Effects of TMS over the Right Prefrontal Cortex on Latency of Saccades and Convergence

Abstract

purpose. The prefrontal cortex (PFC) is known to inhibit unwanted saccades through its connections to the superior colliculus (SC). Indeed, transcranial magnetic stimulation (TMS) of the PFC decreases saccade latency by increasing the rate of express saccades. This study examined whether a similar phenomenon exists for vergence.

methods. In a gap paradigm, six healthy subjects were asked to look at LEDs placed in a horizontal plane and to make lateral saccades, pure convergence along the median plane, and combined eye movements. Eye movements were recorded binocularly. TMS was applied over the right (r)PFC synchronously with the onset of the target. In a control condition, TMS was applied over the motor cortex (MC).

results. TMS over the MC had no effect on the latency of any type of eye movements. In contrast, TMS over the rPFC (1) decreased significantly (P = 0.00367) the latency of contralateral pure saccades, (2) had no effect on the latency of pure convergence, (3) and caused a mild decrease in the latency of both the saccadic and the convergence components of combined eye movements, and the effect was bilateral. Decreased latencies were mainly due to an increase of the rate of express movements.

conclusions. The inhibition exerted by PFC over SC and preventing express movements from occurring is presumably a saccade-specific mechanism. When the saccade is combined with convergence, the express triggering can be transferred to a certain extent to the convergence.

Many studies have been conducted to understand behavioral aspects and the physiological substrate of the human saccadic eye movements. The latency of a saccade, which corresponds to the time between the appearance of a target in the periphery and the initiation of the movement, is a sensitive cognitive indicator. Indeed, during this interval, several processes are believed to occur, such as disengagement of visual attention, disengagement of ocular fixation, computation of parameters of the movement, and decision to move. By manipulating experimentally the temporal relationship between the extinction of the fixation point and the onset of the target, it is possible to favor different types of saccades on the basis of the length of their latency: saccades with latencies between 80 and 120 ms called “express,” fast regular saccades (121–180 ms), and slow regular saccades (181–400 ms). For example, in the gap paradigm, a temporal interval is introduced between the extinction of the fixation point and the appearance of the target. It was demonstrated that the gap paradigm, by comparison with the simultaneous condition, in which the target switches on synchronously with the extinction of the fixation point, favors the occurrence of express saccades. ^{1 2} The factors contributing to the gap effect are still controversial. To some investigators, the fixation offset occurring before the target onset provides a warning signal that facilitates either the detection of the target or the response to it. ^{1 2} Kingstone and Klein ³ pointed out that the fixation offset provides a disengagement of attention. Dorris et al. ⁴ advocated the predictive computation of the metrics of the saccade. Findlay and Walker ⁵ proposed that the gap acts on the competition between excitatory and inhibitory mechanisms: the absence of a foveal stimulus (induced by the temporal gap) strengthens the shift of fixation, thereby facilitating the onset of the saccade. 

Cortical control of reflexive or voluntary saccades has been well studied. Different subcircuits are believed to control the reflexive and volitional saccades. ^{6 7} For instance, short-latency express saccades are believed to be generated by a short occipitoparieto-collicular pathway. In support of this view, Schiller et al. ⁸ demonstrated in monkeys that the ablation of the SC or an injection of muscimol into the SC abolishes the occurrence of express saccades. The prefrontal cortex is believed to inhibit reflexive unwanted saccades by a direct pathway to the superior colliculus. ^{6 7} Transcranial magnetic stimulation (TMS) can be used to investigate cortical control of eye movements. TMS consists of a high-current pulse generator producing a current of high intensity that flows through a stimulating coil and generates a magnetic pulse. When applied on the human scalp, this pulse induces currents in the electrically conductive cortex, resulting in a temporary disruption of the function at this location. ⁹ Müri et al. ¹⁰ used a gap saccade paradigm and applied TMS over the right prefrontal cortex. TMS increased the proportion of contralateral express saccades, whereas it had no effect on the latencies of ipsilateral saccades. 

