Latencies in BOLD response during visual attention processes

Abstract

One well-investigated division of attentional processes focuses on alerting, orienting and executive control, which can be assessed applying the attentional network test (ANT). The goal of the present study was to add further knowledge about the temporal dynamics of relevant neural correlates. As a right hemispheric dominance for alerting and orienting has previously been reported for intrinsic but not for phasic alertness, we additionally addressed a potential impact of this lateralization of attention by employing a lateralized version of the ANT, capturing phasic alertness processes. Sixteen healthy subjects underwent event-related functional magnetic resonance imaging (fMRI) while performing the ANT. Analyses of BOLD magnitude replicated the engagement of a fronto-parietal network in the attentional subsystems. The amplitudes of the attentional contrasts interacted with visual field presentation in the sense that the thalamus revealed a greater involvement for spatially cued items presented in the left visual field. Comparisons of BOLD latencies in visual cortices, first, verified faster BOLD responses following contra-lateral stimulus presentation. Second and more importantly, we identified attention-modulated activation in secondary visual and anterior cingulate cortices. Results are discussed in terms of bottom-up and lateralization processes. Although intrinsic and phasic alertness are distinct cognitive processes, we propose that neural substrates of intrinsic alertness may be accessed by phasic alertness provided that the attention-dominant (i.e., the right) hemisphere is activated directly by a warning stimulus.

Introduction

Human cognitive performance essentially relies on attention capacities. As early as 1990, Posner and Petersen presented a model on attentional subsystems (Posner and Petersen, 1990) which triggered numerous research projects. Attentional functions were divided into three main subsystems: alerting, orienting and executive control (Posner and Raichle, 1994, Fernandez-Duque and Posner, 2001). Neuroimaging studies revealed that brain areas originally attributed to distinct attentional processes overlap, so that these concepts have been elaborated since. Alertness, for example, needed to process high priority signals, was subdivided into intrinsic and phasic alertness, the former depicting a state of general wakefulness while the latter comprises the ability to increase response readiness for a short period of time following an external stimulus (Sturm and Willmes, 2001, Sturm et al., 1999, Sturm et al., 2006). Finally visual orienting refers to the selection of information from sensory input and executive control denominates the process of resolving conflict among responses.

The central role of the dorsal anterior cingulate gyrus (dACC) and the lateral prefrontal cortex for executive control or cognitive control of conflict were stressed (Fan et al., 2005, Bush et al., 2000, MacDonald, et al., 2000), whereas the dorsolateral prefrontal cortex was shown to be predominantly involved in solving conflicts (MacDonald, et al., 2000, Botvinick et al., 2004). While attentional processes involve bilaterally distributed neural networks, core components were revealed to be asymmetrically represented between the two hemispheres. Neural correlates of intrinsic alertness, for instance, originate in the noradrenergic system (Robinson, 1985, Sturm and Willmes, 2001) and comprise the so-called fronto-parietal network, spanning anterior intraparietal areas, inferior parietal and dorso- and ventrolateral frontal regions (Sturm et al., 1999, Sturm and Willmes, 2001) with a strong dominance for the right hemisphere. Brain regions associated with orienting are the bilateral superior posterior parietal lobes and the left temporo-parietal junction (TPJ), the frontal eye field and the posterior lateral thalamus (Posner and Petersen, 1990). In trials focusing on phasic rather than intrinsic alertness on the other hand, where stimuli are preceded by a warning cue, Sturm and Willmes (2001) observed additional activations of the thalamus, the left superior and left ventrolateral frontal gyri along with a more extended right hemispheric network. Authors interpreted the extended right hemisphere activation as a consequence of the extrinsic activation in response to the cue while the left hemispheric recruitment is supposed to be associated with “elementary attention selectivity” needed to actively inhibit responses to the warning stimulus.

For the present functional magnetic resonance imaging (fMRI) study, we expanded previous approaches focusing on the blood oxygenation level dependent (BOLD) contrast magnitude of attention-related neural networks. We employed a modified version of the widely used Attention Network Test (ANT), a combined cue-target and flanker test. The ANT was developed to investigate attentional subsystems (Fan et al., 2002) and to isolate their underlying neural networks (Fan, et al., 2005). When performing the ANT, subjects have to depict the direction of a target arrow which is surrounded by congruent or incongruent flankers. Since imaging and clinical evidence demonstrate hemispheric asymmetries in the attentional networks, we applied a lateralized version of the ANT (Thienel et al., 2009a, Thienel et al., 2009b), with target arrows either presented in the left or right visual field (LVF and RVF, respectively), to study the contribution of each hemisphere. Using the ANT with its advantage of being able to investigate the different attentional components separately, we focused on two novel aspects in the analysis of attention-related networks: (i) the temporal dynamics of the attentional processes as well as (ii) hemispheric lateralization. The behavioral evidence on the latter issue is inconsistent: While their first experiment resulted in a RVF advantage, Greene et al. (2008) did not find differences between the visual fields in a second experiment. To explore the temporal dynamics of attention-related activation, we extended common fMRI analyses approaches and focused on magnitude as well as the latency of event-related BOLD response. While the literature on BOLD latencies is still scarce but growing (Henson et al., 2002, Ford et al., 2005, Steffener et al., 2010), its investigation holds the strong advantage to reveal additional information about the neural/synaptic activity particularly following brief stimulation (Menon et al., 1998, Miezin et al., 2000).

Given our experimental set up and analysis approach, we were able to test four main hypotheses: (1) Latencies in visual processing: as the presentation of a stimulus in either visual field is known to be first processed in the contra-lateral hemisphere, we predicted a longer delay of the BOLD response in the ipsilateral visual cortex. (2) Latencies reflecting alerting: targets that were preceded by a central cue should lead to a faster BOLD response in the bilateral visual cortex, so that we predicted a larger delay in these brain areas for no-cue conditions compared to the cued condition. (3) Latencies reflecting orienting: targets preceded by a spatial cue rather than a central cue should provoke faster responses in contra-lateral hemisphere within the orienting network (“effect of orienting”). (4) Latencies in executive control (conflict): we anticipated that incongruent trials will lead to a larger delay in the ACC than congruent trials (“effect of executive control”).