| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Visual Neuroscience, Division of Neuroscience, Faculty of Medicine, Imperial College London, London, United Kingdom
Submitted 22 July 2005; accepted in final form 26 November 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
), foveal stimuli inhibit eye movements. Such fixation-related inhibition has been popularly indexed by a simple paradigm: Participants make saccades to peripheral stimuli while a fixation stimulus either stays on ("overlap condition") or disappears just before the saccade target appears ("gap condition"). Saccadic latency is reduced in the gap condition, a phenomenon known as the gap effect or fixation-offset effect (Reuter-Lorenz et al. 1991; Saslow 1967). If, in contrast, the fixation stimulus becomes brighter or larger just before a saccade, latency is increased (e.g., Pratt et al. 2000).
The gap effect has become associated with inhibitory connections in the superior colliculus (SC), a midbrain center that makes the major input into the brain stem saccadic generator (e.g., Dorris et al. 1997; Isa 2002; Munoz and Istvan 1998; Munoz et al. 2000; Sparks 2002). "Fixation cells," located in the rostral pole of the SC, are active in response to foveal stimuli and are thought to inhibit "saccade cells" located caudally in the intermediate layers (e.g., Dorris et al. 1997; Munoz and Fecteau 2002; Munoz and Istvan 1998; Munoz and Wurtz 1992). Alternatively or additionally, the SC fixation cells may suppress saccadic initiation via omnipause neurons in the pontine reticular formation (Takahashi et al. 2005).
Cortical areas such as the frontal eye fields (FEF) also seem to play a role. Some FEF cells are active during fixation (e.g., Hanes et al. 1998), whereas other cells show a correlate of disengagement from fixation (Dias and Bruce 1994). The FEF makes a large projection to the intermediate and deep layers of the SC, both directly and via the basal ganglia, and is thought to exert an important regulatory influence on the SC (e.g., Everling and Munoz 2000; Hikosaka and Wurtz 1989; Krauzlis 2005; Schall 1997; Segraves and Goldberg 1987). The FEF also makes direct projections to brain stem nuclei (e.g., Izawa et al. 2005). Similarly, the lateral intraparietal area (LIP in monkey, possibly corresponding to the human parietal eye field) makes considerable projections both to the SC and to other oculomotor nuclei in the brain stem (e.g., Krauzlis 2005; Pare and Wurtz 1997). Thus cortical areas have the potential both to modulate collicular mechanisms and to mediate fixation-related inhibition independently or upstream of the SC.
Although the gap paradigm has been used widely in the context of motor production, little attention has focused on the neural pathways that may supply the visual information to the saccade system. One possibility is that fixation-related inhibition is caused by visual signals reaching the SC directly (and the contribution of cortical areas may be to modulate this collicular mechanism). Alternatively, the gap effect may be caused by visual information reaching cortical oculomotor mechanisms, and the observed patterns of collicular activation may occur downstream. Third, various visual pathways may supply input to two or more inhibitory mechanisms, each of which produces fixation-related inhibition in its own right.
There are several ways in which visual information can reach the SC without involving the cortical eye fields. The most direct way is by the retinotectal pathway, which projects directly from the retina to the superficial layers of the SC. The role of this pathway remains debated, with suggestions including eye movements, attention, and even fear perception (Morris et al. 1999; Rafal et al. 1990, 1991). There are also projections from primary visual cortex to the SC superficial layers (e.g., Sparks 1986). These sensory superficial layers are arranged in a spatial map in which cells responding to foveal stimuli are aligned with the fixation cells in the intermediate layers (Schiller and Stryker 1972), whose activity is thought to cause fixation-related inhibition (e.g., Dorris et al. 1997; Munoz and Istvan 1998). However, it is not yet clear how strong the interlaminar connections are between the sensory cells and the fixation cells (e.g., Isa 2002). The intermediate layers themselves receive projections from extrastriate areas of visual cortex, and also from parietal, frontal, and temporal areas (e.g., Sparks 1986). To take one example, middle temporal and medial superior temporal areas (MT/MST), which are associated with motion perception, make strong collicular projections that are important for pursuit eye movements (e.g., Krauzlis 2005). Because fixating a stationary stimulus and fixating a moving one may share neural mechanisms, this extrastriate projection may contribute to fixation-related inhibition.
The visual signals traveling to the SC by any of the cortico-collicular pathways seem to rely on activity in the magnocellular geniculostriate pathway. Schiller et al. (1979) found that inactivating the magnocellular laminae of the lateral geniculate nucleus (LGN) reduced or eliminated visually driven activity in all collicular cells except those in the superficial layers driven directly by the retinotectal pathway. Inactivating the parvocellular laminae had no effect (Schiller et al. 1979), and converging evidence has shown that the initial sensory activity in SC cells does not distinguish colors (Marrocco and Li 1977; Ottes et al. 1987). Color-specific activity develops later if color-defined stimuli are targets for saccades, presumably following target selection in cortical areas (Ottes et al. 1987). In this sense, color pathways seem to influence the SC only via some intervening oculomotor processing that selects saccade targets.
