Behavioural conditions affecting saccadic eye movements elicited ...

European Journal of Neuroscience, Vol. 11, pp. 2431±2443, 1999

ã European Neuroscience Association

Behavioural conditions affecting saccadic eye movements elicited electrically from the frontal lobes of primates Edward J. Tehovnik, Warren M. Slocum and Peter H. Schiller

Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, E25-634, Cambridge, MA 02139, USA Keywords: behavioural state, electrical stimulation, frontal eye ®elds, oculomotor behaviour, rhesus monkeys, supplementary eye ®elds

Abstract We assessed the effects of varying the time at which electrical stimulation was delivered to the dorsomedial frontal cortex (DMFC) and the frontal eye ®elds (FEF) relative to the onset of a visual target. Monkeys were required to ®xate the visual target to obtain a drop of apple juice as reward. We found that the probability of eliciting saccades increased with increases in the delay of electrical stimulation relative to target onset. Also, the current threshold to evoke saccades decreased as electrical stimulation was delivered later following target onset. There were major differences in the magnitude of this effect with stimulation of the DMFC versus the FEF. The current threshold to evoke saccades from the DMFC was 16 times greater when electrical stimulation was delivered 200 ms after target onset as compared to when it was delayed 200 ms after target offset. In contrast, the current threshold to evoke saccades from the FEFs was only three times greater when stimulation was delivered under similar conditions. These results suggest that the FEF are more closely connected with the saccade generator for the execution of saccadic eye movements than is the DMFC, even though both regions have direct projections to brainstem oculomotor centres.

Introduction The behavioural state of an animal affects saccadic eye movements elicited electrically from many regions of the brain (e.g. Marrocco, 1978; Sparks & Mays, 1983; Schiller & Sandell, 1983; Shibutani et al., 1984; Goldberg et al., 1986; Mann et al., 1988; Kurylo & Skavenski, 1991; Bon & Lucchetti, 1992). For instance, Marrocco (1978) observed that during electrical stimulation of the frontal eye ®elds (FEF), increases in the stimulation current of several times the threshold failed to dislodge the eyes and produce saccadic eye movements when a monkey was performing smooth pursuit. Marrocco concluded that stimulation-elicited responses at the level of the FEF can be overridden by the behavioural state of the animal. The purpose of the present study, therefore, was to investigate the behavioural conditions that affect saccadic eye movements elicited electrically from two regions of the frontal lobe: the dorsomedial frontal cortex (DMFC, which contains the supplementary eye ®elds as ®rst described by Schlag & Schlag-Rey, 1987) and the FEF. Both regions have been associated with the execution of saccadic eye movements. Saccadic eye movements evoked electrically from the DMFC terminate in an orbital position that we call a termination zone (Schlag & Schlag-Rey, 1987; Mann et al., 1988; Schall, 1991b; Bon & Lucchetti, 1992; Tehovnik & Lee, 1993; Tehovnik et al., 1994; Lee & Tehovnik, 1995; Tehovnik & Sommer, 1997; Tehovnik et al., 1998; Sommer & Tehovnik, 1999; but see Russo & Bruce, 1993). Termination zones are represented topographically within the DMFC (Tehovnik & Lee, 1993; Tehovnik et al., 1994; Lee & Tehovnik, 1995; Tehovnik & Sommer, 1997; Tehovnik et al., 1998; Sommer & Tehovnik, 1999). If stimulation is delivered continuously to the

DMFC, it is typical for no more than one saccadic eye movement to be produced (Schlag & Schlag-Rey, 1987; Schall, 1991b; Tehovnik & Lee, 1993; Sommer & Tehovnik, 1999), after which the eyes remain ®xed at the termination zone until the end of stimulation (Tehovnik & Lee, 1993; Tehovnik et al., 1994). In contrast, saccadic eye movements elicited from the FEF exhibit a similar direction and amplitude irrespective of starting eye position (Robinson & Fuchs, 1969; Marrocco, 1978; Bruce et al., 1985; Goldberg & Bruce, 1990; Schall, 1991b; Tehovnik & Lee, 1993). The size of these movements is represented topographically within the FEF (Robinson & Fuchs, 1969; Bruce et al., 1985). Continued stimulation of the FEF produces a staircase of saccadic eye movements that consists of a series of saccades with the same amplitude and direction separated by brief periods of ®xation (Robinson & Fuchs, 1969; Schiller, 1977; Schall, 1991b; Tehovnik & Lee, 1993). That behavioural conditions of an animal affect saccadic eye movements elicited from the DMFC and FEF is known (e.g. Marrocco, 1978; Schiller & Sandell, 1983; Goldberg et al., 1986; Mann et al., 1988; Bon & Lucchetti, 1992). What has not been known, and what we explore in this study, is the extent to which systematic manipulation of the temporal course of visual stimulation and reward relative to the time of electrical stimulation delivery affects saccadic eye movements evoked from the DMFC and FEF. We report here that the probability of eliciting saccades increases as the delay of electrical stimulation is increased relative to target onset, and that this effect is far greater when the DMFC is stimulated.

Materials and methods Correspondence: E. J. Tehovnik, as above. E-mail: [email protected] Received 1 October 1998, revised 1 March 1999, accepted 4 March 1999

Subjects Two adult rhesus monkeys (Macaca mulatta) I and J were used. Throughout this study, food was freely available. The monkeys were

2432 E. J. Tehovnik et al. deprived of water overnight before each day of experimental testing. After testing, they were allowed to drink until sated before being returned to the vivarium. The monkeys were provided for in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and the guidelines of the Massachusetts Institute of Technology Committee on Animal Care. Surgery Monkeys were anaesthetized with pentobarbital intravenously (30 mg/kg), and prepared for aseptic surgery. A scleral search coil was implanted (Judge et al., 1980), and a stainless-steel post, to restrain the head, was secured to the skull with stainless steel or titanium head screws and acrylic cement. Subsequently, a recording chamber was implanted over the DMFC. A second chamber was implanted over the right FEF of monkey I. Behavioural tasks A monkey, with head ®xed, faced a 52 ° horizontal by 39 ° vertical LED board positioned 24 cm away. The animal ®xated an LED (2.0 ° of visual angle) to receive a drop of apple juice. During ®xation, the animal was required to keep his eyes within a 5 3 5 ° window; otherwise the trial was terminated. Stimulation was delivered on 50% of trials after the centre of gaze shifted to the LED. Across trials, stimulation was presented randomly so that the monkey could not predict its presentation. Monkeys performed all tasks in a dimly lit room. Two types of experiments were conducted. In the ®rst set of experiments, all stimulation was delivered after the termination of the ®xation spot (Fig. 1A). A gap was imposed between the offset of the ®xation spot and the onset of electrical stimulation. The gap was varied randomly from 0 to 600 ms. For most experiments of this kind, the duration of ®xation was ®xed at 600 ms and the juice reward was delivered immediately after the termination of the ®xation spot. For two experiments, however, these conditions differed. In one, the ®xation duration was varied from 50 to 600 ms with juice delivery occurring immediately after the termination of the ®xation spot; in another, the ®xation duration was ®xed at 600 ms while the time of juice delivery was varied from 0 to 400 ms after the termination of the ®xation spot. In the second set of experiments, stimulation was delivered at random times before and after the termination of the ®xation spot (Fig. 1B). In these experiments, the duration of ®xation was 600 ms, and the juice reward was delivered at the termination of the ®xation spot. For all tests, a single ®xation position located 26 ° from the centre of gaze in the hemi®eld ipsilateral to the side of stimulation was used. This guaranteed that saccades were elicited from the full rostrocaudal extend of the DMFC as stimulation of DMFC always evokes contraversive saccades from an ipsilateral ®xation position (Tehovnik & Lee, 1993; Tehovnik et al., 1994; Tehovnik & Sommer, 1997; Tehovnik et al., 1998; Sommer & Tehovnik, 1999). All saccades evoked from the DMFC and FEF travelled contralateral to the side of stimulation. The position of the head with respect to the LED board was central and constant between different sessions of data collection. Data collection and analysis A PDP 11/73 computer controlled the presentation of visual stimuli, the delivery of electrical stimulation, the display and collection of eye position (sampled at 200 Hz), and the delivery of juice. The eyes were required to achieve a velocity of at least 100 °/s to qualify as a saccade, and such a movement had to occur during the stimulation period to qualify as a stimulation-elicited saccade.

