Skip to main content
NCBI home page
Human Brain Mapping logoLink to Human Brain Mapping
. 2009 May 11;30(11):3541–3552. doi: 10.1002/hbm.20777

Dissociable frontal controls during visible and memory‐guided eye‐tracking of moving targets

Jinhong Ding 1,2,✉, David Powell 3, Yang Jiang 2,✉
PMCID: PMC2767402  NIHMSID: NIHMS120386  PMID: 19434603

Abstract

When tracking visible or occluded moving targets, several frontal regions including the frontal eye fields (FEF), dorsal‐lateral prefrontal cortex (DLPFC), and anterior cingulate cortex (ACC) are involved in smooth pursuit eye movements (SPEM). To investigate how these areas play different roles in predicting future locations of moving targets, 12 healthy college students participated in a smooth pursuit task of visual and occluded targets. Their eye movements and brain responses measured by event‐related functional MRI were simultaneously recorded. Our results show that different visual cues resulted in time discrepancies between physical and estimated pursuit time only when the moving dot was occluded. Visible phase velocity gain was higher that that of occlusion phase. We found bilateral FEF association with eye‐movement whether moving targets are visible or occluded. However, the DLPFC and ACC showed increased activity when tracking and predicting locations of occluded moving targets, and were suppressed during smooth pursuit of visible targets. When visual cues were increasingly available, less activation in the DLPFC and the ACC was observed. In addition, there was a significant hemisphere effect in DLPFC, where right DLPFC showed significantly increased responses over left when pursuing occluded moving targets. Correlation results revealed that DLPFC, the right DLPFC in particular, communicates more with FEF during tracking of occluded moving targets (from memory). The ACC modulates FEF more during tracking of visible targets (likely related to visual attention). Our results suggest that DLPFC and ACC modulate FEF and cortical networks differentially during visible and memory‐guided eye tracking of moving targets. Hum Brain Mapp, 2009. © 2009 Wiley‐Liss, Inc.

Keywords: fMRI, smooth pursuit, saccade, working memory, visual motion, simultaneous recording

INTRODUCTION

Our brain has an amazing ability to accurately predict the velocity and future position of visual targets in motion such as flying birds. Even when a moving target is temporarily occluded, e.g. a bird flies behind a tree and flies out; the brain is able to estimate the velocity and future position of the bird. Many areas of the brain are involved in tracking a moving target including early visual cortices, human V5 (MT/MST complex), intraparietal regions, and frontal areas. The frontal areas, known as the frontal eye field (FEF), are an important area for controlling eye movements during smooth pursuit [e.g. Fukushima et al.,2006; Krauzlis,2004; Milea et al.,2002; Petit et al.,1997; Tregellas et al.,2002]. Using simultaneous eye tracking and functional magnetic resonance imaging (fMRI), this study examined the cortical mechanisms involved in predicting the location of visible and occluded moving targets under different visual cue conditions.

Smooth pursuit eye movements (SPEM) are accompanied by a sequence of saccades. Executions of both saccadic and smooth pursuit eye movements induce bilateral FEF activation located medially at the junction of the precentral sulcus and the superior frontal sulcus. FEF activation also extends laterally to the precentral gyrus [Petit et al.,1997; Tanabe et al.,2002]. Pursuit‐related areas have been distinguished from saccade‐related areas both in terms of spatial extent and location. Pursuit‐related areas are smaller than their saccade‐related counterparts. Furthermore, the mean location of the pursuit‐related FEF is more inferior and lateral than the location of the saccade‐related FEF [Petit et al.,1997]. Other pursuit‐related areas are consistently posterior to their saccade‐related counterparts [Petit and Haxby,1999]. However, the neural mechanisms that control pursuit and saccadic eye movements are still being debated. The FEF was found to be involved in both smooth eye movements and saccade suppression [Milea et al.,2002]. In addition, it is related to corrective saccades which redirect gaze whenever the actual eye position differs from the desired eye position during oculomotor task [Murthy et al.,2007]. Krauzlis [2004] argued that both the control of pursuit and saccades might in fact stem from a single cascade of sensory‐motor neural signals. Differences between pursuit and saccades may be explained by a hierarchy of control levels as opposed to the presence of separate, independent eye movement control systems. The FEF directly controls pursuit eye movement via a gating mechanism that is regulated by the Superior Colliculus (SC). The FEF also uses parts of the same premotor circuits in the brain stem for regulating and formulating the final motor command [Krauzlis,2004]. Fukushima et al. [2006] reported that the majority of the FEF pursuit neurons code for eye velocity, gaze velocity, and retinal image motion during pursuit. Additionally, research shows the cerebellum is bilaterally activated during an ocular task [Nitschke et al.,2004].

The FEF plays a role in the suppression of reflexive saccades to make information processing more effective [Cornelissen et al.,2002]. Additionally, the FEF activity may represent short‐term memory [Cornelissen et al.,2002] and contribute to spatial working memory [Curtis and D'Esposito,2006]. The reflexive saccades may be less related to higher cognitive functioning (e.g. memory of moving path). Besides this downstream control, there are other frontal areas that modulate the FEF's control of SPEM. The dorsolateral prefrontal cortex (DLPFC) is anatomically close to the FEF. Pierrot‐Deseilligny et al. [2005] found the DLPFC has higher level functions involved in controlling several types of intentional saccades by triggering other parietal and frontal areas, i.e. the parietal eye field (PEF) and the FEF. According to Pierrot‐Deseilligny et al. [2005], the DLPFC also controls short‐term spatial working memory used for forthcoming memory‐guided saccades as well as temporal working memory controlling successive movements. Burke and Barnes [2008] differentiated the regions for different types of eye movements showing there was higher activation in the DLPFC for pursuit and higher activation in the frontopolar region for saccades.

