Efficient Visual field rehabilitation in the homonymous hemianopia by

prism attached to lens

Fernando Palomar Mascaro1 & J. Antonio Aznar-Casanova2

1- Palomar's Optometry Center. Barcelona, Spain.

2- Faculty of Psychology, Universitat de Barcelona, Spain.

Running head: Homonymous Hemianopia Rehabilitation

Correspondence regarding this article should be addressed to:

J. Antonio Aznar-Casanova
Facultad de Psicología, Universitat de Barcelona,
Passeig de la Vall d’Hebron, 171, Barcelona, 08035, Spain.
Phone: +34 933 125 145
Fax: +34 934 021 363
E-mail: jaznar2@ub.edu
Abstract

Background: A relevant problem in Visual Sciences which demands an urgent solution

arises to what extent the rehabilitation and partial recovery of the visual field
corresponding to a hemiretina in complete homonymous hemianopia (CHH) can be
retrieved. Also concerned the question whether the acquisition of compensatory
oculomotor strategies for these patients results (or not) in an improvement in the regular
performance.

Methods: The group of patients had neuro-ophthalmologic symptoms with a sector

field loss, type CHH. We tested the effectiveness of a visual rehabilitation treatment for
CHH, consisting of coupling a prism in band on each crystal glasses. We designed an
experiment in which participants had to point to a target located at a particular distance
and direction in the visual field, using a 'device pointer' inserted on a king size
protractor.

Results: Results from an ANOVA revealed that patients improved the ability to locate
objects in space, after three months applying the treatment, i.e. by means of practice and
adaptation.

Conclusion: The partial recovery of the lost visual field can be explained by two
alternative hypotheses. One based on brain neuroplasticity and another based on
perceptual learning and attentional filtering of two overlapping images.

Keywords: homonymous hemianopia, visual rehabilitation, visual psychophysics,

hemianopic prisms, loss of visual field.

2
Introduction

Patients with normal visual acuity and complete homonymous hemianopia (CHH)
presents spatial orientation difficulties that impact on their quality of life. They can lose
the ability to dress up, ignore routes that are familiar, stumble when walking alone, to
turn collide and cannot read, or work with your computer. In short, they lose personal
autonomy. Paradoxically, it is common that in the beginning of this neurological
condition, the patient is not aware of these symptoms (Egido & Díez-Tejedor, 2003). In
this context, a crucial problem in Visual Sciences, which urges to be answered and
which should dedicate a research effort may arise in the following terms: to what extent
is it possible to achieve some rehabilitation or some partial recovery of the lost of field
of vision (FoV)?
Among the specialists in visual rehabilitation concerning hemianopia (eg Pambakian,
Currie & Kennard, 2005; Kasten, Bunzenthal & Sabel, 2006) there is also some concern
about whether the acquisition of compensatory oculomotor strategies could result (or
not) in an improvement in performance and normal operations in daily life.
Hemianopsia is often classified based on the location of the VF that has been lost
(Cohen & Waiss, 1996). The complete homonymous hemianopia (CHH) can be defined
as absolute or partial loss of vision in the right or left FoV of both eyes (Duke-Elder,
1970). The CHH occur as a result of structural pathological processes affecting
retrochiasmatic visual pathways, and are caused by a variety of lesions and in different
topographies (Dantas, 1984; Duke-Elder, 1970; Harrington, 1993).
There are few articles in the scientific literature (Smith et al 1982;.. Trobe et al, 1983),
reviewing the epidemiological, etiology, clinical and evolution factors of hemianopsias.
As for the etiology, hemianopical defects are among the most common disorders after
strokes. In USA, Rossi, Kheyfets & Reding, (1990) reported that in a year occurred
about 10 million cases of head injuries and 20% of them were associated with brain
injuries, but a third of patients who survived a stroke had complete or incomplete HH.
In UK, Pambakian and Kennard (1997) indicated that approximately one third of
patients who survived a stroke, had complete or incomplete HH, 40% of homonymous
hemianopia being caused by injuries in the occipital lobe, 30% in the parietal lobe, 25%
in the temporal lobe, and 5% in the optic tract and lateral geniculate nucleus.

