B R A I N R E S E A R C H X X ( 2 0 0 6 ) X X X – X X X

BRES-35475; No. of pages: 7; 4C:

ava i l ab l e a t www.sc i enced i rec t . com

www.e l sev i e r. com/ l oca te /b ra in res

ARTICLE IN PRESS

Research Report

Improving residual vision by attentional cueing in patientswith brain lesions

Dorothe A. Poggela,b,c,⁎, Erich Kastena, Eva M. Müller-Oehringa,d,Ulrike Bunzenthala, Bernhard A. Sabela

aInstitute of Medical Psychology, Otto-von-Guericke University Magdeburg, Leipziger Str. 44, 39120 Magdeburg, GermanybCenter for Innovative Visual Rehabilitation, Boston VA Medical Center, Mail Stop 151E, 150 South Huntington Ave., Boston, MA 02130, USAcCenter for Biomedical Imaging, Boston University School of Medicine, Department of Anatomy and Neurobiology, 650 Albany Street, Boston,MA 02118, USAdDepartment of Psychiatry and Behavioral Sciences, Stanford University School of Medicine; 401 Quarry Road; Stanford, CA 94305-5723, USA

A R T I C L E I N F O

⁎ Corresponding author. Center for Innovatition, Boston VA Medical Center-Mail StoHuntington Avenue, Boston, MA 02130, USA.

E-mail address: dapoggel@bu.edu (D.A. Po

0006-8993/$ – see front matter © 2006 Publisdoi:10.1016/j.brainres.2006.04.011

A B S T R A C T

Article history:Accepted 3 April 2006

Visual attention is crucial for almost all processes of visual perception, particularly whenperception is difficult. We were interested in the effects of cueing spatial attention inpatients with cerebral lesions who face difficulties in visual perception in areas of residualvision at the border of visual field defects. In 23 patients with visual field loss due to post-geniculate brain lesions, stimulus detection performance and reaction times were mappedwith high-resolution computer-based perimetry. A cueing procedure using Gestaltcompletion to attract attention to areas of residual vision was implemented in this testand performance compared in attended and unattended conditions. Stimulus detection andreaction times in areas of residual vision improved significantly under attended conditions.The extent of this effect depended on the size of areas of residual visionwithin the cued field.Unexpectedly, facilitation was also observed, though to a lesser extent, in invalid cueingconditions, suggesting an unspecific increase of alertness in unattended areas. Our findingsshow that top-down influences are relevant for visual field testing. Visuo-spatial attentionmay change patterns of neural activation and induce short-term plasticity not only in theintact visual system but also in the presence of visual field loss after brain lesions.Attentional cueing induces a co-activation of the lesioned visual system and (intact)attentional networks in the brain inducing immediate facilitation of visual perception. Thiseffect may be relevant for designing new strategies to permanently improve vision duringneuropsychological rehabilitation.

© 2006 Published by Elsevier B.V.

Keywords:Visual field defectAttention focusConscious perceptionDetection thresholdFunctional recoveryVision restoration

Abbreviations:ARV, area of residual visionHRP, high-resolution perimetry

ve Visual Rehabilita-p 151E, 150 SouthFax: +1 857 364 6645.ggel).

hed by Elsevier B.V.

1. Introduction

Patients with post-geniculate lesions of the visual systemtypically suffer from blindness in circumscribed homonymousregions of the visual field of both eyes. In perimetric tests, theyoften show areas of residual vision (ARVs) at the visual fieldborder (Zihl and von Cramon, 1979; Kasten et al., 1999).Thresholds of light perception are elevated in ARVs, indicated

2 B R A I N R E S E A R C H X X ( 2 0 0 6 ) X X X – X X X

ARTICLE IN PRESS

by a reduced probability of detecting supra-threshold stimuli(Fig. 1a), prolonged reaction times (Fig. 1b), and a lowersubjective quality of perception compared to stimuli in intactareas (Kasten et al., 1999; Poggel, 2002).

In normally sighted subjects, visuo-spatial cueing ofattention enhances performance at the attended location inthe visual field by reducing perceptual thresholds. Focusingattention at a specific position improves detection anddecreases reaction times in that region (Posner, 1980) andenhances neuronal activity in visual brain areas (Moran andDesimone, 1985; Heinze and Mangun, 1995; Gandhi et al.,1999). The facilitating effects are most prominent underconditions of difficult or ambiguous perception (Poggel, 2002).

