5 General discussion and future directions


In sum, the results of the three fMRI studies presented in this dissertation contribute essentially to the understanding of the functional role of the lateral prefrontal cortex in interference processing in dual tasks.

Study 1 showed the generality of dual-task-related activity in the lPFC. The finding of dual-task-specific involvement of the lPFC was extended to stimulus-response pairings with non-overlapping stimulus- and response modalities, thus unambiguously relating dual-task-specific activity changes to the processing of central bottleneck interference. In addition, Study 1 revealed that not only the coordination of the temporal order in dual tasks is related to lPFC activity but also coordination with respect to overlapping central task representations. These findings show that the lPFC has a rather generic function for the processing of interference at different levels. Interference processing in single tasks is only one facet (Miller & Cohen, 2001). Perceptual (Jiang et al., 2004), motor (Herath et al., 2001) and, importantly, response selection interference in dual-task situations require the coordination and control of two relevant task streams according to internal priorities and goals. Study 1 has firstly shown unambiguously that the lPFC is involved in coordinating the processing of two decisional processes that that can not proceed in parallel.


In Study 2, the behavioral and neural dissociability of two control functions involved in dual-task processing in the lPFC was shown. Whereas the control of the task order in dual tasks was associated with lPFC activity, task set maintenance was mainly associated with lateral premotor activity. This result supports the idea that task order control is a separable control mechanism in the lPFC that deals with dual-task interference and that can be dissociated from simple task set maintenance. From a cognitive neuroscience perspective this result supports the conclusion that not all types of cognitive control mechanisms are located in overlapping regions within the lPFC (see also Hester, Murphy, and Garavan, 2004). Instead, it was shown that the specific localization of different cognitive control mechanisms can be distinguished using appropriate experimental designs. In addition, this neural evidence for the involvement of task order control in dual-task processing clearly contradicts cognitive models that assume passive bottleneck processing (Pashler, 1994). Instead, direct evidence is provided for active control mechanism involved in dual-task processing (Logan & Gordon, 2001; Meyer & Kieras, 1997; Sigman & Dehaene, 2006).

The third study of this dissertation showed that the differential interaction of the lPFC with posterior task relevant regions (FFA, VWFA) is one crucial mechanism how the lPFC deals with dual-task interference. Functional coupling of the lPFC with regions relevant for S1- processing was increased at high temporal overlap with a secondary task. This is in accordance with the assumption that active task setting mechanisms focus attention on task 1 to prevent interference in this prioritized task. In contrast, task 2 performance costs might be related to the deficient functional coupling of regions relevant for S2 processing with the lPFC at high temporal overlap. This result greatly extends previous insights on the functional neuroanatomy of dual-task processing which was primarily investigated within the lPFC. Study 3 provides evidence that setting priorities between multiple relevant information streams takes place in interaction with those information streams. Study 3 showed that the integration of relevant sensory information with the internal information on priorities and goals is realised by functional coupling of sensory and prefrontal regions.

Taken together, the functional role of the lPFC in dual-task processing can be described within the framework of cognitive control: the lPFC coordinates multiple task-relevant information streams in accordance with our internal goal hierarchies. The lPFC does so by biasing the processing in posterior brain regions in order to set task priorities and to switch between relevant task representations (Koechlin et al., 2003; Miller & Cohen, 2001).


The finding that goal-oriented mechanisms of task order control can be separated from other dual-task related mechanism like task set maintenance also fits well to conceptions of a hierarchical organization of the frontal cortex (Fuster, 1989; 2001; Koechlin et al., 2003). These models relate the premotor cortex to pure sensory-motor control and more rostral lPFC regions to the context-specific control and control related to specific episodes. Thus, the findings of the dissertation support existing models on cognitive control in the lPFC, providing extensive new empirical evidence for task situations with multiple relevant information streams.

The focus of all three studies presented here lay on the functional role of the lPFC in dual-task processing. Not surprisingly, in all three studies additional regions were involved as indicated in the respective whole brain analyses. The potential role of these regions will be shortly discussed at this point.

