| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Otolaryngology, University of Texas Medical Branch, Galveston, Texas
Submitted 14 December 2005; accepted in final form 5 March 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Schall et al. 1995). Past studies have shown that this region is associated with the initiation of eye movements (Bizzi and Schiller 1970; Bruce et al. 1985; Dias and Segraves 1999; Goldberg et al. 1986; Keating and Gooley 1988; Robinson and Fuchs 1969; Schall 2002; Schiller et al. 1979; Smith 1949; Sommer and Tehovnik 1997; Tehovnik et al. 2000; van der Steen et al. 1986). Whether the FEF participates in generating head movements is not clear. In an early anecdotal observation, Levinsohn (1909, described in Smith 1949) reported that stimulation in the dorsomedial frontal cortex, i.e., the supplementary eye field (SEF) as identified today, evoked head movements that often preceded eye movements. In contrast, this was not observed in the lateral oculomotor region, i.e., the FEF as identified today. Levinsohn noted this unique characteristic as the major difference between dorsomedial and lateral oculomotor regions of the frontal cortex, i.e., the SEF and the FEF.
Past studies have revealed some seemingly conflicting findings regarding whether the FEF is involved in the control of head movements (Bizzi and Schiller 1970; van der Steen et al. 1986). van der Steen et al. (1986) conducted a unilateral FEF lesion study in head-unrestrained monkeys and found that the lesioned monkeys were reluctant to track objects in the contralateral visual field. When the monkeys did track visual targets, the monkeys moved their heads more often than their eyes. When gaze shifts occurred, the accompanying head amplitudes were atypically large, and eye (re head) amplitudes small. At gaze completion, the eye was often counter-rotated rapidly to near the center of the orbit unlike its prelesion behavior. These findings were interpreted by the authors as indicating that FEF lesions led to selective eye-movement deficits.
Bizzi and Schiller (1970) recorded the neuronal activity in the FEF of head-unrestrained monkeys. They found that, in agreement with the role of FEF in eye-movement control, FEF neurons discharged in association with eye movements. However, they also found an unusual type of neuron that discharged exclusively during head movements. Even though the exact characteristics and anatomical connectivity of FEF head neurons have not been identified, the presence of the head-movement-related discharge in FEF neurons remains an enigma in the understanding of FEF function (cf. Knight and Fuchs 2001; cat FEF: Guitton and Mandl 1978).
Recent development of experimental techniques in head-unrestrained monkeys has made it possible for investigators to revisit the issues of motor control in head-unrestrained conditions (Chen and Walton 2005; Collins and Barnes 1999; Corneil et al. 2002; Cullen et al. 2004; Crawford and Guitton 1997; Freedman and Sparks 1997; Gandhi and Sparks 2001; Goffart et al. 1998; Goossens and van Opstal 1997; Guitton et al. 2003; Isa and Sasaki 2002; Martinez-Trujillo et al. 2003; Peterson 2004; Phillips et al. 1995; Sparks et al. 2001; Stahl 2001; Waitzman et al. 2002). A recent study (Tu and Keating 2000) reported that FEF stimulation evoked both eye and head movements. The authors noted that, at low current intensity (25 µA), FEF stimulation evoked eye movements. When the stimulation intensity was increased to 200–300 µA, stimulation evoked gaze shifts accompanied by head movements. However, these head movements were too small (4.3 ± 0.4° in Tu and Keating 2000) to contribute significantly to gaze shifts. These findings raise new questions regarding the exact role of head movements evoked by FEF stimulation.
The present study addressed the following questions. First, does FEF stimulation evoke head movements? If so, do stimulation-evoked head movements contribute to shifting the line of sight (i.e., gaze)? In other words, is the FEF involved in gaze (eye plus head) feedback control? Gaze feedback hypothesis has been postulated to account for the head involvement in large orienting gaze shifts (Guitton et al. 1990, 2003; Laurutis and Robinson 1986). This hypothesis states that the gaze error (difference between current and desired gaze displacements) signal drives both eye and head movements, such that the range of gaze displacement is extended beyond the range accomplished by eye (re head) displacement alone. This explanation accounts for the fact that large visually guided gaze shifts (e.g., gaze amplitude 
20°) usually recruit significant contribution of the head (Freedman and Sparks 1997; Fuller 1992; Galiana and Guitton 1992; Guitton et al. 1990, 2003; Laurutis and Robinson 1986; Tomlinson and Bahra 1986). On the other hand, according to this hypothesis, there exists a temporal coupling between head and gaze velocities. As the gaze error diminishes by the end of gaze shifts, the drive to move the head will diminish and head velocity will begin to decrease (Guitton et al. 1990, 2003; Matsuo et al. 2004). Therefore the head should reach its peak velocity prior to, as opposed to hundreds of milliseconds after, the end of gaze shifts. The present stimulation study provided an opportunity to evaluate these predictions in the FEF.
