| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Department of Physiology and Pharmacology, University of Western Ontario; 2Robarts Research Institute, London, Ontario; 3Department of Psychology and 4Graduate Program in Neuroscience, University of Western Ontario, London, Ontario, Canada
Submitted 5 May 2004; accepted in final form 19 February 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Miller 2000; Miller and Cohen 2001) in this cognitive control of action.
The antisaccade task (Hallett 1978), in which subjects are instructed to look away from a flashed visual stimulus, rather than toward it, is a useful task for investigating the flexible control of movement (Everling and Fischer 1998; Munoz and Everling 2004). Subjects sometimes generate errors on antisaccade trials. These errors usually take the form of a rapid saccade toward the stimulus followed by a secondary correction saccade away from the stimulus. Thus correct performance of the antisaccade task requires at least 2 processes: the suppression of a reflexive prosaccade toward the stimulus and the generation of the antisaccade away from the stimulus to an empty location in the visual field.
Single-neuron recording studies in monkeys performing an antisaccade task indicate that correct task performance requires the suppression of saccade neurons in the frontal eye fields (FEFs) (Everling and Munoz 2000) and the superior colliculus (Everling et al. 1998) before stimulus presentation. It has been proposed that without sufficient inhibition, the incoming visual signal could sum with an elevated prestimulus neuronal activity and directly pass the neurons' threshold for saccade initiation. The result would be a reflexive saccade that moves the eyes toward the stimulus instead of away from it. Several brain areas, mainly in the frontal lobes (Gaymard et al. 1998; Guitton et al. 1985; Pierrot-Deseilligny et al. 2003) and basal ganglia (Briand et al. 1999; Lasker and Zee 1997), have been proposed as putative sources of this top-down signal. Indeed, single-neuron recordings in the monkey supplementary eye fields (SEFs) have found increased neural activity on antisaccade trials during the preparatory period that preceded stimulus presentation (Amador et al. 2003; Schlag-Rey et al. 1997).
Results from recent event-related functional magnetic resonance imaging (fMRI) studies in humans that used long preparatory periods from 4 to 14 s are consistent with monkey electrophysiology data in showing differences between pro- and antisaccade trials during the preparatory periods before stimulus presentation and saccade initiation in 3 areas of the frontal lobe: SEF (Curtis and D'Esposito 2003), FEF (Connolly et al. 2002; Desouza et al. 2003), and dorsolateral prefrontal cortex (DLPFC) (Connolly et al. 2002; Curtis and D'Esposito 2003; Desouza et al. 2003). Of particular interest is the study by Curtis and D'Esposito (2003), which directly compared correct trials and error trials. The authors reported a larger functional activation during the preparatory period before correct antisaccades compared with errors in an area located rostral to the supplementary eye fields to which they refer as presupplementary motor area (pre-SMA). Further, they found stronger functional activations in the FEF and intraparietal sulcus (IPS) during the stimulus-response period on correct antisaccade trials compared with error trials.
Curtis and D'Esposito (2003) used a region of interest analysis in their study and investigated the activations in the SEF, FEF, pre-SMA, and IPS. These 4 regions were identified based on anatomical landmarks and activations in a visually guided saccade task. It therefore remains unknown whether other brain areas also show differences in the preparatory period or stimulus-response period between correct antisaccade trials and error trials. To address this question, we performed voxelwise analyses on the preparatory period and the stimulus-response period using the general linear model (GLM) method. Our results demonstrate that additional cortical areas show differences between correct antisaccade trials and error trials. Of particular clinical relevance might be the DLPFC and the anterior cingulate cortex (ACC).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Ten subjects (7 male, 3 female, mean age of 28 yr) provided informed consent and participated in this study. Two subjects, SE and MB, were coauthors. All were right handed, could see clearly at arm's length without glasses, and reported no history of head injury, epilepsy, or neurological or psychiatric disorder. The experiments were approved by the University of Western Ontario Review Board for Health Sciences Research Involving Human Subjects and are in accordance with the 1964 Declaration of Helsinki.
Experimental task
Each event-related functional scan (Fig. 1A) began with 14 s of fixation on a central white cross (1.3°). A change in the color of the fixation cross conveyed the instruction of whether to generate a pro- or antisaccade on stimulus presentation. The colors (red, green, or blue) used as the pro-/antisaccade instruction were kept constant across functional scans within subjects and were randomized across subjects. After 10 s, the central fixation cross disappeared, and 200 ms of darkness (gap period) were introduced before a peripheral stimulus (white circle, 2°) was flashed for 500 ms, either 10° to the left or 10° to the right of the initial fixation cross. Subjects were required to maintain fixation on the central cross during the intertrial interval (ITI), the preparatory period, and the gap period. The gap period was introduced to increase the number of errors on antisaccade trials (Bell et al. 2000; Fischer and Weber 1997; Forbes and Klein 1996).
|
Each subject performed between 5 and 7 event-related functional scans. Each scan was 302 s long.
Visual display and eye tracking
Visual stimuli were generated using SuperLab Pro 2.0 software (Cedrus, San Padro, CA) and presented using SMI iView-fMRI Eye tracking (SensoMotoric Instruments, Needham/Boston, MA) and Silent Vision SV-40 21 (Avotec, Stuart, FL). This device uses fiber optics housed in dual stalks that are placed in front of a subject's eyes, allowing presentation of visual stimuli (30° horizontally x 23° vertically with a resolution of 800 x 600 pixels) and simultaneous CCD video-based infrared video eye tracking of the right eye at 60 Hz, spatial resolution 0.1°, limited by accuracy to 0.5–1.0°. After subjects were placed in the MR bed, the video display and tracking apparatus were adjusted manually for stimulus viewing and eye tracking. Stimulus projection was binocular, with each LCD display aligned independently along 3 axes for each eye. Once optimal viewing was established for each subject, the infrared video eye tracker was calibrated at center position and 4 eccentric points for the subject's right eye. Additionally, each subject's eye display on the computer monitor during data collection was recorded on digital video and viewed by the experimenters as each event-related trial was completed. Analysis of the eye movement traces was performed off-line.
