REVIEWS .9 一FEF一Ps一TPJ 一P一TP (FEF)an IP).te ng the to 0ne rast to the ICURE s (middle)shows that the TPJ is not active du tion of e of al net ore la ali to the right hemisphe ulaepresenited nde d locations a right ventral fr ed on the res There is al hat was activ d in anothe or durin ng delay.which indi that thi ttentio u he aednpetialreoitent un was NATURE REVIEWSNATURE REVIEWS | NEUROSCIENCE VOLUME 3 | MARCH 2002 | 209 REVIEWS FIGURE 5c (middle) shows that the TPJ is not active dur￾ing search for a motion target (it is actually deactivated). However, it is strongly activated, predominantly in the right hemishere10,21, by target detection (FIG. 5b, regions labelled in blue). When the targets occur at an unexpected location, the activity in this network is further enhanced and even more lateralized to the right hemisphere. FIGURE 6a shows the cortical regions that are most active when stimuli are presented at unattended locations and sub￾jects reorient towards them. They include areas that are centred on the right supramarginal and superior tem￾poral gyrus (or TPJ), and on the inferior frontal gyrus (IFg). There is also activation in the right IPs and FEF10. FIGURE 6b shows a similar right-hemisphere network that was activated in another experiment that involved reorienting of attention to unattended (infrequently stimulated) locations79. An initial hypothesis about this network was that it is involved in spatial reorienting to an unattended loca￾tion. However, it is now clear that activation of the right ventral frontoparietal network is independent of strong right-hemisphere dominance, in contrast to the more bilateral organization of the IPs–FEF system, has important clinical implications for the pathophysiology of unilateral spatial NEGLECT, a common neuropsychologi￾cal syndrome that occurs after injury to the right hemi￾sphere. The existence of a ventral frontoparietal network is supported by a series of recent brain-imaging studies. First, we discuss some of its functional properties, before considering some hypotheses about its role in the control of visual attention. In contrast to the dorsal frontoparietal network, the right ventral frontoparietal network is not engaged by cues that carry advance information about forthcoming stimuli. For example, FIG. 5c (left) contrasts the responses of the IPs, FEF and TPJ to the presentation of a cue that indicates the likely direction of motion of a subsequent stimulus. There is little activation in the TPJ for the cue or during the ensuing delay, which indicates that this network is not activated by generating or maintaining an attentional set. Moreover, again unlike the dorsal fronto￾parietal network, these regions are not activated by the application of this set during stimulus processing. NEGLECT A neurological syndrome (often involving damage to the right parietal cortex) in which patients show a marked deficit in the ability to detect or respond to information in the contralesional field. FEF IPs TPJ IPs TPJ Misses Hits L IPs Detection (H > M) 0 100 Saccades Saccade Covert 200 0 100 200 Distracter Target “ ” Expectation (~4.6 s) Search + detection (~2.3 s) a Search/detect motion target b c Attention to motion direction Detection Search Termination of search Search + detection (~4.6 s) FEF PFC IPs TPJ MT+ IFg Figure 5 | Dorsal frontoparietal network and salience. a | The frontal eye field (FEF) and salience maps (reproduced with permission from REF. 118 © 2001 Massachusetts Institute of Technology). Top: neuronal response in macaque FEF during the detection of oddball targets or distracters in a visual search paradigm. Target detection was signalled by an eye movement to the target (left column) or was covert (right column). Within 120 ms, the single-unit response differentiated between target and distracters. Bottom: FEF as a salience map in which targets defined on the basis of different visual attributes can be selected. b | Human brain activity in the dorsal frontoparietal network during search and detection. Left: subjects see an arrow that cues a direction of motion and then search for (and detect) a moving target in a dynamic noise display. Right: functional magnetic resonance imaging (fMRI) map during search and detection. Areas in red are involved in searching for the target; areas in blue are recruited at or after detection. IFg, inferior frontal gyrus; IPs, intraparietal sulcus; MT+, middle temporal complex; PFC, prefrontal cortex; TPJ, temporoparietal junction. c | Time course of fMRI signals during cueing, search and detection (for the paradigm shown in b). Left: the IPs and FEF respond during the cue period as subjects establish a set for motion, but no activity is observed in the TPJ. Middle: the IPs is active from the onset of the search display, but the TPJ is recruited only when the target is detected. Right: the IPs response to hits (target was detected, H) and misses (target was not detected, M) during search and detection. The signal is initially enhanced on hit trials relative to miss trials, reflecting target detection, but it then falls off, reflecting the termination of search after detection. L IPs, left IPs