Patients with visual field defects face such difficultperceptual situations in ARVs as a result of the brain lesion.We therefore wanted to determine whether focusing visuo-spatial attention at the visual field border would improve theirvisual performance in partially defective regions. Since mostpatients do not have an exact topographical representation oftheir ARVs, a simple verbal instruction cannot reliably inducea precise shift of the attention focus to those locations.Therefore, we adapted a conventional design of attentionalcueing (Posner, 1980) to help patients attend to areas at thevisual field border. Visuo-spatial attention has been success-fully manipulated in normally sighted subjects using directand symbolic cueing (Posner, 1980, 1995). A direct cue appears

Fig. 1 – Stimulus detection probability and reaction times in arperformance in five successive superimposed HRP tests in a pafield, there is an abrupt transition between intact and blind areARV in the lower right part of the visual field. White = intact (stimfive trials); grey = ARV (20%–80% detection). 1b (below): Reactionresponse times were almost equally distributed in the left (intareaction times were markedly slower than in the correspondintimes.

at the location where the target is to be presented after somedelay. This type of cue induces a topographically precise shiftof attention to the target location. However a partially blindpatient may not perceive the cue when it is presented inregions of the visual field which have a reduced stimulusdetection probability and might thus be unable to shiftattention in response to the cue. Symbolic cues, e.g. anarrow pointing towards the left or right, are commonlypresented in the center of the visual field and would thus beeasily perceptible for hemianopic patients. However, thosecues are designed to induce an attention shift to a wholequadrant or hemifield so that the spatial accuracy is notsufficient for moving the focus of attention specifically to theARVs. With these problems in mind, we designed a cue withboth direct and symbolic characteristics that would allow thepatient to shift attention reliably and precisely. We used alarge square frame positioned over the visual field borderwhich attracted the patient’s attention precisely towards theARVs and blind areas bymeans of Gestalt-completion into theblind field (Fig. 2).

2. Results

In 23 patients with visual field defects after post-geniculatebrain lesions, repeated performance in computer-based visual

eas of residual vision (ARVs). 1a (above): Stimulus detectiontient with incomplete right hemianopia. In the upper visualas. In contrast, the visual field border is diffuse with a largeulus detected in all five trials); black = blind (no detection intimes of the same patient, measured in one HRP test. The

ct) field. However, in the ARV in the lower right visual field,g region in the intact field. Shades of gray reflect reaction

Fig. 2 – Single-case results of patients with narrow vs. broad ARVs. Narrow ARV (left column): Five superimposed HRPtests (a) from patient #10 show a narrow transition zone in both the upper and lower visual field. The unattendedcondition (result of a single HRP test, c) and the condition with attentional cueing (result of the attention field testperformed in the same test session, e) show hardly any difference in stimulus detection in the attention field (dark greyframe in upper visual field) or probe field (light grey frame in lower visual field). Broad ARV (right column): Patient #2 hada broad ARV and fuzzy visual field border in the upper visual field (b). Compared to the visual field test without cueing(d), he showed a marked increase of stimulus detection in the attention field test (f), especially in the attended area of theattention field.

3B R A I N R E S E A R C H X X ( 2 0 0 6 ) X X X – X X X

ARTICLE IN PRESS

field tests (high-resolution perimetry, HRP) was comparedwith vs. without employing a visuo-spatial cue. Accuracy andspeed of stimulus detection in areas of residual vision (ARVs)were compared in attended vs. unattended conditions.Stimulus detection was significantly elevated in the attentionfield after presentation of a valid cue (Table 1) as compared toperformance in the same area in uncued (neutral) trials.Moreover, reaction times were markedly decreased afterpresentation of a valid cue as compared to neutral trials(Table 1). In the probe field, i.e. in the invalidly cued condition,costs had been expected (i.e. lower detection rates and longerreaction times) because attention was shifted away from thetarget location in these trials. However, we found a benefiteven when the cue indicated the stimulus position incorrectly(improvement of detection rate by 3% and decreased reactiontimes by 29 ms, see Table 1).