Note that head contribution to gaze shifts is limited to the head displacement between gaze shift onset and gaze shift offset, during which the vestibuloocular reflex (VOR) is suppressed for the eyes and head to move in the same direction (Cullen et al. 2004; Guitton et al. 1990). The total head displacement additionally includes the head movement that may occur prior to the onset of gaze shift (VOR gain 
1) and the head movement that often lasts beyond the end of gaze shift (VOR gain 1) (Bizzi et al. 1971; Chen and Walton 2005; Corneil et al. 2002; Freedman and Sparks 1997; Guitton et al. 1990; Tomlinson and Bahra 1986). Both monkeys used in the present study had previously participated in a SEF study (Chen and Walton 2005) in which stimulation evoked significant contribution of the head to gaze shifts. It is pertinent to know whether differential motor control mechanisms exist in the two eye fields of the same monkeys under the same task controls.
Second, does FEF stimulation evoke head movements independent of gaze shifts? Recent studies have shown that low current stimulation in the superior colliculus (Corneil et al. 2002; Pelisson et al. 2001) or stimulation at contralateral eye positions in the SEF evoked head movements in the absence of gaze shifts (Chen and Walton 2005). In addition, stimulation-evoked postgaze-shift head movements (i.e., the head displacements between gaze shift offset and head offset) in the SEF facilitated re-centering the eyes in the orbits (Chen and Walton 2005). It is pertinent to know under what circumstances FEF stimulation evokes head-alone movements and whether stimulation-evoked postgaze-shift head movements minimized the eye deviation from the center of the orbit (Sparks et al. 2001).
It has been shown that initial eye (IEP) and head (IHP) positions are pertinent variables that determine the metrics of head movements (Chen and Walton 2005; Delreux et al. 1991; Freedman and Sparks 1997; Goossens and van Opstal 1997; Phillips et al. 1995; Tomlinson and Bahra 1986; Volle and Guitton 1993). In this study, IEP and IHP were systematically varied for the sake of assessing FEF stimulation-evoked head movements. We found that FEF stimulation indeed evoked head movements in addition to eye movements; however, the head movements contributed little to shifting the line of sight. Our findings agree with early studies and suggest that head motor control may be independent from eye control in the FEF.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Two juvenile rhesus monkeys (Macaca mulatta, 5–7 kg) served as subjects. They were implanted with a chamber (angled 15° from the midsagittal plane) and head-posts. The details of the surgical implants, chairing setups, and neurophysiological procedures have been described previously (Chen and Walton 2005). All surgical and experimental procedures conformed to the guidelines for the Care and Use of Animals of National Institutes of Health and the Institutional Animal Care and Use Committee.
Eye and head positions were tracked using the search coil technique (Fuchs and Robinson 1966; Judge et al. 1980). A 42-in cubic coil and a phase-angle detection system (CNC Engineering) were used to measure horizontal and vertical position signals of gaze (eye re space) and head coils, sampled at 500 Hz. During recording, a lightweight microlaser (Edmund Scientific, No. M52263) was mounted on the monkey's head to provide the visual feedback of head positions. The signals of gaze and head coils were calibrated under conventional, head-fixed conditions based on the alignment of gaze and head coil signals with the visual targets of known positions. Eye (re head) positions were computed off-line by mathematically converting horizontal and vertical gaze coil signals to unit vectors, rotated with respect to head vectors in Fick coordinates. Movement amplitude was computed following vector rotations. Data were acquired using a Pentium microcomputer running in-house data acquisition software.
Standard microelectrodes (Frederic Haer) were used to penetrate dura, record neuronal signals, and deliver electrical stimulation (for details, see Chen et al. 2001). Neural signals were band-pass (500 to 5 kHz) filtered using a differential BAK amplifier. Microstimulation was carried out using a stimulator (Grass S88) and an optical isolation unit (Grass PSIU6). The stimulation trains consisted of 0.2-ms, monopolar, cathodal pulses. Typical stimulation was 80 µA (range: 50–150 µA), 200 Hz (range: 100–200 Hz), and 300 ms (range: 300–500 ms). Prolonged stimulation ensured that the movements of interest (i.e., head movements) were not truncated prematurely (for review, see Freedman et al. 1996; Graziano et al. 2002). When necessary, stimulation of different parameters was explored. Because the threshold of stimulation-evoked head movements in the FEF was not known, we typically applied stimulation of two to three times the threshold current (i.e., the current that evoked movements in 50% of the trials) for eliciting eye movements. It was difficult to monitor the actual current delivered through the high-impedance electrodes (0.5–1.2 M
measured in saline at 1 kHz). All of the current intensity reported here was taken from the face value of the stimulator.