FIG.1. An animal was required to ®xate at the target for 600 ms (®xation). Immediately after termination of the ®xation spot, a juice reward was delivered (juice). (A) A gap, ranging from e.g. 0 to 600 ms, was imposed between the termination of the ®xation spot and the onset of electrical stimulation (stim). (B) For some experiments, stimulation was delivered before or after the termination of the ®xation spot.

Electrodes Glass-coated platinum±iridium electrodes were constructed with conical tips having impedances between 1.0 and 3.0 MW tested at 1 kHz. After delivering stimulation through a glass-coated platinum± iridium electrode, it is common for its impedance to drop; the impedance at the end of an experimental session was therefore used to compute the surface area of an electrode tip. After passing 50, 400 or 1600 mA (at 0.2-ms pulse durations) through these electrodes, the impedance dropped to 0.5, 0.1 and 0.05 MW, respectively. The estimated (maximum) exposed surface area of these electrodes was 0.0008, 0.008 and 0.02 mm2, respectively, using the formula, surface area = 0.0003/(impedance)1.4 (Tehovnik & Sommer, 1997). The charge densities generated at these electrode tips were 12 500, 10 000 and 16 000 nC/mm2 per phase, respectively. Charge densities exceeding 16 000 nC/mm2 per phase delivered to neural tissue continuously for many hours produced histological damage at the electrode tip (McCreery et al., 1990). The charge densities used in the current study were at or below 16 000 nC/mm2 per phase. The average effective current spread for 50, 400 and 1600-mA pulses (at a 0.2-ms pulse duration) is estimated to be 0.2, 0.6 and 1.1 mm, respectively,

Ó 1999 European Neuroscience Association, European Journal of Neuroscience, 11, 2431±2443

Behavioural conditions affecting elicited saccades

2433

FIG. 2. The top view of one side of the DMFC is illustrated. Each oval shows a region studied by particular investigators using unit recording and electrical stimulation methods on monkeys to probe oculomotor responses. The method of Tehovnik (1995) was used to determine the size and location of an oval situated with respect to the cerebral midline (Ml) and the posterior tip of the arcuate sulcus (Pa). The grey ovals represent the regions investigated in the current report. The posterior tip of the arcuate was unknown in monkey J. The rostrocaudal location of the centre of the DMFC well in monkeys I and J was 33 and 35 mm anterior to the interaural line, respectively. We estimated the rostrocaudal location of the region investigated in monkey J by situating this region with respect to the rostrocaudal location of the region investigated in monkey I. The rostrocaudal location of the centre of the FEF well in monkey I was 28 mm anterior to the interaural line and 20 mm off the midline. Each tick mark along the midline axis (MI) is spaced by 1 mm. The arcuate sulcus (As) and central sulcus (Cs) are indicated.

FIG. 3. On the right is an overhead view of the right DMFC and right FEF of monkey I (A), and an overhead view of the DMFC of monkey J (B). The location of a well (symbolized by a dashed circle) was determined by noting cortical landmarks during surgery and by noting the relative location of each well using stereotaxic coordinates. The cerebral midline (Ml) was visible while mounting the well over the DMFC, and the arcuate sulcus (As) was visible while mounting the well over the FEF. A penetration site from which saccades were elicited, with 50 mA pulses delivered in trains of 200 ms or less is marked by a black dot. The pulse duration was set at 0.2 ms and the frequency set at 250 Hz. All electrical stimulation was delivered 200 ms after termination of the ®xation spot and delivery of juice reward. A penetration site from which saccades could not be elicited using the above parameters is marked by a star. The location of a mark was determined with respect to the coordinate grid used for inserting an electrode. The anterior, posterior, medial and lateral orientations with respect to the midline are indicated in (B). A scale bar, in mm, is also indicated. At the left of the ®gure are penetration sites, numbered from 1 to 68, that correspond to the sites shown on the right.


2434 E. J. Tehovnik et al.

FIG. 4. Examples are shown of saccades evoked from two sites in the DMFC and one site in the FEF (penetration sites 10, 11 and 46 for A±C, respectively; see Fig. 3, left) using train durations ranging from 40 to 400 ms. The ®xation position, which was located 26 ° in ipsilateral craniotopic space, is indicated by a square and the black dots represent the trajectory of the stimulation-evoked saccades. An animal was required to ®xate the ®xation spot for 600 ms. Juice delivery occurred immediately after the ®xation spot was turned off. All stimulation was delivered 200 ms after the termination of the ®xation spot on 50% of trials. The current, pulse duration and pulse frequency were set at 50 mA, 0.2 ms and 350 Hz, respectively. At the bottom, saccades generated when no stimulation was applied are shown (control). A scale bar representing 50 ° of visual angle is shown.

from the electrode tip [Tehovnik, 1996: current spread = (current/K)1/2, average K = 1300 mA/mm2 from Stoney et al., 1968]. Impedance was measured between bouts of stimulation using a 1-kHz, nanoampere tester (Bak Electronics, model A-1B) which was calibrated with a 1.0-MW resistor.

action potentials were ampli®ed (Bak A-1B) and ®ltered (Krohn-Hite 3750). During preliminary mapping, 50-mA pulses (at a 0.2 ms pulse duration) with pulse frequencies of 250 Hz and train durations of 200 ms were used to evoke saccades from the DMFC and FEF.