The decrease of reaction times was more pronounced invalid trials than in invalidly cued trials (see Table 1; WilcoxonTest, Z = 3.186, p = 0.001) indicating that the effect was specific

and that patients had indeed shifted their attention efficientlyto the cued location. However, for detection performance, theimprovement in the attention field was numerically greaterthan in the probe field, but the difference did not reachsignificance. Closer analysis revealed that the gain of stimulusdetection by spatial attention depended on the size of theARVs in the attention field, i.e. the larger the ARVs, the greaterwas the cueing effect on stimulus detection (Spearman’s Rho:ρ = 0.744, p < 0.001). In patients with larger ARVs in theattention field the difference was significant, i.e. the cueingeffect in the attention field was significantly greater than theunspecific effect in the probe field (n = 5; increase of detectedstimuli in attention field: 5.15 ± 3.12 vs. probe field: 2.05 ± 1.53;Mann Whitney Test: Z = 2.453, p = 0.014, two-tailed) but not inpatients with the same size of ARVs in both areas or an evenlarger transition zone in the probe field (n = 18; increase ofdetected stimuli in attention field: 2.65 ± 3.25 vs. probe field:3.32 ± 3.55). Moreover, the difference between the cueing-effect in the attention field (valid trials) vs. probe field (invalid

Table 1 – Group results: Cueing effect in the attention field and probe field

Mean (±S.E.M.)without cue

Mean (±S.E.M.)with cue

Improvement(difference cued-uncued)

Z(Wilcoxon test)

Significance (p)

Detection rate attention field 68% (±1.4) 72% (±1.3) 4% 3.943 p < 0.001Detection rate probe field 71% (±1.8) 74% (±1.7) 3% 3.711 p < 0.001Difference detectionprobe vs. attention field

3% 2% 1%

Reaction time attention field 506 ms (±9.9) 449 ms (±12.0) 57 ms 4.230 p < 0.001Reaction time probe field 495 ms (±12.0) 466 ms (±11.2) 29 ms 3.729 p < 0.001Difference reaction timeprobe vs. attention field

−11 ms 17 ms −28 ms

Detection rates and reaction times in the attention field (valid cue) and probe field (invalid cue) in conditions with vs. without directing attentionat the area marked by the cue.

4 B R A I N R E S E A R C H X X ( 2 0 0 6 ) X X X – X X X

ARTICLE IN PRESS

trials) correlated with a larger difference between the size ofthe ARVs in the attention field vs. probe field (Spearman’s Rho:ρ = 0.541, p < 0.01).

Comparison of results from typical patients illustrated theimportant role of theARVs for the cueingeffect: Patient #10hada verynarrowARV (Fig. 2a). Presentation of the cue in theuppervisual field was followed by only a marginal increase ofstimulus detection performance in the attention field ascompared to uncued trials (from 61.1% to 62.5%, Table 2, leftcolumn), but a marked decrease of reaction times, indicatingthat he had indeed shifted the focus of attention to the cuedarea. The patient also showed some unspecific benefit in the(unattended) probe field at the lower visual field border (Fig. 2cvs. 2e). In contrast, patient #2 had a broad ARV, particularly inthe upper visual field (Fig. 2b). This subject showed a markedfacilitation in valid trials: stimulus detection increased by16.0%and reaction timesdecreasedby 83ms. In theprobe field,an unspecific effect of attention in invalid trials was found, butit amounted to only half that of the specific effect in valid trials(8.3% and 39 ms, respectively, see Table 2).

Results from the complete patient sample were inaccordance with these single-case results. The probabilityto achieve an increase of stimulus detection performanceby cueing was greater in minimally defective areas with ahigher stimulus detection baseline than in strongly defec-tive areas. For stimulus positions in almost blind areas (0-20% detection in the unattended condition) improvementby attentional cueing amounted to 37.34% ± 5.13. In regionswith a baseline detection of 20-40% we found a cue-

Table 2 – Quantitative single_case results of patients with narro

Patient #10: Narrow ARV

Stimulus detection (%) Reaction

HRP all trials 55.1 358AFT all uncued trials 54.6 446AFT attention field uncued 61.1 462AFT attention field valid cue 62.5 368AFT probe field uncued 69.4 464AFT probe field invalid cue 72.2 407

Comparison of stimulus detection performance and reaction times (RT). Dshow the benefit induced by the cue (AFT = attention field test).

induced increase of 58.61% ± 9.32. In areas with 40-60%unattended detection, the increase was 91.67% ± 3.81; andin areas with 60-80% unattended detection we found anincrease amounting to 73.17% ± 6.09 (Kendall’s W = 0.227,p < 0.001).