The cingulate cortex is related to smooth pursuit and corrective saccades [Haller et al.,2008; Kimmig et al.,2008; Murthy et al.,2007]. The anterior cingulate cortex (ACC) has been reported to be involved in immediate motor planning and memory of the timing and trajectory of a target [Schmid et al.,2001]. Pierrot‐Deseilligny et al. [2002] found that the motivation and preparation of all intentional saccades activate the posterior part of the anterior cingulate cortex. Grosbras et al. [2001] also reported that neuronal resources recruited by eye gazing changes with the familiarity of the scan path to be followed. Because the visual system can remember the target velocity, the brain runs its predictive internal models using this memorized visual input [Knowler et al.,1977]. The predictive eye movements can be made based upon the previously viewed velocities [Poliakoff et al.,2005].

Maintaining SPEM in the absence of a visible target activates additional cortical areas, such as the frontal regions associated with prediction, visuo‐spatial attention and transformation, multimodal visuomotor control, and working memory [Lencer et al.,2004]. The FEF and prefrontal cortex (PFC) are involved in reconstitution and prediction, especially during the maintenance of smooth pursuit in the absence of a visual target [Nagel et al.,2006]. It seems obvious that the frontal areas, i.e. FEF, DLPFC, and ACC, interact functionally with each other to govern ocular motor behavior, but the relationships between them are not clear. How would these areas interact to accomplish SPEM with different predictabilities? Is a hierarchical network used when predicting the position and velocity of a target during smooth pursuit? We hypothesized that the DLPFC and ACC modulate the FEF differently depending on the levels of cognitive demand required when pursuing visual or occluded targets. To test this hypothesis, we recorded eye movements and brain responses simultaneously while subjects performed a smooth pursuit task with visible and occluded moving targets.

MATERIALS AND METHODS

Participants

Twelve undergraduate or graduate students between the ages of 18 and 28, eight females and four males, participated in this experiment. All participants provided informed consent for procedures approved by the University of Kentucky Institutional Review Board. The participants were normal, healthy, right‐handed individuals with normal or corrected‐to‐normal visual acuity. Subjects were paid for their participation.

Visual Stimuli and Apparatus

The visual display and response system was controlled by a computer running E‐prime scripts (Psychology Software Tools, Pittsburgh, PA). The visual stimuli were back‐projected onto a screen placed 35 cm away from a small viewing mirror located above the subject's eyes. Avotec's Silent Vision 6011 projection system was used for stimulus projection. The projected background was gray with a dark gray occlusion bar (see Fig. 1). A 0.2° diameter white dot moved from left to right across the screen in the gray region and followed a sinusoidal pattern with a random starting phase until it disappeared behind the dark gray occlusion bar. There were four different conditions. The dot's sinusoidal path was illuminated during three cue conditions: (1) “Full trace” cue present during both the visible and occluded phases (with high predictability); (2) “partial trace” cue present during the visible phase only (with medial predictability), or (3) “no trace” present during neither the visible nor occluded phase (with low predictability). (4) The “fixation only” condition had no moving dot and was used as a baseline measurement. Note the “no trace” and fixation only were two different conditions. Under the “no trace” cue condition, there was no visible sinusoidal cue during both visible and occluded phase. In contrast, the fixation condition was a baseline condition, in which subject fixated a “+” presenting in the center of screen without smooth pursuit eye‐movement.

Figure 1.

Figure 1

Open in a new tab

Visual displays under different visual cues. The baseline condition (A) showed a “+” in the middle of the screen. A white dot moved left to right following a sinusoidal pattern across a screen of uniformly gray background with a darker gray vertical occlusion bar. The unfilled circle “○” indicated the sampling positions of the dot and the filled circle “•” was the real moving dot. No sinusoidal pattern was illuminated during (B) no trace condition under both visible and occluded phases, and during occluded phase of partial trace condition. The sinusoidal pattern was illuminated only during (C) partial trace condition of visible phase and both visible and occluded phases of (D) full trace condition.

The white dot moved with the maximum velocity of 10.4°/s and the minimum velocity of 3.4°/s in the sinusoidal pattern. The display (visible plus occluded) had visual angles of 15.8° horizontally and 7.9° vertically. The display was on for 10 s for each trial. The E‐prime computer was synchronized to the MRI scanner by a TTL pulse that was delivered at the beginning of each EPI volume acquisition. A second computer controlled the ASL eye position recording system (Model 504LRO, Applied Sciences Laboratories) and eye tracking software. The E‐prime computer was synchronized to the eye‐tracking computer via a parallel port cable.

