This study examined the contribution of visual field defect to the so-called "word-length effect" often described in pure alexia (alexia without agraphia), which is usually, but not always, due to an infarction in the domain of the posterior cerebral artery.
One purported characteristic of the reading behavior of people with alexia is that the time required to read a word is a function of the number of letters in the word, also known as the "word-length effect." Given the frequent association of right hemianopia with pure alexia, it is uncertain whether and how much of the word-length effect may be attributable to the hemifield loss.
This study, which involved simulating hemianopia in healthy subjects, found a correlation between mean reading time and the slope of the word-length effect in hemianopic conditions. The findings clarify the magnitude of the word-length effect that originates from hemianopia alone and show that the criteria for a word-length effect indicative of alexia differ according to the degree of associated hemifield loss.
To isolate the contribution of the visual field defect, the authors simulated hemianopia in 13 healthy subjects with a gaze-contingent paradigm during an eye-tracking experiment.
They found a minimal word-length effect of 14 ms/letter for full-field viewing, which increased to 38 ms/letter in right hemianopia and to 31 ms/letter in left hemianopia. The 95 percent upper prediction limits for the word-length effect were 51 ms/letter in subjects with full visual fields and 161 ms/letter with simulated right hemianopia. These limits, which can be considered diagnostic criteria for an alexic word-length effect, were consistent with the reading performance of six patients with diagnoses based independently on perimetric and imaging data: two patients with probable hemianopic dyslexia, and four with alexia and lesions of the left fusiform gyrus, two with and two without hemianopia. Two of these patients also showed reduction of the word-length effect over months, one with and one without a reading rehabilitation program.
The authors say that in addition to quantifying the word-length effect for hemianopia, their data for virtual hemianopia yielded a number of interesting findings. First, equivalent effects were found for both right and left hemianopia. Second, response times were correlated between full-field viewing and hemianopic viewing, right or left. Third, under both virtual hemianopic conditions there was a positive correlation between mean response time and word-length effect, with subjects who took longer to read doing so particularly for longer words. This indicates that mean response time and the word-length effect are not independent effects.
The study's data also suggest that word-length effects of more than 160 ms/letter in hemianopic subjects indicate that they have an additional factor impairing their reading aside from their field defect.