The Journal of Neurophysiology Vol. 80 No. 5 November 1998, pp. 2785-2789
Copyright ©1998 by the American Physiological Society

RAPID COMMUNICATION

Human Head-Free Gaze Saccades to Targets Flashed Before Gaze-Pursuit Are Spatially Accurate

Troy M. Herter and Daniel Guitton

Montreal Neurological Institute and Department of Neurology and Neurosurgery, McGill University, Montreal, Quebec H3A 2B4, Canada

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Herter, Troy M. and Daniel Guitton. Human head-free gaze saccades to targets flashed before gaze-pursuit are spatially accurate. J. Neurophysiol. 80: 2785-2789, 1998. Previous studies have shown that accurate saccades can be generated, in the dark, that compensate for movements of the visual axis that result from movements of either the eyes alone or the head alone that intervene between target presentation and saccade onset. We have carried out experiments with human subjects to test whether gaze saccades (gaze = eye-in-space = eye-in-head + head-in-space) can be generated that compensate for smooth pursuit movements of gaze that intervene between target onset and gaze-saccade onset. In both head-unrestrained (head-free) and -restrained (head-fixed) conditions, subjects were asked to make gaze shifts, in the dark, to the remembered location of a briefly flashed target. On most trials, during the memory period, the subjects carried out intervening head-free gaze pursuit or head-fixed ocular pursuit along the horizontal meridian. On the remaining (control) trials, subjects did not carry out intervening pursuit movements during the memory period; this was the classical memory-guided saccade task. We found that the subjects accurately compensated for intervening movements of the visual axis in both the head-free and head-fixed conditions. We conclude that the human gaze-motor system is able to monitor on-line changes in gaze position and add them to initial retinal error, to program spatially accurate gaze saccades.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

The programming of saccadic eye movements to visual targets involves a potentially simple sensory-motor transformation. Retinocentric models state that the direction and amplitude of saccades are specified strictly by the "retinal error," i.e., the location of a target on the retina relative to the fovea (Schiller and Koerner 1971). However, retinocentric models do not explain how accurate saccades can be programmed to compensate, in the absence of any new visual information, for smooth or saccadic eye movements that occur between the brief presentation of a target and the subsequent targeting saccade (Gellman and Fletcher 1992; Hallett and Lightstone 1976; McKenzie and Lisberger 1986; Ohtsuka 1994; Schiller and Sandell 1983; Schlag et al. 1990; Schlag-Rey et al. 1989; Sparks and Mays 1983; Zivotofsky et al. 1996). Therefore it is believed that the oculomotor system has on-line access to extraretinal eye displacement information that can be added to initial retinal error to yield accurate target position relative to the eyes.

Accurate saccades can also be generated that compensate for movements of the visual axis that result from displacements of the head-in-space with no concurrent movements of the eyes relative to the head (Bloomberg et al. 1988; Israel and Berthoz 1989; Israel et al. 1993; Segal and Katsarkas 1988). Thus it is believed that the oculomotor system has on-line access to vestibular information concerning angular and linear displacements of the head, which can be added to initial retinal error to yield accurate target position relative to the eyes.

Coordinated movements of the eyes and head (gaze shifts) are routinely used to displace the visual axis relative to space. In cats, the gaze-motor system is able to generate accurate gaze saccades, in the dark, that compensate for intervening gaze saccades produced by microstimulation of the superior colliculus (Pelisson et al. 1989). By comparison, it is unknown how well the human gaze-motor system can program gaze saccades that compensate for naturally occurring intervening gaze displacements that necessitate updating of an initial retinal error. To investigate this, subjects were required to make gaze saccades, in the dark, to the remembered location of a target that was briefly flashed prior to gaze pursuit. All subjects consistently generated gaze saccades that accurately compensated for intervening gaze displacements. It is concluded that the human gaze-motor system has accurate on-line access to information regarding coordinated displacements of both the eyes and head.

	METHODS

Abstract Introduction Methods Results Discussion References

Protocols were approved by the Montreal Neurological Institute and Hospital Research Ethics Committee, and all subjects gave informed and voluntary consent before commencement of experimentation. Experiments were conducted with one experienced (TH) and two naive subjects (CD and MS), with an average age of 27. Subjects were seated facing a cylindrical screen located at a constant distance of 55 cm along the horizontal meridian. Eye position relative to head was recorded with the use of bitemporal DC electrooculography (EOG).¹ Head position relative to space was recorded with the use of the magnetic search coil technique. Gaze position, i.e., the position of the visual axis relative to space, was constructed by adding the eye and head position signals.

