Optimizing gaze control in three dimensions

Science

Volumn 281, Issue 5381, 1998, Pages 1363-1366

  Tweed, Douglas ^a,c   Haslwanter, Thomas ^b   Fetter, Michael ^b  

^a UNIVERSITY OF WESTERN ONTARIO   (Canada)

^b UNIVERSITY OF TÜBINGEN   (Germany)

^c UNIVERSITY OF TORONTO   (Canada)

Author keywords

[No Author keywords available]

Indexed keywords

ARTICLE; COMPUTER SIMULATION; EYE MOVEMENT; GAZE; HEAD MOVEMENT; PRIORITY JOURNAL; RETINA IMAGE; VISUAL FIELD;

EID: 0032575728     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.281.5381.1363     Document Type: Article

References (28)

1. H. von Helmholtz, Handbuch der physiologischen Optik (Voss, Hamburg and Leipzig, Germany, 1867).
2. P. A. Merton, J. Physiol. (Proceedings) 132, 25 (1956); E. F. Miller, Acta Oto-Laryngologica 54, 479 (1962).
3. P. A. Merton, J. Physiol. (Proceedings) 132, 25 (1956); E. F. Miller, Acta Oto-Laryngologica 54, 479 (1962).
4. L. Ferman, H. Collewijn, A. V. Van den Berg, Vision Res. 27, 929 (1987).
5. D. Tweed and T. Vilis, ibid. 30, 111 (1990); A. W. H. Minken, A. J. Van Opstal, J. A. M. Van Gisbergen, Exp. Brain Res. 93, 521 (1993); D. Tweed, H. Misslisch, M. Fetter, J. Neurophysiol. 72, 1425 (1994); D. Straumann, D. S. Zee, D. Solomon, A. G. Lasker, D. Roberts, Vision Res. 35, 3321 (1995).
6. D. Tweed and T. Vilis, ibid. 30, 111 (1990); A. W. H. Minken, A. J. Van Opstal, J. A. M. Van Gisbergen, Exp. Brain Res. 93, 521 (1993); D. Tweed, H. Misslisch, M. Fetter, J. Neurophysiol. 72, 1425 (1994); D. Straumann, D. S. Zee, D. Solomon, A. G. Lasker, D. Roberts, Vision Res. 35, 3321 (1995).
7. D. Tweed and T. Vilis, ibid. 30, 111 (1990); A. W. H. Minken, A. J. Van Opstal, J. A. M. Van Gisbergen, Exp. Brain Res. 93, 521 (1993); D. Tweed, H. Misslisch, M. Fetter, J. Neurophysiol. 72, 1425 (1994); D. Straumann, D. S. Zee, D. Solomon, A. G. Lasker, D. Roberts, Vision Res. 35, 3321 (1995).
8. D. Tweed and T. Vilis, ibid. 30, 111 (1990); A. W. H. Minken, A. J. Van Opstal, J. A. M. Van Gisbergen, Exp. Brain Res. 93, 521 (1993); D. Tweed, H. Misslisch, M. Fetter, J. Neurophysiol. 72, 1425 (1994); D. Straumann, D. S. Zee, D. Solomon, A. G. Lasker, D. Roberts, Vision Res. 35, 3321 (1995).
9. A. Roucoux, M. Crommelinck, J. M. Guerit, M. Meulders, in Progress in Oculomotor Research, A. F. Fuchs and W. Becker, Eds. (Elsevier, Amsterdam, 1980), pp. 309-315; V. P. Laurutis and D. A. Robinson, J. Physiol. London 373, 209 (1986).
10. A. Roucoux, M. Crommelinck, J. M. Guerit, M. Meulders, in Progress in Oculomotor Research, A. F. Fuchs and W. Becker, Eds. (Elsevier, Amsterdam, 1980), pp. 309-315; V. P. Laurutis and D. A. Robinson, J. Physiol. London 373, 209 (1986).
11. D. Guitton and M. Volle, J. Neurophysiol. 58, 427 (1987); D. Tweed, B. Glenn, T. Vilis, ibid. 73, 766 (1995).
12. D. Guitton and M. Volle, J. Neurophysiol. 58, 427 (1987); D. Tweed, B. Glenn, T. Vilis, ibid. 73, 766 (1995).
13. Our four participants were healthy, aged 30 to 38, and gave informed consent. They were instructed to look to the target spot using a fast combined motion of eyes and head, with the head ending up 30° CW or 30° CCW. Similar results were obtained for torsional gaze shifts while looking up and for "natural" gaze shifts, in which the person was given no instructions about final head position but simply looked between four targets - up-right, down-right, down-left, and up-left - at 90° eccentricity. These latter gaze shifts were made in pseudorandom order and followed spoken commands.
14. The position of the left eye was recorded at 375 samples per second with the scleral-search-coil technique devised by D. A. Robinson [IEEE Trans. Biomed. Electron. BME-10, 137 (1963)]. Head position was recorded at the same rate with the use of search coils fastened to a lightweight bar molded to fit the person's teeth. Electrical signals from the eye and head search coils were converted to quaternions with the use of algorithms from D. Tweed, W. Cadera, and T. Vilis [Vision Res. 30, 97 (1990)]. For an oculomotor introduction to quaternions, see G. Westheimer, J. Opt. Soc. Am. 47, 967(1957); D. Tweed and T. Vilis, J. Neurophysiol. 58, 832 (1987); or T. Haslwanter, Vision Res. 35, 1727 (1995).