Behavioral Procedures

A “+” appeared on the left side of the screen as a cue and would be replaced by a white dot after 500 ms. Subjects were asked to focus on this “+” and then follow the moving white dot with their eyes, tracking its sinusoidal motion. The participants were asked to press the left‐hand button to indicate when the dot disappeared behind the left side of the dark gray occlusion bar and to press the right‐hand button when they predicted that the moving dot should reappear from the right‐hand side of the occlusion bar. The participants were told that behind the occluded bar, the white dot continued to follow the same sinusoidal pattern. They were asked to press the button at the exact moment they thought the white dot should reappear from behind the occlusion bar. Under the baseline condition, the participants were asked to fixate on a cross “+” in the middle of the screen and press both buttons at the same time. Subjects practiced the task for 15 min prior to going into the scanner.

Event‐Related fMRI Design

Sixty trials in each of the four conditions, i.e. baseline fixation (Fig. 1A), no trace (Fig. 1B), partial trace (Fig. 1C), and full trace (Fig. 1D), were counterbalanced in an event‐related fMRI design. The order of the counterbalance was chosen from simulations using the AFNI (Analysis of Functional NeuroImages) program to achieve the best nonoverlapped fMRI signals for each condition. The 240 trials were broken down into eight fMRI runs. For each trial, the visual phase and occluded phase lasted for 5 s each (Fig. 1E). AFNI software was used to create a mixed event‐related fMRI experiment with a randomized order for the “best” signal detection.

Oculomotor Recording and Analysis

The eye position was monitored in the scanner at 120 Hz with an infrared video‐graphic camera equipped with a telephoto lens (Model 504LRO, Applied Sciences Laboratories). Nine‐point calibrations were performed at the beginning of each session. Eye movement data was scored offline. Two subjects' eye‐tracking data was excluded because of their long eyelashes and excessive head movement. Blinks, drifts, and other artifacts were detected and eliminated from the oculomotor data. The corrective saccades were classified if their amplitude fell within 0.8° and 8° [Haller et al.,2008]. The frequency of corrective saccade varied from 2.00 Hz to 2.50 Hz (see Fig. 2A for an example of a single subject and Fig. 2B for group average). The cumulative saccade amplitude was about 2.00 deg/s on average (see Fig. 2C). No significant difference of corrective saccade frequency and cumulative saccade amplitude was found across conditions (The corrective saccades are marked by circles in eye positions in Fig. 2). The smooth pursuit velocity gain for each condition was calculated in terms of peak velocity of target and eye in the window of 100 ms. The time discrepancy between the time (1) that dot disappeared at the left border of the bar or reappeared at the right side of the bar (known physical moving time) and time (2) that the participant indicated that the dot reached the left side of the bar or reappeared at the right side of the bar (recorded estimated times) were calculated.

Figure 2.

Figure 2

Open in a new tab

Frequency and amplitude of saccades. (A) An example from a single subject in six conditions. The corrective saccades are marked by circles. (B) The frequency of corrective saccade varied from 2.00 Hz to 2.50 Hz on average. (C) The cumulative saccade amplitude was about 2.00 deg/s (averaged from 10 subjects).

Image Acquisition

Anatomical and functional images were acquired on a 3 Tesla Siemens Trio MR scanner at the University of Kentucky's Magnetic Resonance Imaging and Spectroscopy Center. An 8‐channel, receive only head coil was used. A high‐resolution, 3D anatomic image was acquired using a sagittal T‐1 weighted (MP‐RAGE) sequence (TR = 2,100 ms, TE = 2.93 ms, TI = 1,100 ms, flip angle = 12°, FOV = 192 mm × 224 mm × 256 mm, with 1 mm isotropic voxels). The functional images were T2* weighted echo‐planar images. The following parameters were used: TR = 2.5 s; TP = 156; TE = 30 ms; flip angle = 81°; 39 axial slices; 224 mm × 224 mm FOV (field of view); slice thickness = 3.5 mm; 64 × 64 matrix (yielding a 3.5 mm × 3.5 mm × 3.5 mm voxel); bandwidth = 2,056 Hz/Px.

fMRI Data Analysis

For each subject, after preprocessing, the structural 3D data was transformed into Talairach space using AFNI. fMRI data was motion corrected and smoothed (4 mm FWHM), and voxels outside the brain were masked. Each voxel was normalized in each functional run. For each subject, the voxel time series for each trial type (no trace, partial trace, full trace, and fixation) was estimated using AFNI. Using multiple regression analysis, statistical maps were constructed representing the association between the observed time series and a linear combination of regressors for each subject. The hemodynamic delays for the BOLD signals were time locked from the onset of the target presentation. A linear combination of these regressors was used to estimate the evoked hemodynamic delay for each trial type with no assumptions about the shape of the BOLD response. Time courses were estimated for both the visible and occluded conditions. Voxels were selected within each region of interest (ROI) that showed a significant Z value when contrasting the smooth pursuit trials to the fixation trials. Within each structurally defined ROI, only the voxels with a Z score greater than 5.51 for the DLPFC (BA 9/46), ACC (BA 32), and FEF (BA 6) were included in further analyses. These ROIs had at least 20 continuous voxels that showed response to the task regardless of trial type. All of the voxels within the ROI were averaged for each time point creating a single, spatially averaged, time course for each trial type. The percent change in the BOLD signal for each condition was calculated in the ROIs for further analysis using AFNI.