The experimental task was a variation of the flash-pursuit paradigm used previously to study eye displacement signals within the oculomotor system (Gellman and Fletcher 1992; McKenzie and Lisberger 1986; Ohtsuka 1994; Schlag et al. 1990; Zivotofsky et al. 1996). All subjects performed two sets of trials, one head-free and the other head-fixed. As shown at the top of each panel in Fig. 1, each trial began with the presentation of a central fixation target (FT) in an otherwise totally dark room. One thousand milliseconds later, a second target (T) was presented for 50 ms at 1 of 12 randomly chosen positions between +30 and 30° along the horizontal meridian. T was always presented at the same level as FT in the vertical plane such that all eye and head movements were predominately horizontal. After another 400-800 ms, FT was moved to either the left or the right at a constant velocity of 15°/s for a displacement of 15, 20, 25, or 30°. The direction and distance that FT moved were randomly chosen such that subjects could not predict the displacement of FT relative to the remembered location of T. Subjects were instructed to maintain their gaze on FT until it was extinguished; thus subjects were required to carry out gaze pursuit to maintain gaze on FT. The disappearance of FT signaled the subjects to make a targeting gaze shift to the remembered location of T. For each trial, 2,000 ms after FT was extinguished, an overhead fluorescent light, with an inherently fast decay time, was illuminated, indicating the end of the trial. The overhead light remained on for 3,000 ms (~1/3 of the time of each trial) and was extinguished 1,000 ms before the reappearance of FT at the beginning of the next trial. The timing of overhead lighting provided sufficient time for pupillary light accommodation and insufficient time for significant retinal dark adaptation, thereby minimizing changes in retinal potentials and preventing EOG drift.

View larger version (30K): [in this window] [in a new window]   FIG. 1. Raw data traces of flash-pursuit trials of subject CD. A and B: head-fixed trials. C and D: head-free trials. A and C: trials in which the spatial and retinotopic amplitudes are in the same direction but of highly different magnitudes. B and D: trials in which the spatial and retinotopic amplitudes are in opposite directions. The timing of the onset and offset of the fixation target (FT) and saccade target (T) are indicated at the top of each panel. Solid area indicates when FT is stationary. Hashed area indicates when FT is moving. Position of T during its brief presentation of 50 ms is circled in each panel. ···, remembered spatial location of T following flash presentation. -·-·-, remembered retinotopic location of T at gaze-shift onset. - - - - -, FT position. A and B: , eye position. C and D: , gaze position. - - -, head position.  , eye position.

A major difference between our flash-pursuit paradigm and all previously used flash-pursuit paradigms is that, in our experimental task, T was presented before pursuit onset. We chose to present the target before pursuit onset to maximize the amount of gaze displacement intervening between target presentation and the onset of the targeting gaze shift. This allowed for maximal dissociation of gaze-shift amplitudes predicted by retinotopic (absence of compensation) and spatial (presence of compensation) models. A clear, strong dissociation is particularly important because systematic and variable errors are normally associated with memory-guided saccades (Gnadt et al. 1991; White et al. 1994). All previous studies presented T during pursuit, which reduces the dissociation between the saccade amplitudes predicted by retinotopic and spatial models.

In addition to flash-pursuit trials, all subjects completed (control) trials in which FT remained stationary [flash trials; equivalent to the classical memory-guided saccade task of Hikosaka and Wurtz (1983)]. These trials were interleaved with the flash-pursuit trials. This control task allowed us to determine whether, in addition to those errors normally associated with memory-guided gaze shifts, errors in gaze-shift accuracy in the flash-pursuit trials were introduced by pursuing FT. Flash trials were different from flash-pursuit trials in that in the former, FT was extinguished at the point in time at which it would have normally moved. As a result, the average delay (memory period) for flash trials (600 ms) was 1,500 ms shorter than for flash-pursuit trials (2,100 ms). Also, for flash trials, T was reilluminated for a period of 2,000 ms, 2,000 ms after FT was extinguished. This provided the subject with a feedback error that was not present for flash-pursuit trials. The overhead light was illuminated once T was reextinguished.

