| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Psychology, Royal Holloway, University of London, Egham, United Kingdom
Submitted 13 April 2007; accepted in final form 2 July 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
). Macaque pulvinar includes two subregions that are connected to the visual cortex and are retinotopically organized together with at least one other visual region that is not retinotopic. Proposed functions of the visual pulvinar are to orient visual attention and to disambiguate relationships between stimuli within the visual field. In macaques, visually responsive neurons are found that have bilateral receptive fields (Petersen et al. 1985), suggesting that the proposed integrative function acts across hemifields. In humans, such a view is supported by functional magnetic resonance imaging (fMRI) studies showing increased pulvinar activity during tasks that distribute attention between hemifields relative to those that do not (Kastner et al. 2004). Switching attention between bilaterally demanding stimuli results in task-related activity in left pulvinar, and it is the parietal cortex that is thought to control these shifts (Yantis et al. 2002). The posterior parietal cortex makes reciprocal connections with the pulvinar in nonhuman primates (Asanuma et al. 1985; Baleydier and Morel 1992), specifically at the border of the nucleus termed pulvinar dorsomedial (PdM) (Petersen et al. 1985). Units of this region exhibit bilateral receptive fields, fire during saccades to attended locations, and have little (Petersen et al. 1985) or no (Bender 1981) retinotopic organization. The organization of those portions of the human pulvinar that are active in attention-related studies share these characteristics, being reportedly without bias for spatial location of visual stimuli (Kastner et al. 2004; Yantis et al. 2002). This suggests that the pulvinar region activated in these human studies may be connected primarily to posterior parietal cortex and may be homologous to macaque PdM.
More inferior (ventral), retinotopic portions of the macaque posterior pulvinar connect with occipital cortex. Here the emphasis is on visual activity rather than attentional orienting. In contrast to more dorsal regions, a strong spatial organization has been reported. In each macaque pulvinar (left and right), two complete maps of the contralateral hemifield are found (Bender 1981; Shipp 2003). In humans, retinotopic maps have been found in multiple cortical areas (e.g., Engel et al., 1997; Sereno et al. 1995; Silver et al. 2005; Wade et al. 2002) and recently in the lateral geniculate nucleus (LGN) (Kastner et al. 2004), but they have not been reported in the pulvinar. Our objective here is to examine the human pulvinar for retinotopic organization. In macaque, neurons in inferior pulvinar can be driven by passive visual stimulation, but they are also strongly attentionally modulated (Petersen et al. 1985). For this reason, we combined a rich moving stimulus designed to give strong visual drive with the use of a task performed at the same spatial location. We report the existence of a bilateral region in the inferior pulvinar that responds preferentially to contralateral visual stimulation. This result, taken together with primate results, suggests the presence of a retinotopic map.
Parts of this work have previously been published in abstract form (Cotton and Smith 2006).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
A total of 22 participants took part. Of these, all but one (age 38) were between the ages of 19 and 30, and all were in good health. They were screened according to standard procedures and informed consent was obtained. In experiment 1, 14 participants (3 male, 11 female) were scanned. In experiment 2, 14 participants (6 male, 8 female) were scanned, of whom 6 also participated in experiment 1. Experiment 3 was conducted in the same scan session as experiment 1, and so the same 14 participants were used.
Visual stimulation and tasks
Computer-generated visual stimuli were presented via an LCD projector on a screen at the end of the scanner bore and were viewed via a mirror mounted on the headcoil.
EXPERIMENT 1: MOTION DISCRIMINATION. To claim evidence for a retinotopic organization, it is necessary, as a minimum, to show strong contralateral biases in symmetrical locations in the two pulvinar nuclei. The purpose of experiment 1 was to seek evidence for such a pair of pulvinar subregions. Bilateral, retinotopically organized visual activity has not previously been reported in studies of human pulvinar. We used a large stimulus that contains motion as well as spatial structure and is expected to produce strong activation in multiple areas of visual cortex. Reasoning that feedforward and -back connections to the cortex extend the length of the pulvinar, this rich stimulus should promote visual drive in the nucleus by virtue of its connections. Because of the pulvinar's strong association with attention, we presented these stimuli in the context of a demanding task.