Correlations Between Eye Movement and Functional MRI BOLD Signals

We expected that the frontal eye field and high‐level frontal regions were differentially involved in tracking the moving white dot with or without a trace during the occluded period. Pearson correlations were conducted to correlate the velocity gain measured by the eye‐tracker to the BOLD signals measured from the fMRI data in the cortical regions.

RESULTS

Predicted Times

The time difference between the moving target's actual disappearance and reappearance from the occlusion bar versus its predicted disappearance and reappearance was analyzed with a MANOVA. Across all trials, no significant difference was found between the time the target disappeared behind the occlusion bar and the time the subjects registered the disappearance. However, the time discrepancy between the actual and predicted reappearance times under the three different types of trace were significantly different [F(2,22) = 5.72, P < 0.05] (see Fig. 3). The interaction between the target visibility (visible or occluded) and visual cue conditions (full, partial, or no trace) was also significant [F(2,22) = 10.32, P < 0.01]. The simple effect analysis indicated that the error in the subjects' prediction of when the target should reappear from behind the occlusion bar was significantly different under the condition of no trace [F(1,11) = 5.56, P < 0.05]. In addition, there was a significant difference in the estimated amount of time it would take for the moving dot to come out of the occluded bar among trace conditions [F(2,22) = 7.89, P < 0.01]. Thus, the trace condition significantly affected the subjects' response time.

Figure 3.

Figure 3

Open in a new tab

Time discrepancy between the white dot's (target) actual and the subjects' registered arrival times. The filled circle represents the dot's disappearance behind the occlusion bar, the filled squares represents the dot's reappearance from behind the occlusion bar.

Eye Movements: Smooth Pursuit Velocity Gain

We recorded simultaneous eye‐movements and fMRI data for each subject during the task. The velocity gain is defined as a ratio of the peak velocity of the recorded eye movement to the peak velocity of the moving white dot (see Fig. 4). A velocity gain of 1 means the dots' velocity was tracked perfectly. If the gain is smaller than 1, the eye velocity was slower than the moving dot's velocity.

Figure 4.

Figure 4

Open in a new tab

The velocity gains for each condition.

A two‐way repeated ANOVA showed that the difference between the velocity gains in the visible and occluded regions is significant [F(1,9) = 5.29, P < 0.05]. But the main effects of the trace condition and the interaction between visibility (visible or occluded) and the trace condition are not significant.

Differential fMRI BOLD Signals in FEF and Other Frontal Regions

The fMRI responses in several frontal regions are summarized in Figure 5. The percent BOLD signal changes for each trace condition contrasted against the fixation only condition are presented for both of the visibility (visible and occluded) regions in the task. The number of voxels in each observed area were left FEF: 217 (mean volume = 9.30 cm3; Z > 5.51), right FEF: 312 (mean volume = 13.38 cm3; Z > 5.51), left DLPFC: 199 (mean volume = 8.53 cm3; Z > 5.51), right DLPFC: 115 (mean volume = 4.93 cm3; Z > 5.51) and ACC: 27 (mean volume = 1.16 cm3; Z > 5.51).

Figure 5.

Figure 5

Open in a new tab

Average BOLD signal changes in FEF and DLFPC during eye‐tracking of moving targets. (Note: ** stands for P < 0.01). DLPFC, the right DLPFC in particular, communicates more with FEF during tracking of occluded moving targets.

Frontal Eye Field (FEF)

A MANOVA with three factors (trace condition × visibility × hemisphere) was conducted. There was no main effect for visibility. Significances were found for the cue conditions (no trace, partial trace, and full trace) [F(2,22) = 19.07, P < 0.01] and the cue condition × hemisphere interaction [F(2,22) = 5.69, P < 0.01]. Simple effects analysis indicated that the cue trace condition effects are significant in the bilateral FEF [left: F(2,22) = 21.08, P < 0.01; right: F(2,22) = 14.62, P < 0.01]. Post hoc analysis showed that the full trace condition evoked the lowest FEF response. The BOLD signal from the right FEF is significantly stronger than those from the left FEF only during the full trace condition during visible target pursue (Fig. 5 top panel). Under five other conditions, there was no difference between left and right FEF responses.

DLFPC (BA 9/46)

In contrast to FEF, DLFPC showed a different response pattern (Figure 5 bottom panel). The main effects of the three factors are significant [left and right BA 9: F(1,11) = 5.61, P < 0.05; trace cue: F(2,22) = 16.80, p < 0.01; visibility: F(1,11) = 30.06, p < 0.01]. The interaction of trace cue × visibility is not significant. On the other hand, the interactions of hemisphere × cue, area × visibility, and area × trace × visibility are significant respectively [F(1,11) = 4.09, P < 0.05; F(1,11) = 33.16, P < 0.01; F(2,22) = 3.50, P < 0.05]. Compared to the FEF where increased fMRI signals were observed during both visible and the occluded pursue, the DLPFC showed increased activity only when the moving target was occluded. There were also significant hemisphere effects. Compared to the left DLPFC, the right DLPFC showed larger increased responses during occlusion and larger decreased responses during the visible period.

