Reading strategies in Stargardt's disease with foveal sparing

Background Subjects with a ring scotoma can use two retinal loci, a foveal and a peripheral, for reading. Our aim was to investigate the relative use of both retinal loci as a function of the spared foveal area size and the spatial resolution at both retinal loci. Findings Two patients with Stargardt's disease and ring scotomas read through a scanning laser ophthalmoscope a series of letters and words at various character sizes. The number of fixations made using each retinal locus was quantified. The relative use of each retinal locus depended on character size of the stimulus. Both patients used exclusively the eccentric retinal locus to read words of large character sizes. At small character sizes, the central retinal locus was predominantly used. For reading letters or words, once foveal fixation was used, patients did not shift back to the eccentric retinal locus. When spatial resolution allowed deciphering at both the eccentric and the central areas, patients consistently fixated with the eccentric retinal locus. Conclusions Spatial resolution at the eccentric locus appears as a determinant factor to select the retinal area for reading. Reading strategies in patients with Stargardt's disease and a ring scotoma demonstrate a pattern of coordination of both eccentric and central retinal loci, reflecting a high degree of adaptation.


Background
Central scotomas are frequently encountered in low vision populations and induce an impediment to efficient reading. Common causes of central scotomas include age-related macular degeneration (AMD) and Stargardt's disease. Patients with central scotoma use an extrafoveal locus for fixation in place of the fovea [1]. Such a locus is commonly referred to as preferred retinal locus (PRL) [2]. Some patients affected by macular degeneration develop a ring scotoma. Such a condition is of particular interest with regard to its functional implications. As previously reported [3][4][5], subjects with a ring scotoma fixate small stimuli with the central locus, but wider stimuli with the eccentric locus. The alternation between central and eccentric fixations was found in 8 out of 40 eyes of patients with Stargart's disease [6], and it was suggested that this fixation pattern was a phase in the transition from foveal to eccentric fixation [7].
In the present study we investigated reading strategies in patients with Stargardt's disease and a ring scotoma. Our purpose was to assess the relative use of two PRLs, a foveal and a peripheral, as a function of the spared foveal area size and of the spatial resolution at both PRLs. The results showed that reading strategies strongly depended on the spatial resolution at the eccentric retinal locus.

Patients
Two female patients (CC and CN) with bilateral ring scotomas following Stargardt's disease were studied. They were aged 28 (CC) and 36 years (CN). The time since onset of the macular disease was 5 and 2 years respectively. None of the patients had received either low vision rehabilitation or optical magnifiers before data collection. Patients gave their informed consent to the testing procedures that were in accordance with the Declaration of Helsinki.

Experimental set-up and testing procedures
Investigations were conducted on the eye showing the best acuity, according to a method previously described elsewhere [8]. Briefly, the examination procedure consisted in three sessions. In the first session, visual acuity was determined for both eyes with ETDRS charts. To delineate the scotoma, a microperimetry was performed within the central 15 degrees of the retina using the Rodenstock scanning laser ophthalmoscope (SLO), by means of Goldman III targets of 200 cd/m 2 in intensity ( Figure 1). In the second session, patients were asked to read aloud through the SLO single letters and isolated words of eight different character sizes, decreasing by 0.1 logMAR steps, from 1.5 to 0.8 logMAR. A total of 32 isolated letters and words of 2, 5, and 10 letters were presented at random locations and in a random order within the field of view of the instrument. During the reading task, images of the ocular fundus and the superimposed word stimuli were recorded onto videotapes at a frequency of 25 frames per second. In the third session, the spatial resolution at the central and the eccentric PRLs was measured using the Visumetry program of the SLO.

Data analysis
The locations of the central and the eccentric PRLs were determined after the testing, with a computer program developed in our laboratory [9]. Images of the ocular fundus and the superimposed word stimuli were digitised and converted into schematic representations. Eye movements were automatically extracted by comparing the shift of the eye fundus position relative to the word as it is read. Sequences of eye movements during reading were represented as "cartoons" of the ocular fundus and the superimposed word ( Figure 2). The PRLs are defined as being the retinal areas consistently used to fixate and read letters and words.
After locating the PRLs, the number of fixations was manually quantified, through frame-by-frame playback of the video recording. A fixation was defined as a stable retinal position of at least 3 successive frames, i.e., 120 ms. Reading strategies were assessed by the proportions of fixations performed with either the eccentric or the central PRL. The proportions of fixations were calculated for each subject and each letter or word. To determine the relative use of both PRLs as a function of character size and words length, the coefficient PC ) was computed from the number of fixations made with each PRL.