Results

Electrical stimulation Constant-current charge-balanced biphasic pulses were delivered to the brain via a monopolar electrode using a Grass S88 stimulator attached to a pair of constant-current stimulus isolation units (Grass PSIU6B). For each biphasic pulse, a cathodal and anodal pulse followed in immediate succession. Both pulses had the same amplitude and duration. Current was monitored by the voltage drop across a 1000-W resistor that was in series with the return lead of the stimulator. The current was monitored using a Tektronix Oscilloscope (model 5103N, differential ampli®er) and was read as the amplitude of one pulse (cathode or anode) of a biphasic pair. Electrodes were introduced perpendicular to the dural surface with a hydraulic microdrive until action potentials were encountered. The

Stimulation sites The regions investigated in the right DMFC and right FEF of monkey I, and in the right DMFC of monkey J are shown and compared to regions investigated by other researchers studying the DMFC and oculomotor behaviour (Fig. 2). Using a ®xed current of 50 mA, the best stimulation sites in the DMFC were within an 8 mm extent along the rostrocaudal dimension of cortex, between 1 and 8 mm lateral to the cortical midline, and between 0 and 2 mm below the ®rst units encountered on an electrode pass. In monkey I, stimulation of anterior DMFC (e.g. penetration site 3, Fig. 3) evoked saccades that terminated in extreme contralateral craniotopic space; and stimulation of posterior DMFC (e.g. penetration site 29, Fig. 3) evoked saccades



2435

that terminated in ipsilateral craniotopic space (see Tehovnik et al., 1998 for other details). These results are consistent with previous reports (Tehovnik & Lee, 1993; Tehovnik et al., 1994; Lee & Tehovnik, 1995; Tehovnik & Sommer, 1997; Sommer & Tehovnik, 1999). The best stimulation sites in the FEF were located within the anterior bank of the arcuate sulcus between 0 and 5 mm below the ®rst units encountered on an electrode pass. Stimulation of dorsomedial FEF (e.g. penetration site 37, Fig. 3) elicited largeamplitude saccades (i.e. » 27 °), and stimulation of ventrolateral FEF (e.g. penetration site 45, Fig. 3) elicited small-amplitude saccades (i.e. » 2 °). These results concur with previous reports (Robinson & Fuchs, 1969; Bruce et al., 1985). During mapping of the DMFC and FEF, electrical stimulation was delivered 200 ms after termination of the ®xation spot. The animal was required to ®xate the ®xation spot for 600 ms after which a juice reward was delivered and ®xation spot terminated. An effective stimulation site was de®ned as a site from which saccades were induced on at least seven out of 10 stimulation trials. Saccades evoked during the stimulation period on non-stimulation trials for the longest train durations tested (i.e. 200 ms) never occurred on more than four out of 10 trials, and the endpoint of these saccades differed from those evoked during stimulation trials. Trials for which a monkey left the target window before the stimulation period commenced were excluded. Examples of saccades evoked from the DMFC and FEF using a 50 mA current are shown in Fig. 4. The total displacement of the eyes was found to vary with the duration of the pulse train for both the DMFC and FEF. Stimulation of the DMFC typically elicited one saccade, even for long train durations (e.g. Fig. 4A and B, 400 ms). Stimulation of the FEF with train durations of 400 ms typically produced multiple saccades (e.g. Fig. 4C, 400 ms). Examples of saccades evoked spontaneously during nonstimulation trials are shown at the bottom of the ®gure (Fig. 4A±C, control). The endpoints of these saccades were more variable than the endpoints of the stimulation-evoked saccades. The slight variability in the trajectories and endpoints of the stimulation-evoked saccades may have been due to interference by the impending execution of a spontaneously-evoked saccade following termination of the ®xation spot. Saccades evoked following active ®xation The effect of imposing a gap between the termination of the ®xation spot and the onset of electrical stimulation was studied using a 50 mA current (Fig. 1A). For these experiments, the juice reward was delivered immediately after the termination of the ®xation spot, and the duration of visual ®xation was 600 ms. The probability of evoking a saccade is plotted as a function of the gap between the termination of the ®xation spot and the beginning of the period of stimulation for the DMFC and FEF (Fig. 5). The probability of saccades for comparable periods was greater when stimulation was applied than when stimulation was not applied over a range of gap durations. More importantly, this probability increased systematically with greater gap durations, and this increase was always greater than that observed during non-stimulation trials. In the next experiment, a monkey was required to ®xate the ®xation spot for either 50, 200 or 600 ms before the ®xation spot was terminated and a juice reward delivered. Here, juice was always delivered immediately after the termination of the ®xation spot. As before, the probability of saccades immediately after the termination of the ®xation spot was greater for stimulation than non-stimulation trials for both the DMFC (Fig. 6) and FEF (Fig. 7). The ease with which saccades were elicited electrically from the DMFC following

FIG. 5. The probability of eliciting saccades is plotted as a function of the gap between the termination of the ®xation spot and onset of stimulation for the DMFC and FEF. For each of eight plots (A±H), the top curve represents the probability of a saccade during stimulation and the bottom curve represents the probability of a saccade during the same period for non-stimulation trials. At each gap, a data point was determined by noting the number of saccades evoked over 10 stimulation trials or by noting the number of saccades evoked during 10 non-stimulation trials. The train duration for A±D was ®xed at 70 ms, and the train duration for E±H was ®xed at 40 ms. All other parameters of stimulation and behavioural conditions were the same as those described in Fig. 4. Penetration sites 30, 29, 29, 21, 45, 47, 43 and 43 were stimulated for plots A±H, respectively (Fig. 3, left).

the termination of the ®xation spot increased systematically with increases in the duration of visual ®xation. For the DMFC, the longer the duration of ®xation, the shorter was the required gap between the termination of ®xation and stimulation onset to yield a given probability of stimulation-evoked saccades (Fig. 6, top). This effect was less apparent for the FEF (Fig. 7, top). The effect of ®xation duration was greater for the DMFC than FEF when comparing the gap difference between the shortest and longest (50 ms and 600 ms;



FIG. 6. The probability of evoking saccades is plotted as a function of the gap between the termination of the ®xation spot and onset of stimulation for the DMFC at different ®xation durations: 600, 200 and 50 ms. Within each panel (A±D), the top graphs show data for stimulation trials and the bottom graphs show data for the same period for non-stimulation trials. (A) The train duration was 200 ms. (B±D) The train duration was 70 ms. Other parameters of stimulation and behavioural conditions were the same as those described in Fig. 4. Penetration sites 36, 35, 29 and 59 were stimulated for plots A±D, respectively (Fig. 3, left).