Behavioral paradigms and microstimulation
Visual targets were presented on a tangent screen with 49 x 41 tri-state (red, green, yellow) light-emitting diodes (LEDs). The LEDs were equally spaced at 2-in intervals in both horizontal and vertical dimensions. The LED board was placed 72 cm (28.5 in) from the monkeys. The room was dimly lit with a 5-W light bulb placed behind the LED board.
The monkeys were trained in a visually guided gaze shift task that permitted independent control of gaze and head positions (Fig. 1A). The behavioral task consisted of two phases: an initial eye/head alignment-and-dissociation phase and a subsequent visually guided gaze shift phase. The initial phase began with monkeys sitting in head forward positions, aligning the microlaser beam with a visual target (red LED). Later, a second (green) target was illuminated, and the monkeys deviated their eyes to the target while maintaining a stable head position. The green target was usually displayed in 
27° steps horizontally or vertically with respect to the initial red target (e.g., Fig. 1C). 600–800 ms after the onset of the green target, all targets and the microlaser were extinguished for 400–500 ms. The same red and green targets and microlaser beam were re-illuminated briefly (500–700 ms) and then re-extinguished.
|
In the control trials (Fig. 1A), gaze and head positions were constrained under close-loop real-time control within a 5- and 10°-radius "window," respectively, around the designated targets. If either gaze or head stepped outside of the window during the designated periods, the trial was aborted.
In 
50% of the trials within a block, electrical microstimulation trials (Fig. 1B) were carried out. Stimulation trials began with the initial eye-head alignment-dissociation task just like that in the control trials. The electrical stimulation phase began with stimulation 200 ms after the extinction of both visual targets and microlaser. The entire stimulation phase was conducted in darkness without constraint over the eye or head positions. Approximately 800–1,400 ms after the onset of stimulation, the visually guided gaze shift phase began. Reward was contingent on the monkey making a gaze shift to a visual (yellow) target, as in the control trials.
Figure 1C illustrates schematic examples of head, gaze, and eye positions independently controlled in EHD (eye-head dissociation) and EHA (eye-head alignment) trials. In the EHD example trial, the eyes were oriented in upper-right direction, while the head remained centered with respect to the body. In the EHA example trials, the head was oriented leftward with respect to the body, while the eyes remained centered in the orbit.
In some large-saccade sites, we carried out nontask-mode stimulation in which the stimulation train was delivered during the intertrial interval (duration: 500–1500 ms). Nontask-mode stimulation trials were 50% interleaved with task-mode stimulation trials. Unlike the latter, there was no visual target or microlaser illuminated in the former. There was no IEP or IHP constraint throughout the nontask-mode stimulation trials.
Care was taken to exclude the data obtained in stimulating the white matter from the analyses. The electrodes were always advanced deep, and the border between gray and white matter was identified based on the overall diminished unit discharge. Once the electrode was confirmed to have reached the white matter, the electrodes were then slowly withdrawn. Stimulation was carried out >100 µm away from the border of the white and gray matters. Because it was impossible to know whether a given electrode penetration was parallel, orthogonal, or oblique with respect to the cortical surface, we considered the stimulation depth separated by 500 µm as different sites. The movements evoked in different sites were analyzed separately.
Data analyses
Data analyses were performed using an in-house program on a Windows platform. Movement onset and offset were determined based on the velocity criteria (gaze: 80°/s for horizontal and vertical onsets and 60°/s for horizontal and vertical offsets; head: 6°/s for both horizontal and vertical onset and offset). For details of the off-line threshold-filter computation, see Chen and Walton (2005). Movements were displayed 100 ms before and 800 ms after stimulation onset, and measurements were taken strictly based on the velocity criteria. Any movement detected before stimulation onset was removed from further analysis. Throughout this paper, only stimulation-evoked gaze shifts and head movements with onset latency 300 ms were included in the analysis. The criterion of minimal onset latency for gaze shifts and head movements was 20 and 50 ms, respectively.