An event-related fMRI design with a variable ISI was employed. Participants maintained fixation at a central point present throughout data collection while performing a task involving detecting a change in the global motion strength of a random-dot pattern (5° diam). Targets were presented for 3,000 ms and were centered at an eccentricity of 12.5° to either the right or left of fixation. Each consisted of 200 white dots on a dark background. Each dot was 40 arc min in size and moved in a straight path for 133 ms before disappearing and reappearing at a new, random location. Ten percent of dots were repositioned on each frame update. Local motion directions were controlled so as to create a global perception of optic flow. The stimulus was based on Morrone et al. (2000)'s study in which optic flow varies over time, smoothly changing between expansion/contraction and rotation, via intermediate spiral motions. This stimulus activates both low-level motion sensors in areas such as V1 and V2 and also higher-level neurons sensitive to global flow structure, such as those found in primate MST (Duffy and Wurtz 1991; Saito et al. 1986). The strength of the motion percept was manipulated by assigning random directions to a proportion of the dots (noise dots). During the first 1,500 ms of each trial, the signal strength was 50%. Mid-way through each presentation (i.e., after 1,500 ms), the proportion of noise dots was either increased to 70% or decreased to 30%, causing the perception of global motion to either strengthen or weaken. Participants had to maintain fixation throughout and report whether motion strength increased or decreased by pressing one of two keys of a response pad immediately after the end of the trial. The two directions of change were equally probable, and the direction was selected pseudorandomly on each trial. Stimulus location (left or right) was also determined pseudorandomly with equal probability. The location and time of onset of the trials were not precued. The intertrial interval varied (6 s average) according to the "long exponential" distribution of Hagberg et al. (2001). The average intertrial interval was balanced for both target location (right/left) and motion strength change (increase/decrease). A single run lasted 8 min 19 s and consisted of 56 trials. Four runs per session were completed. Participants were instructed to maintain fixation at a simple dot throughout. Fixation itself was not meant to be attentionally demanding, and we purposefully did not introduce a task at fixation in any of our experiments to avoid the confounds this may introduce elsewhere in the visual field (Schwartz et al. 2005).
Prior to the experiment, each participant viewed a practice sequence outside the scanner. The responses were used to adjust the proportion of noise dots, where necessary, such that 
70% of responses were correct. For 11 participants, the standard proportions described in the preceding text were used, whereas the remaining 3 participants required adjustments of 5%. Participants also had a short trial run in the scanner to confirm their ability to perform covert detection at the designated level.
EXPERIMENT 2: LUMINANCE DISCRIMINATION. Experiment 1 involved both strong visual drive and a demanding task. In experiment 2, a demanding task was employed at each of two well-separated locations (left and right) in a similar fashion to experiment 1, but a minimal visual stimulus was used.
An event-related fMRI design, similar to that of experiment 1, was employed. Targets were presented for 1,000 ms at an eccentricity of 10° to either the right or left of the fixation point, centered on the horizontal meridian. The target was a small (1.5°) circle of uniform luminance presented on a uniform background. The circle was brighter than the background and was easily visible. Mid-way through each presentation, the target luminance either increased or decreased by 12%. The participants' task was to detect the direction of the change, while maintaining central fixation, and make a forced-choice response using a keypad. The magnitude of the luminance change was chosen to give a performance level of 70%. On half the trials, determined randomly, luminance increased, and on half, it decreased. Similarly, half were presented on the left and half on the right. The location and time of onset of the trials were again un-cued. Between trials, the display was uniform apart from the fixation spot. The intertrial interval varied in the same way as in experiment 1. A single run consisted of 72 events, lasting 8 min 28 s, and participants completed 
6 such runs per session.
EXPERIMENT 3: PASSIVE VISUAL STIMULATION. In this experiment, motion stimuli similar to experiment 1 were used to give strong visual drive but the task was removed. The objective was to test whether sustained passive visual stimulation is sufficient to generate pulvinar activity and, if so, whether it is contralaterally organized.
To enable us to determine whether visual activity is sustained in nature or transient and related to stimulus onset, long stimulus presentations were used in a block design. Activity was then modeled in two ways. First a conventional block-design analysis was performed to detect sustained activity, and second the data were re-analyzed with a regressor modeling only transient activity at the onset of the block. The optic flow stimulus of experiment 1 was used. The signal component was 70%, the same as the strongest flow component used in experiment 1. Each stimulus was presented continuously for 15 s. Blocks of three types were used: motion in the right hemifield, motion in the left hemifield and fixation blocks with no motion stimulus. The sizes and locations of the stimuli were the same as in experiment 1. Blocks with motion on the right and left were alternated, separated by fixation blocks to provide an inter-stimulus baseline condition. One run contained 7 left and 7 right blocks together with 14 fixation blocks. A buffer of 10 s was used at the beginning and end of each run giving a total duration of 7 min 20 s. Two runs of this passive stimulus were viewed after the event-related scans of experiment 1.
Data acquisition
BOLD activity was measured at 3 Tesla (Magnetom TRIO, Siemens, Erlangen, Germany) using an eight-channel array headcoil and a standard gradient-echo, echoplanar sequence (TR = 2,000 ms, TE = 34 ms, flip angle: 90°). Axial slices were positioned so as to cover the entire thalamus as well as portions of the cerebral cortex. Voxel size was 2.5 mm (isotropic). A high-resolution three-dimensional (3D) anatomical scan (MP-RAGE, 1-mm resolution) was taken at the beginning of each session.
Analysis
The data were analyzed using BrainVoyager QX. Functional data were aligned to high-resolution anatomical scans, motion-corrected and slice-time-corrected. The first three volumes of every run were discarded to ensure a steady state. A high-pass filter was used to remove periodicities greater than three times the mean separation between events.
Both group and individual analyses were performed, each having differing merits. A group analysis with spatial normalization permits increased sensitivity to signals that may be hard to detect in a single scan on a single participant and gives increased confidence that the results will generalize to the population. However, individual analyses avoid the spatial blurring that is inherent to group analyses. The axis of the anterior-posterior commissure was defined in every participant, and both functional and anatomical volumes were aligned with this axis before being subsequently normalized into Talairach space (Talairach and Tournoux 1998) within BrainVoyager. The data were re-sliced using 3-mm voxels.