Likewise, the initial sensory activity in FEF seems to come from magnocellular signals because it occurs with a short latency similar to the activation latency of areas MT and MST (mean 75 ms), and shorter than the mean latency in areas V2 and V4 (82 and 104 ms) (Schmolesky et al. 1998). Importantly, sensory cells in the FEF have not been found to show color or form sensitivity (e.g., Stuphorn and Schall 2002; Thompson and Bichot 2005), despite the known projections to FEF from temporal cortex, for example. This suggests that, similar to the case of the SC, if chromatic signals are to influence FEF processing, they must pass through a process that abstracts color information into signals about targets for oculomotor plans. Fixation-related inhibition may occur as part of this target selection process that turns chromatic visual information into oculomotor information before it reaches the FEF or SC.
Thus although information from all visual pathways may contribute to oculomotor input to the SC and FEF, it seems that visual activity in these sites relies on the magnocellular pathway and the retinotectal pathway. If fixation-related inhibition is mediated within the SC or FEF, based on direct visual input from the fixation stimulus, one might expect such visual information to be transmitted by either the retinotectal pathway or the magnocellular pathway. The aim of our study was to test this hypothesis. We used stimuli that were invisible to both the retinotectal pathway and the magnocellular pathway (Sumner 2005; Sumner et al. 2002, 2004). These stimuli were color changes visible only to the short-wave–sensitive retinal cone receptors (S cones), exploiting the fact that electrophysiological studies have consistently reported that there are no projections to the SC from S cones (de Monasterio 1978; Marrocco and Li 1977; Schiller and Malpeli 1977). The color changes were embedded in a background of spatial and temporal luminance noise, making them additionally invisible to the magnocellular pathway from retina to cortex, because it is not color opponent and receives little if any input from S cones (Calkins 2001; Chatterjee and Callaway 2002; Gouras 1968; Stockman et al. 1991; Yeh et al. 1995). Therefore neither the retinotectal nor the magnocellular pathway should be able to distinguish the S-cone stimuli from the background and fixation-related inhibition mediated by these pathways should not occur.
|
EXPERIMENT 1 |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Methods
PARTICIPANTS. Eight subjects participated (four male, four female, ages 18–29). All subjects had normal or corrected to normal acuity and normal color vision.
APPARATUS. Stimulus presentation was performed by a PC-controlled Cambridge Research Systems (CRS) ViSaGe directly connected to a 21-in. Sony GDM-F520 Trinitron monitor. Stimulus presentation was synchronized with a screen-refresh rate of 100 Hz, and timings were controlled and measured by the CRS clock and thus not subject to the errors produced by normal PC operating systems. The ViSaGe specifies colors with a resolution of 14 bits per gun and the monitor was calibrated using a CRS ColorCal colorimeter. Eye movements were recorded using an Applied Science Laboratories (ASL) model 504 high-speed remote infrared eye-tracker with an ASL 5000 series controller, which samples eye position at 240 Hz (chin and head rest also by ASL). Both vertical and horizontal displacement were measured.
S-CONE STIMULUS CALIBRATION. S-cone stimuli are color transitions that leave unchanged the signals of long-wave (L) and middle-wave (M) sensitive cones, while affecting the signal in short-wave cones. The exact color transitions that do not affect L and M cones differ between individuals, and between retinal locations, because of variations in cone sensitivity, macular pigment, lens optical density, and chromatic aberration. Therefore the S-cone stimuli were calibrated individually for each participant. A standard gray was chosen [MacLeod–Boynton (MB) color coordinates: 0.65, 0.02; luminance: 25 cdm–2] and color transitions from this gray were represented as angles in MB color space. The theoretical S-cone angle is 0° for foveal stimuli (a transition from gray to lilac-blue).
The calibration procedure for each participant took 30–60 min and was as follows. First the luminance of the candidate range of colors was psychophysically matched to the standard gray (25 cdm–2) using the minimum motion procedure introduced by Anstis and Cavanagh (1983). Second, to find the color angle that would not stimulate L and M cones, these color angles and their corresponding luminance values were used in the transient tritanopia procedure developed by Smithson et al. (2003), adapted so that the potential S-cone stimuli were only 0.4° from fixation. In this procedure, colored patches must be detected among gray patches, whereas the S-cone chromatic pathway has been made insensitive by the offset of a yellow adapting field. Thus performance is poorest for a color that cannot be differentiated from the gray using the L–M chromatic pathway. The color angles used ranged from –15 to +15°, first in 5° steps and then 2° steps in a MB color space whose axes were scaled so that threshold distances in the S and L–M directions were approximately equal. The S-cone angle was selected to the nearest degree and finally the minimum motion procedure was used again to check the luminance match of the chosen color to the standard gray.