Imaging and data analysis
All imaging data were acquired on a 4-Tesla whole body MRI system (Varian, Palo Alto, CA; Siemens, Erlangen, Germany). A transmit–receive cylindrical hybrid birdcage head coil was used for radio-frequency transmission and detection of signal. Imaging planes for the functional scans were prescribed from a series of sagittal anatomic images acquired with high gray/white matter contrast (i.e., T1-weighted). The functional planes were 11 contiguous, 6-mm-thick axial-oblique slices oriented in a transverse plane and extending from the superior extent of the cortex to the caudate. These data were collected using blood oxygenation level–dependent (BOLD) images (
-weighted) acquired continuously using an interleaved, 2-segment, optimized spiral imaging protocol (64 x 64 matrix size, TR = 500 ms, TE = 12 ms, flip angle = 30°, 20.0 cm FOV, volume collection time = 1 s). Functional data were superimposed on high-resolution anatomical images (T1-weighted anatomic reference volume acquired along the same orientation as the functional images using a 3-dimensional spiral sequence, 256 x 256 x 128 matrix size, 1.5 mm reconstructed slice thickness, TI = 1,300 ms, TR = 50 ms, TE = 3.0 ms, intersegment delay = 2,550 ms) obtained during each experimental session. Analyses were conducted using BrainVoyager 2000 version 4.8 (Brain Innovation, Maastricht, The Netherlands). Functional runs were resampled into 3 x 3 x 3-mm voxels, superimposed on anatomical scans for each subject, aligned on the anterior commissure–posterior commissure axis, and scaled to the Talairach standard (Talairach and Tournoux 1988). All functional image series underwent motion correction, temporal filtering [linear trend removal, high-pass filter in Fourier domain with cutoff of 6 cycles/run, Gaussian filter in time domain with full width at half-maximum (FWHM) of 2.8 s], and spatial filtering (Gaussian filter in spatial domain with FWHM of 4.00 mm). Each trial was analyzed off-line using the time-locked eye-position traces recorded in the MR scanner and digital video. A trial did not abort if the subject lost fixation, although that trial was excluded from further analysis.
Trials in which subjects broke fixation during either the ITI, 200-ms-gap period, or preparatory period (17 of 744, or 2%) were excluded from further data analysis. To ensure that anticipatory saccades were excluded only saccades with latencies in the normal range (between 80 and 600 ms from target stimulus onset) were included in the analysis. The remaining trials were categorized as correct prosaccades in which the subject generated a saccade toward the peripheral stimulus (n = 243), correct antisaccades in which the subject generated a saccade away from the peripheral stimulus (n = 401), or error antisaccades in which the subject was instructed to look away from the peripheral stimulus but instead generated an initial saccade toward the stimulus (n = 83 of 744, or 11%). Trials in which subjects generated an antisaccade when instructed to make a prosaccade were very rare (n = 18 of 744, or 2%) and so were excluded from further analyses.
Functional data were statistically analyzed using the GLM framework. Data from left and right saccadic eye movements were combined. We compared BOLD activity between correct antisaccades and prosaccades and between correct antisaccades and error antisaccades in 3 epochs. We divided the preparatory period into an early period (lasting from task-cue onset to 4 s into the preparatory period for a total of 4 volumes of functional data acquisition) and a late period (lasting from 5 s into the preparatory period until the end of the preparatory period for a total of 6 volumes of functional data acquisition). Presumably, activation in the early period was more associated with the switch of the color or the central-fixation cross, whereas activation in the late period was more associated with the task-specific preparatory processes. We also defined a stimulus-response period (lasting from the peripheral stimulus onset until 4 s after stimulus onset for a total of 4 volumes of functional data acquisition). Activation in this period constituted a variety of different neural processes, including stimulus presentation, saccade generation, visual feedback after the saccade, fixation, and the return saccade. To this end, we defined, for each of our 10 subjects, 3 sets of 3 predictor functions, including a predictor function for the early preparatory period, late preparatory period, and stimulus/response period for each of the 3 trial types of interest (correct antisaccades, error antisaccades, and prosaccades). This resulted in a total of 9 predictors per subject for a grand total of 90 predictors. To construct each of the predictor curves (Fig. 1C), we defined a series of box-car functions extending over each instance of the appropriate epoch occurring in all functional runs for a given subject, and we then convolved the series of box cars with BrainVoyager's model of the hemodynamic transfer function using parameters tau = 1.25 and delta = 0. BrainVoyager models the hemodynamic function as a gamma curve [formula: h(t) = (t/
)n–1et//(n – 1)!] after Boynton et al. (1996). In addition to the 90 predictors thus defined, BrainVoyager automatically added an offset predictor for each functional run consisting of a constant function defined over the duration of the run in question and zeros for all other runs. BrainVoyager assembled the full set of 90 separate task activation predictors and 61 run offset predictors into a design matrix. This design matrix represented the BOLD activation model. Data were then z-scaled (mean removed and divided by SD) on a run-by-run basis to remove arbitrary signal amplitude effects.