Staircase small-saccade (<10°) gaze shifts were occasionally encountered. These trials were excluded from further analysis. Statistical analyses were performed using Statistica (Statsoft). Throughout this paper, the data are described and plotted as means ± SD.
At the end of the experiments, the monkeys were killed with an overdose of pentobarbital, and the brains were removed for histological examination. Stainless-steel pins were inserted in known coordinates during perfusion to facilitate coordinate reconstruction.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Fukushima et al. 2000; MacAvoy et al. 1991; Russo and Bruce 1993; Tehovnik et al. 2000). This study was aimed at large-saccade sites in the FEF, as large-amplitude gaze shifts were likely to recruit significant head movements (Guitton et al. 1990; Sparks et al. 2001; Tomlinson and Bahra 1986).
In one of the monkeys, the FEF was meticulously mapped along the rostral-caudal dimension (Fig. 2A). Consistent with the notion that the FEF is topographically organized as a saccadic amplitude map, we found that large saccades were evoked in the rostral sites and small saccades and smooth pursuits were evoked in the caudal sites (Bruce et al. 1985; Fukushima et al. 2000; MacAvoy et al. 1991). Stimulation-evoked gaze shift sites stretched 9 mm rostro-medially from the small-saccade and smooth pursuit sites. We exhausted the mapping 2 mm anterior from the most rostral large-saccade sites of the FEF. There was no evidence of site-specific head movement clustering in the FEF.
|
Figure 2C illustrates example traces of the stimulation-evoked movements in a large-saccade site of the FEF. Two main features can be noted. First, after gaze shift onset, the head-position trace deviated from its baseline slowly. It was impossible to detect head movement onset based on visual inspection of the head-position trace. Second, during the gaze shift, the horizontal gaze position almost completely overlapped the horizontal eye (re head) position. Approximately 15 ms before gaze completion, the head started to move above the velocity threshold. Note the eye movement decelerated rapidly toward the end of the gaze shift; hence, the head movement contributed little to the resultant gaze displacement.
The head movement reached its peak velocity (45°/s) 110 ms after the end of gaze shifts. The gaze position after gaze shifts was stable, i.e., the eye counter rotated in the orbit approximately by the same velocity as the head (VOR gain 1). Both examples illustrated in Fig. 2 (B and C) were obtained when IEP was centered in the orbit and IHP was centered with respect to the body.
Our results indicate that the characteristics of stimulation-evoked movements varied depending on IEP and IHP at stimulation onset. Particularly, "eye alone," "eye and head," and (occasionally) "head-alone" movements could be evoked by FEF stimulation at a given large-saccade site. This point will be elaborated in the following text.
Kinematics of stimulation-evoked gaze shifts and head movements
Figure 3 plots the relationship between movement amplitudes and velocities of the stimulation-evoked gaze shifts (A and B) and head movements (C and D) in the FEF. For the sake of data comparison between task conditions, gaze shifts were separated into EHD (left) and EHA (middle) trials (see METHODS). In EHD trials, IEP at stimulation onset ranged –28: 28° horizontally and –25: 28° vertically. In EHA trials, IHP at stimulation onset ranged –32: 32° horizontally and –12: 28° vertically. All stimulation trials of both monkeys were pooled in the analysis. Several kinematic characteristics can be noted.
|
Second, unlike the peak velocity of horizontal gaze shifts that could be as high as 1,000°/s (Fig. 3B), the peak velocity of horizontal head movements never exceeded 100°/s. The peak velocity of horizontal head movements was linearly correlated with horizontal head amplitude (slope = 3.57; r = 0.89; P < 0.01) as has been reported for the head movements occurring during visually guided gaze shifts (Freedman and Sparks 1997; Guitton et al. 1990).
Third, like stimulation-evoked gaze shifts that were primarily contralateral, the vast majority (98.4%; 1,184/1,203) of stimulation-evoked head movements was directed contralaterally with respect to the stimulated side. The remaining had small (–1.7 ± 0.9°) head movements in the ipsilateral direction.
Finally, <1/4 [23% (694/2,973) in EHD trials and 23% (396/1,703) in EHA trials] of stimulation-evoked gaze shifts were accompanied by head movements. In most of the trials, eye-alone movements were observed, i.e., head movements were below the velocity threshold (see METHODS).
Eye alone movements.