ACC (BA 32)

Similar to the DLPFC, the visibility (visible or occluded) effect in the ACC (Figure 6) was significant [F(1,11) = 36.21, P < 0.05]. The increased ACC activity was associated with the occluded region, and decreased BOLD signals were associated with the visible region. Additionally, the main effect of the cue condition [F(2,22) = 7.55, P < 0.01] was significant. The largest ACC activity was found during the no trace or partial trace conditions. However, the interaction between trace condition and visibility was not significant (P > 0.05).

Figure 6.

Figure 6

Open in a new tab

Average BOLD signal changes in ACC during eye‐tracking of moving targets. The ACC modulate FEF during tracking of visible targets.

Correlation Between the FEF, DLPFC, and ACC Responses

The correlations between the different brain areas during the visual and occluded periods were analyzed to better understand their relationship during smooth pursuit (see Fig. 7). The interactions between the FEF, DLPFC, and ACC during the visible or occluded periods are different. Intriguingly, strong and extensive correlations between the bilateral FEF and DLPFC were found when the moving target was visible (r = 0.65–0.87). This correlation indicates the DLPFC is involved in eye movement control when tracking a visible target. When the moving target was occluded, however, the left and right FEF were highly coordinated, presumably to control eye movement r = 0.93. Because the short‐term visual spatial information might be stored in the right hemisphere, the relationship between the right DLPFC and FEF was stronger (from r = 0.67 to r = 0.73) and the relationship between the left DLPFC and FEF was weaker during the occluded period, a significant difference from during the visible period (visible: r = 0.67; occluded: r = 0.43, P > 0.05). A stronger interaction between the left and right DLPFC was required during the occluded period (r = 0.93).

Figure 7.

Figure 7

Open in a new tab

Correlations of fMRI responses among FEF, DLPFC, and ACC.

DISCUSSION

When pursuing a moving target along a 2D sinusoidal trajectory, different visual cues resulted in time discrepancies between physical and predicted pursue time when the moving dot was occluded. The velocity gain during visible phase was higher than that of occlusion phase. Event‐related fMRI showed that bilateral FEF are associated with eye‐movement regardless of whether the moving targets are visible or occluded. In contrast, the DLPFC and the ACC showed increased activity when tracking and predicting locations of occluded moving targets, and were suppressed during smooth pursuit of visible targets. When visual cues were increasingly available, less activation in the DLPFC and the ACC was observed. In addition, there was a significant hemisphere effect in DLPFC, where the right DLPFC showed significantly increased responses over the left one when pursuing occluded moving target.

When predicting the time needed for a moving target to disappear and reappear from behind the occlusion bar, the subjects relied on visual cues. The overall time differences were most accurate for the full trace condition. The current results showed that visual cues (the trace) greatly helped the subjects predict the time the target would reappear from behind the occlusion bar. Specifically, the full trace played an important role during smooth pursuit eye‐movement during the occluded period. In the partial trace trials, the visual cue was available only during the visible phase. This allowed visual memory of the moving dot from the visible period to aid in the prediction of the dot's movement during the occluded period. In fact, the participants made more accurate predictions during “partial trace” trials, where visual cue was available during the visible period, than “no trace” trials. Thus, immediate visual cues or working memory enhances prediction of subsequent target movements including trajectory and velocity.

Boudet et al. [2006] found that three systems are used during the pursuit of a visual target: the smooth pursuit, the saccadic eye‐movement, and the predictive component. The smooth pursuit and the saccadic components are influenced by the predictive component. Our results indicate that the predictive mechanism in the brain works well during the occluded periods if the moving target follows a predictable path. It is consistent with a previous report by de'Sperati and Santandrea [2005] who found that the pattern of eye movements during tracking is almost the same as that during imagery when pursuing a sinusoidally moving target. Pursuit gain may be an index of prediction for immediate learning or memory of movement. One question is whether the visual working memory of learned moving paths is used to predict the future speeds and locations of the moving targets.

The predictions of the target's movements were aided by visually tracking the moving target along a visible path before occlusion. The smooth or small saccadic eye movements keep the images of a moving target in the fovea during a smooth pursuit task. Poliakoff et al. [2005] presented a static cue for 700 ms before the target started to move. When the target was occluded, they found that participants were able to make a predictive eye movement based upon the previously viewed velocity. In our experiment, the visual cue affects the velocity gain of smooth pursuit only during the visible period. When the trace cue was not available, the velocity gain of the moving dot during the visible period was greater than that of the occluded period. These results indicate the eye velocity of smooth pursuit depends upon visible cues, which guide eye movements. This finding is consistent with Heinen et al. [2005] in that predictive signals interact with those used in sensory motor processing during smooth pursuit. How does the brain predict the future position of moving targets? It is well known that the FEF modulates eye‐movements. Do the FEF and other frontal regions play different roles in predicting target movement during the visible and occluded periods?

The FEF

The FEF is involved in controlling smooth pursuit and saccadic eye movements [Krauzlis,2004]. Konen et al. [2005] found that predictable and unpredictable eye‐movements share the same neural network. However, information about the direction of a target's movement reduced the cerebral activity level. Note that the subjects' task was to predict the direction of a moving target rather than to smoothly track it. Our results indicate that the bilateral FEF are evoked during the task of predicting target movements. The BOLD response in the FEF for higher predictability (full trace) is less than that of lower predictability (no trace). If the target is more difficult to predict, the FEF showed an increased BOLD signal. Under the easier tracking condition (full trace), the right FEF is activated more than the left. Interestingly, the FEF is engaged whether the moving target is visible or not. In other words, bilateral FEF are associated with eye movement regardless of whether the target is visible or occluded.