The BOLD activation model was fitted to the functional data using linear regression, which determines the best linear fit of the model to the data, where the "best fit" is the one that minimizes the sum of the squared residuals (fitted model – original data)2. "Fitting" requires scaling the columns of the design matrix by appropriate beta weights (one beta weight per column of the design matrix, where each column contains one predictor function). Separate subject predictors were used to ensure that no one subject could disproportionately influence the beta weighting because each subject's data were used to fit a subject-specific set of 9 activation curve predictors. Statistical comparisons were done using t-test on the beta weights. To do this, a contrast function was defined consisting of the sum of one set of predictors of interest (such as the late preparatory period of all correct antisaccade trials for all subjects) minus the sum of another predictor of interest (such as the late preparatory period of all error antisaccade trials for all subjects). The resulting difference was compared with a null hypothetical value of zero using a t-test, where rejection of the null means that there is a significant difference between the predictors. Computed P values were Bonferroni corrected for multiple comparisons.
The GLM analysis resulted in statistical contrast activation maps for each comparison of interest (P < 0.01, corrected for multiple comparisons; and a cluster threshold size of >50 voxels). Mean BOLD signal time courses for each region identified in the contrast comparisons were computed to illustrate how the BOLD signal evolved over the full course of an event-related trial. Raw data from each trial were transformed into percentage signal change values, [(signal – baseline)/baseline] x 100, where baseline was defined as the average signal over the first 2 s of the trial on a trial-by-trial basis. Ten "within-subject" mean activation curves were computed for each of antisaccade, prosaccade, and error antisaccade tasks. Then, "between-subject" mean curves were computed (mean of mean curves) by taking the mean across all 10 subjects for each of the 3 trial types.
The mean beta weights derived from the GLM were also calculated for each area localized in the statistical comparisons. These were calculated by first extracting the beta weights separately for each of the 10 subjects and then combining these individual subject averages to calculate the group average associated with the GLM contrast for each area localized in this statistical comparison. To investigate the consistency of these group statistical activation maps across individual subjects, post hoc random effects analyses were conducted using paired-samples t-test (P < 0.05) across the individual subject beta weights for each area found to show significant differences in the group GLM contrast.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
All subjects generated errors on a number of the antisaccade trials (i.e., they initially looked toward the peripheral stimulus before they looked away from it). Figure 1B shows horizontal eye position traces from one subject performing antisaccades in the MR scanner during the experiment. Black traces show correct trials in which the subject looked away from a stimulus, whereas gray traces show error trials in which the subject initially looked toward the stimulus before generating an antisaccade away from it. The error rates of individual subjects ranged from 4 to 35% [16 ± 3.4% (mean ± SE), median 11%]. Subjects did not show significant differences in the percentage of errors between leftward (48.1%) and rightward (51.9%) stimulus presentations (paired t-test, t = –0.36, df = 9, P = 0.73). No significant difference in error rates was found between experienced subjects (defined as those subjects who had previously performed an antisaccade task for any experiments) and naïve subjects (defined as those subjects who had never before performed antisaccades in any experimental task) (t-test, t = 0.14, df = 8, P = 0.89).
Figure 2 shows the saccadic reaction time (SRT) histograms for prosaccades, antisaccades, and antisaccade errors. The mean SRT differed significantly between correct antisaccades (355 ms), prosaccades (289 ms), antisaccade errors (326 ms), and the correction saccade on antisaccade error trials (178 ms). All subjects generated correction saccades on antisaccade error trials. The SRT of antisaccade errors indicates they were not express saccades. Correct antisaccades had significantly longer mean SRTs than prosaccades (paired t-test, P < 0.01), antisaccade errors (paired t-test, P < 0.05), and correction saccades on antisaccade error trials (paired t-test, P < 0.001). Prosaccades had significantly shorter mean SRTs than both correct antisaccades (paired t-test, P < 0.01) and antisaccade errors (paired t-test, P < 0.001); however, SRTs for correction saccades on antisaccade error trials were significantly shorter (paired t-test, P < 0.001).
|
We compared the BOLD signal intensities for the early preparatory period, late preparatory period, and saccade period (see Fig. 1C) separately for correct antisaccades versus correct prosaccades and for correct antisaccades versus error trials by using the general linear model (P < 0.01, corrected for multiple comparisons; see METHODS for details). Both the early and late preparatory periods were identical across trial types regarding sensory stimuli presented. Subjects had been instructed by a color cue of the upcoming trial type but had not been presented with a peripheral stimulus and were maintaining fixation on a central cue.
Analysis during the early preparatory period, which began with the onset of the instruction cue to make either a pro- or antisaccade and lasted until 4 s into the instruction period (for a total of 4 volumes of functional data acquisition), showed no significant differences between correct antisaccades and prosaccades or between correct antisaccades and errors.