3
Considering both factors the prevalence of hemianopia resulting from stroke (including
strokes) and longevity of the population, research on the treatment of hemianopia
rehabilitation must be a priority for preventing problems in visual health in the
population.
Few studies (Smith, Weiner & Lucero, 1982; Rossi, Kheyfets & Reding, 1990; Perlin &
Dziadul, 1991) have reported that the rehabilitation by prisms can be efficient, and these
have usually referred to the use of Fresnel prisms. Thus Palomar Petit (1979) describes
a new technique for rehabilitation in patients with hemianopia consisting of a small
strip-shaped prisms placed on to the side with hemianopia and getting the restoration of
the central field. Gottlieb (1988) proposed the use of a sector prism monocularly
adapted to the side of lens corresponding to loss VF. He noted that when the patient's
eyes were in primary position of gaze or directed outside the hemianopic hand, the
monocular prism had no effect on the FoV. However, when the gaze was directed into
the prism, confusion or perception of two different objects in the same situation
occurred. He interpreted that confusion would arise by the appearance and visibility of
an object which would be invisible without the prism. He also noted that diplopia
occurred with the expansion of the resulting FoV, which could be very disorienting and
unpleasant for the patient. These drawbacks could explain the limited success of this
technique.
Rossi et al. (1990) examined recovering of the FoV in 18 patients with stroke and
homonymous hemianopia or unilateral visual neglect to treatment with 15-diopter
plastic press-on Fresnel prisms (n = 18) and they compared these patiens with 21
controls. After 4 weeks, the prism-treated group performed significantly better than
controls on several tasks requiring hand-eye coordination. They concluded that the
treatment with 15-diopter Fresnel prisms improves visual perception test scores but not
in the Barthel ADL test in stroke patients with homonymous hemianopia or unilateral
visual neglect.
Zhil (1995), Kasten y Sabel (1995), y Kasten, Wust, Behrens y Sabel (1998) suggested
that regular training of the blind FoV using visual stimuli similar to those used in a
computer-controlled perimetry test could facilitate recovery next FoV to the midline and
provide an expansion of the FoV.

4
Pakambian y Kennard (1997) addressed the issue of whether it was possible to restore
visual function in patients with CHH. They emphasized the importance of several
rehabilitation treatments, for example psychophysical techniques, to improve care in the
blind half of the FoV. They also suggested the possibility of using optical aids,
hemianopic mirrors and prisms, as well as cognitive techniques for improving eye
movements. They concluded that research on rehabilitation of patients with brain
damage and functional impairment was a very difficult and laborious task. They
recognized that the effectiveness of such treatments were not properly detailed, since
there was insufficient research, and most of the previous studies suffered from some
methodological flaw.
In a comprehensive review, Peli (2000) classified the effects of the instruments used in
the rehabilitation of hemianopia into two groups: those relocating the FoV and other
producing expansion. He argues that the expansion effect of the FoV is preferred
because the simultaneous FoV is wider and allows the patient to control the
environment at all times, which enable him for safer mobility. However, relocation only
changes the position of the lost FoV, or their relative position with regard to the
midline. Peli also holds that the FoV changes when viewed through binoculars sectors.
Since the patient does not see objects in that part of the FoV it is less likely to fixate,
thus making a voluntary eye movement is required. He points out that in addition to
these limitations, have an optical loss of FoV in the center of the FoV caused by the
binocular sectorial prism.

Peli (2008) has also developed a method consisting on a monocular sectorial prism
limited to top, bottom or both peripheral FoV. This prism is placed in the entire width of
the lens, to be effective in all lateral positions of gaze (Patent No.: US 7,374,284 B2).
The prism expands the FoV by promoting peripheral diplopia, producing optically
peripheral exotropia, while maintaining bifoveal alignment. Peli (2000), stated that this
expansion of the FoV can be measured with standard binocular perimetry, because it is
effective in all positions of gaze, including the primary position. He uses 40 diopters
Fresnel prisms, which give a spread of 20° around the midline. However, since the
prism only affects the peripheral vision, one can use other prism of greater power.
Bowers, Keeney y Peli (2008) used glasses with 40 diopters Fresnel prisms, placed in
the hemianpsic top and bottom side crystal. They establish that improvement in mobility
of patients with the use of said prisms.