Vergence eye movements correspond to a change of gaze in depth and require movements of the eyes in opposite directions. Targets aligned along the median plane require a pure vergence, whereas targets located in different depth and direction require combined saccade–vergence movements. Until today, only a few studies have been conducted to investigate the influence of the gap paradigm on the latency of vergence. Takagi et al. ¹¹ found significantly decreased latency of both convergence and divergence in a gap task by comparison with the simultaneous task or an overlap task (a condition in which the fixation LED stays on when the target LED is switched on). This indicates the presence of a gap effect for vergence as well. The gap effect, however, was smaller for vergence (17 ms) than for saccades (41 ms). It is important to note that, for vergence, the gap effect was not mediated by an increase of express latencies. The express vergence is defined arbitrarily as having the same latency as express saccades. This study was limited to three subjects and the question of the strength of the gap effect for vergence should be investigated further. 

The subcortical substrate of vergence has been widely examined in animals. ^{12 13 14} Cortical regions involved in the control of vergence have also been studied in nonhuman primates by a number of recent studies. ^{15 16 17} In contrast, in humans, cortical regions involved in the control of vergence are almost unexplored. Hasebe et al. ¹⁸ used positron emission tomography (PET) in normal humans and studied disparity-driven vergence movements. They reported bilateral activation of the temporo-occipital junction, activation of the left inferior parietal lobule, and activation of the right fusiform gyrus. Kapoula et al. ¹⁹ used a gap paradigm and reported that TMS over the right posterior parietal cortex (PPC) of normal humans causes a prolongation of the latency of both saccades and vergence eye movements, suggesting that the PPC is involved in the triggering of both these movements. 

The specific question investigated in this study was whether the prefrontal cortex (PFC) exerts an inhibitory control on the triggering of vergence and of combined eye movements similar to that of saccades. To our knowledge, this topic has not been explored in animals or humans. Most of the recent studies on the combined eye movements deal with the interaction between saccade and vergence—namely, the acceleration of vergence during a saccade. They address questions related to the oculomotor subsystems at the brain stem level (saccadic or vergence), producing the unequally sized saccades that occur during combined gaze shifts in direction and in depth. ^{20 21 22} To our knowledge, at the behavioral level, the latency of combined eye movements and the effect of the gap has been studied in only three subjects in a study by Takagi et al. ¹¹ They found that the gap effect for combined movements was the same as for saccade and vergence studied alone. Thus, the study of latency of combined eye movements and of the effect of TMS is another unexplored area. 

The goal of the present study was twofold: to reexamine the effect of the gap paradigm on vergence and combined eye movements (involving both vergence and versional movements) in six subjects and to examine the effect of TMS applied on the right (r)PFC on the latency of saccade, vergence, and combined eye movements. We expected TMS to increase the rate of express responses, but only for the type of movements for which express latencies occur naturally in the gap paradigm. 

Methods

Subjects

Six healthy right-handed subjects, three women and three men, participated in the main experiment (TMS over the rPFC); two subjects were naive, having never participated in eye-movement studies. The mean age was 34.7 years (range, 22–42). Four of them participated in a control experiment (TMS over the motor cortex [MC]). Subjects had no ophthalmologic or neurologic symptoms. Binocular vision, assessed by the TNO Random-Dot Stereotest (Richmond Products, Boca Raton, FL), was normal. All subjects gave their written consent to participation in the study, which was approved by the ethics committee and was consistent with the Declaration of Helsinki. 

TMS Device and Location of Cortical Regions

Single-pulse TMS was applied by a magnetic stimulator (model 200; MagStim, Whitland, UK) with a nonfocal stimulation coil of 90-mm diameter. In the main experiment, the PFC of the right hemisphere was stimulated. For that, in a first step, the motor hand area was defined by motor threshold stimulation responsible for slight muscle contractions in the contralateral left-hand muscles. The coil was then moved 5 cm anteriorly, which corresponds to location of the PFC as reported by Pascual-Leone and Hallett. ²³ For the control experiment, the coil was placed at the location of the vertex. Each region was stimulated with 80% of the machine capacity, which corresponds to approximately 10% to 30% above individual motor thresholds and only a segment of the coil was placed tangentially on the scalp, so as to obtain a more focal stimulation of the area. The noise generated by the coil at the time of stimulation was the same in both the experimental and the control conditions. The length of the TMS pulse was 1 ms. 

Visual Display

Each subject faced a horizontal plane with six red LEDs arranged on two isovergence circles. Three LEDs were placed: one at the center and two at ±20° at a distance of 150 cm from the subject’s eyes. The required mean angle of convergence for fixating any of these LEDs was 2.3°. The other three LEDs were placed at a distance of 20 cm: one at the center and two at ±20°. The mean angle of convergence was 17.1° (Fig. 1A) . 