In the case of the group analyses, Talairach-transformed data were spatially smoothed using a 4-mm FWHM Gaussian kernel. Transformed data from all subjects were then entered into a multiple regression analysis (random-effects which tests for between-subject differences in response). For experiments 1 and 2, responses to all same-hemifield presentations were averaged and contrasted with the interstimulus baseline and contrasted to their contralateral counterparts. In addition, the responses to all events were compared with baseline. Baseline was defined as the average activity during a period of 4 s immediately preceding each event, the definition also used to calculate event-related time courses. The criterion for statistical significance employed was P < 0.01 corrected for multiple comparisons (FDR) in all analyses. Pulvinar voxels meeting this threshold were included in the event-related time course plots. In experiment 3, blocks containing stimuli on the right and left were independently contrasted with baseline blocks and were also contrasted with each other. To examine results for individual participants, a separate GLM analysis was conducted on the data without spatial smoothing for each participant in each experiment. These analyses were performed in Talairach space to allow meaningful interpretation of locations, and the contrasts made were the same as those used in the group analyses.
In all experiments, regions of activity were identified within the pulvinar as defined by the stereotaxic atlas of Morel et al. (1997). Activity in other parts of the brain was not examined. For each cluster of active voxels identified, an average time course was calculated reflecting the mean of all voxels in the cluster. For ease of comparison with other studies, the Talairach coordinates of these clusters are quoted where appropriate.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
The results of experiment 1 reveal two small regions of activity in the inferior pulvinar (z = 1), one in each hemisphere, each of which is active only when the participant performs the task in the contralateral hemifield. The existence of these areas is revealed very clearly and reliably in the group analysis involving all 14 participants. Figure 1 shows slices from the standard brain used for normalization with significantly active voxels overlaid in color. The activation is derived from contrasting left events with right in which only differential activation between the two sides is detected; any activity that is present in both sides is discounted. The figure reveals activity in a region of left pulvinar (shown in blue) that is active during events occurring in the right hemifield and another area in right pulvinar (orange) in which activity is correlated with events on the left.
|
). It should be noted that the time courses are derived from the same data as the region of interest itself and that some people argue that this can exaggerate effects by introducing sampling bias. When responses to events in both right and left hemifields were pooled and contrasted with baseline activity, the group analysis showed no significant activity in the inferior pulvinar in either hemisphere. This presumably reflects the fact that only half the events contributed positively to such a contrast, whereas the other half contributed neutrally or even negatively. Significant activation was seen, in both pulvinar (more strongly in right than left), when contralateral events only were contrasted with baseline (not shown). But the clearest results were obtained with the contralateral versus ipsilateral contrast used for Fig. 1.
Because ventrolateral pulvinar is adjacent to the LGN, it is important to avoid mistaking LGN activity for pulvinar activity. Complete reliance on Talairach coordinates is unsafe, particularly in a group analysis. We therefore checked for the existence of a separate peak of activity in the LGN. This was clearly seen in the expected location, based on anatomical expectations together with previously published coordinates for functional activation (Kastner et al. 2004). Figure 2 A, shows the distinct peaks of activity in the two regions. The LGN is centered some 9 mm from the pulvinar, being more lateral as well as more ventral, as expected.
|
|
|
These results show that the use of lateralized complex motion stimuli results in activation of the inferior pulvinar in both hemispheres. They also show total specificity for contralateral stimuli, providing the first evidence consistent with retinotopic organization in the human pulvinar.
In the random-effects group analysis, significant activity was also seen in a more anterior/dorsal location, in the left thalamus only (–10, –17, 14). Unlike the more ventral region discussed in the preceding text, this dorsal region was active when all events were contrasted with baseline and not when right and left events were contrasted with each other. However, this location is probably too anterior to be in the pulvinar and is close to the ventroposterior nucleus (VPL) of Morel (1997), a position that has been associated (based on its connectivity) with motor functions (Johansen-Berg et al. 2005). Given that our participants were responding via a button pad with the right hand and activity was seen only on the left, but for all events, this activity may be related to motor aspects of the task.
Although some previous studies (Kastner et al. 2004; Villeneuve et al. 2005) have reported activity in more superior portions of the pulvinar, only one study (Yantis et al. 2002) has previously reported visually related activity in the inferior pulvinar, and then only on the left. The inferior right pulvinar has not previously been activated by any paradigm to our knowledge. Other studies reporting right pulvinar activity do so only at more superior locations. In our random-effects analysis, we did not see activity in the region of superior pulvinar identified as active during attentional engagement by Kastner et al. (2004). Even with fixed-effects analysis, a less stringent analysis that does not test between-subject variability, only a small cluster was evident, seen when all events were contrasted with baseline (i.e., not retinotopic). Presumably our paradigm is not optimal for activating this region. The relationship of our results to activity shown in previous studies will be considered in the DISCUSSION.