Comparison of correct antisaccades and prosaccades
Figure 3 shows the group statistical activation map of the GLM contrast comparing correct antisaccades to prosaccades during the late preparatory period, which began 5 s after the onset of the instruction cue to make either a pro- or antisaccade and lasted until the end of the instruction period (for a total of 6 volumes of functional data acquisition), combining all 10 subjects. All regions identified showed a significant increase in activation for antisaccade trials over prosaccade trials during this period. Figure 4 shows the average BOLD signal time courses (see METHODS for details) (gray traces, prosaccade; black traces, antisaccade). Mean beta weights (see METHODS for details) are also shown (gray bars, prosaccades; black bars, antisaccade). To investigate the consistency of these group statistical activation maps across individual subjects, post hoc random effects analyses were conducted using paired-samples t-test across the individual subject beta weights for each area found to show significant differences in the group GLM contrast. The differences between the beta weights for each condition were significant (paired t-test, P < 0.05), indicating these results were consistent across subjects. Figure 4 (top) shows the BOLD signal time course averaged across 10 subjects for an area in the right anterior cingulate cortex (ACC). This area was significantly more active for correct antisaccades than correct prosaccades during the late preparatory period, which illustrates the typical activation patterns observed for all significant areas for the late preparatory period comparison. The instruction signal indicating whether to prepare for an antisaccade or a prosaccade, conveyed by a change in the color of the central fixation cross, evoked an increase in the BOLD signal that peaked after about 4 s. The signal then started to decay but built up again by 7–8 s after instruction cue presentation, still before stimulus onset. This buildup of the BOLD signal continued into the stimulus-response period and peaked at around 3 s after the presentation of the peripheral stimulus. It decayed back to the baseline level after about 10 s. It may be noted that a decrease in activation with respect to baseline may reflect the fact that fixation (our baseline condition) is an active process, not a zero activation state.
|
|
|
|
|
Subjects sometimes fail to suppress a reflexive saccade toward the stimulus when instructed to generate an antisaccade. These errors are especially frequent in young children and patients with certain neurological or psychiatric disorders that involve the frontal cortex and/or the basal ganglia (Everling and Fischer 1998; Munoz and Everling 2004). Therefore it has been proposed that these areas provide top-down suppression signals on antisaccade trials before the stimulus is presented. To identify cortical and subcortical areas that may be related to antisaccade task performance, we compared the BOLD signals during the late preparatory period between correct antisaccade trials and error trials in which subjects initially looked toward the stimulus. In both cases, subjects were maintaining fixation on an identical central visual fixation cross during the preparatory period. This comparison revealed differences in only 5 areas, all within the frontal lobe (Figs. 5 and 6 and Table 1). Correct antisaccades were preceded by an increased BOLD signal compared with errors in 3 of these 5 areas. The largest of these areas was located in the right DLPFC. Areas located in ACC and pre-SEF were also identified. In contrast, the other 2 areas localized showed a different pattern of activation, with an increase in the BOLD signal preceding antisaccade errors compared with correct antisaccades. To investigate the consistency of these group statistical activation maps across individual subjects, post hoc random effects analyses were conducted using paired-samples t-test across the individual subject beta weights for each area found to show significant differences in the group GLM contrast. The differences between the beta weights for each condition were significant (paired t-test, P < 0.05), indicating these results were consistent across subjects.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Preparatory period
A recent study by Curtis and D'Esposito compared prosaccades, correct antisaccades, and error trials with fMRI in an event-related design that shared some similarities with ours (Curtis and D'Esposito 2003). The key differences were their use of shorter instruction periods (6 vs. 10 s in our study); the requirement that subjects memorize the instruction during the preparatory period, whereas the instruction cue was present throughout the trial in our study; and the number of potential stimulus locations (8 vs. 2 in our study). Data analysis was also different because the authors investigated preselected, anatomically defined regions of interest [FEF, SEF, presupplementary motor area (pre-SMA), and intraparietal sulcus (IPS)]. Among these areas, they found differences between prosaccades and antisaccades only in the pre-SMA. This area is very close to, and could even be the same as, the areas that we identified here as the pre-SEF based on the more anterior anatomical location and Talaraich coordinates compared with those of other studies (Grosbras et al. 1999). The lack of any differences in oculomotor areas (e.g., FEF, SEF) between pro- and antisaccades during the preparatory period is surprising given the results from other human event-related fMRI studies. Connolly et al. (2002) compared activations in the FEF and IPS on prosaccade and antisaccade trials in gap tasks with 0-, 2-, and 4-s gap periods. They reported a buildup of activation in the FEF but not in the IPS during the gap period. This buildup was larger on antisaccade trials. Our group has recently compared pro-and antisaccades in an event-related design with variable preparatory periods between 6 and 14 s (Connolly et al. 2002; Desouza et al. 2003). Consistent with the present study, we found larger activations in the FEF and DLPFC on antisaccade trials. Activations in the SEF and IPS were not significantly different in the previous study, although they approached significance.
The main goal of the present study was to compare directly the prestimulus activations between correct antisaccades and errors. Contrary to the comparison between pro- and antisaccades, this comparison revealed significant differences in only a few frontal cortical areas (DLPFC, ACC, pre-SEF, superior frontal sulcus) but not in the FEF or IPS. In contrast, Curtis and D'Esposito (2003) found differences between correct trials and error trials in the SEF, bilateral FEF, pre-SMA, and right IPS. It is not clear whether the differences in task design (shorter instruction periods, memorization of the instruction, more potential stimulus locations) can account for these differences between our studies. It is conceivable that the larger number of potential stimulus locations could account for the small differences between pro- and antisaccades in the Curtis and D'Esposito (2003) study. Basso and Wurtz (1998) showed that the preparatory activity of saccade neurons in the superior colliculus (SC) decreases with an increase in the number of potential target locations. In fact, SC saccade neurons exhibit almost no preparatory prestimulus activity in tasks with 8 potential target locations. Saccade preparation therefore might have been larger in our study, which used only 2 target locations, than that in the study of Curtis and D'Esposito (2003). Further, in our task, the instruction stimulus was present throughout the entire trial and subjects did not have to memorize the instruction during a delay period. Consequently, it might be possible that some of the errors in the Curtis and D'Esposito study were failures of subjects to memorize the correct task instruction, rather than to suppress a reflexive saccade.