Analysis of performance in catch trials revealed that thenumber of errors (false positive reactions) did not increasewith increased detection rates, i.e. the decision criterion didnot change (Spearman’s Rho ρ = -0.027, p = 0.822).

Subjective reports of the patients indicated that the cuechanged the quality of perception in the attended area, i.e.stimuli were spontaneously reported as being brighter, largerand/ or having clearer contours than those in uncued trials atthe same position. These phenomena were not reported incatch trials without a target.

3. Discussion

In this study, a paradigm of selective visuo-spatial atten-tion commonly employed in normal subjects (Posner, 1980)was successfully adapted to patients with visual fielddefects. A specifically designed cue joining aspects ofsymbolic and direct cueing helped subjects to direct theirattention focus at the visual field border. Gestalt-comple-tion of the cue-frame was readily achieved- often to thesurprise of the patient- even within the first practice trials.This observation confirmed earlier findings of Gestalt-completion in the blind field of patients with occipital

w vs. broad ARVs

Patient #2: Broad ARV

time (ms) Stimulus detection (%) Reaction time (ms)

54.9 46858.2 51458.3 51474.3 43175.0 46783.3 428

ifferences between cued and uncued trials of the attention field tests

5B R A I N R E S E A R C H X X ( 2 0 0 6 ) X X X – X X X

ARTICLE IN PRESS

lesions (e.g. Gassel and Williams, 1963; Warrington, 1962;Torjussen, 1978; Sergent, 1988; Safran and Landis, 1996).Thus, attention could be shifted specifically to a part of thepatient’s visual field that had not been consciouslyrepresented before.

Instead of the expected costs in performance whenstimuli were presented at an unexpected position in thevisual field (Posner, 1980), i.e. in the probe field, mostsubjects also showed some unspecific benefit in that invalidcue condition. This may be explained as a topographicallyunspecific alerting effect of the relatively infrequent cue inthe attention field test that placed high demands on thesubject’s vigilance due to its long duration. Also, theinterval between cue and target (SOA) was rather long sothat a voluntary shift of attention was triggered which mayhave prevented the long inhibitory periods after cuepresentation typically found for automatic shifts (Nakayamaand Mackeben, 1989). It is also possible that attention wasnot shifted exclusively at the attention field: the focusmight have been diffuse or patients may even have dividedit between the attention field and the probe area (Egly andHoma, 1984). This behavior could have been caused by asubjectively lower cue validity than the objective validity(80%) since in valid trials, the target stimulus was some-times presented in the blind area and could not beperceived by the patient. If in those cases the patientassumed that these trials were invalid or catch trials, thiswould have reduced the subjectively perceived ratio of validand invalid trials. Hence the patient’s motivation toexclusively shift attention to the cued location may havebeen reduced. However, the present design does not allowtesting whether the attention focus had been diffuse or wassplit during cued trials.

Consistent with earlier findings that visuo-spatial cueingcan induce an increase of perceptual sensitivity withoutchanging the subjects’ response criterion (Heinze and Man-gun, 1995) we observed in our patients that the probability ofstimulus detection increased and reaction times decreasedwithout a criterion change. Presumably, shifting attention tothe ARVs with the help of a visuo-spatial cue reducedperceptual thresholds. Especially at “more intact” positionswithin an ARV, the cue provided the amount of activation thatwas necessary to cross the threshold of conscious perception,but it did not improve stimulus detection in the completelyblind zone.

A shortcoming of the present study is the fact that thesubjects’ eye movements were not recorded with an eyetracking device. However, the observation of eye movementsvia a mirror and the elimination of trials contaminated byeye movements yielded plausible results of visual fieldtesting: had the patients simply moved their eyes towardthe cue frame, they could have easily detected all stimulipresented in that area, including the targets in the blindfield. Instead, the improvements were moderate and in nocase showed a completely “intact” performance in theattention field after presentation of the cue. Furthermore,the extent of the cueing effect depended on the intactness ofthe visual field position where the target was presentedwhich cannot easily be explained by either eye movementsor a criterion change.