Experiment 2
In experiment 1, rich visual stimulation was combined with a demanding task to maximize pulvinar activation. This leaves us not knowing whether the activity observed reflected passive visual stimulation, task-related factors, or both. Experiment 2 retained the task requirement of experiment 1, but, in contrast to experiment 1, it minimized visual drive by using a small, uniform, static target. If the results were similar to experiment 1, it would suggest that strong visual drive is not essential for activating the contralaterally organized pulvinar regions we have defined in experiment 1. In fact, the observed activation was much weaker.
As in experiment 1, which also used an event-related design, a group analysis was performed using two different contrasts: all events versus baseline, to detect event-related activity that is not spatially organized, and right events versus left events, to detect activity that is spatially mapped. The right versus left analysis revealed activity in the left posterior/ventral pulvinar, corresponding to the region identified in experiment 1, but it was weaker than in experiment 1 and absent altogether in the right pulvinar.
Figure 4A shows an axial slice at Talairach plane z = 2. Superimposed are active voxels from a random-effects group analysis in which right and left events were contrasted. In the left pulvinar, a small cluster is significantly active at the posterior/medial boundary of the pulvinar (–10, –31, 2). This region exhibits strong responses to contralateral (right) events but suppression for ipsilateral (left) events. This is shown in the associated time courses in Fig. 4B. The peak is close to the region found to be active in experiment 1 (–11, –28, 1), and it is therefore likely that they reflect the same functional unit. As in experiment 1, a suppressive effect is seen during ipsilateral events (blue time course).
|
; Shipp 2003). According to Shipp (2003), the caudomedial to rostrolateral axis forms an "axis of isorepresentation." That is, as one moves along this axis, one moves through the connection zones of different visual areas, with no change of visual field location. This gives rise to the possibility that the extended focus seen in Fig. 4C reflects activity in different connection zones. The results for individual participants mostly show no activity. Activity was present in only one hemisphere at the level of P < 0.01 corrected and in one other at P < 0.02. In experiment 2, the added power of a group analysis was required to detect activity whereas in experiment 1, significant activity was seen more often than not in individuals (see Table 1).
These results confirm the demonstration (experiment 1) of the existence of a contralaterally organized region in inferior pulvinar but do so only in one hemisphere. Activity (measured in terms of cluster size) in the inferior left pulvinar is substantially weaker during the luminance-discrimination task of experiment 2 than the motion-discrimination task of experiment 1. Because the task difficulty is similar for the two tasks, the greater activity in experiment 1 may reflect greater visual drive from the motion stimulus, which was both larger and richer. However, such comparisons are difficult because the pool of participants was not the same and there are various factors affecting sensitivity that may vary from scan to scan. It should be noted that medial inferior pulvinar makes dense reciprocal connections with MT in macaques (Shipp 2001), so it is possible that motion stimuli provide greater activation for this reason.
No reliable activity was seen in superior pulvinar in experiment 2, even when using a fixed-effects analysis. As in experiment 1, activity was seen in left thalamus in the region of VPL (–11, –20, 13), probably related to button-pressing with the right hand.
Experiment 3
In experiment 3, motion stimuli were presented and viewed passively with no task other than fixation. Stimuli were presented in a block design to examine whether visual sustained visual responses occur in the pulvinar in the absence of endogenous attention or other task demands. A group analysis was conducted in which each stimulus block (left or right) in turn was contrasted with baseline blocks. Right and left stimulus blocks were also contrasted with each other.
The results of a random-effects group analysis in which responses were modeled conventionally (a boxcar profile convolved with the hemodynamic response function) provide further strong support for the existence of a map of contralateral visual space within both inferior pulvinar nuclei. Where left-stimulation blocks only are contrasted with baseline (fixation) blocks, right inferior pulvinar is active (coordinates: 19, –25, 1), whereas the left pulvinar is not. Similarly, activity in left inferior pulvinar (–21, –25, 2) is evident when right events are contrasted with baseline, whereas right pulvinar activity is absent. The same bilateral activity emerges strongly when right- and left-stimulation blocks are contrasted with each other. This is shown in Fig. 5. As in experiments 1 and 2, this contrast tests for contralateral specificity. In this case, the peak of the activity is at 16, –28, 2 (right) and –22, –27, 5 (left).
|
The associated time courses (also shown in Fig. 5) show sustained activity, consistent with sustained visual drive rather than a transient alerting response. However, this may be because the model used in the analysis assumed sustained activity and therefore picks out those voxels that show it. To cover the possibility that there is transient activity in some voxels, we repeated the analysis of experiment 3 using a model that contained only an impulse (convolved with a hemodynamic response function) at the onset of each stimulus block. This resulted in significant LGN activity but not pulvinar. The lack of transient pulvinar activity confirms that visual drive in the inferior pulvinar is sustained in nature, at least in response to time-varying optic flow.
In the right pulvinar, the time courses show clear suppression of activity during ipsilateral stimulation (bottom right panel, red trace), as seen in experiments 1 and 2. This is not evident in left pulvinar; however, in this case, the time courses are drawn from a very small cluster of voxels and so lack sensitivity.
In both left and right pulvinar, the location of the active region is similar to experiment 1 (no right pulvinar activity was seen in experiment 2) but slightly more lateral and anterior. It is close to the anterolateral end of the area identified in the fixed-effects analysis of experiment 2 and shown in Fig. 4C.