5
O´Neil, Connell, O´Connor, Brady, Reid y Logan (2011) have proposed the use of
monocular prisms on the side of the complete hemianopia, with bases at the default
address. In this way a peripheral exotropia is produced, which achieve the expansion of
the FoV.
Palomar, Palomar, Cardona & Quevedo (2011a and b), emphasized the importance of
ascertaining the prismatic power. As well as the relevance of the right position of the
binocular prism attached to the lens for the success in the rehabilitation process.
This study was aimed to evaluate the efficacy of a visual rehabilitation treatment for
CHH, consisting of coupling a prism attached in band on each crystal glasses. To this
end we have designed an experiment in which participants were asked for locate a
point-target located at a particular distance and direction in the FoV, using a 'device
pointer' inserted on a king size protractor. Results showed how patients improved the
ability to locate objects in space with practice and time, by using the attached prisms.
The recovery of the FoV could be explained by two alternative hypotheses which are
discussed. One based on brain neuroplasticity and another based on perceptual learning
and attentional filtering of overlapping images.

Experiment

Methods
Participants
A total of twenty patients (six female, fourteen male), ranging from 18 to 63
years of age, and fifteen controls (six female, nine male), ranging from 31 to 66 years of
age participated in the study.
The inclusion criteria for the group of patients with CHH were: 1) All
participants had a Typus CHH from more than one year, assessed by Dicon’s
computerized perimetry (clinical paradigm) and neurologically evaluated by CT or
MRI; 2) all they had 20/20 visual acuity (VA); 3) all they had not been previously

6
treated with any visual rehabilitation. While exclusion criteria for this group were: 1) to
present anosognosia of the hemianopia and evidence of mental disorder or serious
physical impairment; 2) to have a normal or corrected VA lower than 20/20.
The inclusion criteria for the group of normal patients were: 1) to have 20/20
normal or corrected VA; 2) no presence of any FoV defects, verified by computerized
perimetry; 3) does not present any anomaly of binocular vision. Exclusion criteria for
this group were: 1) to present a lower than 20/20 VA; 2) to show a low attitude towards
collaboration.
Participants provided written informed consent after the nature of the study was
explained to them. The Declaration of Helsinki tenets of 1975 (as revised in October in
2008) were followed throughout the study, which received approval from the Ethics
Committee of the University of Barcelona.

Stimuli and apparatus

The stimulus were a point 5 mm in diameter which could be placed on each of
five direction on the visual field (-60º, -30º,0º, +30º and +60º). The experimenter put
this target-point at the eye level in a particular randomized angular location next to the
protractor. To measure precision in location of the visual direction of a target point we
designed a device consisting on a protractor provided of a bar fixated on it (see Fig. 1).

INSERT FIGURE 1 HERE PLEASE

Figure 1. Pointer perimeter device consisting on a protractor provided of a bar fixated

on it which enables one to measuring precision in location of point-targets located in
different direction of the visual field.

7
Treatment description: Palomar vertically attached prism
The overall goal of the rehabilitation is threefold: 1) to reduce visual impairment
resulting from hemianopia, 2) increase the confidence of patients and 3) facilitate their
reintegration into social and professional lives in an autonomous and independently way.
Treatment consists of a sectorial prisms vertically attached to his glasses and with the
bases of the prism oriented towards the anopical area (see Fig. 2). This accessory
facilitates access to the patient’s lost FoV and helping them in their spatial orientation
(Palomar, 2011). For its appropriateness in both distant and near vision the power of the
attached prisms were determined. The prisms had 20-25 diopters for far visual distance
and 15-20 diopters for near vision, and were placed attached to the center of the glass
lenses in vertical stripes. The bases of the prism were oriented toward the side of the
homonymous hemianopic default. Depending on the homogeneity of the loss central
FoV, the prism were usually shift between 1-5.5 mm from the center to the
hemianopical side (Palomar 2013). Checking their efficacy it was done by around 30º
computerized perimetry.

Figure 2. Top view (left panel) and front (right panel) of the spectacle of a patient with
right homonymous hemianopia with the 20 diopters Palomar’s attached prisms, whose
bases are oriented to the right side.