Data were analyzed for all trials unless 1) during fixation or pursuit of FT the subject's gaze moved outside of a 5° window centered on FT; or 2) gaze-shift latency, relative to the offset of FT, was >1,000 ms. For all subjects, ~15% of trials fell into these two categories and were rejected. Approximately 15% of accepted trials had "corrective" gaze shifts, i.e., secondary gaze shifts initiated within 500 ms of the end of the initial targeting gaze shift, that on average improved accuracy. For trials with corrective gaze shifts, only the initial targeting gaze shift was considered for analysis. For all accepted trials, we measured the actual amplitude of each targeting gaze shift and compared it with two measures; the spatial and retinotopic amplitudes. The spatial amplitude was defined as the difference between the actual position where T was presented and the position of gaze at the onset of the orienting gaze shift. The retinotopic amplitude was defined as the average foveal angle subtended by T during its brief presentation.

View larger version (38K): [in this window] [in a new window]   FIG. 2. Scatter plots of flash-pursuit trials of subject CD (A and B), subject MS (C and D), and subject TH (E and F) in head-fixed (×) and head-free () conditions. A, C, and E: gaze-saccade amplitude vs. spatial amplitude. B, D, and F: gaze-saccade amplitude vs. retinotopic amplitude. Linear regression lines are indicated in panels A, C, and E for head-free (- - -) and head-fixed (···) conditions. , unity gain line.

For each subject, linear regression lines and correlation coefficients were calculated for gaze-shift amplitude versus spatial amplitude for both the flash-pursuit and flash trials in both the head-free and head-fixed conditions. Where appropriate, the slopes of the regression lines for the flash-pursuit and flash trials were compared statistically with unity slope ( = 0.05) as well as with each other. In addition, correlation coefficients were compared statistically ( = 0.05) with each other.

	RESULTS

Abstract Introduction Methods Results Discussion References

Figure 1 illustrates four representative traces of position versus time for flash-pursuit trials collected from naive subject CD. Figure 1, A and B, illustrates that humans can generate head-fixed saccades to the remembered spatial location of targets when smooth pursuit eye movements occur during the period between target presentation and saccade onset. This is in agreement with previous results obtained in both humans and monkeys (Ohtsuka 1994; Schlag et al. 1990; Zivotofsky et al. 1996). Although our results agree with these previous studies, we cannot assume that the motor command in our experiment is formulated in exactly the same manner, because we flashed T before pursuit rather than during pursuit.

Figure 1, C and D, shows that human head-free gaze saccades are also aimed at the remembered spatial location of targets, thereby compensating for intervening movements of both the eyes and head that occur during gaze pursuit. Note that, although the gaze trajectory was similar to that seen for head-fixed trials, the pattern of eye and head movements associated with acquiring and pursuing FT was quite complex. When the spatial direction and retinotopic direction were opposite each other as in B and D, >95% of gaze shifts of all three subjects were generated in the correct spatial direction in both head-fixed and head-free conditions.

Figure 2 shows six scatter plots of saccade amplitude versus either spatial amplitude (A, C, and E) or retinotopic amplitude (B, D, and F). A and B, C and D, and E and F are from subjects CD, MS, and TH, respectively. This figure confirms that, for each subject, orienting gaze shifts are normally aimed at the remembered spatial location of the target (A, C, and E) rather than at the remembered retinotopic location of the target (B, D, and F) in both the head-fixed (×) and head-free () conditions.

Table 1 gives, for each subject, the correlation coefficients and linear regression slopes linking gaze-shift amplitude to spatial amplitude for both the flash-pursuit and flash trials in the head-fixed and free conditions. First and foremost, we compared the accuracy of flash-pursuit and flash trials in head-free and head-fixed conditions in all subjects. There was no significant difference between the accuracy of flash-pursuit trials and flash trials in either the head-fixed or head-free conditions in each subject. Because the presence of intervening eye or gaze movements is the major difference between the flash-pursuit and flash trials, this indicates that errors in accuracy of flash-pursuit trials are not the result of intervening eye or gaze movements.

Second, to determine the absolute accuracy of gaze shifts for all conditions, we compared the slopes for all conditions in each subject to a unity gain slope; i.e., perfect mean accuracy. A number of slopes differed significantly from unity gain. It is evident that no one condition differed significantly from unity gain across all subjects; however, it is evident that absolute accuracy varied between subjects. Subject TH was most accurate. For all conditions, gaze-shift accuracy did not differ significantly from unity slope. Comparatively, the accuracy of MS was also generally good. On average, MS significantly undershot the target location on flash trials in the head-free condition; but accuracy was not significantly different from perfect mean accuracy in any of the other three conditions. In contrast, subject CD was less accurate than TH and MS. Except for flash-pursuit trials in the head-free condition, CD on average significantly overshot the target location.