No activity was seen in superior pulvinar in experiment 3.
COMPARISON OF RESULTS ACROSS EXPERIMENTS. Our three experiments reveal activity in broadly the same regions. However, close comparison reveals a systematic variation. This is shown in Fig. 6, which summarizes the results from all experiments. The various locations for activation are remarkably circumscribed in the z dimension (dorsal-ventral), all being within 1 or 2 mm of z = 2. But in both x and y dimensions, the location of activity is much more variable. Figure 6 represents an axial slice at z = 2 showing this variability. When the locations of activity, as determined by random-effects analysis (solid colors) are compared, it is evident that they are aligned on an axis that runs posteriomedial to anterolateral and that the alignment is symmetrical in the right and left pulvinar. Activity from experiment 3 (passive visual stimulation) is the most anterolateral, being located in the portion of the pulvinar that is adjacent to the LGN. Experiment 2, in which there is a demanding task but less visual stimulation, activates the most posteromedial location. Experiment 1, which involves both strong visual drive and a task, gives peak activation at an intermediate location. Interestingly, when the active regions contained in the less-stringent fixed-effects analysis are examined (colored outlines), it is not the case that they are centered on the corresponding solid colored regions. Rather, they tend to spread toward the anterolateral end of the pulvinar. This is most obvious in the left pulvinar, where all three experiments produced activation, but results from the right pulvinar are consistent.
|
, 2003) in the macaque pulvinar has demonstrated an organizational structure running along this same axis. The extreme anterolateral portion, adjacent to the LGN, is strongly connected with cortical area V1. As one proceeds in the posteromedial direction, the cortical connection zones move through an orderly sequence (V1, V2, V3, V4), albeit with substantial overlap. Visual field location is invariant in this dimension being mapped in the two orthogonal dimensions. Thus a line drawn in this axis joins points that represent the same field location but with connections to different cortical visual field maps. Shipp referred to this as an axis of isorepresentation. We propose that the same organization may exist in the human pulvinar. Hypothetical axes of isorepresentation are marked on Fig. 6. However, the evidence is limited, and the hypothesis is tentative. On this interpretation, the systematic sequence of random-effects activations across experiments suggests that passive viewing activates primarily pulvinar tissue that is connected to early visual areas, whereas when attention is emphasized, tissue connected with higher visual areas is most active. The fact that it spreads preferentially toward V1-recipient pulvinar in a less stringent analysis, rather than spreading equally in all directions as is often seen in the cortex, suggests that anterolateral pulvinar is always active in visual tasks, but the degree of spread toward portions that connect with higher visual areas depends on the level of attentional engagement.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
We have provided the first fMRI evidence of contralateral organization in the human pulvinar. At the posterior/ventral margin of the pulvinar, a small region of visual sensitivity exists bilaterally with sensitivity only to stimuli in the contralateral visual hemifield. Ipsilateral stimuli do not drive this region and, indeed, appear to suppress it. We will refer here to this region as the contralateral zone of the puvinar (CZP). Contralateral specificity may indicate the presence of a complete map of visual space as found in posterior/ventral macaque pulvinar. But because of the small size of the CZP, confirmation of the presence of a full map of space may have to await technological improvements in fMRI measurement.
In view of the trend for greater activation in the left than right pulvinar (Table 1 and experiment 2), it is important to question whether the contralaterally organized zone of the pulvinar is truly bilateral and symmetrical. In general, more fMRI studies show left- than right-pulvinar activity, whereas this is not true of PET studies. It is worth mentioning a possible MRI-specific methodological interpretation of this. If the location of the pulvinar varies between participants more on the right than on the left or if the location of a particular zone within the pulvinar varies more on the right, group significance may be lost in the right pulvinar given that the focus of activity is small. In the case of PET, spatial resolution is inferior, and this artifact may be absent because the differences in variance of localization are not resolvable. However, in experiment 1, the left cluster was more prominent than the right in most individuals (Table 1), suggesting that it is not an artifact of group analysis.
One possible explanation of the apparent left-bias in our data is that it does not reflect properties of the contralateral region itself, which is expected to be symmetrical, but instead reflects contamination from more dorsal pulvinar regions that are asymmetric. Consistent nonretinotopic activation of a pulvinar region centered some 7–8 mm more dorsally indicates a separate superior region responsive to both left and right events, i.e., with bilateral fields (Kastner et al. 2004). This should not cause an artifactual asymmetry in the contralateral region due to spread of the BOLD response and/or localization errors relating to normalization because it exists bilaterally. But a third, separate area with some specific role in spatial attention might exist on the left, and this might artificially boost measured activity in the left contralateral area. Despite symmetrical anatomical connections, the contribution of attention to pulvinar drive could result in laterality of function in pulvinar. Even if cortico-pulvinar connections are symmetrical, laterality of parietal function coupled with parietal interconnections with the pulvinar (Asanuma et al. 1985; Baleydier and Morel 1992; Schmahmann and Pandya 1990) may result in asymmetric pulvinar activity. Additionally, the pulvinar is a large nucleus, and, taken together with existing data, our experiments suggest that there are multiple compartments of attentional and visual drive within the pulvinar.