The range of horizontal gaze amplitudes varied significantly as a function of eye position, 0.8: 49° (23 ± 8°; n = 1,154), 0.1: 43° (11 ± 7°; n = 900), and 0.1: 20° (4 ± 3°; n = 212) for IEPi [horizontal IEP (IEPh) ipsilateral to the stimulated side], IEPo (IEPh centered in the orbit), and IEPc (IEPh contralateral to the stimulated side) conditions, respectively. The range of vertical gaze amplitudes was –39: 47° (19 ± 14°), –20: 53° (13 ± 8°), and –14: 39° (8 ± 7°) for IEPi, IEPo, and IEPc conditions, respectively. The orbital effect resembles that observed under head-unrestrained conditions (Russo and Bruce 1993).
Effects of varying horizontal IEP
Figure 4 illustrates the movement traces showing the effects of varying IEPh on horizontal head and gaze velocities after electrical stimulation in a FEF site. The trials were grouped by IEPh conditions. Five traces (2 of high peak velocities, 2 of low peak velocities, and 1 of medium peak velocity) were selected to represent horizontal head- and gaze-velocity profiles for each IEPh condition. In these trials, IHP remained centered with respect to the body (the range of horizontal and vertical IHP: –8: 8°). Only IEPh was systematically varied. The range of IEPh was –28: –15°, –10: 10°, and 15: 28° in IEPi, IEPo, and IEPc conditions, respectively.
|
150 µA (bottom left). In contrast, in IEPo (Fig. 4, middle) and IEPc conditions (Fig. 4, right), stimulation did evoke head movements in 1/3 of the trials. Second, when stimulation did evoke head movements, head movement onset often lagged behind gaze shift onset. Third, head velocities often did not reach their peaks until after gaze shifts were completed. The velocity profile of the late head movements can be best appreciated in IEPc trials. To quantify the effect of varying horizontal IEP on eye and head contributions to gaze shifts, eye-and-head movements were selected for further analyses (Fig. 5; Table 1). Overall, eye-and-head movements represented <5, 21, and 32% of the stimulation-evoked gaze-shift trials in IEPi (n = 1,214), IEPo (n = 1,149), and IEPc (n = 552) conditions, respectively. That is, the probability of evoking head movements in FEF stimulation increased as horizontal IEP was deviated in the contralateral direction (i.e., the direction of the stimulation-evoked gaze shifts).
|
|
Effects of varying horizontal IHP
Figure 6 illustrates the movement traces showing the effect of horizontal IHP (IHPh) on the horizontal head and gaze velocities in a large-saccade site of the FEF. In these experiments, IEP remained within ±8° centered in the orbit, and IHP was systematically varied. These trials were grouped by IHPh [IHPi (ipsilateral IHPh; range: –32: –20°), IHPo (IHPh centered; range: –10: 10°), and IHPc (contralateral IHPh; range: 20: 32°)] conditions. Five traces (2 of high peak velocities, 2 of low peak velocities, and 1 of medium peak velocity) were selected to represent the horizontal head- and gaze-velocity profiles of each IHP group.
|
25°/s) near the end of gaze shifts. The apparent dissociation between head and gaze velocities can be appreciated during two staircase gaze shifts (downward empty triangle; Fig. 6). In both examples, the horizontal gaze displacement was 50° (30 and 20° for the 1st and 2nd gaze shifts, respectively). Note that despite the fact that the head was in motion and thus could easily accelerate, head-velocity profile was not altered during the second stimulation-evoked gaze shift. Finally, head movements often reached their peak velocities >>50 ms after gaze offset (Fig. 6, left and middle).
Figure 7 quantifies the effects of varying IHP on the stimulation-evoked movements of all trials (n = 1,694) of all sites (n = 26; M1 = 19, M2 = 7). The probability for FEF stimulation to evoke eye-and-head movements was 58, 21, and 10% for IHPi (n = 106), IHPo (n = 1,528), and IHPc (n = 60) conditions, respectively (Table 1). The horizontal amplitude of gaze shifts in 90% of these trials was <26°. Horizontal gaze displacement was totally accounted for by horizontal eye displacement (Fig. 7C, 
). Head contribution to horizontal (Fig. 7C, gray *) and vertical (data not plotted) gaze shifts was negligible.
|
Nontask-mode stimulation was conducted in 12 large-saccade sites (M1 = 7; M2 = 5) in the FEF (see METHODS; Fig. 7). In these trials (n = 444), the range of IHP at stimulation onset was –31: 32°. At this IHP range, the range of horizontal IEP was –21: 26° in the orbit.