We separated the preparatory period into 2 parts (early preparatory period, beginning at the onset of the instruction cue and lasting until 4 s into the instruction period; and late instruction period, beginning 5 s after the onset of the instruction cue and lasting until the end of the instruction period). Although a number of areas showed modulation in the BOLD signal activation in response to the change in color of the fixation stimulus, no areas showed any significant differences between pro- and antisaccade trials or between correct and error antisaccade trials. This is consistent with Curtis and D'Esposito's study, in which they also did not find significant differences during the instructional cue period. These findings suggest that the initial neural processes during the preparatory period do not differ between pro- and antisaccades or between correct trials and error trials.
Stimulus-response period
Curtis and D'Esposito (2003) reported larger activations in the FEF, SEF, pre-SEF, and IPS for antisaccade than for prosaccades in the stimulus-response period. This is in line with our recent study (Desouza et al. 2003). However, the differences between pro- and antisaccades disappeared when the preparatory period was used as the baseline, indicating that the larger activation for antisaccades in the stimulus-response period is a carryover effect from the preparatory period (Desouza et al. 2003). Similarly, in the present study, the late preparatory period predictor already accounted for differences that emerged before the stimulus-response period. Together, these findings support the interpretation that the higher activations found for antisaccades versus prosaccades in previous block design imaging studies originate predominantly from increases in activation related to preparatory set for antisaccades and not from increases in response-related processes for anti- over prosaccades. However, in the present study, we found an increased activation for prosaccades over antisaccades in a number of cortical areas during the stimulus-response period. This increased activation for pro- over antisaccades may not be surprising, given the role of many of these areas in visual processing. In a prosaccade task, the subject makes a saccade to a visual target; in an antisaccade the subject makes a saccade to blank visual space. Therefore the different foveal stimulation after prosaccades compared with antisaccades may account for some of the differences. This was different from our previous study in which we presented a stimulus at the target location after the antisaccade and required subjects to fixate there for 12 s (Desouza et al. 2003). In the present study, we also observed an increase during the stimulus-response period for prosaccades in some frontal regions (bilateral superior frontal gyrus and superior frontal sulcus). To our knowledge, these areas have no known role in visual fixation or saccade control.
Neural processes associated with prosaccade and antisaccade trials
The correct performance of randomly interleaved pro- and antisaccade trials involves a number of different neural processes. Our event-related trial design with a long instruction period allowed us to dissociate between processes during the preparatory period and those during the stimulus-response period.
Initially, subjects had to identify the color of the fixation point after it switched color. The absence of any differences between pro- and antisaccades and between correct antisaccades and errors during the early preparatory period suggest that this process was the same for all conditions. Subjects then had to use this information to prepare either a prosaccade or an antisaccade. The location of the stimulus was not known at this point so subjects could not prepare the exact metrics of the saccade. However, because only 2 stimulus locations were used in this study, subjects could prepare 2 potential saccades. Microelectrode recordings in monkeys have demonstrated that the activity during the preparatory or instruction period differs between pro- and antisaccades in the oculomotor system. Some neurons show an increase in activity [fixation neurons in rostral SC (Everling et al. 1999), FEF fixation neurons (Munoz and Everling 2004), SEF fixation (Amador et al. 2003), and saccade neurons (Amador et al. 2003; Schlag-Rey et al. 1997)], whereas other neurons decrease their activity [saccade neurons in caudal SC (Everling et al. 1999), FEF saccade neurons (Everling and Munoz 2000)] on antisaccade trials. Together, these modulations have been interpreted as support for the hypothesis that the correct performance of an antisaccade requires the presetting of the oculomotor system to allow for the preparation of an antisaccade by reducing the probability of a reflexive prosaccade (Munoz and Everling 2004).
It is in this late preparatory period before the stimulus is presented, where we observed the largest differences between pro- and antisaccades and between correct antisaccades and errors in many areas. In particular, the areas that show significant differences for the comparison of correct antisaccades and errors might be involved in some aspect of saccade suppression. For example, it has been shown that patients with discrete lesions of the DLPFC have difficulties suppressing the automatic prosaccade in the antisaccade task (Gaymard et al. 2003; Pierrot-Deseilligny et al. 1991, 2003; Walker et al. 1998). Further, a recent fMRI study that compared BOLD signals associated with antisaccade task performance in schizophrenic patients and healthy controls found an increased BOLD signal in the right DLPFC for control subjects but not for schizophrenic patients (McDowell et al. 2002), suggesting that abnormalities in the DLPFC may play an important role in impaired antisaccade task performance.
The finding of mainly right-sided areas in the frontal lobe showing increased activation for correct versus error antisaccades in the late preparatory period is consistent with previous studies that have demonstrated varying degrees of right-hemisphere lateralization during antisaccades (Desouza et al. 2003; Ettinger et al. 2005; McDowell et al. 2002; Walker et al. 1998) and other tasks requiring the withholding of a prepotent motor response (Garavan et al. 1999).
A recent case study suggests that the direct corticotectal pathway may be involved in saccade suppression (Gaymard et al. 2003). Similarly, a role of the SEF in the suppression of reflexive saccades has been suggested by single-neuron recordings in monkeys (Amador et al. 2003). Many SEF neurons with fixation-related activity showed an increased activation on antisaccade trials compared with prosaccade trials during the preparation periods before stimulus presentation. Moreover, the activity of these neurons was lower during the prestimulus period on trials in which the monkey failed to suppress a reflexive prosaccade on an antisaccade trial. In contrast, studies with patients with SEF lesions have not found differences between correct trials and error trials (Gaymard et al. 1990; Husain et al. 2003).