Because of the historical association of the pulvinar with attention and the existence of a more dorsal pulvinar subregion with clear dependence on spatial attention (Kastner et al. 2004), we have examined the influence of task demands that include a likely requirement for covert spatial attention on the CZP as well as documenting visual drive. Attentional facilitation is known to occur retinotopically in cortex (Slotnick et al. 2003; Tootell et al. 1998) and has recently been shown to be specific to the size of the attended target (Bles et al. 2006), so specificity for location and size may be expected for attentional effects as well as for visual drive. We tentatively suggest that of the two influences, visual drive may be dominant in CZP, provided a suitable visual stimulus is employed. A rich, complex global-flow stimulus gives strong visual drive, whether it is attended (experiment 1) or not (experiment 3). But when visual drive is minimized to isolate effects of attention (experiment 2), activation is weaker and (at least for our particular luminance discrimination task) only detectable in the left pulvinar. It is even possible that task demands made little or no contribution to the CZP activity we report despite an undeniable role elsewhere in posterior pulvinar. After all, our minimal stimulus did have a spatial form, an onset, an offset, and a luminance transition in the middle of the presentation. However, this question is far from resolved. Further experiments will be required to probe the roles of attention and visual drive more directly, perhaps by pitting them against each other in the same experiment.
The sustained time courses seen in response to 15-s "block" stimuli (experiment 3; Fig. 5) show that the CZP does not respond in a solely transient manner. If CZP activity does, to some extent, reflect spatial attention, then this activity largely reflects sustained attention, rather than a transient orienting signal. The electrophysiology of primate pulvinar also suggests sustained visual response time courses (Petersen 1985).
Relation to other studies
Several studies have previously documented visual responses in human pulvinar, but none has previously succeeded in finding the zone of retinotopic organization that is expected based on primate studies. Possibly this is due in part to some aspect of data acquisition or to the use of insufficient numbers of participants. But more likely it is due largely to the stimulation protocol. Our success is probably attributable to the use of a stimulus that is richer, in both spatial and temporal domains, than has been used before in this context. When this stimulus was not used (experiment 2), only weak, unilateral activity was seen. Kastner et al. (2004) presented flickering checkerboard stimuli to each hemifield in turn, with passive viewing, and reported no activation in the pulvinar. When they presented similar stimuli to both hemifields and alternated attention between the two sides, they found bilateral attention-related activity in a dorsal region of posterior pulvinar (average z coordinate = 10). We found only weak signs of activity at this location. Two stimulus differences may explain the difference between the results of their study and ours. First, by exclusively stimulating the peripheral visual field, we expect to selectively activate units with well-separated receptive field locations. Kastner et al. used checkerboards spanning 1.5–12.5°; in this case, a between-hemifields contrast may preclude differential activation because units with receptive fields near the vertical meridian may be activated by both right and left stimuli. But the key reason for the different result may be simply that inferior pulvinar is inefficiently driven by flickering checkerboard stimuli, but strongly driven by optic flow stimuli, possibly reflecting its known strong connections with the MT complex.
The presence of CZP only in the left hemisphere in experiment 2 is reminiscent of a previous report of activity lateralized to left pulvinar during attentional switching between right and left (Yantis et al. 2002). The concurrence of Talairach coordinates between the two experiments is notable. This is one of only two fMRI studies, to our knowledge, to have reported pulvinar activity as ventrally as z = 2. Our stimuli were all unilateral, so it is difficult to compare our results directly with theirs. However, two factors make us wonder whether we and they are in fact documenting different functional regions despite the similarity of coordinates. Given that location can vary to some extent across brains, this is not impossible. The first factor is that their time courses for shift to left and shift to right are similar; this is their basis for asserting lack of spatial selectivity in the left pulvinar. This is far from the case in our CZP. Time-courses for left and right stimuli (Figs. 1, 3, 4, and 5) are quite different, one being excitatory and the other inhibitory. The second is the fact that they report transient rather than sustained activity. This latter property clearly does not pertain in the lateral portion of our CZP during experiment 3, which, as previously noted, shows that activity is sustained (Fig. 5). It may be significant that whereas the response time course for right CZP is quite flat, that in left CZP could be interpreted as the sum of a sustained response, similar to that seen on the right, and an early transient component. If this is the case, it is possible that our experiment 3 and the Yantis et al. study do indeed identify different functional regions but that our left CZP cluster contains some voxels belonging to that of Yantis et al. (2002). Against this, such voxels ought to be excluded by a left versus right contrast if their time courses are similar for both. Also, it is odd (on this interpretation) that there is no sign of positive-going transient activity for ipsilateral stimulation; this shows suppression from the onset of the stimulus. We therefore think that such an interpretation is unlikely.
One other fMRI study has reported activity in the posterior pulvinar. Fan et al. (2005) used a task designed to separate different aspects of attention. They found activity relating to executive control while performing a visual judgment in the left pulvinar, at a lateral location (–22, –27, 3) consistent with the more lateral portion of left CZP. The task is so different, that again it is hard to know whether this is truly the same functional zone. Several PET studies are also relevant to our work, although they do not allow reliable distinction between the ventral and dorsal zones we have discussed. Some have associated the pulvinar with attention and arousal (Bundesen et al. 2002; LaBerge and Buchsbaum 1990; Sturm et al. 1999) but, interestingly, two have found strong responses to motion stimuli with a global structure (Beer et al. 2002; Villeneuve et al. 2005). The left bias evident in some fMRI studies is not prominent in the PET literature; indeed some studies report greater activity in right pulvinar than left (Barrett et al. 2001; Sturm et al. 1999).
