45693$ $$$8 Hogrefe & Huber Publishers Ð EXPPSY 52/01/2005 Ð 1. Bel. Ð 10-09-04 11:50:34 Ð Rev 16.04x Components of Attentional Set-switching M. F. S. Rushworth, R. E. Passingham, and A. C. Nobre Department of Experimental Psychology, University of Oxford, England Abstract. A series of distinct event-related potentials (ERPs) have been recorded from the scalp of human subjects as they switch from one task to another. It is possible that task switching may depend on different mechanisms depending on whether the switch requires a change in attentional set, in other words the redirecting of attention to different aspects of a sensory stimulus, or whether it requires a change in intentional set, in others words a change in the way that responses are selected. To address this issue, the current study recorded ERPs while subjects switched between attentional sets and the results were compared with those of a previous investigation in which subjects switched between intentional sets. Subjects selected stimuli according to two conflicting attentional sets, each emphasizing one visual stimulus dimension (colour, shape). Pairs of stimuli, only one of which was to be attended, were presented for between eight and seventeen trials then either a switch or a stay cue was shown. The switch cue instructed subjects to switch from the current attentional set to the other set, while the stay cue instructed subjects to maintain the current set. Comparing ERPs time-locked to the switch and stay cues revealed neural correlates of the initiation of a task switch. Comparing the ERPs time locked to the first stimuli after either stay or switch cues identified neural correlates of the implementation of a task switch. A similar modulation over parietal electrodes was seen when subjects were switching between either attentional or intentional sets. While an intentional set switch began with a medial frontal modulation, attentional set switching began with a lateral frontal modula- tion. Implementing a new attentional set was associated with modulation of relatively early visual potentials, while imple- menting a new intentional set was associated with modulation of later response-related potentials. The results confirm that task switching consists of a number of constituent processes which may be taxed to different degrees depending on whether a task-switch paradigm requires subjects to change the way in which they select stimuli or responses. Key words: attention, intention, prefrontal cortex, task switching, event-related potential (ERP) Introduction Neuropsychologists have emphasized the successful performance of prefrontal patients on a wide variety of cognitive tasks, but commented on the difficulty they have in managing and switching between dif- ferent types of tasks (Burgess, Veitch, de Lacy Cos- tello, & Shallice, 2000; Shallice & Burgess, 1991). Lateral prefrontal lesions in monkeys cause difficul- ties in switching between tasks that require attention to be allocated to different visual aspects of stimuli (Dias, Robbins, & Roberts, 1996; Passingham, 1972). In order to explain such results, it has been argued that executive processes associated with the prefrontal cortex normally select the currently rele- We gratefully acknowledge the assistance of Carlo Miniussi, Anling Rao, Ed Wilding, Ivan Griffin & Heather Jordan. Supported by the Royal Society and the Wellcome Trust. DOI: 10.1027/1618-3169.52.1.x ” 2005 Hogrefe & Huber Publishers Experimental Psychology 2005; Vol. 52(1): 1Ð16 vant task set so that appropriate actions are made and relevant stimuli are attended. The cognitive mechanisms of task switching have recently been subject to considerable scrutiny and, as a result, the degree to which task switching reflects an active executive processes has been questioned. On the one hand, a prospective process of “task-set reconfiguration” prior to task performance has been invoked because the cost of switching, as measured by RT increase, is lower when subjects have greater opportunity for advanced preparation (Rogers & Monsell, 1995). Other authors, however, have em- phasized the degree to which switching costs are just the consequence of the persisting activation of pro- cesses related to the first task (Allport, Styles, & Hsieh, 1994; Allport & Wylie, 1999; Wylie & All- port, 2000). According to this theory, advance prepa- ration reduces switching costs because it allows time for interfering “task set inertia” from the previous task to dissipate. Moreover, regardless of their posi- tion in this debate most researchers are agree that