The dorsal convexity of the human frontal and parietal lobes forms a network that is crucially involved in the selection of sensory contents by attention. This network comprehends cortex along the intraparietal sulcus, the inferior parietal lobe, and dorsal premotor cortex, including the frontal eye field. These regions are richly interconnected with recurrent fibers passing through the superior longitudinal fasciculus. The posterior parietal cortex has several functional characteristics—such as feature-independent coding, enhancement of activity by attention, representation of task-related signals, and access to multiple reference frames—that point to a central role of this region in the computation of a feature- and modality-independent priority map of the environment. The priority map integrates feature information elaborated in sensory cortex and top-down representations of behavioral goals and expectations originating in the dorsolateral prefrontal and premotor cortex. This review presents converging evidence from single-unit studies of the primate brain, functional neuroimaging, and investigations of neuropsychological disorders such as Bálint syndrome and spatial neglect for a decisive role of the frontoparietal attention network in the selection of relevant environmental information.

Background: In previous studies, the authors found that patients with spontaneous confabulation differ from those with nonconfabulating amnesia by 1) temporal context confusion (TCC) in memory based on an inability to suppress intrusions of currently irrelevant memory traces into ongoing thinking; and 2) lesions involving the orbitofrontal cortex, basal forebrain, or amygdala and perirhinal cortex. Objectives: To study the long-term clinical course of spontaneous confabulations, determine whether TCC in memory also parallels the clinical course of spontaneous confabulations, and study the impact of lesion site on clinical course. Methods: Eight patients with spontaneous confabulation were re-examined 18 months after onset. Tests of memory and executive functioning and measurement of TCC in memory were again applied. MRI according to a standard protocol was performed to determine areas of permanent damage. Results: Seven patients eventually stopped confabulating. TCC, but not common memory or executive tests, precisely paralleled the course of spontaneous confabulations. Patients with isolated, less extensive, orbitofrontal lesions stopped confabulating first and had the best neuropsychological outcome. Patients with basal forebrain lesions continued to confabulate for several months and remained amnesic. One patient with extensive orbitofrontal damage and perirhinal cortex damage continues to confabulate after more than 3 years, continuing to confuse memory traces. Conclusions: Temporal context confusion in memory is not only the sole feature reliably separating patients with spontaneous confabulation from those with nonconfabulating amnesia in the acute stage, it is also the only feature that precisely parallels the clinical course of spontaneous confabulations. Most patients eventually stop confabulating but duration of confabulations depends on the lesion site.

Experimental neurophysiology and functional neuroimaging have identified a dorsal attention network that encodes neural signals related to the behavioral significance of a stimulus. The core anatomical areas of this network are the frontal eye fields and the posterior parietal cortex, which are interconnected by the superior longitudinal fasciculus. Here, we show that damage or disconnection of this network predicts the extent to which task-relevant stimuli capture attention of human stroke patients with spatial neglect. Healthy volunteers, right-hemisphere-damaged control participants, and patients with left neglect reacted to peripheral targets defined by their color, which were preceded by a brief distracter stimulus. The position of the distracter and its relevance for the current trial were independently varied. In neglect patients with damage including the frontal eye fields and the superior longitudinal fasciculus, ipsilesional distracters impaired orienting into contralesional space regardless of their relevance for the current task. In contrast, patients with sparing of these regions were only impaired when distracters were task-relevant. These findings indicate that the dorsal attention network controls spatial orienting by modulating the saliency of distracter stimuli according to current action goals.