Third, we compared the accuracy of the head-free and head-fixed conditions for flash-pursuit and flash trials in all subjects. The slope of the head-free condition was always less and tended to be more accurate than that for the head-fixed. However, the slopes of the head-free and head-fixed conditions were only significantly different for flash-pursuit trials in subject CD and flash trials in subject MS. To further investigate the significance of this difference, we pooled the data of all subjects (see Table 1). This procedure revealed that for flash-pursuit trials, head-free gaze shifts were significantly more accurate than those made head fixed. For the flash trials, the head-free gaze shifts were again more accurate than those made head fixed, but the difference was only marginally significant. This marginal significance for flash trials is quite likely due to the smaller number of data points. Note that both the flash-pursuit and flash trials were on average perfectly accurate in the head-free condition and that both differed significantly from mean perfect accuracy in the head-fixed condition.

Finally, to determine the consistency of the mean size of errors, we compared statistically the correlation coefficients for all conditions in each subject. The variability did not differ significantly for any two conditions in each subject. This indicates that the average size of the errors was always the same for all conditions.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

We have shown that head-free humans are able to consistently aim gaze saccades at the remembered spatial location of a target, although combined movements of both the eyes and head intervened between target presentation and gaze-saccade onset. Errors in flash-pursuit trials are similar to errors in flash trials where no intervening movements occurred; thus intervening gaze or eye movements are entirely compensated for by head-free gaze saccades and head-fixed ocular saccades. The head-free results are particularly remarkable given that each subject often used complex patterns of eye and head movements to pursue the moving FT.

It may be tempting to assume that head-fixed and head-free trials are physiologically similar, i.e., that gaze pursuit is identical to head-fixed pursuit, and that the vestibuloocular reflex (VOR) subtracts from the pursuit signal an amount exactly equal to what is added due to head movement. However, as the VOR gain is less than unity during gaze pursuit, such a simplifying assumption is incorrect (Cullen et al. 1991; Wellenius et al. 1997). Further evidence against this assumption is also seen in our study in that head-free gaze saccades were more accurate than head-fixed ocular saccades in our subjects.

Given that our experiments were performed with the body fixed relative to space, we cannot determine whether target position in our subjects was coded in a spatial or a body-centric reference frame. Previous experiments showing that head-fixed saccades can compensate for displacements of the whole body relative to space indicate that saccade targets may not be represented in a body-centric frame (Bloomberg et al. 1988; Israel and Berthoz 1989; Israel et al. 1993; Segal and Katsarkas 1988).

Considerable evidence indicates that during the generation of head-free gaze saccades, continuous feedback of gaze displacement is used to null gaze-motor-error, the difference between desired and actual gaze positions (reviewed by Guitton 1992). Our subjects could not predict the changes in gaze position that intervened between target presentation and gaze-shift onset during any given trial. Hence intervening changes in gaze position must have been monitored in an on-line manner. This indicates that during the planning of a head-free gaze shift, feedback of gaze displacements is used by the gaze-motor system to continuously update initial retinal error. As a result, when a gaze shift is triggered, the gaze-motor error at the start of the movement corresponds to that required to bring the visual axis onto target.

	ACKNOWLEDGEMENTS

  This research was supported by the Medical Research Council of Canada (MRC). T. Herter was the recipient of a MRC studentship.

	FOOTNOTES

¹   EOG was used because the duration of the experimental session was considerably longer than the maximum recommended time for use of the scleral search coil technique, and because we were only interested in horizontal eye movements that are accurately detected by EOG.

  Address for reprint requests: T. Herter, Montreal Neurological Institute, 3801 University St., Montreal, Quebec H3A 2B4, Canada.