Finally, one neuropsychological study is relevant. Ward et al. (2002) documented the effects of circumscribed unilateral damage to the dorsal anterior boundary of the right pulvinar, which produced an attentional deficit restricted to the lower left visual field. This supports the presence of contralateral topography and indeed suggests further subdivision of the map into quadrants and presumably a full map; however, the lesion in this patient appears to be rather more anterior than CZP. An upper-lower quadrant differentiation is also, more loosely, supported by fMRI studies of the "filling in" phenomenon, in which regions in both right and left pulvinar are active during discrimination in the upper hemifield to a greater extent than in the lower hemifield (Mendola et al. 2006). How the organization of CPZ might represent upper and lower space is the subject of current work in our lab.
Conclusion
We have described a small, bilateral region of the human inferior pulvinar that responds exclusively to contralateral visual stimuli. We provisionally call it CZP (contralateral zone of pulvinar). It is distinct from previously reported attention-related regions of pulvinar activity, and it is characterized by visual sensitivity, both in the presence and in the absence of attention. We also present tentative evidence for an organizational structure within CZP that shares commonalities with the retinotopic zone of macaque inferior pulvinar. That we have been able to demonstrate contralateral specificity in human pulvinar is most likely due to the use of stimuli that provide very strong visual drive. Our data support the existence, within the posterior human pulvinar, of functional compartmentalization on a scale of a few millimeters.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: A. Smith, Dept. of Psychology, Royal Holloway, University of London, Egham, Surrey, TW20 0EX, UK (E-mail: a.t.smith{at}rhul.ac.uk)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Baleydier C, Morel A. Segregated thalamocortical pathways to inferior parietal and inferotemporal cortex in macaque monkey. Vis Neurosci 8: 391–405, 1992.[ISI][Medline]
Barrett NA, Large MM, Smith GL, Michie PT, Karayanidis F, Kavanagh DJ, Fawdry R, Henderson D, O Sullivan BT. Human cortical processing of color and pattern. Hum Brain Mapp 13: 213–225, 2001.[CrossRef][ISI][Medline]
Beer J, Blakemore C, Previc FH, Liotti M. Areas of the human brain activated by ambient visual motion, indicating three kinds of self-movement. Exp Brain Res 143: 78–88, 2002.[CrossRef][ISI][Medline]
Bender D. Retinotopic organization of macaque pulvinar. J Neurophysiol 46: 672–693, 1981.
Bles M, Schwarzbach J, De Weerd P, Goebel R, Jansma BM. Receptive field size-dependent attention effects in simultaneously presented stimulus displays. Neuroimage 30: 506–511, 2006.[CrossRef][ISI][Medline]
Bundesen C, Larsen A, Kyllingsbaek S, Paulson OB, Law I. Attentional effects in the visual pathways: a whole-brain PET study. Exp Brain Res 147: 394–406, 2002.[CrossRef][ISI][Medline]
Cotton P, Smith AT. Spatial and attention-related activity in the human pulvinar. Org Hum Brain Mapp Abstr 487-WAM, 2006.
Duffy CJ, Wurtz RH. Sensitivity of MST neurons to optic flow stimuli. I. A continuum of response selectivity to large-field stimuli. J Neurophysiol 65: 1329–1345, 1991.
Engel SA, Glover GH, Wandell BA. Retinotopic organization in human visual cortex and the spatial precision of functional MRI. Cereb Cortex 7: 181–192, 1997.
Fan J, McCandliss BD, Fossella J, Flombaum JI, Posner MI. The activation of attentional networks. Neuroimage 26: 471–479, 2005.[CrossRef][ISI][Medline]
Hagberg G, Zito G, Patria F, Sanes J. Improved detection of event-related functional MRI signals using probability functions. Neuroimage 14: 1193–1205, 2001.[CrossRef][ISI][Medline]
Johansen-Berg H, Behrens TEJ, Sillery E, Ciccarelli O, Thompson AJ, Smith SM, Matthews PM. Functional-anatomical validation and individual variation of diffusion tractography-based segmentation of the human thalamus. Cereb Cortex 15: 31–39, 2005.
Karnath H-O, Himmelbach M, Rorden C. The subcortical anatomy of human spatial neglect: putamen, caudate nucleus and pulvinar. Brain 125: 350–360, 2002.
Kastner S, Oconnor DH, Fukui MM, Fehd HM, Herwig U, Pinsk MA. Functional imaging of the human lateral geniculate nucleus and pulvinar. J Neurophysiol 91: 438–448, 2004.
LaBerge D, Buchsbaum M. Positron emission tomographic measurements of pulvinar activity during an attention task. J Neurosci 10: 613–619, 1990.[Abstract]
Mendola JD, Conner IP, Sharma S, Bahekar A, Lemieux S. fMRI measures of perceptual filling-in in the human visual cortex. J Cog Neurosci 18: 363–375, 2006.