GAP PARADIGM PROCEDURE. The experimental paradigm is illustrated in supplementary movie 1.1 The task was to detect small dark gray squares appearing 8° to the left or right of fixation, and make a saccade as quickly as possible to these targets. Before each target occurred, participants were required to fixate a central square that differed from the background in color or luminance. The background was a patchwork of gray squares (0.8° across) that created an environment of luminance noise. The luminance of each square changed every 50 ms throughout each trial, to a value chosen at random between 24.5 and 25.5 cdm–2. The luminance fixation stimulus was created by shifting the mean luminance of the central square to 30 cdm–2 (with random flicker range between 29.5 and 30.5 cdm–2). The S-cone fixation stimulus was created by shifting the chromaticity of the central square in the direction of the S-cone color angle calibrated as described above, creating an 80% increase in S-cone signal (with luminance flicker still occurring in the range between 24.5 and 25.5 cdm–2).
The saccade target appeared on each trial at a time chosen randomly between 500 and 900 ms after presentation of the background squares and fixation square. The target was the darkening to 15 cdm–2 of the square 8° either to the left or right of fixation. There were four types of fixation condition: 1) luminance overlap, in which a luminance fixation stimulus remained present for the duration of the trial; 2) luminance gap, in which the fixation stimulus disappeared (i.e., returned to randomly flickering gray between 24.5 and 25.5 cdm–2) 200 ms before the onset of the target; 3) S-cone overlap, in which an S-cone fixation stimulus remained present; and 4) S-cone gap, in which the S-cone stimulus disappeared 200 ms before target onset.
Participants were informed that the fixation stimulus would flicker, and sometimes disappear and sometimes stay on, and were instructed to ignore this and just try to detect the target onset. The screen stopped flickering 500 ms after the target appeared, and the next trial began 1,000 ms later with presentation of the fixation stimulus, at which time the patchwork of squares began to flicker again. After a short period of demonstration and practice (40 trials), 480 trials were completed by each subject, with breaks every 40 trials to allow rest and free eye movements. Thus there were 120 trials of each of the four fixation conditions—with equal numbers of each associated with left and right saccade targets—presented in a random order.
ANALYSIS. The criterion for saccade detection was a velocity of 60° s–1, and saccadic onset was defined by a velocity >22° s–1. Eye-position traces were inspected for all trials to check that the custom Matlab routine had correctly located saccades. Saccades with latencies between 80 and 500 ms were accepted as responses to the target, and trials were discarded in which fixation was not maintained preceding target presentation. That is, trials were included in the latency analysis only if the first saccade >0.5° occurred between 80 and 500 ms after target presentation and approached the target with an accuracy of within 2°. This was achieved in >90% of trials in all experiments. Trials were counted as errors if the first saccade after target presentation was in the opposite direction to the target, or if no response was made (<2% of trials in all experiments).
Results
The results showed that latencies for gap trials were considerably shorter than those for overlap trials—the standard gap effect. There was little difference between the effect of a luminance offset at fixation or an S-cone offset at fixation (see Fig. 1). This was confirmed with a two-way ANOVA with factors of gap (gap vs. overlap) and color (luminance vs. S cone). There was a main effect of gap [F(1,7) = 35.8, P < 0.001] but no effect of color and no interaction between the gap effect and the color of the fixation stimulus [F(1,7) = 2.1, P = 0.19].
|
|
EXPERIMENT 2 |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Klein et al. 1995; Taylor et al. 1998). However, several studies show that there is also a component related specifically to the presence or absence of a visual-fixation stimulus, indexing true fixation-related inhibition. For example, although a highly predictive warning tone reduces saccadic latency, a nonpredictive (temporally varied) fixation offset reduces latency even further (Pratt et al. 2000; Reuter-Lorenz et al. 1995; Taylor et al. 1998). This additional effect occurs for offsets at or near fixation but not for visual offsets 
2° from fixation (Fendrich et al. 1999; Taylor et al. 1998). An alternative to using a warning tone is to have the fixation stimulus undergo a change that supplies visual warning without an offset. For example, Pratt et al. (2000) found that abruptly reducing the fixation stimulus in size reduced latency and an increase in size prolonged latency, even though both manipulations provided visual warning. The term "fixation-offset effect" is often used to refer exclusively to the nonwarning component of the gap effect (e.g., Pratt et al. 2000). In experiment 2 we tested whether the S-cone fixation stimulus provides fixation-related inhibition over and above any warning effect. To this end, we used a change at fixation on every trial, rather than just on gap trials (see supplementary movie 2). If the S-cone gap effect resulted from a warning, and any true fixation-related inhibition was restricted to retinotectal or magnocellular sensory signals, then it should not matter whether a luminance fixation stimulus disappears or whether it changes into an S-cone stimulus (i.e., one that differs from the background only in the S-cone color direction). Neither the retinotectal pathway nor the magnocellular pathway should differentiate between these two conditions because they would not signal the difference between the S-cone stimulus and the flickering background squares. Both the luminance decrement and warning element are identical in each condition. If, on the other hand, S-cone stimuli cause true fixation-related inhibition, we predict that latencies should be longer when a luminance fixation offset is accompanied by an S-cone fixation stimulus onset, compared with when the luminance fixation stimulus simply disappears. All aspects of the apparatus, calibration, procedure, and analysis were identical to experiment 1, except where specified below.