t3 = 5.1, P < 0.025; Bonferroni adjustment), the shortest and intermediate (50 ms and 200 ms; t3 = 5.1, P < 0.025; Bonferroni adjustment), and the intermediate and longest (200 ms and 600 ms; t3 = 3.2, P < 0.05; Bonferroni adjustment) ®xation durations to yield a 70% probability of evoking a saccade. In the following experiment, a monkey was required to ®xate the ®xation spot for 600 ms, after which the ®xation spot was terminated. Afterwards, juice was delivered either at the termination of the ®xation spot, at 200 ms after the termination of the ®xation spot, or at 400 ms after the termination of the ®xation spot. As long as the monkey remained on the ®xation spot for 600 ms, it received a juice reward. The monkey was not penalized for leaving the ®xation spot between the termination of the ®xation spot and time of juice delivery in cases where juice delivery was delayed by 200 or 400 ms. As before, the probability of saccades immediately after the termination of the ®xation spot was greater for stimulation than non-stimulation trials for both the DMFC (Fig. 8) and FEF (Fig. 9). When the delivery of juice reward was delayed after the termination of the ®xation spot, the saccades triggered by electrical stimulation of the DMFC were also delayed. The longer the delay of the juice reward, the greater was the required gap between the termination of ®xation and stimulation onset to yield a given probability of stimulation-elicited saccades (Fig. 8, top). This effect was less evident for the FEF (Fig. 9, top). The effect of juice delay was greater for the DMFC than FEF when comparing the gap difference between the shortest and longest (0 ms and 400 ms; t3 = 4.4, P < 0.05; Bonferroni adjustment), the shortest and intermediate (0 ms and 200 ms; t3 = 5.0, P < 0.025; Bonferroni adjustment), and the intermediate and longest (200 ms and 400 ms;

t3 = 3.2, P < 0.05; Bonferroni adjustment) juice delays to yield a 70% probability of evoking a saccade. In summary, imposing a gap between the termination of the ®xation spot and the onset of stimulation increased the ease with which saccades were elicited from the DMFC and FEF. For the DMFC only, the ease with which saccades were evoked increased with increases in ®xation duration and decreased with delays in reward delivery. Saccades evoked during active ®xation In preceding experiments, stimulation was always delivered at or after the ®xation spot was turned off. The next experiment compared saccades elicited from the DMFC before or after the ®xation spot was terminated (Fig. 1B). Monkeys were required to maintain ®xation of a ®xation spot for 600 ms. At the end of ®xation, a juice reward was delivered after which the monkey was free to leave the ®xation window. A 400-ms train of stimulation was delivered 200 ms after the termination of the ®xation spot (Fig. 10A, left), 0 ms after the termination of the ®xation spot (Fig. 10B, left), 200 ms before the termination of the ®xation spot (Fig. 10C, left), or 400 ms before the termination of the ®xation spot (Fig. 10D, left). When stimulation was delivered at or after the termination of the ®xation spot, 100% (85/85) of the saccades were evoked during stimulation. When stimulation was delivered 200 ms before the end of ®xation, 54% (29/54) of the saccades were evoked during stimulation. When stimulation was delivered 400 ms before the end of ®xation, 4% (2/54) were evoked during stimulation. The amplitude of the elicited saccades was also shorter when stimulation started during



2437

FIG. 7. The probability of evoking saccades is plotted as a function of the gap between the termination of the ®xation spot and onset of stimulation for the FEF at different ®xation durations: 600, 200 and 50 ms. Within each panel (A±D), the top graphs show data for stimulation trials and the bottom graphs show data for the same period for non-stimulation trials. (A) The train duration was 40 ms. (B) The train duration was 50 ms. (C and D) The train duration was 70 ms. Other parameters of stimulation and behavioural conditions were the same as those described in Fig. 4. Penetration sites 47, 43, 46 and 39 were stimulated for plots A±D, respectively (Fig. 3, left).

®xation than when stimulation started after ®xation. Finally, the latency of these saccades was shorter when stimulation commenced after ®xation (Fig. 10A, right) than when stimulation commenced at the end of or during ®xation (Fig. 10B and C, right). Typically, the stimulation-elicited saccades were delayed until after the ®xation spot was terminated (Fig. 10B and C, left). Thus, active ®xation decreased the probability of stimulation-elicited saccades from the DMFC, and it decreased saccadic amplitude and increased saccadic latency. To assess the effect of active ®xation on saccades evoked from the DMFC and FEF, the current strength to evoke saccades on 70% of stimulation trials was determined. A 200-ms train of stimulation was delivered 200 ms after termination of the ®xation spot, 0 ms after termination of the ®xation spot, 200 ms before the termination of the ®xation spot, or 400 ms before the termination of the ®xation spot (Fig. 11, left). In these experiments, juice delivery always occurred at the termination of the ®xation spot. For the DMFC, the current threshold to elicit saccades ranged from 43 to 100 mA when stimulation was delivered 200 ms after ®xation (Fig. 11A, stimulation onset time, 200 ms), the current threshold ranged from 245 to 438 mA when stimulation was delivered at termination of ®xation (Fig. 11A, stimulation onset time, 0 ms), the current threshold ranged from 533 to 1000 mA when stimulation was delivered 200 ms before the end of ®xation (Fig. 11A, stimulation onset time, ±200 ms), and the current threshold ranged from 672 to 1500 mA when stimulation was delivered 400 ms before the end of ®xation (Fig. 11A, stimulation onset time, ±400 ms). For the FEF, the current threshold to elicit saccades ranged from 6 to 30 mA when stimulation was delivered 200 ms after ®xation (Fig. 11B, stimulation onset time, 200 ms), the current threshold ranged from 6 to 55 mA when stimulation was

delivered at termination of the ®xation (Fig. 11B, stimulation onset time, 0 ms), the current threshold ranged from 22 to 65 mA when stimulation was delivered 200 ms before the end of ®xation (Fig. 11B, stimulation onset time, ±200 ms), and the current threshold ranged from 24 to 86 mA when stimulation was delivered 400 ms before the end of ®xation (Fig. 11B, stimulation onset time, ±400 ms). To compare changes in the current threshold observed for the DMFC and FEF, the average current threshold observed 200 ms after ®xation was set to unity, and all other thresholds were expressed as a multiple of this threshold. This yielded a normalized threshold function (Fig. 11C). For both the DMFC and FEF, the average threshold to evoke saccades decreased systematically as stimulation was delivered later in ®xation and thereafter. For the DMFC, the current threshold at the beginning of ®xation was 15.6 times higher (on average) when compared to the current threshold observed 200 ms after termination of the ®xation spot; for the FEF, the current threshold at the beginning of ®xation was only 3.4 times higher (on average) when compared to the current threshold observed 200 ms after termination of the ®xation spot. The elevation in threshold early in ®xation was greater for the DMFC than it was for the FEF (t5 = 3.05, P < 0.025). Thus, the current strength required to evoke saccades during active ®xation was increased appreciably more for the DMFC than FEF.

Discussion Although a number of studies have reported that thresholds for eliciting saccadic eye movements are higher in both the DMFC and FEF when electrical stimulation is applied while a monkey ®xates a



FIG. 8. The probability of eliciting saccades is plotted as a function of the gap between the termination of the ®xation spot and onset of stimulation for the DMFC at different juice delays: 0, 200 and 400 ms. Within each panel (A±D), the top graphs show data for stimulation trials and the bottom graphs show data for the same period for non-stimulation trials. (A and B) The train duration was 200 ms. (C and D) The train duration was 100 ms. All other parameters of stimulation and behavioural conditions were the same as those described in Fig. 4. Penetration sites 16, 16, 17 and 51 were stimulated for plots A±D, respectively (Fig. 3, left).

