Statistics from Altmetric.com
Barton and Black indicated in a recent report that hemianopic subjects without unilateral neglect may demonstrate a horizontal line bisection bias towards the scotomatous field.1 Their study was prompted by their impression that the influence of visual field defects on spatial tasks has been little studied. The authors unfortunately seem not to have been aware of studies by Kerkhoffet al 2-4 that not only replicated Barton and Black’s findings but indicate that such findings have been recorded as long ago as 1894. It would be useful if Barton and Black could indicate whether any aspects of their study are distinguished from the preceding reports.
The authors reply:
Mark is right in noting that Kerkhoffet al have mentioned this phenomenon of contralateral bisection bias in hemianopia. We will first examine the data provided by Kerkhoff et al, then peruse the literature he quotes, to answer Mark’s question.
Firstly, two of the three papers quoted by Mark do not provide data on hemianopic bisection bias. In one,1-1 the phenomenon is mentioned in a single sentence in the methods section; in the other,1-2 a German review contrasting hemianopia and hemineglect published last year, there is a short discussion about bisection without experimental data. The third paper studied bisection of vertical or horizontal lines in a series of six patients1-3: however, all had bilateral cerebral damage and complex bilateral field defects, not hemianopia. Extrapolating from such unusual patients to those with more common unilateral hemianopia is not always appropriate.
Kerkhoff et al state that others have made this observation before.1-1-1-3 He quotes two works from the English literature. One is the paper by Schenkenberget al.1-4 This report analysed by side of brain damage without respect to hemianopia. Furthermore, hemianopic bias cannot be deduced from their tabulated data: there is bisection data for only six patients with hemifield defects without neglect, five with left hemispheric lesions, and the type of field defect is not stated. An informal t test on the small sample of five patients with left hemispheric lesions shows no difference from the controls.
Another citation is a monograph by Teuber et al.1-5 This was also puzzling, because line bisection did not seem to be reported in the methods or results. We believe that Kerkhoff et al may have been referring to patients with lateral or vertical shift of fixation (“pseudo-fovea”, p 79). It does not necessarily follow that a shift of fixation leads to a shift in bisection judgments, though. Teuberet al do state that this might occur (p 82), and refer to another work of theirs,1-6 a case study of a man with a right occipital wound and left hemianopia. However, his bisection errors were biased rightward, as with neglect, rather than leftward, as we find with hemianopia.
Kerkhoff et al are more accurate in noting precedents in the older German literature. We became aware of these when we had a German monograph of Werth translated recently.1-7 Axenfeld1-8 does seem to be the first to make this finding, in his report of a man with left hemianopia and leftward bisection errors. Apart from precedence, though, few conclusions are permitted from a single case study without controls.
The most important work we discovered is that by Liepmann and Kalmus.1-9 These authors documented contralateral bisection bias in 10 patients with homonymous hemianopia and one with bitemporal hemianopia. Interestingly, they noted that this bias was not present with very small or very large lines. They dismissed Axenfeld’s hypothesis that the error was related to a coexistent gaze palsy,1-8 but speculated that it was related to increased numbers of saccades and fixations into the blind hemifield, an explanation which resembles the idea that the bias may be secondary to an adaptive shift of attention. (Note, though, that we and others have provided evidence that the spatial representation within normal hemifields can itself generate this type of bias.1-7 1-10)
We appreciate the opportunity to publicise the prior observations of Axenfeld1-8 and Liepmann and Kalmus,1-9 and the prescience of the latter’s speculations. Among our colleagues, “hemianopic bisection error” does not seem to be well known. We would also suggest that its scientific foundation in the German literature could be improved. We have provided better evidence for both the existence of the bias and the conclusion that it is caused by hemianopia, with the use of statistical analysis, normal controls, and non-hemianopic brain-damaged controls, none of which were used in the older German work. We hope that our brief initial report and this correspondence increase awareness of “hemianopic bisection error”.