Oculomotor Paradigm

The gap paradigm was used to elicit short-latency reflexive eye movements. For each trial, the fixation LED was always at the far center LED. After a 2500-ms fixation period, this central LED was switched off, and a target LED appeared 200 ms later. The target LED stayed on for 1500 ms. The TMS pulse occurred at the same time as the appearance of the target (Fig. 1B) . 

Three types of movements were studied: pure saccades to the left or to the right, pure convergence along the median plane, and combined movements to the left or to the right. The required saccade amplitude was 20° for both pure saccades and combined movements. The required convergence movement was 14.8°, both along the median plane and for the combined gaze shift. Each block contained 40 trials. In each block, the three types of eye movements were interleaved randomly. Herein, the term “pure” (saccade or vergence) is used in the technical sense, to describe the properties of the visual stimuli. Because the lateral LEDs were placed on an isovergence circle, these stimuli called for a pure saccade. Conversely, targets aligned along the median plane called for pure vergence. Nevertheless, as will be mentioned in the Results section, this does not imply that subjects effectively performed pure movements—that is, saccades without transient vergence and vergence without saccadic intrusions (Fig. 2) . Divergence was not studied, to keep the duration of the session short. 

The instruction given to the subjects was to look at the target LED as accurately and as rapidly as possible. The two naive subjects performed three blocks for the two conditions: without TMS and with TMS over the rPFC. Experienced subjects performed between three and five blocks in both conditions. Four of the subjects performed a control experiment (in three blocks) in which TMS was delivered over the MC. For each condition with TMS, TMS was performed in 95% of the trials. A few trials with no TMS were introduced to reduce expectancy effects. We found no difference between latencies in that 5% of trials and in those in the blocks without TMS. Nevertheless, the results of the 5% of non-TMS trials were excluded from the analysis. 

Eye-Movement Recording

Eye movements were recorded binocularly with an eye-tracking device (IRIS; Skalar, Delft, The Netherlands). Calibration was conducted before and after each block of trials. A back head support and a medical collar stabilized the head. Eye position signals were low-pass filtered with a cutoff frequency of 200 Hz and digitized with a 12-bit analog-to-digital converter. Each channel was sampled at 500 Hz. 

Analysis of Eye Movement Data

After calibration of the individual eye signals, we calculated two derived signals (Fig. 2) : the saccadic or conjugate signal, corresponding to the average of the two eyes signals [(left + right eye position)/2]; and the vergence signal, which was the difference between the two signals (left − right eye position), so that convergence was positive and divergence was negative. Analysis was conducted on these two derived signals. Eye velocity was computed with a digital filter designed to differentiate up to 64 Hz. Markers for the initiation of saccade were placed mostly automatically. The onset was taken at the point when the conjugate eye velocity exceeded 45 deg/sec. For the initiation of vergence component, markers were placed at the moment when the vergence velocity exceeded 5 deg/sec, which was the criterion also used by others. ^{11 19} For all eye movements, latency was measured as the time between the onset of the target and the initiation of the movement. Because the main hypothesis of this study concerned the latency of the movements, only latencies were analyzed and presented in the Results section. 

Eye movements in the wrong direction or contaminated by blinks were rejected (3.79% of all trials; range, 0.5%–12.5%). The occurrence of blinks was not related to TMS but was subject dependent. Eye movements with latencies shorter than 80 ms (0.57% of all items; range, 0%–1.15%) or longer than 400 ms (0.78% of all items; range, 0%–4.33%) were also rejected. In total, 5.14% (range, 0.5%–11.4%) of original recorded items were not taken into consideration. 

For all movements, both individual and general descriptive and parametric inferential analysis (Student’s t-test or analysis of variance) was conducted to evaluate both the disparities between subjects and the global tendency. 

Figure 2 shows examples of the three types of eye movements. Note the presence of a transient vergence during a pure saccade (Fig. 2A) and the presence of saccadic intrusions during a pure convergence (Fig. 2B) . These phenomena are consistent with prior reports. ^{24 25}  

Results

Pure Eye Movements

Natural Latency of Pure Movements.