The patient simultaneously receives the images from the FoV of the left and right eye,
projected onto the functional hemiretinas. Then, images corresponding to the FoV of the
non-functional hemiretina were captured through the prisms (Fig. 3). Upon receiving these
two different overlapped images, probably the patient must sequentially process these
images. Consequently, visual system must perform a reconstruction of the visual space
subtended for each eye, combining (merging) both reconstructed spaces. Thus, using a
campimeter (perimenter) an assesment of the restored central FoV is possible by
comparing the spatial localization accuracy in both conditions, executed with/without the
aid of the attached prisms. Thus, it is possible to quantify the beneficial effect of treatment.

8
Figure 3. Diagram illustrating the functioning of the attached prisms in the case of a
patient with right CHH.

Procedure
The participants were tested for precision in the localization of one point placed
to a constant distance (30 cm in radius) for five visual directions in their visual field. Of
course, that visual direction in which the patient is facing the front was the 0º visual
direction and ±30º and ±60º deviation to the left or right, were also presented in random
order. Obviously, patients with homonymous hemianopia in the left visual field will
show difficulties locating point-targets on the right visual field and vice versa patients
with homonymous hemianopia in the right visual field will show difficulties locating
point-targets on the left visual field.
The participants were instructed to rotate a bar to pointing towards the target
point as precisely as possible and we register the rotation (visual direction) of the bar
pointing to the target fixating the bar on a protractor (see Figure 1). To this aim we used
a magnitude production psychophysical method. Five trials (repettions) were conducted
by using the pointing-bar task one for each visual direction. These 25 trials were did
under three visual conditions (two under monocular vision: LE and RE and another
under binocular vision). In addition, the participants run this test three times: a) at
baseline (first session); b) after a month; and c) after three months.

9
 Therefore, each participant performed a total of 225 pointing-trials per session (5
repetitions  5 Directions  3 visual conditions  3 times). Each participant took around
40 minutes to complete the 225 trials/session. So, the total duration of the tests in the
three sessions was around 120 minutes.

Results
First, we calculated the pointing errors (deviation or difference in rotation) for
each trial in the group of patients with hemianopia. Then the mean of the errors were
submitted to a mixed factorial ANOVA with “Directions” (-60º, -30º,0º, +30º and +60º),
“Visual Condition” (Left Eye; Right Eye and Binocular), and “Time” (baseline, 1
month, 3 month) as repeated measures and “Hemianopia” as between-subject factor.
The Greenhouse-Geisser correction for non sphericity was applied where necessary and
is indicated by adjusted degrees of freedom.
The results from the 5 x 2 x 3 x 3 (Directions x Hemianopia x Condition x Time)
ANOVA showed that the main effects of “Directions” [F(2.097, 37.75)= 42.468; p <
.001; η2p= .702; pow= 1.0] and “Hemianopia” [F(1,18)= 12.830; p < .001; η2p= .536;
pow= .991] were statistically significant. While the main effects of “Condition” and
“Time” were not significant. Errors increased as the deviation of the frontal line (visual
direction) did it (see Fig 4). But errors also depended on the lateralization of the
hemianopia (to the left or to the right). Thus, patients with left-hemianopia did greater
errors (-.643º) on the right-visual-field and vice versa patients with right-hemianopia did
greater errors (+.451) on the left-visual-field.

INSERT FIGURE 4 HERE PLEASE

10
Figure 4. Mean of the errors in the pointing task according to “Visual direction” and the
best fit function.

The two-way interaction of “Directions x Hemianopia” [F(2.097, 37.75)= 6.934;

p < .001; η2p= .278; pow= .992], and “Directions x Time” [F(3.34, 60.117)= 47.302; p
< .001; η2p= .724; pow= 1.0], were significant. Simple effects analysis revealed that
errors in baseline and at the first month were significantly greater for contralateral
directions to the hemianoptic retina but they were null for the ipsilateral retina (Fig. 5)
and differences were not significant. Besides, the error in these directions decreased
significantly with time in such a way that results of running the test at the third month,
showed that the error was almost null (Fig. 5) and differences were not significant.
But also the two-way “Hemianopia x Time” interaction was significant [F(1.14,
20.52)= 24.584; p < .001; η2p= .577; pow= .990]. Both errors that corresponding to the
left hemianopia and that caused by damage in the right hemianopia decreased
significantly as time passes.