The large number of frontal and parietal cortical areas that exhibited differences between pro- and antisaccades, however, makes it unlikely that these differences can be attributed solely to a reduced saccade activity and an increased fixation activity on anti- compared with prosaccade trials. The activation in some areas might be more related to other antisaccade-specific task requirements. Correct antisaccades require not only the suppression of a prosaccade but also the inversion of the saccade target vector. This process is not required for prosaccades because the visual stimulus activates for these saccades the correct population of saccade neurons in the SC and FEF. In the antisaccade task, the initial visual activity must be suppressed, and saccade neurons must be activated in the opposite hemifield. Single-neuron recordings in monkeys suggest that the lateral intraparietal (LIP) area seems to be involved in this process. Zhang and Barash (2000) identified a paradoxical type of response among some visual neurons in LIP for memory-delayed antisaccades. These visual LIP neurons that code for the saccade vector were activated at about 50 ms after the visual neurons on the opposite side of the brain. The authors argue that this activation might represent a remapped visual response through a nonstandard input pathway that was activated by "some context-categorization process." The BOLD activations in the IPS during the late preparatory period might represent the preparation of this pathway.
Single-neuron recordings in monkey SEF have shown that many SEF neurons have larger stimulus-related and saccade-related responses for antisaccades than for prosaccades (Amador et al. 2003; Schlag-Rey et al. 1997). This increased saccade activity might be needed to compensate for the reduced saccade activity in the SC and FEF for antisaccades. The larger activation in the pre-SEF in the late preparatory period might be related in part to the preparation of this area for the generation of a voluntary saccade to an empty location in space.
The actual stimulus-response period constitutes many different neural processes: stimulus detection, vector inversion for antisaccades, saccade generation, fixation, new visual feedback, generation of a correct antisaccade after an error, and generation of a return saccade back to center. fMRI does not provide the temporal resolution that is needed to dissect these processes, many of which occur in parallel on the timescale of only tens or a few hundreds of milliseconds. Most of the differences that we found between pro- and antisaccades during the stimulus-response period may simply be related to new foveal stimulation after prosaccades. The comparison of correct antisaccades and errors is even more difficult during this period because subjects generated 2 saccades on correct antisaccade trials (antisaccade and refixation saccade) and 3 saccades on error trials (erroneous prosaccade, antisaccade, refixation saccade). Surprisingly, we did not find significantly greater activations in any of the saccade-related areas on error trials. The only area that exhibited significantly greater activation on error trials than on correct antisaccade trials was the ACC. The same area showed lower BOLD signal intensity during the preparatory period on error trials.
There is growing consensus that the ACC has a role in performance monitoring. The ACC region that was identified in this study (x = 1, y = 10, z = 43) is very close to the ACC region recently identified in a Stroop task (x = 4, y = 1, z = 43) in which subjects were presented with colored words in either compatible or incompatible ink (MacDonald III et al. 2000). The ACC was more active during the response period when subjects had to name the ink colors compared with when they were instructed to read the words. The ACC did not show any differential activation during the preparatory period between the word-reading and ink color–naming conditions. In contrast, the DLPFC was more active in that study during the preparatory period of the print color–naming condition but did not show any differences during the response condition. The authors therefore hypothesized that the DLPFC provides top-down control to support the task-appropriate behavior, whereas the ACC is involved in conflict monitoring.
Clearly, in our present study both the DLPFC and the ACC were more active during the preparatory period on correct antisaccade trials than on error trials. This suggests that both areas may have a role in top-down control. Our finding of an increased activation of the ACC during the stimulus-response period on error trials, however, supports a role of the ACC in error monitoring (Botvinick et al. 2001). Psychophysical studies have shown that response errors are associated with a sharp negative deflection in the event-related potential (ERP) with a frontocentral distribution that peaks at about 80 ms after the incorrect response (Falkenstein et al. 1991; Gehring et al. 1995). Nieuwenhuis et al. (2001) demonstrated that this error-related negativity is also present after response errors in an antisaccade task. They further showed that this ERP component is similar for perceived and unperceived direction errors in this task, whereas a later error positivity, which might be related to a P3 component, is present only for perceived direction errors. It thus seems likely that the error-related negativity found in ERP studies originates predominantly in the ACC. Recent single-unit recordings in monkeys challenge this hypothesis by demonstrating an error signal in the SEF (Stuphorn et al. 2000) but not in the ACC (Ito et al. 2003) in a saccadic countermanding paradigm (Hanes and Schall 1995). We found some activation in the SEF for the stimulus-response period comparison of errors and correct responses, but the peak activation was clearly located in the ACC. It remains to be determined whether this is related to differences between the 2 species, the tasks, or the techniques.
Taken together, our data suggest that the preparation of an antisaccade activates a large frontal and parietal network that may be involved in presetting the oculomotor system for the antisaccade task. This requires a strengthening of fixation-related activity, a decrease of saccade-related activity, and the preparation for the upcoming saccade vector inversion and voluntary saccade generation before the stimulus is presented (Munoz and Everling 2004). It will be interesting to test whether high error rates in the antisaccade task in schizophrenia and other psychiatric disorders can also be linked to different activation levels during the preparatory period in any of these areas.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: S. Everling, The Centre for Brain and Mind, Robarts Research Institute, 100 Perth Drive, London, Ontario N6A 5K8, Canada (E-mail: severlin{at}uwo.ca)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Basso MA and Wurtz RH. Modulation of neuronal activity in superior colliculus by changes in target probability. J Neurosci 18: 7519–7534, 1998.