FIG. 9. The probability of eliciting saccades is plotted as a function of the gap between the termination of the ®xation spot and onset of stimulation for the FEF at different juice delays: 0, 200 and 400 ms. Within each panel (A±D), the top graphs show data for stimulation trials and the bottom graphs show data for the same period for non-stimulation trials. (A,C) The train duration was 40 ms. (B,D) The train duration was 70 ms. All other parameters of stimulation and behavioural conditions were the same as those described in Fig. 4. Penetration sites 47, 44, 46 and 39 were stimulated for plots A±D, respectively (Fig. 3, left). Ó 1999 European Neuroscience Association, European Journal of Neuroscience, 11, 2431±2443


2439

FIG. 10. Shown are the horizontal eye traces of the saccades evoked from the DMFC and their corresponding saccadic latencies. The left panels show saccades produced when stimulation was delivered 200 ms after the termination of the ®xation spot (A); 0 ms after the termination of the ®xation spot (B); 200 ms before the termination of the ®xation spot (C); and 400 ms before the termination of the ®xation spot (D). All other behavioural conditions were the same as those described in Fig. 4. The top grey bar indicates the duration of ®xation (®x), and the black bars indicate the onset and duration of stimulation for the four conditions. All saccades are aligned to the onset of the ®xation spot. Current, pulse duration, pulse frequency and train duration were 400 mA, 0.1 ms, 150 Hz and 400 ms, respectively. Penetration site 3 was stimulated for all plots (Fig. 3, left). The right panels show saccadic latency distributions aligned to the onset of stimulation.

visual target (Goldberg et al., 1986; Schlag & Schlag-Rey, 1987; Bon & Lucchetti, 1992), none has noted differences between these regions. This study reveals two major differences between the DMFC and FEF. First, in the DMFC, when stimulation was delivered early during ®xation, the current threshold to evoke a saccadic eye movement was 16 times higher than when stimulation was delivered after termination of the ®xation spot; by contrast, the threshold increased only threefold when the FEF was stimulated under comparable conditions. Parameters such as ®xation duration and time of juice delivery were also found to have a greater effect on the probability of eliciting saccades from the DMFC than from the FEF. By imposing a gap between the termination of a ®xation spot and the onset of electrical stimulation, the ease with which saccades were evoked from the DMFC and FEF increased, which probably re¯ects an animal's release from active ®xation. It was only for the DMFC, however, that electrical stimulation elicited saccades more easily as ®xation duration was increased or time of juice delivery following the termination of the ®xation spot shortened. Second, higher currents and longer train durations were required to evoke saccadic eye movements from the DMFC than from the FEF. For the DMFC, effective current levels ranged from 43 to 1500 mA (at a 0.2 ms pulse duration) and effective train durations ranged from 70

to 400 ms; by contrast, for the FEF, current levels ranged from 6 to 86 mA and train durations ranged from 40 to 70 ms. Effective current for evoking saccades from the DMFC and FEF Previous studies show considerable variation in the effectiveness with which saccadic eye movements could be elicited from both the FEF and DMFC. Currents from 40 to 2000 mA have been used to evoke saccadic eye movements from the FEF (Robinson & Fuchs, 1969; Marrocco, 1978; Schiller & Sandell, 1983). Using 50 mA currents or less (at a 0.2 ms pulse duration), Bruce and Goldberg found that eye movements could be elicited only from a restricted region of the arcuate sulcus within which resides the topographic map for the evocation of saccadic eye movements (Bruce et al., 1985). Some investigators suggested that the size of the FEF can best be de®ned by limiting electrical stimulation to 50 mA (e.g. Goldberg et al., 1986; Schall, 1991b; Tehovnik & Sommer, 1997; current study). Similar current levels were used by Schlag & Schlag-Rey (1987) to de®ne the region in the DMFC from which eye movements could be evoked. Tehovnik & Lee (1993), however, found that in order to obtain reliable saccadic eye movements to describe a topographic map in the DMFC, currents > 50 mA were required. It has been shown that the



FIG. 11. The current threshold for eliciting saccades on 70% of stimulation trials is plotted as a function of stimulation onset time for the DMFC (A) and FEF (B). For the DMFC, curves (a±f) represent the current threshold for evoking saccades from penetration sites 17, 5, 7, 4, 12 and 6, respectively (Fig. 3, left). For the FEF, curves (a±h) represent the current threshold for evoking saccades from penetration sites 42, 45, 48, 41, 39, 43, 44 and 49, respectively (Fig. 3, left). Stimulation was delivered 200 ms after the termination of the ®xation spot (200 ms), 0 ms after the termination of the ®xation spot (0 ms), 200 ms before the termination of the ®xation spot (±200 ms), or 400 ms before the termination of the ®xation spot (±400 ms). A normalized threshold ratio is plotted as a function of stimulation onset time for the DMFC and FEF (C). The ratio was computed by dividing the mean current required to evoke saccades at a given stimulation onset time by the mean current required to evoke saccades at a stimulation onset time of 200 ms. Standard errors are indicated. The left panel shows the method. The top grey bar represents the duration of ®xation (®x), which was 600 ms, and the black bars represent onset and duration stimulation which was 200 ms. Juice was delivered immediately after the ®xation spot was terminated. Pulse duration and pulse frequency were 0.2 ms and 200 Hz, respectively.

optimal stimulation parameters for eliciting saccades from the DMFC differ from those required to elicit saccades from the FEF (Tehovnik & Sommer, 1997). It had been proposed that the relatively large size of the eye ®eld area reported by Tehovnik & Lee (1993) may have been due to the high current levels used (Tanji, 1994). The ®ndings we report here now provide new clues as to why in the various studies such large differences had been found in the current levels needed to elicit saccadic eye movements from the DMFC. Some experimental conditions used to study the DMFC favoured saccade production, whereas others inhibited saccades. In the experiments of Schlag & Schlag-Rey (1987), electrical stimulation was often delivered while a monkey was not actively ®xating. This procedure, as shown in our present study, favours saccade production, and explains why Schlag and Schlag-Rey routinely produced saccades using currents of 50 mA. In the experiments of Russo & Bruce (1993), electrical stimulation

was delivered after a monkey ®xated a stimulus for 1±2 s; and, after stimulation dislodged the eyes from the ®xation spot, the animal was given time to reacquire the ®xation spot to obtain a reward. The lengthy ®xation time and the lack of a requirement to maintain ®xation during stimulation leads to lax ®xation on the part of the animal and easy evocation of a saccade. By contrast, in the experiments of Tehovnik & Sommer (1997), stimulation was delivered 200 ms after initiation of ®xation, and an animal was not permitted to reacquire the ®xation spot to get a reward once stimulation dislodged the eyes. The shorter ®xation time and more stringent behavioural requirement imposed in these experiments explains the necessity for higher currents to elicit saccades. As illustrated in the current report, something as seemingly inconsequential as decreasing the ®xation period or increasing the reward delay can radically alter the probability of evoked saccadic eye movements from the DMFC when low stimulation currents (i.e.