Methods
PARTICIPANTS. Eight subjects participated (three male, five female, ages 18–26). All subjects had normal or corrected to normal acuity and normal color vision and none had participated in experiment 1.
PROCEDURE. The overlap conditions in experiment 1 were replaced with "fixation change " conditions, in which the fixation square changed from a luminance stimulus to an S-cone stimulus or vice versa, 200 ms before target onset. Thus there were four conditions: 1) luminance offset, as in experiment 1; 2) S-cone offset, as in experiment 1; 3) luminance–to–S-cone change; and 4) S-cone–to–luminance change (see supplementary movie 2). Again, participants were informed that the fixation square would flicker and change, but to try to ignore this and simply detect the target and move their eyes to it as quickly as possible.
Results
Because there was a change at fixation 200 ms before every target, the warning effect in all conditions should be equivalent. The only exception might be the S-cone offset condition—the S-cone pathway is known to be slower than luminance pathways (Smithson and Mollon 2004) and to have lower temporal resolution (Brindley et al. 1966), potentially leading to a weaker warning effect and longer latencies. Thus the warning explanation predicts equal latencies for all conditions containing a luminance change at fixation (i.e., the two change conditions and the luminance offset condition). Figure 2 shows that this was not the case.
|
It is also interesting to compare the two change conditions. If the luminance and S-cone stimuli produced equal fixation-related inhibition then latencies should be equivalent whether the change in fixation was from luminance to S cone or vice versa. In fact the S-cone–to–luminance change produced significantly longer latency than the luminance–to–S-cone change (t = 3.8, df = 7, P < 0.01). Thus the S-cone stimulus produced less fixation-related inhibition than did the luminance stimulus, consistent with the idea that signals reaching the SC directly might cause a greater effect.
|
EXPERIMENT 3A |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Dacey et al. 2003). The exact projections of this color pathway become less clear in extrastriate cortex, but it is assumed to project to V4 and from there to the temporal cortex. However, S cones may make some contribution to the magnocellular pathway (Calkins 2001; Chatterjee and Callaway 2002; Stockman et al. 1991) and, although this signal should be masked by luminance noise, it remains possible that the luminance noise in our paradigm was insufficient. It also remains possible that there are unknown S-cone projections in the retinotectal pathway. Third, the luminance levels used in these experiments would not prevent all rod activity, which might allow the S-cone stimuli to activate the magnocellular and retinotectal channels (although any contribution of rod activity to "S-cone" stimuli should have been minimized by having a psychophysical, rather than computational, calibration procedure, and rod intrusion should also be masked by the luminance noise).
Thus it remains possible that the S-cone stimuli may cause fixation-related inhibition by the same route as luminance stimuli. If this is the case, the S-cone stimuli should simply behave in the same way as luminance stimuli of low (but unknown) contrast. The key to the logic of experiment 3 is that this "equivalent-luminance-contrast" of the S-cone stimuli should be fixed if the S-cone fixation stimuli cause their inhibitory effect in the same way as the luminance fixation stimuli. If, however, the S-cone stimuli cause their effect by a distinct route, their relationship to the luminance stimuli need not be fixed because luminance and color pathways are known to be nonlinear in different ways (e.g., Kaplan and Shapley 1986) and adaptation to luminance contrast and color contrast can occur independently in postreceptoral pathways (e.g., Pugh and Mollon 1979; Webster and Mollon 1993).
Rather than exhaustively measure the exact "equivalent-luminance-contrast" of fixation-related inhibition caused by S-cone stimuli in different situations, the aim of experiment 3 was to take advantage of the likely nonlinearity in the luminance pathways (e.g., Kaplan and Shapley 1986) to show that the relative "strength" of an S-cone stimulus and a given luminance stimulus can reverse from one simple condition to another. This would prove that the S-cone stimuli do not have a consistent "equivalent-luminance-contrast," and must therefore be transmitted by a different pathway.
Experiment 3A measured the relative inhibitory effects of the S-cone and low-luminance stimuli compared with a high luminance stimulus. In two of the conditions in experiment 3A we used fixation changes from low to high luminance, or vice versa. It was expected that the high-luminance stimulus would cause the most fixation-related inhibition, and thus saccade latencies would be greater after a low–to–high-luminance change at fixation compared with the high-to-low condition. The difference in mean latency between these conditions would be the measure of the relative difference in fixation-related inhibition caused by the two stimuli. We also used fixation changes from the S-cone stimulus to the high-luminance stimulus, or vice versa. It was again expected that the high-luminance stimulus would cause the most fixation-related inhibition, and thus after the S–to–high-luminance change at fixation, saccadic latencies would be longer than those in the high-to-S condition. The key question was whether the difference in fixation-related inhibition between the high-luminance and S-cone stimuli was greater or smaller than the inhibition difference between the high- and low-luminance stimuli. If it was smaller, then the S-cone stimulus would seem to create more fixation-related inhibition than the low-luminance stimulus (Fig. 3Ai). If it was larger, the S-cone stimulus would seem to create less inhibition than the low-luminance stimulus (Fig. 3Aii).
|
Methods
PARTICIPANTS. Eight subjects participated (three male, five female, ages 22–24). All subjects had normal or corrected to normal acuity and normal color vision, and none had served in experiment 1 or experiment 2.