Half (51%; n = 227/444) of these were eye-alone movements, the remaining were eye-and-head movements. In the latter movements, the maximal horizontal gaze amplitude was 31° (Fig. 7D), whereas the maximal horizontal head amplitude was 16° (Fig. 7E; Table 1). However, given comparable metrics, head contribution to stimulation-evoked horizontal gaze shifts remained negligible (Fig. 7F, bottom, +). Horizontal gaze displacement was entirely accounted for by horizontal eye displacement (Fig. 7F, bottom, ).
Visually guided gaze shifts
Figure 8 illustrates the metrics of visually guided eye-and head movements (n = 2,931; M1: 1,557, M2: 1,374). These visually guided movements were obtained following the yellow target in the control and stimulation trials (see METHODS). There were three major differences in the eye-head coordination between visually guided and stimulation-evoked movements.
|
; Phillips et al. 1995; cat: Guitton et al. 1990). When horizontal gaze shifts were 25–35° in IEPo condition, the average horizontal amplitude of the head was 17 ± 8° (n = 301; Fig. 8A, top). In contrast, given comparable horizontal gaze amplitudes, the average horizontal amplitude of FEF stimulation-evoked head movements was only 5 ± 3° (n = 60). The difference was highly significant (F = 130, P < 0.001). Second, for visually guided gaze shifts, small, but significant vertical head movements were often observed. For example, when the vertical amplitude of gaze shifts was 25–35° in IEPo condition, the average vertical amplitude of the head was 11 ± 8° (n = 401, Fig. 8B, top). In contrast, given comparable vertical gaze amplitudes, the average vertical amplitude of FEF stimulation-evoked head movements was only 1.3 ± 1.2° (n = 20). The difference was significant (F = 34, P < 0.001).
Third, head contribution to visually guided horizontal gaze shifts was a function of horizontal gaze amplitudes and horizontal IEP in the orbit (Fig. 8A, bottom). For example, when the horizontal amplitude of gaze shifts was 50° in IEPo condition, the contribution of the head was as large as 13°. In contrast, we never observed FEF stimulation-evoked gaze shifts as large as 50° in horizontal amplitude in IEPo condition. When horizontal amplitude of visually guided gaze shifts was 25–35° in IEPo condition, the average contribution of the head was 4 ± 2° (n = 301; Fig. 8A, bottom). In contrast, given comparable horizontal gaze amplitude, the average head contribution to FEF stimulation-evoked gaze shifts was only 0.6 ± 0.5° (n = 60). The difference was highly significant (F = 139, P < 0.001).
Timing difference between head movement onset and gaze shift onset
Head movement onset varied systematically as a function of horizontal IEP (Fig. 9A) and IHP (B), whereas gaze shift onset remained consistent across EHD and EHA trials. There was a significant correlation between head movement onset and IEPh (Fig. 9A) or IHPh (B). There was no significant correlation between gaze shift onset and IEPh (Fig. 9A) or IHPh (B).
|
Peak velocity latencies of head movements
Figure 10 quantifies the latency difference between gaze shift offset and the peak velocities of head movements. Only the trials of head amplitude 2° were selected for analysis. More than 90% of stimulation trials [98% (567/576) in EHD trials and 97% (300/311) in EHA trials] exhibited positive values (Fig. 10, A and B; H 2, ). This indicates that the vast majority of head movements reached their peak velocities after gaze completion. The average peak velocity of head movements of all stimulation (EHD and EHA) trials combined was 136 ± 91 ms, and the mode of the distribution fell on the 121- to 130-ms bin.
|
) plots the distributions when only the stimulation-evoked head movements of 5° were considered. The average peak velocity of head movements of all stimulation (EHD and EHA) trials combined was 150 ± 102 ms, and the distribution peak fell on the same 121- to 130-ms bin. In other words, regardless of whether the head movements were small or large, the distribution exhibited peaks at comparable latencies.
The next question is under what circumstances do visually guided gaze shifts recruit head movements similar to those observed by FEF stimulation. We analyzed the visually guided movements with comparable horizontal gaze amplitudes (42°) and total head amplitude (2°; Fig. 10C, ). The average peak head velocity latency of these visually guided movements was 69 ± 85 ms (n = 2,631), which was significantly different from that evoked by FEF stimulation (t = 14; P < 0.001).
Further analysis indicates that different subsets of visually guided gaze shifts recruited head movements of different peak velocity latencies. When the horizontal contribution of the head was 2° (Fig. 10C, ), the average latency of the peak head velocity (re gaze offset) was 24 ± 45 ms (n = 1,192). When the horizontal contribution of the head was 1° (Fig. 10C, 
), the average latency of the peak head velocity was 129 ± 96 ms (n = 781). The difference between the two distributions was highly significant (t = 33; P < 0.001). The latter distribution was not significantly different from that of stimulation-evoked eye-head movements (t = 1.5; P > 0.10).