Behavioural conditions affecting elicited saccades 50 mA) are delivered following target extinction. This evidence supports the conclusion that such manipulations delay a monkey's release from active ®xation. If one is attempting to map the underlying eye ®eld in the DMFC, it is necessary to use suf®cient electrical current to override active ®xation; otherwise, one is primarily probing the behavioural state of the animal. Changes in current threshold during active ®xation For the DMFC, currents as high as 1500 mA were required to evoke saccadic eye movements when stimulation was delivered early in active ®xation; for the FEF, currents as high as 86 mA were required to evoke saccades when stimulation was delivered early in active ®xation. According to our current spread estimates (see Materials and methods), a 1500 mA current (at a 0.2 ms pulse duration) directly activates elements located within 1.1 mm from the electrode tip, and a 86 mA current directly activates elements located within 0.26 mm from the electrode tip. Previously, we have argued that higher currents are required to evoke saccades from the DMFC as compared to the FEF because not all neurons in the DMFC are dedicated to the execution of saccadic eye movements (Tehovnik & Sommer, 1997). Many neurons within the DMFC are also modulated during the execution of skeletomotor responses (Mann et al., 1988; Schall, 1991a; Bon & Lucchetti, 1992; Mushiake et al., 1996; Chou & Schiller, 1997a,b). A second factor contributing to the need for higher currents when stimulating the DMFC is that active ®xation has a greater effect on saccades evoked from the DMFC as compared to the FEF. Currents at or below 100 mA were suf®cient to evoke saccades from the DMFC after the termination of the ®xation target. We believe that the differential effect of active ®xation on the current threshold to evoke saccadic eye movements from the DMFC and FEF is not occurring at the site of direct stimulation. The chronaxies of directly-stimulated elements in the DMFC and FEF for eliciting saccades (i.e. 0.1±0.24 ms) are similar to those of pyramidal tract neurons (i.e. 0.1±0.22 ms) (Stoney et al., 1968; Asanuma et al., 1976; Tehovnik & Lee, 1993; Tehovnik & Sommer, 1997), which suggests that the directly-stimulated elements in the DMFC and FEF are comprised of the output neurons from these regions. [A strength±duration curve, from which a chronaxie value is derived, is determined by delivering various levels of current over a range of pulse durations to neural elements to evoke a given response which can be, e.g. an action potential (Stoney et al., 1968), neurotransmitter release (Farber et al., 1997), self-stimulation (Matthews, 1977), or a saccadic eye movement (Tehovnik & Lee, 1993; Tehovnik & Sommer, 1997). The chronaxie is the duration value on the strength±duration curve at twice the rheobase current, which is the amount of current that evokes a response at long-pulse durations. The chronaxie is an estimate of the time constant of a directlystimulated element (Ranck, 1975). The more excitable an element, the shorter its chronaxie, such that axons have shorter chronaxies than cell bodies (axons: 0.03±7 ms; cell bodies: 7±31 ms) when the cell bodies are activated intracellularly (Ranck, 1975; see Nowak & Bullier, 1998a for details), and large myelinated axons have shorter chronaxies than small non-myelinated axons (large: 0.03±0.7 ms; small: > 1 ms)(Ranck, 1975; Li & Bak, 1976; West & Wolstencroft, 1983). The chronaxie of an axon is negatively correlated with its conduction velocity (West & Wolstencroft, 1983; Nowak & Bullier, 1998a) and positively correlated with its refractory period (Shizgal et al., 1991).] The synaptic inputs to pyramidal neurons are most numerous and effective at the level of the cell body and initial segment (Sloper & Powell, 1979; Farinas & DeFelipe, 1991a,b; Swadlow, 1992). Because electrical stimulation can bypass these inputs by stimulating the axon beyond the initial segment (Hikosaka & Wurtz, 1985, 1986;

2441

Nowak & Bullier, 1998a,b), it is highly likely that the changes in current threshold observed during active ®xation for both the DMFC and FEF occur outside the site of direct stimulation. What, then, can account for the differential effect of active ®xation on saccadic eye movements evoked from the DMFC and FEF? Saccades can be elicited from the FEF with high frequency stimulation (e.g. 1000 Hz), but this is not so for the DMFC (Tehovnik & Lee, 1993; Tehovnik & Sommer, 1997). We have argued that this difference in frequency may be related to relatively unexcitable terminal branches intervening between the DMFC and the saccade generator (Tehovnik & Sommer, 1997). If true, then stimulationtriggered input to the saccade generator from the DMFC would be less effective at evoking a saccade when other regions of the behaving brain, e.g. the FEF, are competing for access to the generator. Indeed, the behavioural state of the animal had much less of an effect on inhibiting saccades when stimulation was delivered to the FEF. Role of the DMFC and FEF in the execution of saccadic eye movements Even though both the DMFC and FEF send direct projections to the saccade generator in the brainstem (Schiller, 1977; Keating et al., 1983; Keating & Gooley, 1988; Stanton et al., 1988; Shook et al., 1988; Huerta & Kaas, 1990; Shook et al., 1990; Tehovnik et al., 1994), the FEF has more direct access to the generator and is more involved in saccade execution. This conclusion is supported by four pieces of evidence: First, the timing of the neuronal burst prior to saccade initiation is less variable and more time-locked to saccade initiation for cells in the FEF than the DMFC (Hanes et al., 1995). Second, currents for evoking saccades from the FEF are lower than the currents for evoking saccades from the DMFC under comparable behavioural conditions (current study; Tehovnik & Sommer, 1997). Third, active ®xation has a smaller effect on the current threshold for eliciting saccades from the FEF than the DMFC (current study). Finally, lesions of the FEF disrupt the execution of saccadic eye movements, but lesions of the DMFC have negligible effects (Sommer & Tehovnik, 1997; Schiller & Chou, 1998; Sommer & Tehovnik, 1999). Effect of active ®xation on the evocation of saccadic eye movements Active ®xation had a greater effect on saccadic eye movements evoked from the DMFC as compared to the saccadic eye movements evoked from the FEF. Changing the parameters of active ®xation, i.e. by shortening the ®xation duration or delaying the delivery of the juice reward, affected saccades evoked from the DMFC more than it affected saccades evoked from the FEF. Also, the change in current strength to evoke saccades during active ®xation was much greater for the DMFC than FEF. Accordingly, when electrical stimulation is delivered to the DMFC, regions outside the DMFC that mediate active ®xation (e.g. the FEF, posterior parietal lobe, superior colliculi, etc.: Bizzi, 1968; Lynch et al., 1977; Bruce & Goldberg, 1985; Munoz & Wurtz, 1993; Hanes et al., 1998) override the effect of such stimulation. Yet, when stimulation is delivered to the FEF, regions outside of the FEF that mediate active ®xation (e.g. the DMFC, posterior parietal lobe, the superior colliculus, etc.: Lynch et al., 1977; Bon & Lucchetti, 1992; Schlag et al., 1992; Munoz & Wurtz, 1993; Lee & Tehovnik, 1995) are less effective at collectively overriding the effect of stimulation. This difference concurs with a recent ®nding in humans. Magnetic stimulation of the DMFC was not at all effective at disrupting saccadic eye movements, whereas such stimulation of the FEF always disrupted such movements (Terao et al., 1998).