For pure saccades, results are presented in Figure 3 . Shown is the latency of pure leftward saccades (mean latency, 174 ms, range, 146–223; Fig. 3A , left) and rightward saccades (mean latency, 170 ms, range, 133–211; Fig. 3A , right). As expected, the gap paradigm induced express saccades with latencies between 80 and 120 ms. The group mean rate of express latencies was 3.60% (range, 0%–10%) for saccades to the left, and 8.33% (range, 0%–22.73%) for saccades to the right. There was no difference between left and right (t < 1). 

Figure 3B shows the baseline latency of pure convergence (mean, 257 ms; range, 212–303). Convergence latency was significantly longer than saccade latency in either direction (saccades to the left, F_1,10 = 17.3, P < 0.01; saccades to the right, F_1,10 = 17.7, P < 0.01). For pure convergence, no express latency was observed. 

Effect of TMS over rPFC on the Latency of Pure Movements.

Pure Saccades. Figure 4A shows differences in latency between the experimental condition (TMS over rPFC) and the reference condition (without TMS) for pure saccades. For pure saccades to the left, the TMS over the rPFC, was responsible for a change in latency of −25 ms. The group analysis showed a significant difference (F_1,5 = 26.3, P < 0.01). On the contrary, no statistical difference was found between the experimental and reference conditions for rightward saccades (mean difference, −7 ms; F < 1). 

TMS over the rPFC increased the rate of express contralateral saccades for all subjects. The group rate increased from 3.60% to 17.18%. The difference between the rates in the two conditions was marginally significant (t ₁₀ = 2.13, P = 0.059). To determine whether the decrease in latency was due only to the increased rate of express saccades, we conducted a further analysis from which was excluded the additional express saccades (14.18%, i.e., 17.78%–3.60%). Thus, in the experimental condition, there were again 3.60% express saccades, similar to the reference condition. After such exclusion, the mean latency difference between the two conditions was no longer statistically significant (F < 1). This confirmed that it was indeed the proportion of additional express saccades in the TMS condition that was responsible for the latency decrease. For ipsilateral saccades, TMS over rPFC had no significant effect on the occurrence of express saccades. The group mean rate of express latencies was 8.33% and 7.46%, respectively, without and with TMS. 

When the MC control condition was subtracted from reference condition (without TMS), the mean difference over the group was +7 ms for both leftward and rightward saccades (data not shown). There was no difference, either for leftward saccades (F_2,13 = 1.10, P = 0.360) or for rightward saccades (F < 1). When the MC control condition was subtracted from the experimental condition (rPFC), there was a difference only for contralateral leftward saccades (−11 ms for the group), which was comparable to that observed between the experimental condition and the reference condition. 

Pure Convergence.

Figure 4B shows the difference between the experimental condition (TMS over rPFC) and the reference condition (without TMS) for pure convergence. The group mean difference was −3 ms (range, −16 to +9). There was no significant difference individually or group. No express convergence latencies were observed in the experimental condition. 

We also compared convergence latency from the control condition (TMS over MC) with the reference condition (without TMS) and, by contrast, convergence latency in the two TMS conditions (TMS over MC and TMS over rPFC). No statistically significant difference, either individual or for the group of subjects, was found between any of these conditions. 

To summarize, our use of the gap paradigm elicited express eye movements only for saccades in both directions, but not for pure convergence. When the TMS was applied over rPFC, increasing the occurrence of express movements significantly reduced latency of contralateral saccades. In contrast, TMS over rPFC had no effect on latencies of ipsilateral saccades or on latencies of convergence along the median plane. 

Combined Eye Movements

Natural Latency of Combined Movements.

Saccadic Component. Figure 5A shows the latencies of the saccadic component of combined eye movements. The mean latency was 212 ms for both leftward (range, 187–261) and rightward (range, 183–245) movements. By comparison with pure saccades, the group mean latencies of the saccadic component of combined movements were significantly longer than those of pure leftward saccades (F_1,5 = 23.5, P < 0.01) and those of pure rightward saccades (F_1,5 = 28.1, P < 0.01). Contrary to observations for pure saccades, the gap paradigm did not elicit express latencies for the saccadic component of combined movements. 

Convergence Component.

Data are shown in Fig. 5B . The mean latencies were 261 ms for the convergence component of leftward movements (range, 232–339) and 250 ms for that of rightward movements (range, 226–307). No significant difference was found between combined convergence and pure convergence latencies for leftward (F_1,5 = 2.83, P = 0.153) or for rightward movements (F < 1). No express latency was observed for the convergence component of combined eye movements. 