In a related study using the same cueing design (Poggel,2002; Poggel et al., in prep.) normally sighted subjects showeda decrease of reaction times in the cued area when they weretested with supra-threshold stimuli (perfect detection perfor-mance in attended and unattended trials), and improvementof detection and reaction timeswhen tests were repeatedwithnear-threshold stimuli. These findings suggest that attention-al cueing is most beneficial when neuronal activation is justbelow the threshold of conscious perception, either after alesion in the visual system, or due to physical stimulusattributes.

ARVs in patients with visual field defects can beconsidered a functional representation of partially survivingneuronal structures in or near the damaged regions of thebrain (Sabel, 1999; Kasten et al., 1999) with reduced neuronalactivation. Single-cell recordings (e.g., Gilbert, 1998; Moranand Desimone, 1985), EEG-experiments (e.g., Heinze andMangun, 1995) and neuroimaging studies (e.g., Sengpiel andHubener, 1999; Gandhi et al., 1999) indicate that attentionplays a crucial role in modulating early sensory processing ofvisual stimuli. Presumably the cue increased neuronalactivity in partially lesioned brain areas representing theARVs. Visuo-spatial attention also triggers changes ofreceptive field size in cortical neurons (Gilbert, 1998).Attentional cueing at the visual field border may lead to anacute expansion of receptive fields into the blind region sothat stimuli would be detected even when they arepresented further away from the intact area.

3.1. Conclusions

Focusing spatial attention at ARVs in patients with visualfield defects increases the probability of stimulus detectionand reduces reaction times. Both effects depend on the sizeof the ARVs. The cue reduces visual thresholds andfacilitates conscious perception in the ARVs, presumablyby inducing short-term plasticity in partially lesionedcortical regions. This may be the basis for a new strategyusing attention to induce long-term plasticity and topermanently restore visual function in partially blindpatients. In fact, in a previous study (Poggel et al., 2004)we have already shown that the efficacy of vision restora-tion therapy (Kasten et al., 1998) can be significantlyenhanced in a long lasting manner when it is combinedwith a cueing task.

4. Experimental procedure

4.1. Subjects

Twenty-three patients with chronic homonymous visualfield defects (Table 3) caused by post-geniculate brainlesions (visual cortex and/ or optic radiation) were recruitedfor a training study (Poggel et al., 2004) from patient files inhospitals. The sample has been described in detail byPoggel (2002). Patients were included only when the lesionage was at least six months (average time since lesion:35.5 months, median: 24.4 months, range 6.6−184.3 months).All participants had normal or corrected to normal vision in

Table 3 – Description of patient sample

Case Gender Age Side of Lesion Etiology

1 Female 31 Left Infarction2 Male 59 Right Infarction3 Male 36 Right Vascular4 Male 39 Left Infarction5 Female 23 Right Infarction6 Male 36 Right Infarction7 Male 67 Left Infarction8 Female 40 Right Infarction9 Female 60 Left Infarction10 Male 20 Left Infarction11 Male 61 Left Infarction12 Female 40 Left Infarction13 Female 30 Right Trauma14 Female 33 Right Vascular15 Male 35 Right Vascular16 Female 36 Right Infarction17 Male 58 Left Infarction18 Male 50 Left Infarction19 Male 37 Right Infarction20 Male 41 Left Infarction21 Female 25 Left Infarction22 Female 37 Right Tumor23 Male 48 Left Infarction

6 B R A I N R E S E A R C H X X ( 2 0 0 6 ) X X X – X X X

ARTICLE IN PRESS

their spared visual field. Volunteers with other forms ofvisual impairment (based on ophthalmologic exams), cog-nitive/ attention deficits (based on a variety of neuropsy-chological tests, see Poggel, 2002, and Poggel et al., 2004, fordetails), a history of psychiatric disorders, deterioration orimprovement of vision of more than 2% during baselinevisual field tests, as well as minors, were excluded fromparticipation. Written consent was obtained from all sub-jects. The study design had been approved by the localethics committee and was in accordance with the Declara-tion of Helsinki.