Both behavioral and eye movement results demonstrate that a visual cue for a dot's path improves prediction of future movement, visible or occluded. A monkey single‐unit recording study [Chou and Lisberger,2004] and a human transcranial magnetic stimulation (TMS) study [Gagnon et al.,2006] showed that the FEF (frontal pursuit area, FPA) is related to the gain of visual‐motor responses. Our present results on the correlation between pursuit velocity gain and FEF response also confirm this connection. When the moving target is occluded under the partial trace condition, the eye velocity gain is significantly correlated to the BOLD response in the bilateral FEF. It is likely that the trajectory of the moving target is freshly learned and subsequently facilitates predicting the dot's movement when it continued to move behind the occlusion bar. Working memory may be related to mechanisms that also control the gain of SPEM. If this is the case, it is logical to postulate that memory‐related regions should show sustained responses when the moving target is not visible during the occluded period.

The DLPFC and ACC During Smooth Pursuit of Invisible Target

We found that the DLPFC (BA9/46) cortex and the posterior part of the anterior cingulate (BA 32) showed significantly increased BOLD responses when the moving target was occluded versus when the target was visible. In contrast, the FEF, DLPFC, and ACC had decreased responses when the target was visible compared to occluded. The DLPFC has been associated with the spatial working memory, task difficulty, response selection, preparation for response, and the control of saccadic eye‐movements [Kawawaki et al.,2006; Pierrot‐Deseilligny et al.,2005]. The ACC has been linked to attention‐related executive control, conscious effort, conflict, and pain [Brown and Braver,2005; Crottaz‐Herbette and Menon,2006; Davis et al.2005; Fan et al.,2007; Kerns,2006; Mulert et al.,2005]. It has been shown that the ACC is involved in saccades as well [Ford et al.,2005]. In our study, participants' eyes tracked a moving dot with or without a trace. During the occluded periods with the no trace and partial trace conditions, participants used their spatial working memory of the dot's sinusoidal path from the visible period. The tracking of an occluded moving target is demanding because it requires attention, working memory, motor control, motor learning, and visual motor imagery. Our results showed that the FEF had similar levels of responses during the SPEM of visible or occluded moving targets. However, both the bilateral DLPFC and ACC were needed when predicting the movement of occluded moving targets. The more predictable (different cue conditions) the moving target was, the less activation there was in the DLPFC and ACC. These results indicate that when limited visual cues are present, these frontal regions are more involved in retrieving information and making decisions about target movement.

DLPFC also showed an interesting hemisphere effect, where right DLPFC responded more strongly than the left during the tracking of the occluded target. We reported a similar hemispheric effect in the early visual cortex while eye‐tracking occluded visual targets [Jiang et al.,2008]. These results lead to a plausible hypothesis that right DLPFC may engage in top‐down modulation to right early visual cortex during pursuit of occluded moving targets. Future studies should examine the effective connectivity and interplay among the regions involved in the task.

Interactions Among Frontal Regions

When examining the interplay among frontal areas, we found that correlations between the bilateral FEF and ACC were significant only during visible period (r = 0.64). This indicates the ACC monitored the FEF during visual attention but not covert attention (occluded region). Because attention may play an important role in the maintenance of smooth pursuit eye movements, participants needed to concentrate on the moving target when it was visible so as to pursue it smoothly [Hutton and Tegally,2005; Kerzel and Ziegler,2005]. In contrast to the ACC, the BOLD signals in DLPFC during occlusion were stronger than those during visible period. That is, working memory played a bigger role in monitoring the visual tracking of occluded moving targets. In addition, BOLD signals in the FEF showed a stronger correlation to the DLPFC than to the ACC during the tracking of occluded targets, which was likely to be guided by visual memory.

Using simultaneous eye‐tracking and event‐related fMRI, we found evidence that the frontal regions FEF, DLPFC, and ACC played distinct roles when tracking visible or occluded moving targets with various visual cues. Recent reports indicate that FEF stimulation induced a topographically specific pattern of enhancement and suppression in early visual areas, but only in the presence of a visual stimulus [Ekstrom et al.,2008]. The current paradigm will allow us to further examine top‐down and bottom‐up processing during eye‐tracking of visible and occluded moving targets in the future.

Supporting information

Additional Supporting Information may be found in the online version of this article.

Supplementary Movie: The animation illustrates the visual displays under different cues, i.e. (1) No Trace Cue, (2) at 18 s: Fixation only; (3) at 30 s: Partial Trace Cue; (4) Full Trace Cue.

Click here for additional data file. (10.3MB, mpg)

Acknowledgements

We thank A. Lawson for assistance in part of the data collection, H. Nolan and S. Kiser for their help with editing.

Contributor Information

Jinhong Ding, Email: dingjh@mail.cnu.edu.cn.

Yang Jiang, Email: yjiang@uky.edu.

