returned in the restituted visual field area, the http://jnnp.bmj.com/ patients. In some cases, however, colour dis- visual capacities were sufficient for normal vision crimination was also affected by the cerebral (visual orientation; form discrimination, reading). lesion. These patients (cases 5, 6, 7, and 8) showed This improvement in vision is strongly dependent a low discrimination profile without any specific on the training period and is thus not the result deficit. After the training period the discrimination of spontaneous recovery. The interocular transfer profile of these patients was normal according to of the increase in contrast sensitivity indicates that the profile obtained from subjects with normal the training effect occurs at a central site in the

vision (Farnsworth, 1943). For cases 5 and 10, the visual system. on September 29, 2021 by guest. mean of error scores was markedly reduced after The observation that loss or impairment of training (from nine errors to four errors). visual function can be diminished with practice is in agreement with observations in primates. After Discussion a partial lesion of the striate cortex in monkeys recovery of the corresponding scotoma was ob- Our purpose in this study was to explore the tained by systematic external stimulation in the possibility of restitution of vision in patients with area between the intact and blind parts of the lesions of the geniculostriatal visual system. The visual field. This improvement in sensitivity was results indicate that visual field size in these not the result of spontaneous recovery since post- J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

Restitution of visual function in patients with cerebral blindness 319 2.5 - A 8- B 4 4.0 - 12 - 6.4 - 20 - Fig. 7 Light-difference threshold and a - 32 - before after 10- r .~~~single training session (A, B) X 1 6- 50 - _ l fl and course of contrast 25 - 80 - jsensitivity during periods of w and intervals without < 40- 128 - LI LI LI training train ng (C, D). Increase in < 64- 200 - contrast sensitivity is limited 100 - 320 - to the training session; the 5 threshold after each session 1 2 3 4 5 2 3 4 S (bladk bars) is the same as DAY before the next training DAY session on thle next day (white bars). In A, threshold was 0.3 - 0.3- E C D mneasured along the O° meridian - 1.0- at 20 eccentricity (case 6), in 1.0- B at 100 eccentricity along the 3.2 - 3.2- same meridian. In intervals 10- 10- without training (C, D dark 32- 32- Irnes at the bottom) no increase w 100- in contrast sensitivity was 0 100 - observed (C, case 6; D, case 9). _- 320- - 320- Meridian and eccentricity same

as in A and B. Protected by copyright. t 20 40 60 80 100 t 20 40 60 75 DAYS DAY S

Table 4 Spatial and temporal resolution before and The increase in visual field size is limited to the after training.__Values for the fovea trained region and is not an overall effect of stimulation. The visual field region within which im- Case Spatial resolution Temporal resolution provement was also found becomes larger towards Before After Increase* Before After Increaset the periphery of the visual field. The diameter of (min (mi (%) (Hz) (Hz) (%) arc) arc) this region, however, remains approximately the same if corrected by the cortical magnification it 12 7 37 31 34 14.7 factor. The correspondence of the enlargement of 2 12 12 - 30 31 3.2 3$ 7 7 - 28 32 12.5 the visual field with the known central repre- 4+ 12 12 - 29 31 6.5 sentation of the visual field in primates (Daniel 5+ 18 7 60 24 30 20.0 6 12 7 37 31 36 13.9 and Whitteridge, 1961; Rolls and Cowey, 1970)

7+ 112 7 37 26 35 25.7 and the retinotopic dependence of the training http://jnnp.bmj.com/ 8+ 7 7 - 30 34 11.7 9$+ 18 7 60 29 33 12.1 effect indicate that the increase in visual field 10 - 69§ - 35§ size depends on the primary visual cortex. For the 11 18 7 60 26 30 10.3 prestriate areas of the monkey, a retinotopically 12 12 12 - 26 30 13.3 organised representation of the visual field was no *Increase in % of relative acuity. longer found (Zeki, 1971). tCFF after training taken as 100% The neuronal mechanism underlying the ob- +Corrected for hypermetropia or myopia. §After visual field was intact for contrast sensitivity. tained restitution of visual functions is not yet

clear. Lashley (1938) has pointed out that restitu- on September 29, 2021 by guest. tion of function in the cerebral cortex depends operative testing was performed two and a half strongly upon the preservation of some part of the years after removal of the cortical tissue. If a system affected by the lesion. A critical amount visual field defect is produced by a retinal lesion, of tissue must remain intact for the recovery of no recovery either spontaneously or with practice the diminished or depressed function. The tran- was observed by Cowey (1967). This result cor- sition zone between the intact visual field and the relates with the observed increase in visual field scotoma may be correlated with parts of intact size in our patients usino a similar psychophysical tissue at the border of the lesion. It may be sug- method of training. gested, therefore, that the transition zone (ex- J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