PROCEDURE. All aspects of the apparatus, calibration, procedure, and analysis were identical to experiments 1 and 2, except where specified below. In all trials there was a change in the fixation stimulus 200 ms before the onset of the saccade target. There were two types of block: in luminance blocks, the fixation stimulus changed from high luminance (mean 35 cdm–2, randomly flickering over a range of 1.0 cdm–2; see Methods of EXPERIMENT 1) to low luminance (mean 28 cdm–2) or vice versa; in S-cone blocks, the fixation stimulus changed from high luminance (mean 35 cdm–2) to an S-cone stimulus (mean 25 cdm–2), or vice versa. As before, the S-cone stimuli were individually calibrated for each participant to differ from the flickering gray background squares only in the S-cone color direction, creating an 80% increase in S-cone signal. Participants performed 120 trials per block, and three blocks of each type. Again, participants were informed that the fixation square would flicker and change, but were instructed to ignore this and simply detect the target and move their eyes to it as quickly as possible.
Results
Figure 4 A shows the results for experiment 3A. As expected, fixation changes from low to high luminance caused longer saccadic latency than changes from high to low luminance—the difference was 29 ms. Also as expected, fixation changes from S-cone to high-luminance stimuli caused longer saccadic latency than changes from high-luminance to S-cone stimuli, although crucially the difference, 16 ms, was smaller than the difference produced by low- and high-luminance stimuli [ANOVA interaction: F(1,7) = 6.5, P < 0.05]. This indicates that the S-cone stimuli caused more fixation-related inhibition than the low-luminance stimuli. That is, the "equivalent-luminance-contrast" of the S-cone stimuli was greater than that of the low-luminance stimuli.
|
|
EXPERIMENT 3B |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Methods
PARTICIPANTS. The same eight subjects in experiment 3A participated in experiment 3B.
PROCEDURE. All aspects of the apparatus, calibration, procedure, and analysis were identical to experiment 3A, except that the fixation stimulus changed from low-luminance (mean 28 cdm–2) to the S-cone stimulus, or vice versa. Participants performed three blocks of 120 trials each. Again, participants were informed that the fixation square would flicker and change, but to try to ignore this and simply detect the target and move their eyes to it as quickly as possible.
Results
Figure 4B shows the results for experiment 3B, which are opposite to those predicted by the results of experiment 3A (see Figs. 3, A and B and 4A). Fixation changes from S-cone stimuli to low-luminance stimuli produced saccadic latencies that were 13 ms longer than changes from low luminance to S cone (t = 3.1, df = 7, P < 0.05). In other words, the low-luminance stimuli created more fixation-related inhibition than did the S-cone stimuli, despite the fact that in experiment 3A, the S-cone stimuli appeared to be more similar to the high-luminance stimuli than did the low-luminance stimuli. This result is inconsistent with the idea that the S-cone stimuli achieve their inhibitory effect by the same route as that of the luminance stimuli. If this were the case there should be a consistent relationship between the effects of each stimulus.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Dacey et al. 2003) and invisible to the retinotectal and magnocellular sensory pathways. To be sure that the S-cone stimuli cause fixation-related inhibition by a route that is really distinct from the route by which luminance stimuli cause fixation-related inhibition, we conducted experiment 3. We found that S-cone stimuli do not have a consistent "equivalent-luminance-contrast"—that is, they do not consistently behave like any luminance stimulus of a fixed, although unknown, luminance contrast. S-cone stimuli cannot be ranked in efficacy alongside luminance stimuli because the S-cone stimulus may in one condition appear to produce more fixation-related inhibition than a certain luminance stimulus (experiment 3A), but in another condition produce less inhibition (experiment 3B). These apparently paradoxical results can be explained if the luminance and S-cone signals cause inhibition by separate pathways and the luminance pathway is highly nonlinear. The nonlinearity means that relatively small luminance changes can have nearly as much effect as larger luminance changes and thus the signal from the luminance pathway may be quite similar in all three of our conditions containing luminance increments or onsets: the low–to–high-luminance condition, S-cone–to–high-luminance condition, and the S-cone–to–low-luminance condition. Any large difference in effect between these conditions must then be explained by signals in another pathway. In this case offset signals in the S-cone pathway oppose the effect of the luminance onset in two of the conditions, but not for the low–to–high-luminance change, which had the greatest effect. Thus to explain the results there must be at least two pathways at play with different response properties.