2442 E. J. Tehovnik et al. The current threshold to evoke saccadic eye movements from V1 and the posterior parietal cortex is much higher than the current threshold to evoke saccades from the FEF and superior colliculi (Schiller, 1977; Schiller & Stryker, 1972; Shibutani et al., 1984). Unlike the FEF and superior colliculi, which have direct access to the saccade generator, V1 and the posterior parietal cortex have indirect access via the superior colliculus (Schiller, 1977; Keating et al., 1983; Keating & Gooley, 1988). Based on these differences, we predict that active ®xation will have a greater effect on saccades evoked from V1 and the posterior parietal cortex as compared to saccades evoked from the FEF and superior colliculi. In conclusion, active ®xation has a greater effect on saccadic eye movements evoked electrically from the DMFC than it does on saccades evoked electrically from the FEF. This indicates that the DMFC, as compared to the FEF, is less well connected to the saccade generator for the execution of saccadic eye movements. The exact reasons for this difference remain to be determined.

Acknowledgements This work was supported by NIH EY08502 to P. H. Schiller. We would like to thank Tirin Moore, Marc Sommer and Andreas Tolias for taking the time to review this work. We also thank I-han Chou for the use of monkey J.

Abbreviations DMFC, dorsomedial frontal cortex; FEF, frontal eye ®elds.

References Asanuma, H., Arnold, A. & Zarzecki, P. (1976) Further study on the excitation of pyramidal tact cells by intracortical microstimulation. Exp. Brain Res., 26, 443±461. Bon, L. & Lucchetti, C. (1992) The dorsomedial frontal cortex of the macaca monkey: ®xation and saccade-related activity. Exp. Brain Res., 89, 571± 580. Bizzi, E. (1968) Discharge of frontal eye ®eld neurons during saccadic and following eye movements in unanesthetized monkeys. Exp. Brain Res., 6, 69±80. Bruce, C.J. & Goldberg, M.E. (1985) Primate frontal eye ®eld. I. Single neurons discharging before saccades. J. Neurophysiol., 53, 603±635. Bruce, C.J., Goldberg, M.E., Bushnell, M.C. & Stanton, G.B. (1985) Primate frontal eye ®elds. II. Physiological and anatomical correlates of electrically evoked eye movements. J. Neurophysiol., 54, 714±734. Chou, I. & Schiller, P.H. (1997a) Encoding of arm and eye movements in the dorsomedial frontal cortex (DMFC). Neural Control Mov. Abstr., 23, 1151. Chou, I. & Schiller, P.H. (1997b) Single cell responses in dorsomedial frontal cortex (DMFC) to eye and arm movements. Soc. Neurosci. Abstr., 23, 1552. Farber, S.A., Bogdanov, M., Marshall, D.L. & Tehovnik, E.J. (1997) Excitability of neural elements within the rat corpus striatum. J. Neurosci. Meth., 76, 93±104. Farinas, I. & DeFelipe, J. (1991a) Patterns of synaptic input on corticocortical and corticothalamic cells in the cat visual cortex. I. The cell body. J. Comp. Neurol., 304, 53±69. Farinas, I. & DeFelipe, J. (1991b) Patterns of synaptic input on corticocortical and corticothalamic cells in the cat visual cortex. II. The axon initial segment. J. Comp. Neurol., 304, 70±77. Goldberg, M.E. & Bruce, C.J. (1990) Primate frontal eye ®elds. III. Maintenance of spatially accurate saccade signal. J. Neurophysiol., 64, 489±508. Goldberg, M.E., Bushnell, M.C. & Bruce, C.J. (1986) The effect of attentive ®xation on eye movements evoked by electrical stimulation of the frontal eye ®elds. Exp. Brain Res., 61, 579±584. Hanes, D.P., Patterson, W.F. & Schall, J.D. (1998) Role of frontal eye ®elds in countermanding saccades: visual, movement, and ®xation activity. J. Neurophysiol., 79, 817±834. Hanes, D.P., Thompson, K.G. & Schall, J.D. (1995) Relationship of presaccadic activity in frontal eye ®eld and supplementary eye ®eld to saccade initiation in macaque: Poisson spike train analysis. Exp. Brain Res., 103, 85±96.

Hikosaka, O. & Wurtz, R.H. (1985) Modi®cation of saccadic eye movements by GABA-related substances. I. Effect of muscimol and bicuculline in monkey superior colliculus. J. Neurophysiol., 53, 266±291. Hikosaka, O. & Wurtz, R.H. (1986) Saccadic eye movements following injection of lidocaine into the superior colliculus. Exp. Brain Res., 61, 531± 539. Huerta, M.F. & Kaas, J.H. (1990) Supplementary eye ®eld as de®ned by intracortical microstimulation: connections in macaque. J. Comp. Neurol., 293, 299±330. . Judge, S.J., Richmond, B.J. & Chu, F.C. (1980) Implantation of magnetic search coils for measurement of eye position: an improved method. Vision Res., 20, 535±538. Keating, E.G. & Gooley, S.G. (1988) Disconnection of parietal and occipital access to the saccadic oculomotor system. Exp. Brain Res., 70, 385±398. Keating, E.G., Gooley, S.G., Pratt, S. & Kelsey, J.E. (1983) Removing the superior colliculus silences eye movements normally evoked from stimulation of the parietal and occipital eye ®elds. Brain Res., 269, 145± 148. Kurylo, D.D. & Skavenski, A.A. (1991) Eye movements elicited by electrical stimulation of area PG in the monkey. J. Neurophysiol., 65, 1243±1253. Lee, K.M. & Tehovnik, E.J. (1995) Topographic distribution of ®xationrelated units in the dorsomedial frontal cortex of the rhesus monkey. Eur. J. Neurosci., 7, 1005±1011. Li, C.L. & Bak, A. (1976) Excitability characteristics of the A- and C-®bres in peripheral nerve. Exp. Neurol., 50, 67±79. Lynch, J.C., Mountcastle, V.B., Talbot, W.H. & Yin, T.C.T. (1977) Parietal lobe mechanisms for directed visual attention. J. Neurophysiol., 40, 362± 389. Mann, S.E., Thau, R. & Schiller, P.H. (1988) Conditional task-related responses in monkey dorsomedial frontal cortex. Exp. Brain Res., 69, 460± 468. Marrocco, R.T. (1978) Saccades induced by stimulation of the frontal eye ®elds: interaction with voluntary and re¯exive eye movements. Brain Res., 146, 23±34. Matthews, G. (1977) Neural substrate for brain stimulation reward in the rat: cathodal and anodal strength-duration properties. J. Comp. Physiol. Psych., 91, 858±875. McCreery, D.B., Agnew, W.F., Yuen, T.G.H. & Bullara, L. (1990) Charge density and charge per phase as cofactors in neural injury induced by electrical stimulation. IEEE Trans. Biomed. Eng., 37, 996±1001. Munoz, D.P. & Wurtz, R.H. (1993) Fixation cells in monkey superior colliculus. I. Characteristics of cell discharge. J. Neurophysiol., 70, 559± 575. Mushiake, H., Fujii, N. & Tanji, J. (1996) Visually guided saccade versus eyehand reach: contrasting neuronal activity in the cortical supplementary and frontal eye ®elds. J. Neurophysiol., 75, 2187±2191. Nowak, L.G. & Bullier, J. (1998a) Axons, but not cell bodies, are activated by electrical stimulation in cortical gray matter. I. Evidence from chronaxie measurements. Exp. Brain Res., 118, 477±488. Nowak, L.G. & Bullier, J. (1998b) Axons, but not cell bodies, are activated by electrical stimulation in cortical gray matter. II. Evidence from selective inactivation of cell bodies and axon initial segments. Exp. Brain Res., 118, 489±500. Ranck, J.B. (1975) Which elements are excited in electrical stimulation of mammalian central nervous system: a review. Brain Res., 98, 417±440. Robinson, D.A. & Fuchs, A.F. (1969) Eye movements evoked by stimulation of frontal eye ®elds. J. Neurophysiol., 32, 637±648. Russo, G.S. & Bruce, C.J. (1993) Effect of eye position within the orbit on electrically elicited saccadic eye movements: a comparison of the macaque monkey's frontal and supplementary eye ®elds. J. Neurophysiol., 69, 800± 818. Russo, G.S. & Bruce, C.J. (1996) Neurons in the supplementary eye ®eld of rhesus monkeys code visual targets and saccadic eye movements in an oculocentric coordinate system. J. Neurophysiol., 76, 825±848. Schall, J.D. (1991a) Neuronal activity related to visually guided saccadic eye movements in the supplementary motor area of rhesus monkeys. J. Neurophysiol., 66, 530±558. Schall, J.D. (1991b) Neuronal activity related to visually-guided saccades in the frontal eye ®elds of rhesus monkeys: a comparison with supplementary eye ®elds. J. Neurophysiol., 66, 559±579. Schiller, P.H. (1977) The effect of superior colliculus ablation on saccades elicited by cortical stimulation. Brain Res., 122, 154±156. Schiller, P.H. & Chou, I. (1998) The effect of frontal eye ®eld and dorsomedial frontal cortex lesions on visually guided eye movements. Nature Neurosci., 1, 248±253.