To summarize, combined eye movements showed no express latency. Latency of the saccade was longer than that of corresponding pure saccades. Latency of convergence was always long and similar for pure and combined movements. 

Effect of TMS over rPFC on the Latency of Combined Movements.

Saccadic Component. Figure 6A shows the difference in latency between the experimental condition (TMS over rPFC) and the reference condition (without TMS). For combined movements to the left, the group mean difference was −13 ms but it failed to reach significance (F_1,5 = 5.13, P = 0.073). For the saccadic component of combined eye movements to the right, the group mean decrease was unexpected: −19 ms, which was statistically significant (F_1,5 = 7.58, P < 0.05). 

In the experimental condition, express latencies appeared for the saccadic component at the following rates: 6.85% (range, 0%–33.33%) for leftward movements, and 8.97% (range, 0%–18.18%) for rightward movements. To determine whether the mean decrease in latency (−13 ms for the left and −19 ms for the right) was due entirely to the increase of the rate of express saccades, we conducted a further analysis in which were excluded from the experimental condition all express latencies (recall that in the reference condition, there were no express latencies at all). For leftward movements, the mean difference decreased from −13 to −4 ms, and, for rightward movements, the mean difference decreased from −19 to −10 ms. Neither of these differences was statistically significant. Thus, it can be concluded that the decrease in latency of saccades of combined movements induced by TMS over rPFC was due to the appearance of express saccades for both leftward and rightward combined movements. 

Convergence Component.

Figure 6B shows the difference in latency of the convergence component of combined eye movements between the experimental (TMS over rPFC) and reference (without TMS) conditions. For the convergence combined with a saccade to the left, the group mean decrease was −20 ms and was statistically significant (F_1,5 = 9.55, P < 0.05). For the convergence component combined with a saccade to the right, the group mean difference was also −20 ms, but failed to reach statistical significance (F_1,5 = 5.07, P = 0.074) because of larger intersubject variability. Interindividual significant differences were found in three of six subjects. 

Although we found no express movements in the reference condition without TMS, the new finding is that in the experimental condition (TMS over rPFC) express convergence was observed. The rates were 2.70% (range, 0%–22.22%) for convergence combined with leftward saccades and 1.35% (range, 0%–7.69%) for convergence combined with rightward saccades. For leftward movements, when the express latencies of the convergence component were excluded, the mean difference decreased from −20 to −13 ms. This difference was no longer significant (F < 1), suggesting that the decreasing in latency induced by the TMS was due to the appearance of express latencies. For rightward movements, the mean difference decreased from −20 to −19 ms after exclusion of express latencies. Even though not statistically significant, this result suggests for rightward movements the involvement of another possible mechanism—for example, an increase of the occurrence of fast regular latencies (121–180 ms). 

TMS over the MC (control condition) had no effect on the latency of any component of the combined movements. 

In summary, the gap paradigm did not elicit express latencies for either the saccadic or the convergence component of the combined eye movements. When TMS was applied over rPFC, the latencies of both these components were reduced. The decrease of latencies was due to the occurrence of express movements for the saccadic component in either direction and also for the convergence component of leftward movements. 

Discussion

The main findings of this study are the following: (1) The gap paradigm elicited express latencies for pure saccades only, but not for convergence, either pure or combined with a saccade. (2) TMS over the rPFC reduced the latency of contralateral pure saccades. (3) TMS over the rPFC had no effect on the latency of pure convergence. (4) For combined movements, TMS produced a mild bilateral decrease in latency of both saccadic and convergence components. All these effects were region specific. Stimulation of the MC had no effect on any type of eye movements. 

As mentioned, we used the term “pure” movements in the technical sense. Both prior studies and ours indicate the presence of transient vergence during a “pure” saccade and some saccadic intrusions during a “pure” vergence. The origin (central or peripheral) of such a phenomenon is not clear. According to Collewijn et al., ²⁴ the transient vergence during saccade is central, whereas Zee et al. ²¹ attributed it to muscular differences and/or differences of delay in the innervation between the lateral and the medial recti. Whatever are the causes, it is very unlikely that these transient intrasaccadic vergence and saccadic intrusions during vergence are triggered by the physical properties of lateral targets and targets in depth. In our experiment, lateral targets called for the activation of the saccadic oculomotor subsystem, whereas targets along the median plane called for the activation of the vergence subsystem. These two systems have distinct structures, at least at the brain stem level. ²⁶ With these considerations, we now discuss the findings of the study. 