Fig. 3 – Time course and design of attention field test. In neutral/without a cue, i.e. physically identical to conventional high-resolwas presented for 200 ms in the upper visual field (solid frame imean SOA of 1000 ms (randomized between 750 and 1250ms). Inattention field at the upper visual field border, but stimuli were prnot visible for patients). These three trial types as well as catch tpresented in randomized order.

4.2. Campimetric procedures

4.2.1. Visual field testComputer-based high-resolution perimetry was used for theassessment of visual field size and determination of ARVs(HRP, Nova Vision AG, Magdeburg). Subjects were seated infront of a 17”-computer screen (see Poggel, 2002, and Poggelet al., 2004, for details) in a darkened room, the head beingstabilized on a chin rest. Stable fixation was monitored bya central task (detection of color change of the fixationpoint, Poggel, 2002; Poggel et al., 2004) and by observinggaze position via a mirror. Patients had to press the spacebar on the computer keyboard when they detected a target,i.e. small white light stimuli (duration = 150 ms) presentedin random order at 474 positions on the screen. Testduration was approximately 20 minutes. Hits, misses, falsepositives, and reaction times were recorded and graphicallydisplayed (Fig. 1). HRP was repeated five times, and thedetection rate at each visual field position was determined.ARVs were defined as the areas at the visual field bordercontaining stimulus positions with a detection rate be-tween 20% and 80%. Patients with large vs. small ARVswere categorized according to a median split with respectto the size of the ARVs. More details about the size andGestalt of ARVs in the patient sample as well as a graphicdocumentation of visual field tests can be found in Poggel(2002).

4.2.2. Attention field testIn order to test visual performance in ARVs under attendedconditions, a conventional paradigm of visuo-spatial cueing(Posner, 1980) was adapted and implemented into the HRP testto help patients focus attention at the visual field borderduring visual field testing.

A large square frame (12° × 12°), positioned over thevisual field border, served as a cue. It covered parts ofintact and blind areas, as well as a defined region of the

uncued trials (screenshots on the left) stimuli were presentedution perimetry (HRP). In valid trials (middle column), the cuendicating the attention field), followed by a stimulus after ainvalid trials (right column), subjects were cued to the

esented in the probe field (dashed frame in lower visual field,rials without stimulus presentation after the cue were

7B R A I N R E S E A R C H X X ( 2 0 0 6 ) X X X – X X X

ARTICLE IN PRESS

ARVs. Hence, a part of the frame was “hidden” in the blindfield and was completed to the Gestalt of a square (seePoggel, 2002, for details) to attract attention not only to theintact but also to the defective parts of the visual fieldenclosed by the cue frame. Thus, the frame acted like asymbolic cue, but since it included the target position, itwas also a direct cue (Posner, 1980).

This cue was implemented into the HRP test in a cueingprocedure structured according to adesigndescribedbyPosner(1980) (Fig. 3). Inneutral trials, no cuewaspresentedprior to thetarget, i.e. the physical characteristics of stimulation were thesame as in standard HRP. In valid trials, the cue-frame wasshown for 200 ms, and after a randomized interval (meanstimulusonset asynchrony, SOA=1000ms±250ms), the targetwas displayed in the attended field. In invalid trials, the cuewas presented at the same location as in valid trials, but thetarget appeared in a probe field located opposite the attentionfield (Fig. 3). Additionally, catch trials, i.e. presentation of a cuenot followed by a target, were included in order to test forpossible shifts of the subjects’ decisional criterion. Patientswere instructed to press the space bar on the keyboard upondetection of a target on the screen. Valid (144 trials), invalid (36trials), neutral (474 trials), and catch trials (20) were presentedin random order. The ratio of valid to invalid trials was 4:1.

The attention field (validly cued targets) was always locatedat the upper visual field border, the probe field (invalidly cuedtargets) was placed approximately opposite the attention fieldin such a way that the baseline rate of stimulus detection aswell as proportions of blind, intact, and partially defectiveareas enclosed by the probe fieldwere approximately the sameas in the attention field. However, due to the naturaltopography of the visual field borders, this was not alwayspossible. All trialswith eyemovementswere eliminated onlineand repeated at the end of the test. Total test duration wasapproximately 30 minutes. Each patient performed threeattention field tests in the course of baseline measurements.

Data were acquired before the training procedure (seePoggel et al., 2004) was started, i.e. in the present study onlythe pre-training baseline measurements were included sothat the test results were not biased by potential processesof visual system plasticity.