Postgaze–shift head displacement
Horizontal postgaze-shift head displacement (Hpgh) was defined as the head displacement between gaze completion and the moment when the head stopped moving. Hpgh was positively correlated with horizontal eye position at gaze offset in EHD and EHA trials (Fig. 11). The data of the two trial types overlapped each other (ANCOVA, homogeneity-slope model; F = 3, P > 0.07). However, compared with visually guided gaze shifts, FEF stimulation-evoked head movements were smaller. When the eye position following gaze completion was 20°, the average Hpgh in visually guided movements was 10°, whereas the average FEF stimulation-evoked Hpgh was significantly smaller (4°). Overall, there was a significant difference between FEF stimulation-evoked Hpgh and visually guided Hpgh (ANCOVA, homogeneity-slope model, F = 810, P < 0.001).
|
0° and 1 trial at IEPh 13°) had a relatively smaller horizontal amplitude of the head (2.4 ± 0.2°) compared with those obtained in IEPc condition. The displacement of head-alone movements as a function of eye position at head movement onset (
) is superimposed on Fig. 11. It can be noted that, when eye positions were >10° contralateral to the stimulated side, the data points of head-alone movements fell on top of those of EHD and EHA trials. Compared with visually guided gaze shifts, head-alone movements contributed less to minimizing the eye deviation from the center of the orbit (ANCOVA, homogeneity-slope model; F = 578, P < 0.001). |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
1/3 of stimulation-evoked head movements occurred in the absence of gaze shifts. Finally, postgaze-shift head displacement was positively correlated with the eye position in the orbit, suggesting that these late head movements did contribute to minimizing the eye deviation from the center of the orbit. This process did not involve a change of gaze. Some alternative interpretations of these findings are discussed in the following text. Stimulation-evoked nonvolitional movement
One may argue that some forms of task-associated movement suppression (or promotion) were involved in the findings of the present study. Specifically, the demands in this study for controlling initial eye and head positions might have contributed to the apparent late onset of stimulation-evoked head movements. This possibility is unlikely for several reasons. First, comparable results were observed in nontask-mode as well as task-mode stimulation trials. In the former task, there was no feasible promotion or suppression of a given type of movement, yet the stimulation-evoked gaze shifts recruited negligible contribution of the head. Second, FEF stimulation was conducted in darkness, 200 ms after the extinction of visual targets. This procedure minimized the influence of visual fixation-induced motor suppression (Goldberg et al. 1986; Tehovnik and Slocum 2000). Third, the reward was not contingent on any movements during or following stimulation. Rather, the reward was contingent on the visually guided gaze shifts at the end of the trials (see METHODS). Fourth, the direction of stimulation-evoked head movements was not random but primarily contralateral with respect to the stimulated side, suggesting that these head movements were likely evoked by stimulation as opposed to initiated volitionally. Fifth, significant contribution of the head was observed during visually guided gaze shifts (Fig. 8), indicating that the animals were capable of performing coordinated eye-head movements. Finally, stimulation-evoked head movements in the FEF were dramatically different from those in the SEF in which stimulation-evoked gaze shifts recruited significant contribution of the head (Chen and Walton 2005). These results were obtained in the same animals under the same task demands. Therefore it seems that the lack of head contribution to stimulation-evoked gaze shifts was unique for the FEF.
The metrics of FEF stimulation-evoked head movements in this study were in general agreement with the recent study of Tu and Keating (2000). First, low current stimulation in the FEF evoked primarily eye-alone movements. High current stimulation evoked gaze shifts accompanied by head movements that were otherwise undetectable [200–300 µA at 250 Hz in Tu and Keating (2000); 150 µA at 200 Hz in this study]. Second, head movement onset often lagged behind gaze shift onset (Fig. 2 in Tu and Keating 2000) (Fig. 9). Third, the average horizontal amplitude of stimulation-evoked head movements was modest [4.3 ± 0.4° in Tu and Keating (2000); 4.3° ± 2.9° in this study]. Fourth, these head movements contributed little to shifting the line of sight (Fig. 2 in Tu and Keating 2000) (Figs. 2, 5, 7, and 9; Table 1). However, note that head contribution to gaze shifts and head amplitude were described interchangeably in Tu and Keating (2000); hence, they concluded that head contribution (amplitude in their measurement) was significant in the report (cf. Scudder et al. 2002).