15. The position of the left eye was recorded at 375 samples per second with the scleral-search-coil technique devised by D. A. Robinson [IEEE Trans. Biomed. Electron. BME-10, 137 (1963)]. Head position was recorded at the same rate with the use of search coils fastened to a lightweight bar molded to fit the person's teeth. Electrical signals from the eye and head search coils were converted to quaternions with the use of algorithms from D. Tweed, W. Cadera, and T. Vilis [Vision Res. 30, 97 (1990)]. For an oculomotor introduction to quaternions, see G. Westheimer, J. Opt. Soc. Am. 47, 967(1957); D. Tweed and T. Vilis, J. Neurophysiol. 58, 832 (1987); or T. Haslwanter, Vision Res. 35, 1727 (1995).
16. The position of the left eye was recorded at 375 samples per second with the scleral-search-coil technique devised by D. A. Robinson [IEEE Trans. Biomed. Electron. BME-10, 137 (1963)]. Head position was recorded at the same rate with the use of search coils fastened to a lightweight bar molded to fit the person's teeth. Electrical signals from the eye and head search coils were converted to quaternions with the use of algorithms from D. Tweed, W. Cadera, and T. Vilis [Vision Res. 30, 97 (1990)]. For an oculomotor introduction to quaternions, see G. Westheimer, J. Opt. Soc. Am. 47, 967(1957); D. Tweed and T. Vilis, J. Neurophysiol. 58, 832 (1987); or T. Haslwanter, Vision Res. 35, 1727 (1995).
17. The position of the left eye was recorded at 375 samples per second with the scleral-search-coil technique devised by D. A. Robinson [IEEE Trans. Biomed. Electron. BME-10, 137 (1963)]. Head position was recorded at the same rate with the use of search coils fastened to a lightweight bar molded to fit the person's teeth. Electrical signals from the eye and head search coils were converted to quaternions with the use of algorithms from D. Tweed, W. Cadera, and T. Vilis [Vision Res. 30, 97 (1990)]. For an oculomotor introduction to quaternions, see G. Westheimer, J. Opt. Soc. Am. 47, 967(1957); D. Tweed and T. Vilis, J. Neurophysiol. 58, 832 (1987); or T. Haslwanter, Vision Res. 35, 1727 (1995).
18. The position of the left eye was recorded at 375 samples per second with the scleral-search-coil technique devised by D. A. Robinson [IEEE Trans. Biomed. Electron. BME-10, 137 (1963)]. Head position was recorded at the same rate with the use of search coils fastened to a lightweight bar molded to fit the person's teeth. Electrical signals from the eye and head search coils were converted to quaternions with the use of algorithms from D. Tweed, W. Cadera, and T. Vilis [Vision Res. 30, 97 (1990)]. For an oculomotor introduction to quaternions, see G. Westheimer, J. Opt. Soc. Am. 47, 967(1957); D. Tweed and T. Vilis, J. Neurophysiol. 58, 832 (1987); or T. Haslwanter, Vision Res. 35, 1727 (1995).
19. Our best performing participant showed a range of 16° CCW to 18° CW, tying the world record for human ocular torsion. A. P. Petrov and G. M. Zenkin [Vision Res. 13, 2465 (1973)] also had a study participant who reached 18°, but by a different mechanism: Their participant was rotated torsionally in the dark; the eye movements were not gaze shifts to visual targets but vestibular responses to the head rotation, and they opposed the head's motion; whereas in our study the eye rolled the same way as the head and before it. R. Balliet and K. Nakayama [Invest. Ophthalmol. 17, 303 (1978)] trained people to rotate their eyes torsionally while keeping their heads still. The training, which involved torsional visual feedback, allowed them to increase their torsional range by about 0.8°/hour, so that after 35 hours of training, the best performing person could move over a range of 26.5°. That result showed that the oculomotor system has torsional plasticity, whereas the present findings show that circuits for torsional gaze control are active in normal untrained humans. Our participants practiced the gaze task in Fig. 1 at most five times (that is, for only a few seconds) before recording began, and they showed the full torsional range of eye motion, or very close to it, on the first recorded trial.