We suggest that many pathways and more than one mechanism may be involved in fixation-related inhibition. Because many studies have linked fixation-related inhibition to inhibitory connections in the superior colliculus (for a review see Munoz et al. 2000), it seems likely that the direct retinotectal pathway would contribute. The magnocellular pathway is also known to be essential for collicular function (Schiller et al. 1979) and is likely to contribute both through direct collicular projections from primary visual cortex and through projections via MT/MST and the cortical eye fields (for a review see Krauzlis 2005). In addition the S-cone stimuli contribute by a different cortical route, which may perhaps be via V4 and the posterior part of the inferotemporal cortex (area TEO), and from there to the frontal eye fields (FEF) (Schall et al. 1995).
One possibility is that all fixation-related inhibition still occurs in the superior colliculus and that the S-cone signal is simply delivered to the SC by a less-direct cortical pathway than the luminance signal. However, if this were the case, we would expect there to be chromatic sensory responses in the SC, which have not been found. The initial sensory activity in SC cells does not distinguish colors, and differential activity develops only if color-defined stimuli are relevant for saccades (Marrocco and Li 1977; Ottes et al. 1987). It seems that when eye-movement targets are defined by properties to which the SC is not directly sensitive, spatial information about these targets is conveyed to the SC only after some target selection has occurred elsewhere.
Oculomotor target selection processes are likely to involve the FEF, which contain both visually driven cells and oculomotor cells, and have inhibitory connections between cells with response fields in different parts of the visual field (e.g., Hanes et al. 1998; Schall 1997; Schiller and Tehovnik 2003; Schlag-Rey et al. 1992; Thompson and Bichot 2005). In general, patients with lesions encompassing the FEF seem to have difficulty inhibiting saccades to inappropriate stimuli (Braun et al. 1992; Guitton et al. 1985; Machado and Rafal 2004a). More specifically for fixation-related inhibition, such patients can show abnormally small fixation-offset effects on their contralesional side (Machado and Rafal 2004b). Stimulation in the FEF can suppress saccades (e.g., Izawa et al. 2005) and some cells show fixation-related activity (e.g., Hanes et al. 1998), whereas activity in other cells correlates with disengagement from fixation (Dias and Bruce 1994). Thus the FEF may mediate the gap effect upstream or independently of the SC. Alternatively, rather than being responsible for creating automatic fixation-related inhibition, the FEF has been suggested to provide goal-directed modulation of a collicular mechanism. For example, Machado and Rafal (2004b) found that informatively precueing target location modulated the gap effect in healthy participants but not in FEF lesion patients. Thus FEF involvement in fixation-related inhibition seems likely, although its exact role remains uncertain.
However, the sensory cells in FEF, like those in SC, also seem not to be selective for color or form (e.g., Stuphorn and Schall 2002; Thompson and Bichot 2005), so target selection must involve additional areas that are selective for these properties. We suggest that S-cone fixation stimuli create inhibition in a cortical target selection process incorporating areas such as TEO as well as the cortical eye fields. Thus overall, although it is well established that inhibitory connections in the SC are likely to cause fixation-related inhibition, and that the FEF may play an important role, our results imply that there is also a further cortical inhibitory mechanism. For this mechanism to produce S-cone fixation-related inhibition when warning effects are equated (e.g., experiment 2), this mechanism must be specifically related to fixation stimuli rather than generalized warning and motor readiness.
The wider implications of this are that low-level effects thought to represent collicular function may also represent cortical function and, furthermore, phenomena interpreted as cortical control over colliculus may instead represent interactions within cortex. For example, manipulating the strategic set of saccade expectancy, using warning tones, or by cueing the target location or increasing the proportion of trials requiring saccades, can reduce the fixation-offset effect (Machado and Rafal 2000a,b; Rafal et al. 2000; Taylor et al. 1998) and this effect can be disrupted in patients with frontal lesions that include the FEF (Machado and Rafal 2004b). This has been interpreted as an example of how FEF can exert strategic control over the SC, with the lesions disrupting such control. However, this assumes that the fixation-offset effect represents collicular function. If there is also a cortical source, the strategic control and the effect of the lesions may both reflect intracortical interactions rather than cortico-collicular modulation. Likewise, we would not need to assume any collicular imbalance to explain asymmetrical fixation-offset effects in patients with unilateral cortical damage (e.g., Machado and Rafal 2004a,b).
One further point for discussion is whether the fixation-offset effect is caused by the absence of a fixation stimulus during saccade preparation, as is often implied by descriptions of the effect, or whether it is caused by the occurrence of the offset itself, as might be expected from sensory pathways with nonlinear and transient response properties. In other words, is it the nature of the fixation stimulus present during saccade preparation that is important or is it the nature of the previous stimulus change that is important? Pratt et al. (2000) found that a 20-pixel reduction in fixation size had an appreciably different effect depending on whether any fixation stimulus remained: removing a 20-pixel fixation stimulus produced much shorter latencies than reducing a 40-pixel fixation stimulus to 20 pixels, and the latter had almost no effect when paired with a warning tone, whereas the former still had a large effect. This may be taken to suggest that the presence or absence of a remaining fixation stimulus is the crucial factor rather than the size of the visual change. However, in this example it was absolute change that was equated, whereas relative change may be more relevant. Our results in experiment 3 imply that the nature of the stimulus change at fixation is important in addition to the nature of the stimulus present when a saccade is required. For example, compare the high–to–low-luminance condition with the S-cone–to–low-luminance condition. In both conditions the same fixation stimulus was present by the time the saccade target appeared, but latency was greater in the latter condition than the former, presumably because the latter condition contained a signal increment at fixation rather than a decrement.