4.2.3. Statistical analysisData were analyzed using the SPSS program (SPSS Inc.,Chicago, IL). Wilcoxon test, Mann-Whitney test, Kendall’s W,and Spearman’s Rho correlation coefficient were used tocompare data. A (two-tailed) alpha of 0.05 was set for all tests.

Acknowledgments

Thanks to Dr. S. Brandt and Prof. K. Einhäupl, Charité, HUBerlin, and to Prof. P. Marx and Prof. F. Hoffmann, UKBenjamin Franklin, FU Berlin, for help with recruitment andproviding test rooms. We thank Andreas Bohne for program-ming. Supported by grants from the DFG (Graduiertenkolleg

"Biologische Grundlagen von Erkrankungen des Nervensys-tems") and the state of Saxony-Anhalt.

R E F E R E N C E S

Egly, R., Homa, D., 1984. Sensitization of the visual field. J. Exp.Psychol.: HPP 10, 778–793.

Gandhi, S.P., Heeger, D.J., Boynton, G.M., 1999. Spatial attentionaffects brain activity in human primary visual cortex. Proc.Natl. Acad. Sci. U. S. A. 96, 3314–3319.

Gassel, M.M., Williams, D., 1963. Visual function in patients withhomonymous hemianopia: III. The completion phenomenon,insight and attitude to the defect, and visual functionalefficiency. Brain 86, 229–260.

Gilbert, C.D., 1998. Adult cortical dynamics. Physiol. Rev. 78,467–485.

Heinze, H.J., Mangun, G.R., 1995. Electrophysiological signs ofsustained and transient attention to spatial locationsNeuropsychologia 33, 889–908.

Kasten, E., Wüst, S., Behrens-Baumann, W., Sabel, B.A., 1998.Computer-based training for the treatment of partialblindness. Nat. Med. 4, 1083–1087.

Kasten, E., Poggel, D.A., Müller-Oehring, E.M., Gothe, J., Schulte,T., Sabel, B.A., 1999. Restoration of vision II: Residualfunctions and training-induced visual field enlargement inbrain-damaged patients. Restor. Neurol. Neurosci. 15,273–287.

Moran, J., Desimone, R., 1985. Selective attention gates visualprocessing in the extrastriate cortex. Science 229,782–784.

Nakayama, K., Mackeben, M., 1989. Sustained and transientcomponents of focal visual attention. Vision Res. 29, 1631–1647.

Poggel, D.A., 2002. Effects of visuo-spatial attention on therestitution of visual field defects in patients with cerebrallesions. Shaker Verlag, Aachen.

Poggel, D.A., Kasten, E., Sabel, B.A., 2004. Attentional cueingimproves vision restoration therapy in patients with visualfield defects. Neurology 63, 2069–2076.

Posner, M.I., 1980. Orienting of attention. Q. J. Exp. Psychol. 32,3–25.

Posner, M.I., 1995. Attention in cognitive neuroscience: Anoverview. In: Gazzaniga, M.S. (Ed.), MIT Press, Cambridge,Massachusetts, pp. 615–624.

Sabel, B.A., 1999. Restoration of vision I: Neurobiologicalmechanisms of restoration and plasticity after braindamage—A review. Restor. Neurol. Neurosci. 15, 177–200.

Safran, A.B., Landis, T., 1996. Plasticity in the adult visualcortex: Implications for the diagnosis of visual fielddefects and visual rehabilitation. Curr. Opin. Ophthalmol.7, 53–64.

Sengpiel, F., Hubener, M., 1999. Visual attention: spotlight on theprimary visual cortex. Curr. Biol. 9, 318–321.

Sergent, J., 1988. An investigation into perceptual completion inblind areas of the visual field. Brain 111, 347–373.

Torjussen, T., 1978. Visual processing in cortically blindhemifields. Neuropsychologia 16, 15–21.

Warrington, E.K., 1962. The completion of forms acrosshemianopic field defects. J. Neurol. Neurosurg. Psychiat. 25,208–217.

Zihl, J., von Cramon, D., 1979. Restitution of visual function inpatients with cerebral blindness. J. Neurol. Neurosurg.Psychiat. 42, 312–322.