REFERENCES

  1. Boudet C, Bocca ML, Dollfus S, Denise P ( 2006): The saccadic component of ocular pursuit is influenced by the predictability of the target motion in humans. Exp Brain Res 168: 294– 297. [DOI] [PubMed] [Google Scholar]
  2. Brown JW, Braver TS ( 2005): Learned predictions of error likelihood in the anterior cingulate cortex. Science 307: 1118– 1121. [DOI] [PubMed] [Google Scholar]
  3. Burke MR, Barnes GR ( 2008): Brain and behavior: A task‐dependent eye movement study. Cereb Cortex 18: 126– 135. [DOI] [PubMed] [Google Scholar]
  4. Chou I, Lisberger SG ( 2004): The role of the frontal pursuit area in learning in smooth pursuit eye movements. J Neurosci 24: 4124– 4133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Cornelissen FW, Kimmig H, Schira M, Rutschmann RM, Maguire RP, Broerse A, Boer JAD, Greenlee MW ( 2002): Event‐related fMRI responses in the human frontal eye fields in a randomized pro‐ and antisaccade task. Exp Brain Res 145: 270– 274. [DOI] [PubMed] [Google Scholar]
  6. Crottaz‐Herbette S, Menon V ( 2006): Where and when the anterior cingulate cortex modulates attentional response: Combined fMRI and ERP evidence. J Cognit Neurosci 18: 766– 780. [DOI] [PubMed] [Google Scholar]
  7. Curtis CE, D'Esposito M ( 2006): Selection and maintenance of saccade goals in the human frontal eye fields. J Neurophysiol 95: 3923– 3927. [DOI] [PubMed] [Google Scholar]
  8. Davis KD, Taylor KS, Hutchison WD, Dostrovsky JO, McAndrews MP, Richter EO, Lozano AM ( 2005): Human anterior cingulate cortex neurons encode cognitive and emotional demands. J Neurosci 25: 8402– 8406. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. de'Sperati C, Santandrea E ( 2005): Smooth pursuit‐like eye movements during mental extrapolation of motion: The facilitatory effect of drowsiness. Cognit Brain Res 25: 328– 338. [DOI] [PubMed] [Google Scholar]
  10. Ekstrom LB, Roelfsema PR, Arsenault JT, Bonmassar G, Vanduffel W ( 2008). Bottom‐up dependent gating of frontal signals in early visual cortex. Science 321: 414– 417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Fan J, Kolster R, Ghajar J, Suh M, Knight RT, Sarkar R, McCandliss BD ( 2007): Response anticipation and response conflict: an event‐related potential and functional magnetic resonance imaging study. J Neurosci 27: 2272– 2282. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Ford KA, Goltz HC, Brown MRG, Everling S ( 2005): Neural processes associated with antisaccade task performance investigated with event‐related fMRI. J Neurophysiol 94: 429– 440. [DOI] [PubMed] [Google Scholar]
  13. Fukushima J, Akao T, Kurkin S, Kaneko CRS, Fukushima K ( 2006): The vestibular‐related frontal cortex and its role in smooth‐pursuit eye movements and vestibular‐pursuit interactions. J Vestibular Res 16: 1– 22. [PMC free article] [PubMed] [Google Scholar]
  14. Gagnon D, Paus T, Grosbras M, Pike GB, O'Driscoll GA ( 2006): Transcranial magnetic stimulation of frontal oculomotor regions during smooth pursuit. J Neurosci 26: 458– 466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Grosbras M, Leonards U, Lobel E, Poline J, LeBihan D, Berthoz A ( 2001): Human cortical networks for new and familiar sequences of saccades. Cereb Cortex 11: 936– 945. [DOI] [PubMed] [Google Scholar]
  16. Haller S, Fasler D, Ohlendorf S, Radue EW, Greenlee MW ( 2008): Neural activation associated with corrective saccades during tasks with fixation, pursuit and saccades. Exp Brain Res 184: 83– 94. [DOI] [PubMed] [Google Scholar]
  17. Heinen SJ, Badler JB, Ting W ( 2005): Timing and velocity randomization similarly affect anticipatory pursuit. J Vision 5: 493– 503. [DOI] [PubMed] [Google Scholar]
  18. Hutton SB, Tegally D ( 2005): The effects of dividing attention on smooth pursuit eye tracking. Exp Brain Res 163: 306– 313. [DOI] [PubMed] [Google Scholar]
  19. Jiang Y, Ding JH, Gold BT, Powell D ( 2008). The hemispheric asymmetries in tracking occluded moving targets with the mind's eye: Simultaneous event‐related fMRI and eye‐movement recording. Brain Imaging Behav 2: 300– 308. [Google Scholar]
  20. Kawawaki D, Shibata T, Goda N, Doya K, Kawato M ( 2006): Anterior and superior lateral occipito‐temporal cortex responsible for target motion prediction during overt and covert visual pursuit. Neurosci Res 54: 112– 123. [DOI] [PubMed] [Google Scholar]
  21. Kerns JG ( 2006): Anterior cingulate and prefrontal cortex activity in an FMRI study of trial‐to‐trial adjustments on the Simon task. NeuroImage 33: 399– 405. [DOI] [PubMed] [Google Scholar]