  Received 9 March 1998; accepted in final form 13 July 1998.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• BLOOMBERG, J., MELVILL JONES, G., SEGAL, B., MCFARLANE, S., SOUL, J. Vestibular-contingent voluntary saccades based on cognitive estimates of remembered vestibular information. Adv. Oto-Rhino-Laryngol. 41: 71-75, 1988.[Medline]
• CULLEN, K. E., BELTON, T., MCCREA, R. A. A non-visual mechanism for voluntary cancellation of the vestibulo-ocular reflex. Exp. Brain Res. 83: 237-252, 1991.[Medline]
• GELLMAN, R. S., FLETCHER, W. A. Eye position signals in human saccadic processing. Exp. Brain Res. 89: 425-434, 1992.[Medline]
• GNADT, J. W., BRACEWELL, R. M., ANDERSEN, R. A. Sensorimotor transformation during eye movements to remembered visual targets. Vision Res. 31: 693-715, 1991.[Medline]
• GUITTON, D. Control of eye-head coordination during orienting gaze shifts. Trends Neurosci. 15: 174-179, 1992.[Medline]
• HALLETT, P. E., LIGHTSTONE, A. D. Saccadic eye movements towards stimuli triggered by prior saccades. Vision Res. 16: 99-106, 1976.[Medline]
• HIKOSAKA, O., WURTZ, R. H. Visual and oculomotor functions of monkey substantia nigra pars reticulata. III. Memory-contingent visual and saccade responses. J. Neurophysiol. 49: 1268-1284, 1983.[Free Full Text]
• ISRAEL, I., BERTHOZ, A. Contribution of the otoliths to the calculation of linear displacement. J. Neurophysiol. 62: 247-263, 1989.[Abstract/Free Full Text]
• ISRAEL, I., FETTER, M., KOENIG, E. Vestibular perception of passive whole-body rotation about horizontal and vertical axes in humans: goal-directed vestibulo-ocular reflex and vestibular memory-contingent saccades. Exp. Brain Res. 96: 335-346, 1993.[Medline]
• MCKENZIE, A., LISBERGER, S. G. Properties of signals that determine the amplitude and direction of saccadic eye movements in monkeys. J. Neurophysiol. 56: 196-207, 1986.[Abstract/Free Full Text]
• OHTSUKA, K. Properties of memory-guided saccades toward targets flashed during smooth pursuit in human subjects. Invest. Ophthalmol. Vis. Sci. 35: 509-514, 1994.[Abstract/Free Full Text]
• PELISSON, D., GUITTON, D., MUNOZ, D. P. Compensatory eye and head movements generated by the cat following stimulation-induced perturbations in gaze position. Exp. Brain Res. 78: 654-658, 1989.[Medline]
• SCHILLER, P. H., KOERNER, F. Discharge characteristics of single units in superior colliculus of the alert rhesus monkey. J. Neurophysiol. 34: 920-936, 1971.[Free Full Text]
• SCHILLER, P. H., SANDELL, J. H. Interactions between visually and electrically elicited saccades before and after superior colliculus and frontal eye field ablations in the rhesus monkey. Exp. Brain Res. 49: 381-392, 1983.[Medline]
• SCHLAG, J., SCHLAG-REY, M., DASSONVILLE, P. Saccades can be aimed at the spatial location of targets flashed during pursuit. J. Neurophysiol. 64: 575-581, 1990.[Abstract/Free Full Text]
• SCHLAG-REY, M., SCHLAG, J., SHOOK, B. Interactions between natural and electrically evoked saccades. I. Differences between sites carrying retinal error and motor error signals in monkey superior colliculus. Exp. Brain Res. 76: 537-547, 1989.[Medline]
• SEGAL, B. N., KATSARKAS, A. Goal-directed vestibulo-ocular function in man: gaze stabilization by slow-phase and saccadic eye movements. Exp. Brain Res. 70: 26-32, 1988.[Medline]
• SPARKS, D. L., MAYS, L. E. Spatial localization of saccade targets. I. Compensation for stimulation- induced perturbations in eye position. J. Neurophysiol. 49: 45-63, 1983.[Free Full Text]
• WELLENIUS, G. A., ROY, J. E., CULLEN, K. E. Characterization of eye/head coordination during head-free gaze pursuit in the monkey using Rashbass step-ramp target trajectories. Soc. Neurosci. Abstr. 23: 754, 1997.
• WHITE, J. M., SPARKS, D. L., STANFORD, T. R. Saccades to remembered target locations: an analysis of systematic and variable errors. Vision Res. 34: 79-92, 1994.[Medline]
• ZIVOTOFSKY, A. Z., ROTTACH, K. G., AVERBUCH-HELLER, L., KORI, A. A., THOMAS, C. W., DELL'OSSO, L. F., LEIGH, R. J. Saccades to remembered targets: the effects of smooth pursuit and illusory stimulus motion. J. Neurophysiol. 76: 3617-3632, 1996.[Abstract/Free Full Text]

0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society