In conclusion, the sensory-motor phenomenon of fixation-related inhibition has been widely studied in the context of motor control, but the pathways that supply the sensory information have not previously been explicitly tested. Although fixation-related inhibition has been strongly associated with the SC, our data show that it does not rely on signals in either of the sensory pathways on which visual input to the SC is thought to depend. Further, we suggest that S-cone fixation-related inhibition is generated in a cortical mechanism that is not simply within the FEF. These results may necessitate some reassessment of previous data interpreted in the context of cortico-collicular modulation.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
1 The Supplementary Material for this article (two movies) is available online at http://jn.physiology.org/cgi/content/full/000781.2005/DC1. 
Address for reprint requests and other correspondence: P. Sumner, Dept. of Visual Neuroscience (Room 10L15a), Division of Neuroscience, Faculty of Medicine, Imperial College London, St Dunstan's Road, London W6 8RP, UK (E-mail: p.sumner{at}imperial.ac.uk)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENT 1EXPERIMENT 2EXPERIMENT 3AEXPERIMENT 3BDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Braun D, Weber H, Mergner T, and Schultemonting J. Saccadic reaction-times in patients with frontal and parietal lesions. Brain 115: 1359–1386, 1992.
Brindley GS, Du Croz JJ, and Rushton WAH. The flicker fusion frequency of the blue-sensitive mechanism of colour vision. J Physiol 183: 497–500, 1966.
Calkins DJ. Seeing with S cones. Prog Retin Eye Res 20: 255–287, 2001.[CrossRef][ISI][Medline]
Chatterjee S and Callaway EM. S cone contributions to the magnocellular visual pathway in macaque monkey. Neuron 35: 1135–1146, 2002.[CrossRef][ISI][Medline]
Chatterjee S and Callaway EM. Parallel colour-opponent pathways to primary visual cortex. Nature 426: 668–671, 2003.[CrossRef][Medline]
Dacey DM, Peterson BB, Robinson FR, and Gamlin PD. Fireworks in the primate retina: in vitro photodynamics reveals diverse LGN-projecting ganglion cell types. Neuron 37: 15–27, 2003.[CrossRef][ISI][Medline]
de Monasterio FM. Properties of ganglion cells with atypical receptive-field organisation in the retina of macaques. J Neurophysiol 41: 1435–1449, 1978.
Dias EC and Bruce CJ. Physiological correlate of fixation disengagement in the primate's frontal eye field. J Neurophysiol 72: 2532–2537, 1994.
Dorris MC, Pare M, and Munoz DP. Neuronal activity in monkey superior colliculus related to the initiation of saccadic eye movements. J Neurosci 17: 8566–8579, 1997.
Everling S and Munoz DP. Neuronal correlates for preparatory set associated with pro-saccades and anti-saccades in the primate frontal eye field. J Neurosci 20: 387–400, 2000.
Fendrich R, Demirel S, and Danziger S. The oculomotor gap effect without a foveal fixation point. Vision Res 39: 833–841, 1999.[CrossRef][ISI][Medline]
Gouras P. Identification of cone mechanisms in monkey ganglion cells. J Physiol 199: 533–547, 1968.
Guitton D, Buchtel HA, and Douglas RM. Frontal lobe lesions in man cause difficulties in suppressing reflexive glances and in generating goal-directed saccades. Exp Brain Res 58: 455–472, 1985.[ISI][Medline]
Hanes DP, Patterson WF, and Schall JD. Role of frontal eye fields in countermanding saccades: visual, movement, and fixation activity. J Neurophysiol 79: 817–834, 1998.
Hess WR, Burgi S, and Bucher V. Motor function of tectal and tegmental area. Monatsschr Psychiatr Neurol 112: 1–52, 1946.
Hikosaka O and Wurtz RH. The basal ganglia. In: The Neurobiology of Saccadic Eye Movements, edited by Wurtz RH and Goldberg ME. Amsterdam: Elsevier, 1989, p. 257–282.
Isa T. Intrinsic processing in the mammalian superior colliculus. Curr Opin Neurobiol 12: 668–677, 2002.[CrossRef][ISI][Medline]
Izawa Y, Suzuki H, and Shinoda Y. Initiation and suppression of saccades by the frontal eye field in the monkey. Ann NY Acad Sci 1039: 220–231, 2005.
Kaplan E and Shapley RM. The primate retina contains two types of ganglion cells, with high and low contrast sensitivity. Proc Natl Acad Sci USA 83: 2755–2757, 1986.