320 J. Zihl and D. von Cramon pressed in terms of the course of light-difference is known to modulate light-difference threshold in threshold) represents the functional level of still the visual field (Zihl et al., 1979). The improve- intact cortical neurones. Thus, a critical degree of ment in sensitivity depends strongly on attentional functional level is necessary to mediate sensation processes. The relationship between light-difference in a partially damaged visual system. Diurnal threshold and attention is shown in Fig. 8. The variation in visual field and spontaneous recovery diminution of contrast sensitivity is marked when have been observed in patients with a rather the patient does not know where the target will gradual decrease of contrast sensitivity in this appear and thus can not direct his attention to area (Zihl et al., 1977a,b). Furthermore, in patients that area of the visual field. The neurophysiologi- with such a gradual transition a fairly good cal mechanisms underlying the increase or decrease restitution was obtained by training, whereas in in sensitivity in relation to attention are not patients with an abrupt transition between the known. A modulation dependent on reticular visual field and the scotoma, restitution was activation might be envisaged, since reticular negligible. stimulation leads to facilitatory effects in genicu- Restitution or "reorganisation" of function late and striatal neurones (Singer, 1973; Singer may be understood in terms of the interaction of et al., 1976). It has been suggested that this effect the remaining parts of the affected system may be caused by a shift of attention modulating (Lashley, 1938). This means that restricted lesions the excitability of the neuronal substrate (Bartlett do not necessarily abolish the functional organ- and Doty, 1974). Recently it has been shown that isation of a neuronal network. Since even other collicular pathways mediate the modulation of visual functions reappeared in the restituted visual sensitivity in the visual field, probably by shifting field area it may be assumed that the mode of per- visual attention (Singer et al., 1977). This is in ceptual processing was not affected by the lesion. line with electrophysiological observations in the

In addition, the recovery of non-trained visual superior colliculus of monkeys (Wurtz and Gold- Protected by copyright. functions suggests an associative representation berg, 1972). Furthermore, other cerebral struc- of these functions (P6ppel et al., 1978). tures-for example, the parietal lobes-may be It is also an open question which factors modu- involved in the mechanism for directed attention late neuronal sensitivity and thus mediate the (Lynch et al., 1977; Mountcastle, 1978). The rela- effect of training observed in our patients. One tionship between the improvement of depressed important factor could be selective attention which visual function and attention indicates that the

0.3 am Fig. 8 Dependence of light- IN difference threshold on E attention. When the patient - a is instructed to direct his 1.0 co attention to the area along the a a>__ o horizontal axis, where the o -WI target will appear, contrast w c threshold is lower (broken U 3.2 * N lines) than when the target is http://jnnp.bmj.com/ z presented randomly (solid ,, 1 5°- / lines).\^ OThe ascending method z / \ of limits was used in both I 10 15° conditions to determine D light-difference threshold. The -J \ effect is mainly limited to the area between intact and blind w 32- * parts of the visual field 0 (RVF=right visual field). In on September 29, 2021 by guest. I- this area, contrast threshold differed by about one log 100 - unit for the two conditions. I I I I theTheregiondifferencenearisthegreaterscotoma.for 50 30 1 ° 10 30 50 In the intact left visual field (LVF) the difference in LV F RVF contrast sensitivity is about 0.1-0.2 log units. ( aINTACT*-".-T'SIDECll-C \) J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