  22. Kerzel D, Ziegler NE ( 2005): Visual short‐term memory during smooth pursuit eye movements. J Exp Psychol 31: 354– 372. [DOI] [PubMed] [Google Scholar]
  23. Kimmig H, Ohlendorf S, Speck O, Sprenger A, Rutschmann RM, Haller S, Greenlee MW ( 2008): fMRI evidence for sensorimotor transformations in human cortex during smooth pursuit eye movements. Neuropsychologia 46: 2203– 2213. [DOI] [PubMed] [Google Scholar]
  24. Knowler E, Murphy BJ, Steinman RM ( 1977): Velocity matching during smooth pursuit of different targets on different backgrounds. Vision Res 18: 603– 605. [DOI] [PubMed] [Google Scholar]
  25. Konen C, Kleiser R, Seitz R, Bremmer F ( 2005): An fMRI study of optokinetic nystagmus and smooth‐pursuit eye movements in humans. Exp Brain Res 165: 203– 216. [DOI] [PubMed] [Google Scholar]
  26. Krauzlis RJ ( 2004): Recasting the smooth pursuit eye movement system. J Neurophysiol 91: 591– 603. [DOI] [PubMed] [Google Scholar]
  27. Lencer R, Nagel M, Sprenger A, Sprenger A, Zapt S, Erdmann C, Heide W, Binkofski F ( 2004): Cortical mechanisms of pursuit eye movements with target blanking: An fMRI study. Eur J Neurosci 19: 1430– 1436. [DOI] [PubMed] [Google Scholar]
  28. Milea D, Lobel E, Lehéricy S, Duffau H, Rivaud‐Péchoux S, Berthoz A, Pierrot‐Deseilligny C ( 2002): Intraoperative frontal eye field stimulation elicits ocular deviation and saccade suppression. Neuroreport 13: 1359– 1364. [DOI] [PubMed] [Google Scholar]
  29. Mulert C, Menzinger E, Leicht G, Pogarell O, Hegerl U ( 2005): Evidence for a close relationship between conscious effort and anterior cingulate cortex activity. Int J Psychophysiol 56: 65– 80. [DOI] [PubMed] [Google Scholar]
  30. Murthy A, Ray S, Shorter MS, Priddy GE, Schall DJ, Thompson GK ( 2007): Frontal eye field contributions to rapid corrective saccades. J Neurophysiol 97: 1457– 1469. [DOI] [PubMed] [Google Scholar]
  31. Nagel M, Sprenger A, Zapf S, Erdmann C, Kompf D, Heide W, Binkofski F, Lencer R ( 2006): Parametric modulation of cortical activation during smooth pursuit with and without target blanking: An fMRI study. NeuroImage 29: 1319– 1325. [DOI] [PubMed] [Google Scholar]
  32. Nitschke MF, Binkofski F, Buccino G, Posse S, Erdmann C, Kömpf D, Seitz RJ, Heide W ( 2004): Activation of cerebellar hemispheres in spatial memorization of saccadic eye movements: An fMRI study. Hum Brain Mapp 22: 155– 164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Petit L, Haxby JV ( 1999): Functional anatomy of pursuit eye movements in humans as revealed by fMRI. J Neurophysiol 81: 463– 471. [DOI] [PubMed] [Google Scholar]
  34. Petit L, Clark VP, Ingeholm J, Haxby JV ( 1997): Dissociation of saccade‐related and pursuit‐related activation in human frontal eye fields as revealed by fMRI. J Neurophysiol 77: 3386– 3390. [DOI] [PubMed] [Google Scholar]
  35. Pierrot‐Deseilligny CH, Ploner CJ, Müri RM, Gaymard B, Rivaud‐Péchoux S ( 2002): Effects of cortical lesions on saccadic eye movements in humans. Ann N Y Acad Sci 956: 216– 229. [DOI] [PubMed] [Google Scholar]
  36. Pierrot‐Deseilligny CH, Müri RM, Nyffeler T, Milea D ( 2005): The role of the human dorsolateral prefrontal cortex in ocular motor behavior. Ann N Y Acad Sci 1039: 239– 251. [DOI] [PubMed] [Google Scholar]
  37. Poliakoff E, Collins CJS, Barnes GR ( 2005): Attention and selection for predictive smooth pursuit eye movements. Cognit Brain Res 25: 688– 700. [DOI] [PubMed] [Google Scholar]
  38. Schmid A, Rees G, Frith C, Barnes G ( 2001): An fMRI study of anticipation and learning of smooth pursuit eye movements in humans. Neuroreport 12: 1409– 1414. [DOI] [PubMed] [Google Scholar]
  39. Tanabe J, Tregellas J, Miller D, Ross RG, Freedman R ( 2002): Brain activation during smooth‐pursuit eye movements. NeuroImage 17: 1315– 1324. [DOI] [PubMed] [Google Scholar]
  40. Tregellas JR, Tanabe JL, Miller DE, Freedman R ( 2002): Monitoring eye movements during fmri tasks with echo planar images. Hum Brain Mapp 17: 237– 243. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Additional Supporting Information may be found in the online version of this article.

Supplementary Movie: The animation illustrates the visual displays under different cues, i.e. (1) No Trace Cue, (2) at 18 s: Fixation only; (3) at 30 s: Partial Trace Cue; (4) Full Trace Cue.

Click here for additional data file. (10.3MB, mpg)

Articles from Human Brain Mapping are provided here courtesy of Wiley

RESOURCES