Morel A, Magnin M, Jeanmonod D. Multiarchitectonic and stereotactic atlas of the human thalamus. J Comp Neurol 387: 588–630, 1997.[CrossRef][ISI][Medline]
Morrone MC, Tosetti M, Montanaro D, Fiorentini A, Cioni G, Burr DC. A cortical area that responds specifically to optic flow, revealed by fMRI. Nat Neurosci 3: 1322–1328, 2000.[CrossRef][ISI][Medline]
Pessoa L, Ungerleider L. Neural correlates of change detection and change blindness in a working memory task. Cereb Cortex 14: 511–520, 2004.[CrossRef]
Petersen SE, Robinson DL, Keys W. Pulvinar nuclei of the behaving rhesus monkey: visual responses and their modulation. J Neurophysiol 54: 867–886, 1985.
Saito H-A, Yukie M, Tanaka K, Hikosaka K, Fukada Y, Iwai E. Integration of direction signals of image motion in the superior temporal sulcus of the macaque monkey. J Neurosci 6: 145–157, 1986.[Abstract]
Schmahmann J, Pandya D. Anatomical investigation of projections from thalamus to posterior parietal cortex in the rhesus monkey: a WGA-HRP and fluorescent tracer study. J Comp Neurol 295: 299–326, 1990.[CrossRef][ISI][Medline]
Sereno MI, Dale AM, Reppas JB, Kwong KK, Belliveau JW, Brady TJ, Rosen BR, Tootell RBH. Borders of multiple visual areas in humans revealed by functional magnetic resonance imaging. Science 268: 889–893, 1995.
Schwartz S, Vuilleumier P, Hutton C, Maravita A, Dolan RJ, Driver J. Attentional load and sensory competition in human vision: modulation of fMRI responses by load at fixation during task-irrelevant stimulation in the peripheral visual field. Cereb Cortex 15: 770–786, 2005.
Sherman S, Guillery R. Functional organization of thalamocortical relays. J Neurophysiol 76: 1367–1395, 1996.
Shipp S. Corticopulvinar connections of areas V5, V4, and V3 in the macaque monkey: a dual model of retinal and cortical topographies. J Comp Neurol 439: 469–490, 2001.[CrossRef][ISI][Medline]
Shipp S. The functional logic of cortico-pulvinar connections. Philos Trans R Socf Lond B Biol Sci 358: 1605–1624, 2003.[CrossRef]
Silver MA, Ress D, Heeger DJ. Topographic maps of visual spatial attention in human parietal cortex. J Neurophysiol 94: 1358–1371, 2005.
Slotnick SD, Schwarzbach J, Yantis S. Attentional inhibition of visual processing in human striate and extrastriate cortex. Neuroimage 19: 1602–1611, 2003.[CrossRef][ISI][Medline]
Smith AT, Williams AL, Singh KD. Negative BOLD in the visual cortex: Evidence against blood stealing. Hum Brain Mapp 21: 213–220, 2004.[CrossRef][ISI][Medline]
Sturm W, Simone AD, Krause BJ, Specht K, Hesselmann V, Radermacher I, Herzog H, Tellmann L, Müller-Gärtner H-W, Willmes K. Functional anatomy of intrinsic alertness: evidencefor a fronto-parietal-thalamic-brainstem network in theright hemisphere. Neuropsychologia 37: 797–805, 1999.[CrossRef][ISI][Medline]
Talairach J, Tournoux P. Co-planar Sterotaxic Atlas of the Human Brain. New York: Thieme, 1998.
Tootell R, Hadjikhani N, Hall E, Marrett S, Vanduffel W, Vaughan J, Dale A. The retinotopy of visual spatial attention. Neuron 21: 1409–1422, 1998.[CrossRef][ISI][Medline]
Villeneuve MY, Kupers R, Gjedde A, Ptito M, Casanova C. Pattern-motion selectivity in the human pulvinar. Neuroimage 28: 474–480, 2005.[CrossRef][ISI][Medline]
Wade AR, Brewer AA, Rieger JW, Wandell BA. Functional measurements of human ventral occipital cortex: retinotopy and colour. Philos Trans R Socf Lond B Biol Sci 357: 963–973, 2002.[CrossRef]
Ward R, Danziger S, Owen V, Rafal R. Deficits in spatial coding and feature binding following damage to spatiotopic maps in the human pulvinar. Nat Neuroscie 5: 99–100, 2002.[CrossRef]
Yantis S, Schwarzbach J, Serences JT, Carlson RL, Steinmetz MA, Pekar JJ, Courtney SM. Transient neural activity in human parietal cortex during spatial attention shifts. Nat Neurosci 5: 995–1002, 2002.[CrossRef][ISI][Medline]
This article has been cited by other articles:
|
|
 |
|
  I. Arend, R. Rafal, and R. Ward Spatial and temporal deficits are regionally dissociable in patients with pulvinar lesions Brain, August 1, 2008; 131(8): 2140 - 2152. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||
). In each case, the associated time courses are shown. These show suppressive responses to ipsilataral stimuli as well as excitatory responses to contralateral stimuli.