Restitution of visual function in patients with cerebral blindness 321 functional level of the visual cortex does not Farnsworth, D. (1943). The Famnsworth-Munsell 100- depend only on the specific visual activity of the hue and dichotomous tests for colour vision. Journal neurones but also on other factors such as atten- of the Optical Society of America, 33, 568-578. motivation 1938). The Koerner, F., and Teuber, H.-L. (1973). Visual field tion and (Lashley, 1931, defects after missile injuries to the geniculo-striate reduction of these non-visual functions as a pathway in man. Experimental Brain Research, 18, secondary effect of the lesion leads probably to a 88-113. functional reduction of the preserved tissue. Thus Lashley, K. S. (1931). Mass action in cerebral func- loss of visual function may also be caused by re- tion. Science, 73, 245-254. duction of non-visual factors. An improvement Lashley, K. S. (1938). Factors limiting recovery after of the functional level of the preserved tissue central nervous lesions. Journal of Nervous and demands an increase of visual attention. Mental Disease, 88, 733-755. In conclusion, there is evidence now that areas Lynch, J. C., Mountcastle, V. B., Talbot, W. H., and of blindness can be restored in patients with Yin, T. C. T. (1977). Parietal lobe mechanisms for directed visual attention. Journal of Neurophysi- lesions of the central visual pathways. The in- ology, 4, 362-389. crease in visual field size as the result of a specific Mountcastle, V. B. (1978). Brain mechanisms for training procedure leads to a remarkable improve- directed visual attention. Journal of the Royal ment in visual functions such as spatial orientation Society of Medicine, 71, 14-28. or reading. The restitution of visual sensory Poppel, E., Held, R., and Frost, D. (1973). Residual functions may, therefore, be considered as an visual function after brain wounds involving the important part of the rehabilitation of patients central visual pathways in man. Nature, 243, 295- with damage to the central nervous system. It 296. remains to be clarified, however, which factors Poppel, E., Brinkmann, R., von Cramon, D., and Association and of can be used for predicting the amount of the Singer, W. (1978). dissociation visual functions in a case of bilateral occipital lobe Protected by copyright. restitution of visual sensory functions and the infarction. A rchiv fur Psychiatrie und Nerven- relevance of this restitution for visual behaviour krankheiten, 225, 1-21. in these patients. Poppelreuter, W. (1917). Die psychischen Schadi- gungen durch Kopfschuss im Kriege 1914/16. Bd. This work was supported by Deutsche Forschungs- I: Die Storungen der niederen und hoheren gemeinschaft. Sehleistungen durch Verletzungen des Okzipitalhirns, pp. 25-72. Leopold Voss: Leipzig. References Riddoch, G. (1917). Dissociation of visual perception due to occipital injuries, with special reference to (1972). Visual perimetry. appreciation of movement. Brain, 40, 15-57. Aulhorn, E., and Harms, H. Rolls, E. T., and Cowey, A. (1970). Topography of In Handbook of Sensory Physiology, VII/4: Visual the retina and striate cortex and its relationship to Psychophysics, pp. 102-145. Edited by D. Jameson visual acuity in rhesus monkeys and squirrel and L. M. Hurvich. Springer: Berlin, Heidelberg, monkeys. Experimental Brain Research, 10, 298- New York. 310. Bartlett, J. R., and Doty, R. W. (1974). Influence of Singer, W. (1973). The effect of mesencephalic mesencephalic stimulation on unit activity in striate reticular stimulation on intracellular potentials of http://jnnp.bmj.com/ cortex of squirrel monkeys. Journal of Neuro- cat lateral geniculate neurons. Brain Research, 61, physiology, 37, 642-652. 35-54. Bender, M. B., and Teuber, H. L. (1946). Phenomena Singer, W., Tretter, F., and Cynader, M. (1976). The of fluctuation, extinction and completion in visual effect of reticular stimulation on spontaneous and perception. Archives of Neurology and Psychiatry evoked activity in the cat visual cortex. Brain Re- (Chicago), 55, 627-658. search, 102, 71-90. Bergmann, P. S. (1957). Cerebral blindness. Archives Singer, W., Zihl, J., and Poppel, E. (1977). Sub- of Neurology and Psychiatry (Chicago), 78, 568- cortical control of visual thresholds in humans: evi- 584. dence for modality-specific and retinotopically on September 29, 2021 by guest. Cowvey, A. (1967). Perimetric study of visual field organized mechanisms of selective attention. Ex- defects in monkeys after cortical and retinal abla- perimental Brain Research, 29, 173-190. tions. Quarterly Journal of Experimental Psy- Sloan, L. L. (1971). The Tubinger perimeter of chology, 19, 232-245. Aulhorn and Harms. Archives of Ophthalmology, Cowey, A., and Weiskrantz, L. (1963). A perimetric 86, 612-622. study of visual field defects in monkeys. Quarterly Symonds, C., and Mackenzie, J. (1957). Bilateral loss Journal of Experimental Psychology, 15, 91-115. of vision from cerebral infarction. Brain, 80, 415- Daniel, P. M., and Whitteridge, D. (1961). The repre- 455. sentation of the visual field in the cerebral cortex Teuber, H. L. (1974). Recovery of function after of monkeys. Journal of Physiology, 159, 203-221. lesions of the central nervous system: history and J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.42.4.312 on 1 April 1979. Downloaded from

322 J. Zihl and D. von Cramon prospects. In Functional Recovery after Lesions of postretinal lesions. Experimental Brain Research, the Nervous System. Neurosciences Research Pro- 27, 245-249. gram Bulletin, 12/2, pp. 197-209. Edited by E. Zihl, J., von Cramon, D., Brinkmann, R., and Back- Eidelberg and D. G. Stein. MIT Press: Cambridge. mund, H. (1977b). The course and prognosis of Teuber, H. L. (1975). Recovery of function after brain scotomas in patients with cerebrovascular disorders. injury in man. In Outcome of Severe Damage to Nervenarzt, 48, 219-224. the Central Nervous System, Ciba Foundation Zihl, J., von Cramon, D., and Poppel, E. (1978). Re- Symposium 34 (new series), pp. 159-186. Elsevier: habilitation of sensory functions in patients with Amsterdam, Oxford, New York. post-chiasmatic visual disorders. Nervenarzt, 49, Wurtz, R. H., and Goldberg, M. E. (1972). The pri- 101-111. mate superior colliculus and the shift of visual Zihl, J., von Cramon, D., Poppel, E., and Singer, W. attention. Investigative Ophthalmology, 11, 441-450. (1979). Interhemispheric modulation of light Zeki, S. M. (1971). Cortical projections from two difference threshold in the periphery of the visual prestriate areas in the monkey. Brain Research, 34, field. In Structure and Function of the Cerebral 19-35. Commissures. Edited by J. S. Russell. Macmillan: Zihl, J., Poppel, E., and von Cramon, D. (1977a). London. In press. Diurnal variation in visual field size in patients with Protected by copyright. http://jnnp.bmj.com/ on September 29, 2021 by guest.