R828 Dispatch

Vision research: Losing sight of eye dominanceDavid P. Carey

Most people prefer to use their right eye for viewing.New evidence reveals that this dominance is much moreplastic than that for one hand or foot: it changes fromone eye to the other depending on angle of gaze.Remarkably, sighting dominance depends on the handbeing directed towards the visual target.

Address: Neuropsychology Research Group, Department ofPsychology, University of Aberdeen, Kings College, Old AberdeenAB24 2UB, UK. E-mail:d.carey@abdn.ac.uk

Current Biology 2001, 11:R828–R830

0960-9822/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved.

Most people are perfectly aware of their preferred hand forskilled activities like handwriting, and many are also awareof their preferred foot for activities like kicking a ball orwriting with their toe in sand. A lesser-known fact is thatwe also have a favoured eye for sighting. In our species,but no other, these biases are all rightwards: handedness~90%, footedness ~80% and ‘eyedness’ ~70%. Much likethree tosses of a two-headed coin, it is possible that thesethree population-biases are all rightwards by chance.Nevertheless, these asymmetries have been linked to thespecialisation of the left hemisphere of the brain forlanguage, speech and motor control [1,2].

Because sighting dominance is not perfectly correlatedwith handedness — or indeed with other types of oculardominance such as binocular rivalry — interest in sightingdominance has waned. Adding further insult to injury,experiments by Ono and colleagues [3,4] suggested thatsighting dominance might be an artefact, the ‘dominant’eye being just the one closest to the perceived egocentre,which does not perfectly coincide with the body midlinein most people. However, such arguments do not explainwhy the egocentre is biased to the right more than the leftin most people.

A recent study [5] may ultimately lead to answers to thislatter question. In most demonstrations of sighting domi-nance, participants have to align a target in peripersonalspace with a more distant point while looking straightahead. Khan and Crawford [5] turned this typical situationon its head, by examining sighting as a function of eccen-tricity of gaze. In their two experiments, participants wererequired to fixate a small target through a ring placed atarm’s length from their face. They then reached out andgrasped the ring and drew it rapidly towards their facewhile keeping the distant target within the ring’s aperture.

This procedure (in central vision) results in the partici-pants drawing the circle towards the dominant eye withoutexplicit awareness of doing so [6].

In these studies [5], head position was fixed and, for thefirst time, the ring and target pairs were presented acrossthe hemispace at various eccentricities to the right and leftof the head/body midline. Their results showed the usualright-eye dominance in the majority of their participantsfor the central ring–target pair. Remarkably, they switchedeye preference as the stimuli were moved into lefthemispace (Figure 1, blue functions). Similarly, the threeleft-dominant participants (Figure 1, red functions) shiftedto right dominance when targets were presented past athreshold point in the right hemispace. (A somewhatsimilar side-related bias in determining binocularalignment in depth has been reported by Erkelens et al. [7].

Why might sighting dominance change with the directionof gaze? In the Khan and Crawford [5] study, modulationby gaze direction did not depend on occlusion of theopposite hemifield by the nose: even when targets arevisible and easily within the line of sight of the ‘typically’

Figure 1

Results from the ten participants studied by Khan and Crawford [5] (fortwo of the right-eyed participants the data overlap, so only ninefunctions are shown). The target-ring position, and thus gaze direction,in left and right hemispace is shown across the X-axis. Data pointsabove the 50% line indicate right-eyed sighting. The percentagenumber of trials where the right eye was used appears along the Y-axis.Notice how the left dominant participants (red lines) become right-eyedwhen targets are presented 20–30° to the right of the midline, andconversely the right dominant participants (blue lines) become left-eyedwhen targets were 10–20° to the left of the body midline. (Modifiedfrom [5].)

50º 40º 30º 20º 10º 0º 10º 20º 30º 40º 50º

Gaze direction

RightLeft

Per

cent

rig

ht-e

ye d

omin

ant 100

50

0

Current Biology

Dispatch R829

dominant eye, participants shifted to their non-dominanteye at some point in the opposite hemifield. Crawford andKhan [5] themselves suggest that the superior viewafforded by the eye on the same side of space as the targetmay be responsible for the shifts of dominance from oneeye to the other.