Head movements and the role of FEF in orienting gaze shifts
Gaze feedback hypothesis has been proposed to account for the head involvement in large orienting gaze shifts (Guitton et al. 1990, 2003; Laurutis and Robinson 1986). The hypothesis states that a gaze motor command provides the drive to move the eyes and head, such that the range of gaze displacement is extended beyond the range accomplished by eye (re head) displacement alone (Freedman and Sparks 1997; Fuller 1992; Guitton et al. 1990, 2003; Laurutis and Robinson 1986; Tomlinson and Bahra 1986). As the FEF is anatomically located upstream from most of the oculomotor structures, the question then is whether the FEF issues a gaze motor command that coordinates eye and head movements, which are presumably controlled independently at the brain stem level. Interestingly, based on the following observations, this possibility appears not supported. First, FEF stimulation-evoked gaze shifts did not recruit significant contribution of the head, albeit stimulation-evoked gaze shifts were 20–42° in horizontal amplitude (horizontal IEP centered in the orbit). In contrast, visually guided gaze shifts with comparable gaze displacements always recruit significant contribution of the head (e.g., Freedman and Sparks 1997; Fig. 8). Second, head velocities appeared dissociated from the occurrence of gaze shifts (e.g., stimulation-evoked large staircase gaze shifts in Fig. 6). Unlike the typical visually guided coordinated gaze shifts, stimulation-evoked head movements did not accelerate rapidly during gaze shifts (Figs. 2, 4, and 6). Third, visually guided coordinated eye-head gaze shifts often began to decelerate by the end of gaze shifts, whereas FEF stimulation-evoked head movements did not (cf. Guitton et al. 1990, 2003). Instead, the head often continued to accelerate and reached its peak velocity >100 ms following gaze offset (Fig. 10, A and B). That is, the head-velocity profile was temporally dissociated from the presumed gaze drive. The results all point to the same conclusion: FEF stimulation-evoked head movements were not driven by the gaze drive signal. It seems that the FEF neither generated a gaze motor command nor coordinated eye and head movements. Note head movements with long peak velocity latencies have been shown embedded in some natural, visually guided gaze shifts (Phillips et al. 1995). Our analysis indicates that this unique type of head movements contributed little to shift the line of sight (Fig. 10C, ).
The notion of head motor control independent from eye control in the FEF is consistent with the findings of past studies. It has been found that some FEF neurons discharged exclusively during head movements (Bizzi and Schiller 1970). A recent brief report indicated that the motor-burst discharge of some FEF neurons lasted beyond the end of gaze shifts; the duration of the motor burst was positively correlated with the duration of head movements (Knight and Fuchs 2001). These studies suggest that the FEF neuronal discharge encodes some forms of head motor commands that are dissociated from the neural processes that generate gaze shifts.
If the head motor control is indeed independent from the eye control in the FEF, one would predict that under some circumstances FEF stimulation may evoke head movements in the absence of gaze shifts. This prediction was indeed observed. Head-alone movements were evoked by FEF stimulation, particularly when IEP was deviated contralaterally at stimulation onset. It is possible that the independent head control mechanism in the FEF offers the flexibility for better coordinating eye and head movements (Bizzi et al. 1971; Chen and Walton 2005; Goossens and van Opstal 1997; Morasso et al. 1973; Phillips et al. 1995; van der Steen et al. 1986).
One may wonder the exact role of the FEF in the control of head movements. Recent studies have shown that SEF stimulation-evoked head movements facilitated re-centering the eyes in the orbits (Chen and Walton 2005; Sparks et al. 2001). Our findings indicate that FEF stimulation-evoked postgaze-shift head displacements were positively correlated with eye positions, suggesting that these head movements may contribute to minimizing the eye deviation from the center of the orbit (Fig. 11). However, stimulation-evoked postgaze-shift head displacements were relatively small in the FEF as compared with those evoked by SEF stimulation or visually guided gaze shifts. It seems that the FEF may play a role in, but is unlikely the main source of re-centering the eyes in the orbits.
Comparison with the SEF
A recent stimulation study indicates that the SEF contained mechanisms of head control independent from those of gaze control (Chen and Walton 2005). Interestingly, this study also suggests that the FEF contained independent eye and head control mechanisms. However, SEF stimulation e
: G 
: G 
and 
in C, head movement onset and offset, respectively. Stimulation in site B (M1): 80 µA and 200 Hz. Stimulation in site C (M2): 100 µA and 200 Hz.