Behavioural conditions affecting elicited saccades Schiller, P.H. & Sandell, J.H. (1983) Interactions between visually and electrically elicited saccades before and after superior colliculus and frontal eye ®eld ablations in the rhesus monkey. Exp. Brain Res., 49, 381±392. Schiller, P.H. & Stryker, M. (1972) Single-unit recording and stimulation in superior colliculus of the alert rhesus monkey. J. Neurophysiol., 35, 915± 924. Schlag, J. & Schlag-Rey, M. (1987) Evidence for a supplementary eye ®eld. J. Neurophysiol., 57, 179±200. Schlag, J., Schlag-Rey, M. & Pigarev, I. (1992) Supplementary eye ®eld: in¯uence of eye position on neural signals of ®xation. Exp. Brain Res., 90, 302±306. Shibutani, H., Sadata, H. & Hyvarinen, J. (1984) Saccade and blinking evoked microstimulation of the posterior parietal association cortex of the monkey. Exp. Brain Res., 55, 1±88. Shizgal, P., Conover, K. & Schindler, D. (1991) Medial forebrain bundle units in the rat: dependence of refractory period estimates on pulse duration. Behav. Brain Res., 42, 151±160. Shook, B.L., Schlag-Rey, M. & Schlag, J. (1988) Direct projection from the supplementary eye ®eld to the nucleus raphe interpositus. Exp. Brain Res., 73, 215±218. Shook, B.L., Schlag-Rey, M. & Schlag, J. (1990) Primate supplementary eye ®eld: I. Comparative aspects of mesencephalic and pontine connections. J. Comp. Neurol., 301, 618±642. Sloper, J.J. & Powell, T.P.S. (1979) A study of the axon initial segment and proximal axon of neurons in the primate motor and somatic sensory cortices. Phil. Trans. R. Soc. Lond. Series B, 285, 173±197. Sommer, M.A. & Tehovnik, E.J. (1997) Reversible inactivation of macaque frontal eye ®eld. Exp. Brain Res., 116, 229±249. Sommer, M.A. & Tehovnik, E.J. (1999) Reversible inactivation of macaque dorsomedial frontal cortex: effects on saccades and ®xations. Exp. Brain Res., 124, 429±446. Sparks, D.L. & Mays, L.E. (1983) Spatial localization of saccade targets. I. Compensation for stimulation-induced perturbations in eye position. J. Neurophysiol., 49, 45±63. Stanton, G.B., Goldberg, M.E. & Bruce, C.J. (1988) Frontal eye ®eld efferents

2443

in the macaque monkey: II. Topography of terminal ®elds in midbrain and pons. J. Comp. Neurol., 271, 493±506. Stoney, S.D., Thompson, W.D. & Asanuma, H. (1968) Excitation of pyramidal tact cells by intracortical stimulation: effective extent of stimulating current. J. Neurophysiol., 31, 659±669. Swadlow, H.A. (1992) Monitoring the excitability of neocortical efferent neurons to direct activation by extracellular current pulses. J. Neurophysiol., 68, 605±619. Tanji, J. (1994) The supplementary motor area in the cerebral cortex. Neurosci. Res., 19, 251±268. Tehovnik, E.J. (1995) The dorsomedial frontal cortex: eye and forelimb ®elds. Behav. Brain Res., 67, 147±176. Tehovnik, E.J. (1996) Electrical stimulation of neural tissue to evoke behavioural responses. J. Neurosci. Meth., 65, 1±17. Tehovnik, E.J. & Lee, K.M. (1993) The dorsomedial frontal cortex of the rhesus monkey. Topographic representation of saccades evoked by electrical stimulation. Exp. Brain Res., 96, 430±442. Tehovnik, E.J., Lee, K.M. & Schiller, P.H. (1994) Stimulation-evoked saccades from the dorsomedial frontal cortex of the rhesus monkey following lesions of the frontal eye ®elds and superior colliculus. Exp. Brain Res., 98, 179±190. Tehovnik, E.J., Slocum, W.M., Tolias, A.S. & Schiller, P.H. (1998) Saccades induced electrically from the dorsomedial frontal cortex: evidence for a head-centered representation. Brain Res., 795, 287±291. Tehovnik, E.J. & Sommer, M.A. (1997) Electrically evoked saccades from the dorsomedial frontal cortex and frontal eye ®elds: a parametric evaluation reveals differences between areas. Exp. Brain Res., 117, 369±378. Terao, Y., Fukuda, H., Ugawa, Y., Hikosaka, O., Hanajima, R., Furubayashi, T., Sakai, K., Miyauchi, S., Sasaki, Y. & Kanazawa, I. (1998) Visualization of the information ¯ow through human oculomotor cortical regions by transcranial magnetic stimulation. J. Neurophysiol., 80, 936±946. West, D.C. & Wolstencroft, J.H. (1983) Strength-duration characteristics of myelinated and non-myelinated bulbospinal axons in the cat spinal cord. J. Physiol., 337, 37±50.