Gap-Effect on the Latencies of Pure Saccades, Pure Convergence, and Combined Eye Movements

The gap paradigm allowed us to elicit two types of saccade latency: express and regular. The phenomenon of express saccades has been widely studied. ^{2 5} It could be mediated by multiple hypothetical mechanisms: the offset of the fixation point would produce a general warning signal and facilitate disengagement of attention, thereby shortening saccade latency. Other investigators have thought that the offset influences the competition between fixation and moving activity taking place during saccade preparation (see the introduction). Most likely, several of these mechanisms are involved, ⁵ each having its own neural basis. ²⁷  

The new interesting finding is that there was no express latency in vergence eye movements in any of the six subjects studied. The mean convergence latency was 257 ms—that is, typical of regular movements. This observation consolidates the report on three human subjects in the study of Takagi et al. ¹¹ who compared gap with both simultaneous and overlap tasks. Indeed, they reported shorter latencies for pure vergence in the gap paradigm by comparison with what they obtained in the overlap condition. Nevertheless, they did not observe express vergence (with latency between 80 and 120 ms). As mentioned in the introduction, express vergence is considered to have the same latency as express saccade. This is arbitrary but consistent with studies that reported express latencies for disparity vergence ²⁸ or for pursuit ²⁹ (discussed later). The results in both our study and the study of Takagi et al. ¹¹ were in contrast to findings in monkeys of Busettini et al., ²⁸ who observed latencies of approximately 80 ms for vergence stimulated by disparity steps of large textured scenes. Our findings are also partially in contrast with recent observations of Kapoula et al., ¹⁹ who reported short latencies for vergence in two subjects, although no detailed analysis of the rate of express vergence was reported. 

Similarly, no express latencies were found in our study for combined movements, either for the saccade or for the convergence component. This is in agreement with the study of Takagi et al. ¹¹ Most likely, the existence of express latencies for vergence (pure or combined) is not a stereotypical finding, such as is known for saccades. Rather, it seemed to be a subject-dependent phenomenon, which is consistent with research on the pursuit oculomotor subsystem. Merrison and Carpenter ²⁹ observed express latencies in humans performing smooth pursuit in a condition in which the subjects were presented an auditory cue 160 ms before the onset of target motion. Nevertheless, Krauzlis and Miles ³⁰ failed to observe such express latency for smooth pursuit while using a gap paradigm in monkeys. They proposed that there was no evidence for express smooth pursuit, suggesting that the express initiation pursuit reported by Merrison and Carpenter ²⁹ was due to anticipatory responses induced by the auditory cue. Current research has also suggested that the mechanism for express is not universal for all types of movements. ³¹ Taken together, our results with results in other studies suggest that the gap paradigm, although the most favorable condition, is not sufficient to elicit express latencies for all types of movements. It is sufficient to elicit express pure saccades, but not for pure convergence or convergence combined with saccades. 

Effect of TMS over the rPFC on the Latencies of Saccade, Pure Convergence, and Combined Eye Movements

TMS applied over rPFC increased significantly the rate of express contralateral saccades, whereas it had no effect on the latency of ipsilateral saccades. This result is consistent with the study of Müri et al., ¹⁰ who reported similar effects. The hypothetical mechanism by which this effect is produced is the following: TMS interferes transiently with the prefrontal inhibitory function, which normally prevents the express saccades from being triggered by the direct projections from the PFC to the SC. This interpretation is also compatible with studies in patients that report that lesions in the prefrontal cortex are responsible for increased rates of express saccades. ^{26 32}  

For pure convergence, we found no TMS effect on mean latencies. By comparison with saccadic eye movement generation, which is an automatic routine, pure convergence in our study was slower to trigger and perhaps more volitional. The absence of a significant effect of TMS over the rPFC on the latency of convergence could be due to two reasons: (1) TMS over rPFC could cause express latency only for movements that are already reflexive and of short latency—the idea being that only reflexive movements would activate a circuit involving the PFC. ^{6 7} (2) TMS over rPFC could influence only contralateral, therefore directional, movements. In other words, the functional inhibitory circuitry described between PFC and SC would be exclusively devoted to the control of contralateral saccades. The prediction of the former interpretation would be that TMS of the rPFC could decrease latency of convergence as well under optimal spatiotemporal conditions allowing short-latency expresslike convergence to occur. We and others mostly failed to elicit such express responses for vergence, even when the instructions given to the subjects encouraged them to perform eye movements as fast as possible. Nevertheless, this failure to find express responses does not mean these responses for vergence are not possible. It just indicates that we have not yet found the experimental configuration allowing such responses to occur. Such expectation is correlated with the assumption that the PFC could play a role in the triggering of vergence—namely, the existence of vergence-related cells in the PFC projecting to the SC. To our knowledge, this is an unexplored issue in electrophysiology. In contrast, the prediction of the second interpretation would be that TMS of the rPFC would have no effect on convergence, even for eventual short-latency express like convergence. Our results for combined eye movements, which will be discussed in the following paragraphs, seem to be in line with the second interpretation. 