In normal sighting tasks, of course, central vision is used toprovide equivalent opportunity for either eye to be used toalign a nearby target with a distant one. Indeed, previousinvestigators who have struggled with the definition ofsighting dominance have pointed out that using one handor the other as the near target in some sighting tasks can‘corrupt’ categorising a participant as left-eyed or right-eyed. This so-called ‘nuisance’ variable probably deservesanother look, given the peculiar results of Khan and Craw-ford’s [5] second experiment. They had participants com-plete their ‘grasp and pull’ task with one hand then theother in separate blocks of trials. Remarkably, the thresh-olds were biased by the hand used. Right-hand graspingshifted the functions shown in Figure 1 to the left: targetshad to be further in left hemispace for the participants toshift from sighting with the right eye. Similarly, left-handgrasping shifted the 50–50 thresholds to the right(Figure 2).

We should keep in mind that the story in real environ-ments is inevitably more complicated. In reaching andgrasping without constraining the head, the low inertia ofthe eyes gives a large lead time relative to the head andhand. In such movements, the head follows the eyes andthe vestibulo-ocular reflex ensures that the eyes swingback toward primary position as the hand is arriving ontarget, maintaining fixation. What might this orderingmean for the Khan and Crawford [5] story? Rememberthat participants show consistent sighting preference forone eye when their head is pointing towards the centralring and target pair. The most likely state of affairs wouldbe that peripheral targets would be initially ‘sighted’ —what this means in the absence of a peripersonal target foralignment when the hand is just starting its movement weshall leave for the time being — with the eye on that same,ipsilateral side of the body, but by the time the headmovement is completed the normal, conventionallydefined sighting eye would be aligned with the target.

The difficult problem would be identifying the sightingeye threshold within a trial, keeping in mind that thesighting eye can only be identified unambiguously whennear and far targets need to be aligned. A solution mightpresent itself by varying target position in depth, so thatvergence and versional portions of the movements of eacheye can be recorded. There is a suggestion in the literature([8,9] for example) that the non-sighting eye producesmore of the required vergence in such tasks, implying that

it is being actively realigned by vergence change with thesaccading sighting eye.

So the eye used for the ‘initial’ sighting may depend onthe hand moving towards the target, but (for contralateraltargets, relative to the dominant eye) sighting will switchto the other eye at some point in the head movement.Why switch if the non-dominant eye has already ‘cap-tured’ the target? An interesting clue about the possiblelink between eye and hand comes from the claim that theposition of the dominant eye is preferentially monitoredusing some sort of feedback signal related to eye proprio-ception [10]. The utility of eye proprioception in humanshas been doubted, given the scarcity of spindle receptorsin the ocular musculature and the generally poor percep-tion of stationary eye position in darkness (for reviews see[11,12]). Nevertheless, there is some evidence that feed-back eye position signals play some role in the recalibra-tion of sensorimotor localisation systems at the end ofsaccades [13].

The possible usefulness of calibration biased towards afavoured eye remains unknown. The early literaturesuggests that advantages in sensorimotor performancewhen using the dominant eye are typically obtained intasks that require ballistic movements of the hands, suchas target-directed aiming (for example [14], see [15] forreview). In the experiments of Khan and Crawford [5],participants initiated their movements when already fixat-ing on the specified target through the ring. A moredynamic variant of the task that would require coordination

Figure 2

Sighting eye thresholds are moderated by the hand that is reaching.The data for left-hand reaching are shown by the red line, and those forright-hand reaching are shown by the blue line. Note how thenormalised composite threshold functions are shifted away from theside of the reaching hand. (Modified from [5].)

Current Biology

Right hand

Left hand

Left Right

Gaze direction

Per

cent

rig

ht-e

ye d

omin

ant 100

0

of rapid saccadic eye movements and reaching or graspingmight exacerbate the difference between eye preferencechange thresholds as a function of the hand used.