Kingstone A and Klein RM. Visual offsets facilitate saccadic latency: does predisengagement of visuospatial attention mediate this gap effect? J Exp Psychol Hum Percept Perform 19: 1251–1265, 1993.[CrossRef][ISI][Medline]
Klein RM, Taylor TL, and Kingstone A. Against a role for attentional disengagement in the gap effect: a friendly amendment to Tam and Stelmach (1993). Percept Psychophys 57: 573–577, 1995.[ISI][Medline]
Krauzlis RJ. The control of voluntary eye movements: new perspectives. Neuroscientist 11: 124–137, 2005.[Abstract]
Machado L and Rafal RD.Control of eye movement reflexes. Exp Brain Res 135: 73–80, 2000a.[CrossRef][ISI][Medline]
Machado L and Rafal RD. Strategic control over saccadic eye movements: studies of the fixation offset effect. Percept Psychophys 62: 1236–1242, 2000b.[ISI][Medline]
Machado L and Rafal RD. Control of fixation and saccades during an anti-saccade task: an investigation in humans with chronic lesions of oculomotor cortex. Exp Brain Res 156: 55–63, 2004a.[CrossRef][ISI][Medline]
Machado L and Rafal RD. Control of fixation and saccades in humans with chronic lesions of oculomotor cortex. Neuropsychology 18: 115–123, 2004b.[CrossRef][ISI][Medline]
Marrocco RT and Li RH. Monkey superior colliculus: properties of single cells and their afferent inputs. J Neurophysiol 40: 844–860, 1977.
Morris JS, Ohman A, and Dolan RJ. A subcortical pathway to the right amygdala mediating "unseen " fear. Proc Natl Acad Sci USA 96: 1680–1685, 1999.
Munoz DP, Dorris MC, Pare M, and Everling S. On your mark, get set: brainstem circuitry underlying saccadic initiation. Can J Physiol Pharmacol 78: 934–944, 2000.[CrossRef][ISI][Medline]
Munoz DP and Fecteau JH. Vying for dominance: dynamic interactions control visual fixation and saccadic initiation in the superior colliculus. Prog Brain Res 140: 3–19, 2002.[ISI][Medline]
Munoz DP and Istvan PJ. Lateral inhibitory interactions in the intermediate layers of the monkey superior colliculus. J Neurophysiol 79: 1193–1209, 1998.
Munoz DP and Wurtz RH. Role of the rostral superior colliculus in active visual fixation and execution of express saccades. J Neurophysiol 67: 1000–1002, 1992.
Ottes FP, Van Gisbergen JAM, and Eggermont JJ. Collicular involvement in a saccadic colour discrimination task. Exp Brain Res 66: 465–478, 1987.[CrossRef][ISI][Medline]
Pare M and Wurtz RH. Monkey posterior parietal cortex neurons antidromically activated from superior colliculus. J Neurophysiol 78: 3493–3497, 1997.
Pratt J, Bekkering H, and Leung M. Estimating the components of the gap effect. Exp Brain Res 130: 258–263, 2000.[CrossRef][ISI][Medline]
Pugh EN and Mollon JD. A theory of the p1 and p3 color mechanisms of Stiles. Vision Res 19: 2923–2312, 1979.
Rafal R, Machado L, Ro T, and Ingle H. Looking forward to looking: saccade preparation and control of the visual grasp reflex. In: Control of Cognitive Processes: Attention and Performance, edited by Monsell S and Driver J. Cambridge, MA: MIT Press, 2000, p. 155–174.
Rafal RD, Henik A, and Smith J. Extrageniculate contributions to reflex visual orienting in normal humans: a temporal hemifield advantage. J Cogn Neurosci 3: 322–328, 1991.
Rafal RD, Smith J, Krantz J, Cohen A, and Brennan C. Extrageniculate vision in hemianopic humans: saccade inhibition by signals in the blind field. Science 250: 118–121, 1990.
Reuter-Lorenz PA, Hughes HC, and Fendrich R. The reduction of saccadic latency by prior offset of the fixation point: an analysis of the gap effect. Percept Psychophys 49: 167–175, 1991.[ISI][Medline]
Reuter-Lorenz PA, Oonk HM, Barnes LL, and Hughes HC. Effects of warning signals and fixation point offsets on the latencies of pro- versus antisaccades: implications for an interpretation of the gap effect. Exp Brain Res 103: 287–293, 1995.[ISI][Medline]
Saslow MG. Effects of components of displacement-step stimuli upon latency for saccadic eye movements. J Opt Soc Am 57: 1024–1029, 1967.[Medline]
Schall JD. Visuomotor areas of the frontal lobe. Cereb Cortex 12: 527–638, 1997.
Schall JD, Morel A, King DJ, and Bullier J. Topography of visual-cortex connections with frontal eye field in macaque—convergence and segregation of processing streams. J Neurosci 15: 4464–4487, 1995.[Abstract]
Schiller PH and Malpeli JG. Properties and tectal projections of monkey retinal ganglion cells. J Neurophysiol 40: 428–445, 1977.