Bell AH, Everling S, and Munoz DP. Influence of stimulus eccentricity and direction on characteristics of pro- and antisaccades in non-human primates. J Neurophysiol 84: 2595–2604, 2000.
Botvinick MM, Braver TS, Barch DM, Carter CS, and Cohen JD. Conflict monitoring and cognitive control. Psychol Rev 108: 624–652, 2001.[CrossRef][Web of Science][Medline]
Boynton GM, Engel SA, Glover GH, and Heeger DJ. Linear systems analysis of functional magnetic resonance imaging in human V1. J Neurosci 16: 4207–4221, 1996.
Briand KA, Strallow D, Hening W, Poizner H, and Sereno AB. Control of voluntary and reflexive saccades in Parkinson's disease. Exp Brain Res 129: 38–48, 1999.[CrossRef][Web of Science][Medline]
Connolly JD, Goodale MA, Menon RS, and Munoz DP. Human fMRI evidence for the neural correlates of preparatory set. Nat Neurosci 5: 1345–1352, 2002.[CrossRef][Web of Science][Medline]
Curtis CE and D'Esposito M. Success and failure suppressing reflexive behavior. J Cogn Neurosci 15: 409–418, 2003.[CrossRef][Web of Science][Medline]
Desouza JF, Menon RS, and Everling S. Preparatory set associated with pro-saccades and anti-saccades in humans investigated with event-related fMRI. J Neurophysiol 89: 1016–1023, 2003.
Duncan J. An adaptive coding model of neural function in prefrontal cortex. Nat Neurosci Rev 2: 820–829, 2001.
Ettinger U, Antonova E, Crawford TJ, Mitterschiffthaler MT, Goswani S, Sharma T, and Kumari V. Structural neural correlates of prosaccade and antisaccade eye movements in healthy humans. Neuroimage 24: 487–494, 2005.[CrossRef][Web of Science][Medline]
Everling S, Dorris MC, Klein RM, and Munoz DP. Role of primate superior colliculus in preparation and execution of anti-saccades and pro-saccades. J Neurosci 19: 2740–2754, 1999.
Everling S, Dorris MC, and Munoz DP. Reflex suppression in the anti-saccade task is dependent on prestimulus neural processes. J Neurophysiol 80: 1584–1589, 1998.
Everling S and Fischer B. The antisaccade: a review of basic research and clinical studies. Neuropsychologia 36: 885–899, 1998.[CrossRef][Web of Science][Medline]
Everling S and Munoz DP. Neuronal correlates for preparatory set associated with pro-saccades and anti-saccades in the primate frontal eye field. J Neurosci 20: 387–400, 2000.
Falkenstein M, Hohnsbein J, Hoormann J, and Blanke L. Effects of crossmodal divided attention on late ERP components. II. Error processing in choice reaction tasks. Electroencephalogr Clin Neurophysiol 78: 447–455, 1991.[CrossRef][Web of Science][Medline]
Fischer B and Weber H. Effects of stimulus conditions on the performance of antisaccades in man. Exp Brain Res 116: 191–200, 1997.[CrossRef][Web of Science][Medline]
Forbes K and Klein RM. The magnitude of the fixation offset effect with endogenously and exogenously controlled saccades. J Cogn Neurosci 8: 344–352, 1996.
Garavan H, Ross TJ, and Stein EA. Right hemispheric dominance of inhibitory control: an event-related functional MRI study. Proc Natl Acad Sci USA 96: 8301–8306, 1999.
Gaymard B, Francois C, Ploner CJ, Condy C, and Rivaud-Pechoux S. A direct prefrontotectal tract against distractibility in the human brain. Ann Neurol 53: 542–545, 2003.[CrossRef][Web of Science][Medline]
Gaymard B, Pierrot-Deseilligny C, and Rivaud S. Impairment of sequences of memory-guided saccades after supplementary motor area lesions. Ann Neurol 28: 622–626, 1990.[CrossRef][Web of Science][Medline]
Gaymard B, Ploner CJ, Rivaud S, Vermersch AI, and Pierrot-Deseilligny C. Cortical control of saccades. Exp Brain Res 123: 159–163, 1998.[CrossRef][Web of Science][Medline]
Gehring WJ, Coles MG, Meyer DE, and Donchin E. A brain potential manifestation of error-related processing. Electroencephalogr Clin Neurophysiol Suppl 44: 261–272, 1995.[Medline]
Grosbras MH, Lobel E, Van de Moortele PF, LeBihan D, and Berthoz A. An anatomical landmark for the supplementary eye fields in human revealed with functional magnetic resonance imaging. Cereb Cortex 9: 705–711, 1999.
Guitton D, Buchtel HA, and Douglas RM. Frontal lobe lesions in man cause difficulties in suppressing reflexive glances and in generating goal-directed saccades. Exp Brain Res 58: 455–472, 1985.[Web of Science][Medline]
Hallett PE. Primary and secondary saccades to goals defined by instructions. Vision Res 18: 1279–1296, 1978.[CrossRef][Web of Science][Medline]
Hanes DP and Schall JD. Countermanding saccades in macaque. Vis Neurosci 12: 929–937, 1995.[Web of Science][Medline]
Husain M, Parton A, Hodgson TL, Mort D, and Rees G. Self-control during response conflict by human supplementary eye field. Nat Neurosci 6: 117–118, 2003.[CrossRef][Web of Science][Medline]
Ito S, Stuphorn V, Brown JW, and Schall JD. Performance monitoring by the anterior cingulate cortex during saccade countermanding. Science 302: 120–122, 2003.