INSERT FIGURE 5 HERE PLEASE

11
Figure 5. Mean of the errors in the pointing task for each type of hemianopia as a
function of the“visual direction” and according to “Time”. Bars indicate the standard
error of the mean (95% confidence level).

Finally, the three-way “Directions x Hemianopia x Time” interaction was

significant [F(3.34, 60.117)= 8.378; p < .001; η2p= .318; pow= .994], suggesting that
for participants with hemianopia errors occurred when the stimulus were located at the
contralateral side in the visual field respect to the damaged hemianopia. These errors
were larger for ±30º (p < .03) and even more for ±60º (p < .03) compared with errors in
0º visual direction. However, these effects diminished as the time elapses, disappearing
errors after three month using the treatment of prism attached to the glasses. Figure 5
shows mean of error in the pointing task corresponding to every type of hemianopia as a
function of the visual direction, each panel informing about a particular “Time”. Taken
together these results indicate that Participants with hemianopia improved the
performance in the pointing task, locating the target-points (stimuli) in the affected
region of the visual field and this improvement was achieved in about three months of
treatment with the prisms prism attached to the glasses.

By last, we calculated the pointing errors (deviation or difference in rotation) for

each trial in the Control group (people without hemianopia). Then the mean of the
errors in the trials were submitted to a three factor mixed ANOVA with “Directions” ((-
60º, -30º,0º, +30º and +60º), “Condition” (Left Eye; Right Eye and Binocular), and
“Time” (baseline, 1 month, 3 month). “Directions”, “Condition” and “Times” were
taken as repeated measures. The results from the 5 x 3 x 3 (Directions x Condition x
Time) ANOVA showed that none of the main effects were significant. Moreover,
neither the two-ways interactions nor the three-way one were significant. Therefore, no
significant differences in error in the pointing task were observed.
Errors between the two groups (patients and control) for every “visual direction”
in the last session (after three month of treatment) and only in the binocular visual
condition were compared by a two-way mixed ANOVA: 2 (Groups) x 5 (Visual
Direction), with Group as the between subject variable and “visual directions” as the
within-subjects variable. The results showed again that none of the main effects for

12
“Groups” (p < .812) and “Visual Direction” (p < .186) were significant. Moreover, the
two-way “Groups x Visual Direction” interaction was not significant (p < .187).
Therefore, no significant differences in error between the two groups (after the
treatment) in the pointing task were observed.

Discussion
A major problem in Visual Sciences demanding an urgent response arises to
what extent a rehabilitation and partial restoration of the FoV in patients with CHH is
possible. It also concerned the question of whether the acquisition of compensatory
oculomotor strategies results or not in an improvement in performance and normal
functioning for these patients. The overall goal of the rehabilitation treatment of
hemianopia was to reduce disability resulting from this loss of FoV, to increase the
confidence of the patients and facilitate their reintegration into social and professional
life autonomously and independently. The aim of the present study consisted on
investigating the efficacy of a treatment for homonymous hemianopia, in which the
patients wear prism attached to the lens for three months. More specifically, the
treatment was applied to couple a prism attached in sides on the patient’s glasses
(Palomar-Mascaró, et al., 2011a, b).
To test the efficacy of the treatment we designed an experiment in which
participants had to locate a number of points sequentially presented in random order.
The points were located at a particular distance and direction in the visual field and to
locate the patients used a 'device pointer' transported inserted in a larger size. The
analysis of the observations showed that patients had monocularly recovered the loss
FoV. Thus, patients improved the ability to locate objects in space by using the attached
prisms and with practice and spatial visual adaptation (perceptual learning in eye-hand
coordination, etc.). With the use of the attached prisms, in three cases under treatment
having more than five years of evolution, they there were a recovery of between 5 and 8
degrees around the central FoV. The prism had been displaced between 1 and 2 mm
towards the loss-side FoV. This recovery has been clinically evaluated by computerized
perimetry.

