Orienting attention in time: behavioural and neuroanatomical distinction between exogenous and endogenous shifts
Introduction
We have recently begun to investigate the ability to orient attention selectively towards particular moments in time [7], [39], [40]. Expectancies about when an event will occur can be used to optimise behavioural responding, in an analogous fashion to expectancies about where the event will occur [7]. In a previous neuroimaging study using both positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) [7], we compared and contrasted the neural systems involved in temporal and spatial orienting of attention, with adaptations of the classic attentional cueing paradigm [46]. These visual tasks required detection of peripheral targets, guided by informative cues regarding when and/or where they would occur. Both spatial and temporal orienting activated a common network of frontoparietal regions with most consistent foci in lateral and medial premotor areas and intraparietal sulcus. However, the overlap between spatial and temporal orienting networks was not complete, and the two showed opposite hemispheric lateralisation. The right posterior parietal cortex was associated with spatial orienting, while the left posterior parietal cortex was associated with orienting attention in time. We suggested that the common premotor and parietal areas represent sensorimotor circuits which support attentional functions generally (see [50]). The right hemispheric lateralisation for spatial orienting was consistent with neuropsychological evidence for the right-hemisphere dominance for spatial attention in humans [17], [23], [38], [62]. The left hemispheric lateralisation for temporal orienting was hypothesised to reflect a left-hemisphere dominance for a critically involved cognitive function such as motor intention or preparation [33], [54] or fine temporal discriminations [37], [57].
One of the main purposes of the present experiment was to verify whether the pattern of brain activations for temporal orienting would remain consistent in the absence of any spatial information in the task. It could be argued that the appearance of peripheral targets in the previous experiment introduced an element of reflexive spatial orienting. The common activation of frontoparietal areas might therefore have been a consequence of spatial orienting in all task conditions. This consideration is especially important in light of the hypothesised role of posterior parietal areas in spatial representations (see [1], [6] for reviews). To address this possibility, all stimuli in the present task were presented foveally.
The second objective of the experiment was to identify the brain areas activated by some of the multiple factors involved in temporal orienting. We used event-related fMRI to examine brain activations linked to specific types of trials, in which length of the cue-target delay and the validity of the cue-target contingency were varied. In a previous experiment, we identified brain areas preferentially activated by being invalidly-cued to a target location or temporal interval [40]. Violations of spatial or temporal expectancies were accompanied by selective engagement of the orbitofrontal cortex bilaterally. In addition, increased activation occurred in the lateral premotor and posterior parietal foci of attentional orienting network. The orbitofrontal cortex may be engaged during invalid trials either because of the switch in stimulus contingencies that drive responses [12], [13], [16], [25] or because of changes in emotional states linked to interference with expectancies or levels of task performance ([4], [10], [52]; see also [14]). Enhancement of activity in premotor and parietal areas during invalid trials suggested increased levels of cognitive functions linked to attentional orienting generally, such as sensorimotor preparation [50] or disengaging and shifting the attentional focus [17], [47].
The present experiment builds upon these findings, within the context of a purely foveal temporal orienting task. The cues were foveal concentric circles that predicted the appearance of a foveal target at the pre-specified interval (80% validity). A factorial design was used to test the effects of trial validity (valid, invalid cues) and cue-target interval (600, 1400 ms). Fig. 1 shows a schematic of the experimental design and task parameters. Furthermore, the event-related fMRI methodology provides a means to define the brain areas differentially engaged in single trials (events). Therefore, a more detailed analysis of invalid trials is possible than in our previous blocked design PET experiment [40] in which invalid trials were intermixed with valid trials during a scan. Using event-related fMRI, we can measure not only invalid trials in isolation from valid trials, but also distinct types of invalid trial.