Lasker AG and Zee DS. Ocular motor abnormalities in Huntington's disease. Vision Res 37: 3639–3645, 1997.[CrossRef][Web of Science][Medline]
MacDonald AW III, Cohen JD, Stenger VA, and Carter CS. Dissociating the role of the dorsolateral prefrontal and anterior cingulate cortex in cognitive control. Science 288: 1835–1838, 2000.
McDowell JE, Brown GG, Paulus M, Martinez A, Stewart SE, Dubowitz DJ, and Braff DL. Neural correlates of refixation saccades and antisaccades in normal and schizophrenia subjects. Biol Psychiatry 51: 216–223, 2002.[CrossRef][Web of Science][Medline]
Miller EK. The prefrontal cortex: no simple matter. Neuroimage 11: 447–450, 2000.[CrossRef][Web of Science][Medline]
Miller EK and Cohen JD. An integrative theory of prefrontal cortex function. Annu Rev Neurosci 24: 167–202, 2001.[CrossRef][Web of Science][Medline]
Munoz DP and Everling S. Look away: the anti-saccade task and the voluntary control of eye movement. Nat Rev Neurosci 5: 218–228, 2004.[CrossRef][Web of Science][Medline]
Nieuwenhuis S, Ridderinkhof KR, Blom J, Band GP, and Kok A. Error-related brain potentials are differentially related to awareness of response errors: evidence from an antisaccade task. Psychophysiology 38: 752–760, 2001.[CrossRef][Web of Science][Medline]
Pierrot-Deseilligny C, Muri RM, Ploner CJ, Gaymard B, Demeret S, and Rivaud-Pechoux S. Decisional role of the dorsolateral prefrontal cortex in ocular motor behaviour. Brain 126: 1460–1473, 2003.
Pierrot-Deseilligny C, Rivaud S, Gaymard B, and Agid Y. Cortical control of reflexive visually-guided saccades. Brain 114: 1473–1485, 1991.
Schlag-Rey M, Amador N, Sanchez H, and Schlag J. Antisaccade performance predicted by neuronal activity in the supplementary eye field. Nature 390: 398–401, 1997.[CrossRef][Medline]
Stuphorn V, Taylor TL, and Schall JD. Performance monitoring by the supplementary eye field. Nature 408: 857–860, 2000.[CrossRef][Medline]
Talairach J and Tournoux P. Co-Planar Stereotaxic Atlas of the Human Brain: 3-Dimensional Proportional System: An Approach to Cerebral Imaging. Stuttgart, Germany: Thieme Medical Publishers, 1988.
Walker R, Husain M, Hodgson TL, Harrison J, and Kennard C. Saccadic eye movement and working memory deficits following damage to human prefrontal cortex. Neuropsychologia 36: 1141–1159, 1998.[CrossRef][Web of Science][Medline]
Zhang M and Barash S. Neuronal switching of sensorimotor transformations for antisaccades. Nature 408: 971–975, 2000.[CrossRef][Medline]
This article has been cited by other articles:
|
|
 |
|
  K. Velanova, M. E. Wheeler, and B. Luna Maturational Changes in Anterior Cingulate and Frontoparietal Recruitment Support the Development of Error Processing and Inhibitory Control Cereb Cortex, November 1, 2008; 18(11): 2505 - 2522. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. GODIJN and A. F. KRAMER The effect of attentional demands on the antisaccade cost Atten Percept Psychophys, July 1, 2008; 70(5): 795 - 806. [Abstract] [PDF]
|
|
|
|
 |
|
 C. E. Curtis and J. D. Connolly Saccade Preparation Signals in the Human Frontal and Parietal Cortices J Neurophysiol, January 1, 2008; 99(1): 133 - 145. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.F.W. Neggers, W. Huijbers, C. M. Vrijlandt, B.N.S. Vlaskamp, D.J.L.G. Schutter, and J. L. Kenemans TMS Pulses on the Frontal Eye Fields Break Coupling Between Visuospatial Attention and Eye Movements J Neurophysiol, November 1, 2007; 98(5): 2765 - 2778. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. A. Clementz, S. B. Brahmbhatt, J. E. McDowell, R. Brown, and J. A. Sweeney When Does the Brain Inform the Eyes Whether and Where to Move? an EEG Study in Humans Cereb Cortex, November 1, 2007; 17(11): 2634 - 2643. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. R. G. Brown, T. Vilis, and S. Everling Frontoparietal Activation With Preparation for Antisaccades J Neurophysiol, September 1, 2007; 98(3): 1751 - 1762. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Newsham, P. C. Knox, and R. W. I. Cooke Oculomotor Control in Children Who Were Born Very Prematurely Invest. Ophthalmol. Vis. Sci., June 1, 2007; 48(6): 2595 - 2601. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. S. Manoach, K. N. Thakkar, M. S. Cain, F. E. Polli, J. A. Edelman, B. Fischl, and J. J. S. Barton Neural Activity Is Modulated by Trial History: A Functional Magnetic Resonance Imaging Study of the Effects of a Previous Antisaccade J. Neurosci., February 14, 2007; 27(7): 1791 - 1798. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Johnston and S. Everling Monkey Dorsolateral Prefrontal Cortex Sends Task-Selective Signals Directly to the Superior Colliculus J. Neurosci., November 29, 2006; 26(48): 12471 - 12478. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Boxer, S. Garbutt, K. P. Rankin, J. Hellmuth, J. Neuhaus, B. L. Miller, and S. G. Lisberger Medial versus lateral frontal lobe contributions to voluntary saccade control as revealed by the study of patients with frontal lobe degeneration. J. Neurosci., June 7, 2006; 26(23): 6354 - 6363. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