20. Our best performing participant showed a range of 16° CCW to 18° CW, tying the world record for human ocular torsion. A. P. Petrov and G. M. Zenkin [Vision Res. 13, 2465 (1973)] also had a study participant who reached 18°, but by a different mechanism: Their participant was rotated torsionally in the dark; the eye movements were not gaze shifts to visual targets but vestibular responses to the head rotation, and they opposed the head's motion; whereas in our study the eye rolled the same way as the head and before it. R. Balliet and K. Nakayama [Invest. Ophthalmol. 17, 303 (1978)] trained people to rotate their eyes torsionally while keeping their heads still. The training, which involved torsional visual feedback, allowed them to increase their torsional range by about 0.8°/hour, so that after 35 hours of training, the best performing person could move over a range of 26.5°. That result showed that the oculomotor system has torsional plasticity, whereas the present findings show that circuits for torsional gaze control are active in normal untrained humans. Our participants practiced the gaze task in Fig. 1 at most five times (that is, for only a few seconds) before recording began, and they showed the full torsional range of eye motion, or very close to it, on the first recorded trial.
21. The eye's return motion, when it came back toward Listing's plane while staying locked on the visual target, likely was vestibularly driven; but the initial outbound torsion, which anticipated the upcoming head turn, was not. In the outbound stage, the three-dimensional gaze system rotated the eye to its desired position in space and then locked on by activating the vestibuloocular reflex. This reflex uses vestibular information to monitor head velocity and counterrotate the eye, holding it steady in space as the head completes its motion (5, 6).
22. P. Radau, D. Tweed, T. Vilis J. Neurophysiol. 72, 2840 (1994).
23. For model equations and predictions see D. Tweed, ibid. 77, 654 (1997). In this model, eye and head are driven by two separate feedback loops to goal positions defined in space-fixed coordinates.
24. The loss of acuity caused by horizontal and vertical image slip depends on the spatial frequency of the image, which means that regions with a lot of fine detail are usually degraded more than coarse patches; see R. H. S. Carpenter, The Movements of the Eyes (Pion, London, 1988). Something analogous likely holds for the effects of torsional image slip on orientation detectors; radially oriented contours are degraded more than circumferential contours, so that the spokes of a spinning wheel are hard to make out, while the rim is clear.
25. J.-R. Duhamel, C. Colby, and M. Goldberg [Science 255, 90 (1992)] reported that some visual cells in the parietal cortex adjust their receptive fields in anticipation of gaze shifts. Such cells will, for example, shift their receptive fields to the right immediately before rightward eye movements. Based on our results, one might also expect to find cells that adjust their receptive fields in the torsional dimension. For instance, an orientation-sensitive cell that normally responds best to horizontal lines may prefer CW-rotated lines in the final milliseconds before a CW gaze shift.
26. Advantages of rezeroing torsion are that the eye muscles probably require less force to hold an eye position near Listing's plane and that the eye, at the center of its torsional range, is optimally placed for the next gaze shift, which may go either CW or CCW [see K. Hepp. Vision Res. 35, 3237 (1995) and D. Tweed, ibid. 37, 1939 (1997)].
27. Advantages of rezeroing torsion are that the eye muscles probably require less force to hold an eye position near Listing's plane and that the eye, at the center of its torsional range, is optimally placed for the next gaze shift, which may go either CW or CCW [see K. Hepp. Vision Res. 35, 3237 (1995) and D. Tweed, ibid. 37, 1939 (1997)].
28. We thank J. Dichgans, S. Ferber, V. Happe, C. Hawkins, M. Niemeier, and T. Vilis for comments on the manuscript. This work was supported by the Medical Research Council of Canada (grant MT-12847) and the Deutsche Forschungsgemeinschaft (grant SFB 307-A10).