Different types of invalid trials during temporal orienting may involve different mechanisms of breaking and redirecting expectations. The bimodal nature of the cue-target intervals (short, long) guarantees that if a target does not appear at the short interval it must appear at the long interval. Omission of an expected target at the short interval signals a breach in the cue-target contingency and permits subjects to re-orient attention voluntarily to the later time-point. Analysis of the behavioural data during temporal orienting in our previous experiments [7], [39], [40] supported the presence of additional shifts of attention during such invalid trials. The disadvantage of being invalidly cued was significantly smaller for trials in which the target occurred later than expected compared to those in which the target appeared sooner than expected. Trials in which targets appear later than predicted afford time for subjects to re-orient attention, and may therefore involve top-down, voluntary control of attentional shifts. In contrast, the appearance of an unexpected premature target interrupts the active attentional focus and draws attention reflexively. Trials in which targets appear sooner than predicted may therefore emphasise bottom-up, automatic grabbing of attention. Voluntary and automatic orienting of attention have been hypothesised to have distinct neural bases and have usually been measured using central and peripheral cues respectively [45]. We distinguish between being validly cued to orient attention towards a particular time point, and being invalidly cued which necessitates a shift of attention to a different time point. In our foveal task, we differentially measured voluntary and automatic shifting of attention through the use of unexpectedly short or long invalid trials. Though this approach is novel, we believe that it captures exogenous and endogenous attentional mechanisms at work, analogous to the way in which they may operate during spatial orienting of attention.
The length of the interval between a cue and a target in an attentional orienting task is also a critical variable. In the context of temporal orienting, different cue-target intervals may involve different degrees of motor preparation and anticipation. The event-related fMRI methodology allowed us to examine the contribution of cue-target interval to brain activations associated with temporal orienting. We were interested in distinguishing the effects of cue-target interval from those of trial validity, and in examining how these two factors may interact.
Section snippets
Subjects
Six healthy, right handed volunteers (mean age=26.8, four male) took part in the experiment. Subjects were physically fit, and none were taking medication. The experimental protocol was approved by the local hospital ethics committee, and written informed consent was obtained prior to the study.
Cognitive tasks
The task required detection of a visual target presented after an informative cueing stimulus. All stimuli were presented at the centre of gaze (foveally). Fig. 1(b) shows a schematic of the task. Visual
Behavioural data
Subjects performed the task with a high degree of accuracy. Anticipation errors, where subjects responded prior to target appearance accounted for less than 2% of the responses. Subjects failed to respond to targets on less than 1% of the trials. Reaction times for valid trials were shorter than those for invalid trials [F(1,10)=25.04, p<0.001] (Table 1) demonstrating appropriate allocation of attention to temporal cues. A significant interaction between validity and SOA indicated that the cost
Discussion
This study complements and extends previous research indicating a role for left frontoparietal areas in temporal orienting of attention [7]. Specifically, the present study used foveal (non-spatial) presentation of both cue and targets to disambiguate the role of spatial processing in temporal orienting of attention. Furthermore, the use of event-related fMRI allowed us to model the neural substrates of different types of trials separately. It was possible to identify brain regions sensitive to
Acknowledgements
We thank Dr Matthew Rushworth for invaluable discussion and comments. The experiments were funded by The Wellcome Trust.
References (63)
- et al.
Differential activation of right superior parietal cortex and intraparietal sulcus by spatial and nonspatial attention
Neuroimage
(1998) - et al.
Monitoring for target objects: activation of right frontal and parietal cortices with increasing time on task
Neuropsychologia
(1998) - et al.
The trouble with cognitive subtraction
Neuroimage
(1996) - et al.
Event-related fMRI: characterizing differential responses
Neuroimage
(1998) - et al.
Hemispheric assymetry of movement
Current Opinion in Neurobiology
(1996) - et al.
Functional anatomy of GO/NO-GO discrimination and response selection — a PET study in man
Brain Research
(1996) - et al.
Subregions within the supplementary motor area activated at different stages of movement preparation and execution
Neuroimage
(1999) - et al.
Cognitive conjunction: a new approach to brain activation experiments
Neuroimage
(1997) - et al.
The functional organization of the lateral frontal cortex: conjecture or conjuncture in the electrophysiology literature?
Trends in Cognitive Sciences
(1998) - et al.
The left parietal cortex and motor attention
Neuropsychologia
(1997)