13
Location tasks has been used to asses cortical blind hemifield. So, Zihl & von Cramon
(1980) tested localization in the cortically blind fields of their patients, and found that
performance improved with practice when pointing (Bridgeman and Staggs, 1982) or
saccadic responses were required (Zihl, 1995; Zihl and von Cramon, 1980; Zihl and
Werth, 1984). It is interesting in the context of perceptual learning that the practice
effect transferred to stimuli of lower contrast in the patient tested by Bridgeman and
Staggs (1982).
Our results raise the question about how (through what monocular mechanisms) there
has been such spatial reconstruction. Two possible storylines could explain these
results. One explanatory hypothesis is based on the recovery of perception
corresponding to the central- side FoV of the loss half-retina. This hypothesis assumes
cerebral neuroplasticity processes induced in the patient, which enable recovering of
alternative visual processing pathways. One obvious way to test this guess would be
designing future studies and recording neuroimaging of brain regions involved in
processing the target location when it is localized in positions corresponding to the loss
FoV. Indeed, by fMRI (Functional Magnetic Resonance Images) or PET (Positron
Emission Tomography), recorded prior to and after the treatment, during the execution
of the task, they would allow confirm or ruling out any neural activity in the visual
pathways in the brain. Alternatively, there is another more plausible explanation based
on the spatial reconstruction of the visual scene. This reconstruction would operate
monocularly and would produce some recovery in FoV, which could be explained by
the action of a selective attentional filter, that would be temporarily applied to each of
the two superimposed images that the attached prisms send to the healthy hemiretina.
As a result of treatment (prisms attached in stripes to crystals, Palomar-Mascaro, et al.,
2011a & b) in the retina of the subject two overlapping images are projected. One of
them corresponds to their usual field of registration. The other is overlapped on the first,
due to the diversion of the image that the prisms cause and corresponding visual field
damage. That is why when starting treatment patients reported experiencing spatial
displacement in the restored half-field, not referring any diplopia or confusion. After a
process of sensory adaptation, the patients must learn to separately and sequentially pay
attention to each of these two images. In our case, the improved performance of the
patients in the location task by using a device pointer does not seems this enhancement
be due to an oculomotor use of compensatory strategies. But rather it can be attributed
to some learning by selective access to each of the two images projected on healthy

14
hemiretina and further processing of these images, allowing a reconstruction of the
frontal visual space. We think that the gain in the amplitude of the loss FoV occurred
because to the patient took some practice in the visual selection (attentional) from each
of the two overlapping images, although separable. Attentive processing is voluntary at
the beginning and automatically after some time. The increase in the ability to separate
the two superimposed images on one retina, after a period of adaptation and practical,
allows patients to reconstruct an image with greater horizontal extension, integrating the
two overlapping images.
This perceptual grouping has previously been shown with images of binocular rivalry
by Kovács et collaborators (Kovács, Papathomas, Yang, Fehér, 1996). They used pairs
of dissimilar images with incoherent patterns, breaking the coherency of conventional
stimuli and replacing them by complementary patchworks of intermingled rival images.
In this way they showed that brain unscramble the pieces of the patchwork arriving
from different eyes to obtain coherent percepts. Indeed, the pattern coherency in itself
can drive perceptual alternations, and the patchworks are reassembled into coherent
forms by most observers. Certainly, the role of retinal disparity and diplopia has been
considered from long time (e.g., Wheatstone, 1938, Mitchell, 1966; Howard & Rogers,
2002). However, very little attention has been paid to ‘overlapped images’ in scientific
studies. An exception is Pepperell & Anja Ruschkowski (2013) who tested whether the
inclusion of double images in two-dimensional pictures could enhance the illusion of
three-dimensional space. They concluded that double images could be added to the list
of depth cues available to the list of pictorial depth cues.
To conclude this study has enabled underlying principles to be identified more clearly
an explanation for the recovery of the visual field in patients affected by HHC. We
advocate that attention and automatic processing of images in retina is mainly the
mechanism that supports such rehabilitation. Further research is required in order to
demonstrate the ability to dissociate automatically two images overlapped in retina.

Acknowledgements
This work was supported by a grant from the Spanish Ministry of Economy and
Competitiveness (MECOM; Ref. PSI-2012-35194).

15