Indeed, when combined movements are made, even with latencies rather long and similar to those of pure convergence, there is an effect on the latency that transfers to the convergence component (see Fig. 6 ). Our observations are reminiscent of the conceptual model proposed by Takagi et al. ¹¹ The decision to generate a combined movement would first be common and influenced by the fixation conditions (e.g., gap, overlap). Once this decision is made, activity would be initiated in two independent trigger circuits, one for the saccade and the other for the vergence system. Our findings on combined movements suggest that the control of such movements is similar to that of saccades involving the PFC-SC circuitry. The effect of TMS—transient disruption of the inhibitory activity of PFC—would transfer to the vergence component as well, thereby shortening its latency. Indeed, there was an emergence of express movements, particularly for contralateral combined movements that could explain the mild decrease of latencies for such movements. 

In conclusion, this study showed that in our spatiotemporal arrangement, express movements could occur for pure saccades only. When TMS was applied over the rPFC, the rate of express contralateral saccades was increased, thereby leading to a decrease in latency. An important finding was that TMS over rPFC had no effect at all on pure convergence. Although TMS over the rPFC caused a bilateral mild decrease in the latency of both components of the combined movements, that seemed to be produced mainly by the emergence of a few express latencies, even for such movements. We conclude that the functional inhibitory circuitry between the PFC and the SC involved in the control of eye movements primarily deals with the contralateral saccadic system but may transfer to the vergence system in the case of combined eye movements in direction and depth. 

 Supported by the French Minister of Research and Technology (MRT, Contract Cognitique).

 Submitted for publication February 25, 2002; revised June 20, 2002; accepted July 31, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(A) Six LEDs were positioned on a horizontal plane at eye level. The fixation point was lighted at the center at 150 cm from the eyes of the subject. The target LED could be (1) one of the two lateral LEDs at 150 cm (pure saccade); (2) the center at 20 cm (pure convergence); or (3) one of the two lateral LEDs at 20 cm (combined movement). (B) After a 2.5-second fixation period, a temporal gap of 200 ms was introduced before the appearance of the target. The TMS was applied simultaneously with the onset of the target. The length of the TMS pulse was 1 ms.

Figure 2.

 

Trajectory of a pure saccade (A), a pure convergence (B), and combined eye movement (C). Represented in graph are the saccadic signal ([left eye – right eye]/2) and the vergence signal (left eye − right eye). Arrows: onset of each movement; dotted lines: location of the target. For the combined movement the target eccentricity and target depth are both shown.

Figure 3.

 

Individual mean latencies of pure movements (A, saccades; B, convergence) in the reference condition without TMS. All: Group mean latencies. Bars: standard deviation; n: range of number of trials.

Figure 4.

 

Individual mean latency differences between the experimental condition with TMS over rPFC and the reference condition without TMS (A, saccade; B, convergence). Statistical analysis was conducted between the two conditions within each subject (Student’s t-test) and for the group (ANOVA). *Statistically significant differences; ns: no significant difference. Other notations are as in Figure 3 .

Figure 5.

 

Mean latencies of combined eye movements for the reference condition without TMS. Mean latencies are reported for each subject and for the group for the saccade (A) and for the convergence component (B). *Significant difference when compared with the corresponding pure movements shown in Figure 3 . Other notations are as in Figure 3 .

Figure 6.

 

Individual latency differences obtained by subtracting the latencies in the reference condition without TMS from those in the experimental condition with TMS over rPFC. (A) Saccadic and (B) convergence components of combined movements. Statistical analysis was conducted between the two conditions within each subject (Student’s t-test) and for the group (ANOVA). Other notations are as in Figure 3 .