Info

Note: lg = interval longer than 1 sec; sh = interval shorter than 1 sec; n_mt = interval not defined by movement; mt = interval defined by movement; disc = interval measured noncontinuously (discretely); seq = interval measured continuously (sequentially); CB lat. = lateral cerebellum; F. pole = frontal pole; basal G. = basal ganglia; operc. = frontal operculum; inf. par. = inferior parietal lobe; S. par. = superior parietal lobe; occip. = occipital lobe; S. temp. = superior temporal lobe; M1 = primary motor cortex; S1 = primary sensory cortex.

a The table is explained in detail in the text.

Perhaps the most important observation to make regarding these results is that the patterns seen when studies are divided based on combinations of task characteristics produce a more coherent picture than when all studies are averaged together. If these studies truly all draw on the same time measurement mechanism, then we might expect a stronger consensus regarding the areas involved than what is shown in row 1. Because different networks appear to be activated by tasks with different combinations of characteristics, this meta-analysis strongly supports the possibility of duplicitous mechanisms for time measurement.

Looking more closely at the specific areas activated, we see that several pre-frontal regions believed to contain flexible cognitive modules (Duncan, 2001) are associated with the cognitively controlled tasks, but remain inactive during automatic tasks. These include the DLPFC, VLPFC, IPS, and to a large extent, inferior parietal. Also interesting is the observation that many regions of the motor system (the SMA, sensorimotor cortex, left basal ganglia, right PMC, and right cerebellum) commonly activate during automatic tasks. This pattern supports the hypothesis that what we have termed automatic timing may rely upon mechanisms located within the motor system itself. That some of these areas (bilateral SMA and right PMC) are also commonly activated in association with the cognitively controlled tasks suggests that use of the cognitively controlled system does not preclude involvement of modules from the automatic system. Before reading too much into these patterns, however, it is important to consider whether the observed activity is all truly associated with timing mechanisms, or whether some of it might be due to non-timing-related confounders.

21.2.3.2 Possibility of Confounds

Because we have reported the most inclusive contrast from each study in our analysis, much of the activity we describe may be due to movement or other task-related nontiming behaviors. Observations that the auditory, visual, and primary sensorim-otor cortices are frequently activated in association with automatic timing tasks, for instance, should not necessarily be interpreted as support for the direct involvement of these areas in time measurement, because auditory or visual stimuli and movement in the tasks may have elicited this activity. Based upon the analysis presented thus far, it is impossible to determine whether activities are due to temporal processing or confounding factors. By looking more closely at some of the studies reviewed, however, we can begin to address this question.

If regions of the motor system are active even in those studies of timing where very little movement or movement preparation (or in some cases, no movement or movement preparation at all) occurred during scanning, then we can safely conjecture that their involvement is not merely motor associated, although we cannot rule out the possibility that motor imagery may be involved. This is the case for activity in the right cerebellar hemisphere (Belin et al., 2002; Jueptner et al., 1996; Larasson et al., 1996; Roland et al., 1981; Sakai et al., 1999; Schubotz and von Cramon, 2001), SMA (Gruber et al., 2000; Larasson et al., 1996; Schubotz and von Cramon, 2001), and left basal ganglia (Larasson et al., 1996; Parsons, 2001; Schubotz and von Cramon, 2001) during tasks requiring covert decisions, memory encoding, memory rehearsal of rhythms, or detection of oddballs but not movement. Because this activity is not due to movement, it may be genuinely linked to timing.

Likewise, several studies have described activity in the temporal cortex during time measurement tasks involving no auditory cues (Coull et al., 2000; Larasson et al., 1996; Rao et al., 2001). Others have shown auditory activity during task phases that come after the cessation of auditory cues, such as continuation of tapping after auditory synchronization (Rao et al., 1997) or memory encoding after presentation (Sakai et al., 1999). It has been suggested (Rao et al., 1997) that this activity may be associated with auditory imagery used for the task, so the observation that the right hemispheric superior temporal cortex is one of the most commonly activated areas during tasks that would be expected to draw on the automatic timing system may well mean that the timing of these intervals frequently draws on auditory imagery. By contrast, the lack of studies in which the occipital cortex is activated in response to tasks that do not involve visual stimuli makes it unlikely that the activity observed here is associated with temporal processing.

Because the tasks associated with the cognitively controlled system are quite different from those associated with the automatic system, it could be argued that activity unique to these tasks is due to some form of confounder. Looking carefully at the literature, however, we see that these regions activate even when a more complete cognitive subtraction is used (Lewis and Miall, 2002, in preparation; Rao et al., 2001); hence, their involvement very likely relates directly to temporal processing. Because these areas include regions known for involvement in both working memory (DLPFC) and attention (IPS and inferior parietal lobe), this observation conforms to predictions regarding the cognitively controlled system (for further details concerning activations specific to interval timing, see Hinton, this volume; Hinton and Meck, 1997; Morell, 1996; Pouthas, this volume).

0 0

Post a comment