What about sighting dominance and cerebral lateralisation?Many authors have dismissed any possible links (see [15])because the sensory inputs from the two eyes are notcrossed in the same way as those from hands and feet are. Itis the contralateral half of the retina that is ultimately seenby each hemisphere, not the contralateral eye. Such analy-ses, nevertheless are biased towards how each hemisphereprocesses sensory inputs. Is their any evidence that outputsrelated to eye movements are handled differently by thetwo hemispheres? The answer to the former question is yes,in the sense that in most eye fields in the cerebral cortex,stimulation produces eye movements in a contralateraldirection (albeit these are the same in both eyes). Unfortu-nately few neurophysiologists have been bothered withcerebral asymmetries in this type of control, as asking thesame research question in both hemispheres of the sameanimal would be at the expense of asking other questions.

In neuroimaging of human primates, the potential forstudying the lateralisation of eye movement functions isvast. Some provocative hints have already appeared whichsuggest that further studies on this topic may beworthwhile. For example, in a sample of right-handedparticipants who were making small right-hand fingermovements, gazing to the right produced greater activityincreases in the contralateral hemisphere than gazing tothe left [16]. Unfortunately, sighting dominance was notreported and left-hand movements were not assessed.Another group [17] has claimed increased left hemisphereactivation with monocular viewing (by either eye) in asample of right handers, presumably the majority of whomwould have been right-eyed.

Future studies on the neurobiology of sighting dominancemight take advantage of Khan and Crawford’s [5] thresh-old technique to quantify the degree of sighting prefer-ence rather than just its direction. As handedness can bequantifed in a number of different ways, the potential forstudying interactions between sighting and hand use hasincreased exponentially — some lateral thinking aboutsighting has pointed in a number of interesting directions,well worth keeping an eye on.

References1. Annett M: Predicting combinations of left and right asymmetries.

Cortex 2000, 36:485-505.2. Kimura D: Neuromotor Mechanisms in Human Communication.

Oxford: Oxford University Press, 1993.3. Ono H, Barbeito R: The cyclopean eye vs. the sighting-dominant

eye as the center of visual direction. Percept Psychophys 1982,32:201-210.

4. Barbeito R, Simpson TL: The relationship between eye position andegocentric visual direction. Percept Psychophys 1991, 50:373-382.

5. Khan AZ, Crawford JD: Ocular dominance reverses as a function ofhorizontal gaze angle. Vis Res 2001, 41:1743-1748.

6. Crider BA: A battery of tests for the dominant eye. J Gen Psychol1944, 31:179-191.

7. Erkelens CJ, Muijs AJM, van Ee R: Binocular alignment in differentdepth planes. Vis Res 1996, 36:2141-2147.

8. Enright JT: Monocularly programmed human saccades duringvergence changes. J Physiol 1998, 512:235-258.

9. van Leeuwen AF, Collewijn H, Erklens CJ: Dynamics of horizontalvergence movements: interaction with horizontal and verticalsaccades and relation with monocular preferences. Vis Res 1998,38:3943-3954.

10. Velay JL, Roll R, Lennerstrand G, Roll JP: Eye proprioception andvisual localization in humans: Influence of ocular dominance andvisual context. Vis Res 1994, 34:2169-2176.

11. Steinbach MJ: Proprioceptive knowledge of eye position. Vis Res 1987, 10:1737-1744.

12. Donaldson IML: The functions of the proprioceptors of the eyemuscles. Phil Trans R Soc London Ser B 2000, 355:1685-1754.

13. Li W, Matin L: Visual direction is corrected by a hybrid extraretinaleye position signal. Ann New York Acad Sci 1992, 656:865-867.

14. Lund FH: The dependence of eye-hand coordinations upon eyedominance. Am J Psychol 1932, 44:756-762.

15. Porac C, Coren S: The dominant eye. Psychol Bull 1976,83:880-897.

16. Baker JT, Donoghue JP, Sanes JN: Gaze direction modulates fingermovement activation patterns in human cerebral cortex.J Neurosci 1999, 19:10044-10052.

17. Richter H, Lee JT, Pardo JV: Cerebral dominance and eyedependence of regional cerebal blood flow (rCFB) in humanvisual cortex demonstrated by PET. Invest Ophthalmol Vis Sci1999, 40:S58.

R830 Current Biology Vol 11 No 20