Patients' clinical characteristics
Both patients used two distinct PRLs for reading, one central and the other eccentric. The central PRL was at or close to the fovea. The eccentric PRL was always located at the same area on the retina, above the macular lesion, i.e., below the scotoma in the visual field. The central and the eccentric PRLs were separated by approximately 5 degrees. Patients' clinical characteristics are shown in Table 1. As the spared central area size was 0.75°and 1.3°in diameter for patients CC and CN respectively, it could fit letters up to 1.1 logMAR (CC) and 1.3 logMAR (CN). At the smallest character size   Spatial resolution at the eccentric PRL, as determined by the visumetry test, was 1.2 logMAR and 1.0 logMAR for patients CC and CN respectively. As could be expected from the EDTRS tests, the spatial resolution at the central PRL was better than 0.8 logMAR for both patients, which is the SLO's program spatial resolution limit.

Reading strategies
Both patients used similar reading strategies. At the appearance of the word, fixations were made with one or several eccentric loci of the retina, before the word be fixated with the PRL (Figure 3 and Figure 4). The following fixations were performed using the eccentric or the central PRL. Once starting to read with the central PRL, they did not shift back to the eccentric PRL. Occasionally, stimuli were transiently shifted into the scotoma area during reading.

Discussion
In both patients with Stargardt's disease and a ring scotoma reading strategies involved the combined use of  both the eccentric PRL and the central PRL. The relative use of each PRL depended on character size. For reading words with large character sizes, patients used exclusively the eccentric PRL. For smaller character sizes, patients combined benefits of the eccentric PRL, i.e. a wider visual span and presumably a better spatial word localisation, with those of the preserved foveal zone, i.e. a higher spatial resolution. The use of multiple PRLs has been attributed to several factors. Whittaker [10] noticed that multiple PRLs were more likely to occur with large central scotoma above 20°, while Crossland [11] reported higher occurrence in patients with recent vision loss (<12 weeks).
The selection and the location of the PRLs also depend on the task requirements, including visual acuity, visual span, stimulus and background luminance, correspondence in binocular vision, attention load [12][13][14][15][16]. Alternation between multiple PRLs with complementary functional advantages might improve performance of the visual task.
Our findings emphasised the importance of the spatial resolution at the eccentric PRL as one of the determinant factor in selecting a fixation locus for reading. Interestingly, when characters could be deciphered in both PRLs, patients consistently used the eccentric locus. Although patients had a better spatial resolution at the central than at the eccentric PRL, they did not shift fixation to the central PRL as soon as one letter could fit within it, but only when character size was below the resolution of the eccentric PRL.
In contrast to character size, word length did not appear to influence the relative use of the central or the eccentric PRL. Although the central area could accommodate more than 1 or 2 letters, patients selected the eccentric PRL as long as its spatial resolution enabled reading.
Another interesting finding was that once foveal fixation was used, patients did not shift back to the eccentric PRL. This phenomenon can be attributed to the fact that using the eccentric PRL implies the development of new oculomotor skills, involving shift of the oculomotor reference from the fovea to a non foveal locus [17]. Subsequently, eye movements need to be recoded in another coordinate system as soon as changes in PRLs occur, which is a constraining task, requiring a major plastic reorganisation in the oculomotor control.

Conclusions
Reading strategies in patients with Stargardt's disease and a ring scotoma demonstrate a pattern of coordination of both eccentric and central PRLs. The spatial resolution at the eccentric PRL was a determinant factor in selecting a fixation locus. When spatial resolution allowed deciphering in both eccentric and central areas, patients consistently fixated with the eccentric PRL. More studies with greater patients' samples are needed to confirm these preliminary results.