The medial prefrontal cortex was involved in the comparison of dual tasks and single tasks in Study 1 as well as in the SOA contrast in Study 3. The medial prefrontal cortex, including the pre-supplementary motor area (pre-SMA) and the anterior cingulate cortex (ACC), has been found to be involved in performance monitoring and the detection of errors and conflicting response tendencies (Botvinick et al., 2001; Ridderinkhof, Ullsperger, Crone, & Nieuwenhuis, 2004). Monitoring and detection of interference is another important aspect of cognitive control which is also present in the processing of dual-task interference, being necessary for the initiation of interference resolution. While the focus of the present studies was on control mechanism involved in the resolution of interference in the lPFC, the interactions of lPFC and medial PFC might also be of interest for future studies. This might also shed further light on the specific contributions of lPFC regions to the processing of different types of interference on the hand and to the resolution of this interferenceon the other hand. There may be dissociable regions in the lPFC related to the bottleneck itself (Study 1; Dux et al., 2006; Herath et al., 2001; Jiang, 2004) and regions related to the resolution of interference, e.g. the active control of bottleneck processing (Study 2 & 3, Szameitat et al., 2002; 2006), showing different interactions with the medial PFC. Note that effects of task order control in the present studies were always present in the plPFC but also more anteriorly in middle portions of the IFS. It may be that these regions have different functional roles in interference processing in dual task. For example, the inferior frontal junction region in the plPFC was has been consistiently associated with the updating of task representations as one aspect of cognitive control (Brass, Derrfuss, Forstmann, & von Cramon, 2005). However, additional analyses of Study 3 showed that also in mid-IFS regions, the effect of comparable SOA-effects related to dual-task and single-task processing was present as indicated by the whole-brain analysis testing for interaction effects. Thus, activity patterns of these regions related to a prototypical interference manipulation in dual tasks (SOA) do not reveal different functional roles of mid- and posterior lPFC. Future analyses might specifically test the connectivity patterns with medial PFC to elucidate the specific contributions of these lPFC regions to interference processing in dual tasks.


Furthermore, in the parietal cortex, in particular regions along the intraparietal sulcus (IPS) were consistently activated in the comparison of dual tasks and single tasks (Study 1) as well as in the contrasts for task order control (Study 2) and SOA (Study 3). This finding converges with findings of other studies suggesting a role of the parietal cortex in the spatial coordination and control of motor sequences as it is also present in dual-task situations (Andersen, Essick, & Siegel, 1987; Schubert, von Cramon, Niendorf, Pollmann, & Bublak, 1998; see Culham & Kanwisher, 2001 for a comprehensive review). Also, the IPS has been related to the actual implementation of attentional processes that are initiated by the lPFC (Corbetta & Shulman, 2002; Hopfinger et al. 2000). The specific contribution of the superior parietal cortex to efficient dual-task processing might be another topic for future studies.

On a more abstract level, the aPFC has been associated with the weighting and integration of information in the pursuit of higher behavioral goals (Christoff & Garbieli, 2000; Pollmann, 2001; Ramnani & Owen, 2004). A typical paradigm associated with the aPFC is the prospective memory paradigm where participants are required to perform a pre-defined action upon the presentation of a delayed cue presented within a continuous task performance (Burgess, Veitch, de Lacy Costello, & Shallice, 2000; Burgess, Scott, & Frith, 2003). The prospective memory paradigm seems to be related to the dual-task paradigm with two tasks being performed more or less simultaneously. However, no consistent aPFC activity was found in the present studies which might be related to some fundamental differences between the prospective memory paradigm and the PRP paradigm. Most importantly, the information to be responded to in the PRP paradigm is always externally defined – that is perceivable stimuli that require a distinctive motor response in every task trial. Control mechanisms serve the coordination of this externally presented information. In contrast, carrying out an intended action upon the presentation of the prospective memory cue is more internally guided - it requires the consideration of multiple conditions for correct performance. Although there was no supra-threshold activity in the aPFC in the present studies, investigating the transition between the two types of goal-directed multi-tasking might be worth to be further investigated.

Altogether, the present dissertation contributes essentially to the understanding of how dual-task processing is realized by the human brain. Based on the obtained conclusions on the functional role of the lPFC in dual-task processing, further network-oriented approaches might help to elucidate the interplay of the lPFC with medial and anterior PFC and the superior parietal cortex. Besides effective connectivity measures, the high temporal resolution of evoked activity and the neuronal oscillatory synchronization as measured in electro- and magnet encephalography may further extend the gained insights by specifying the temporal dynamics of cognitive control in dual-task processing (see for example Swainson et al., 2003).

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated: