Where arousal meets attention: a simultaneous fMRI and EEG recording study

Abstract

In this fMRI study, we looked for the regions supporting interaction between cortical arousal and attention during three conditions: detection, observation, and rest. Arousal measurements were obtained from the EEG low-frequency (LF) power (5–9.5 Hz) recorded continuously together with fMRI. Whatever the condition, arousal was positively correlated with the fMRI signal of the right dorsal-lateral prefrontal and superior parietal cortices, closely overlapping regions involved in the maintenance of attention. Although the inferior temporal areas also presented a correlation with arousal during detection, path analysis suggests that this influence may be indirect, through the top-down influence of the previously mentioned network. However, those visual-processing areas could account for the correlation between arousal and performances. Lastly, the medial frontal cortex, frontal opercula, and thalamus were inversely correlated with arousal but only during detection and observation so that they could account for the control of arousal.

Introduction

Vigilance or sustained attention is the ability to maintain a performance over a period of time. It is classically explored using tasks that demand detection of changes occurring at random intervals (Mackworth, 1957). Vigilance is an important function in everyday life, as its deficit is said to account for accidents (Haraldsson et al., 1990) and is observed in disorders, for example, schizophrenia (Green, 1998), hyperactivity, epilepsy, or closed head injury (Parasuraman et al., 1998). Fluctuations of performances in vigilance tasks are, at least partially, caused by interaction between attention and arousal (Parasuraman et al., 1998). But little is known about the actual relationship between them, and this is what this study attempts to investigate: does arousal modulate the whole cortex? Or does it only affect specific regions? And if so, which component of the attentional process is affected? Indeed, attention involves two orienting systems that permit to direct it and various sites where it influences information processing. The sites where attention biases information processing are those areas devoted to the processing of the type of information attended to, for example, the inferotemporal areas in the case of visual stimuli. Attention can be oriented by two systems: the top-down and the bottom-up networks (Corbetta and Shulman, 2002). The first is the endogenous orienting system that permits to focus voluntarily on specific events. It is a controlled process that relies on a network involving the dorsal-lateral prefrontal cortex [namely, the superior frontal sulcus for spatial items and inferior frontal sulcus for objects (Pardo et al., 1991)] and the parietal cortex (the intraparietal sulcus), subsequently called the top-down network. The second is the exogenous orienting system which permits incoming information to catch attention especially when it is biologically salient, for example, a novel event. It is an automatic process that relies on the inferior frontal gyrus and the temporal-parietal junction principally on the right side (Downar et al., 2000). This network will be subsequently called the bottom-up network. The question is if arousal affects only specific areas, where does it meet attention: on its sites, in the bottom-up or in the top-down network?

To localize these orienting networks, they were dissociated during a rare target detection task: subjects were presented with a succession of visual stimuli including frequent distractors (

) and never-repeated oddballs; half of which were letters and the other half were pictures. The seven participants were presented series of intermingled pictures and letters randomly distributed among frequent distractors. They were requested to detect, in turn, either only letters or only pictures (detection blocks) or to passively look at the series of stimuli (observation blocks). Between the detection and observation blocks, the subjects could relax with their eyes closed (resting blocks). Orienting to exogenous events was supposed to be activated by rare stimuli whatever their type is, that is, letter or picture, and no matter which task the subjects had to perform: press a button to signal the detection of a rare stimulus or ignore it, in the detection blocks or passively look at it in the observation blocks. Accordingly, the regions commonly activated by these six conditions (two types of stimuli × three tasks) or, in other words, their conjunction (Price and Friston, 1997) ought to correspond to the bottom-up network. The paradigm was planned in such a way that the predicted event-related and block-related BOLD responses were roughly decorrelated Chawla et al., 1999, Donaldson et al., 2001, Otten et al., 2002, Visscher et al., 2003. This permitted to identify periods of endogenous maintenance of attention without oddball presentation. Consequently, the conjunction of letter detection and picture detection relative to the observation blocks should activate the top-down network.

To address the issue of arousal measurements, previous works relied either on EEG Danos et al., 2001, Lindgren et al., 1999, self-rating, or pharmacological manipulation Coull et al., 1997, Portas et al., 1998. Experiments in which EEG was used as a means to measure cortical arousal relied on the inverse correlation of arousal with the power of low EEG frequencies during awakening. Although some studies focused on the alpha band as the marker of arousal (Danos et al., 2001) (8 to 12 Hz), the 5- to 9.5-Hz band seems to be the most sensitive: it increases with the first sleep stage (Niedermeyer, 1999a), it increases after sleep deprivation (Aeschbach et al., 2001), it is proportional to subjects' self-rating of arousal or sleepiness Lafrance and Dumont, 2000, Strijkstra et al., 2003, and it is correlated with poorer performances Daniel, 1967, Makeig and Jung, 1996 and especially with their degradation as a function of time, that is, the vigilance decrement Horvath et al., 1976, Makeig and Inlow, 1993. EEG can provide an evaluation of arousal with an accuracy of 1 s, which is about the time scale of fMRI sensitivity. In previous studies, arousal was defined as a constant value over the scanning time defining it as a global or average level. This concept, further referred to as “arousal level,” is generally manipulated by means of sleep deprivation or drugs. However, there are indications that faster fluctuations may be even more relevant Doran et al., 2001, Makeig et al., 2000. These fluctuations are spontaneous and probably related to dynamic interaction between different components of the arousal system. Indeed, self-rating of the arousal level does not correlate as well with the mean low-frequency (LF) power, as it does with the standard deviation (SD) (i.e., amplitude of power fluctuations) (Oken et al., 1995). In other words, a low arousal level is better described as larger arousal fluctuations than as an average decrease in arousal Daniel, 1967, Doran et al., 2001, Makeig et al., 2000. Combining EEG with fMRI permits to correlate short-term changes in cortical arousal with BOLD signal variations.

Besides these endogenous fluctuations, arousal systems, such as the noradrenergic locus coeruleus and the cholinergic nucleus of the basal forebrain, also present a transient increase in activity following the presentation of a rare stimulus Aston-Jones et al., 1999, Rajkowski et al., 1994. Such exogenous stimulation of arousal goes together with a decrease in EEG low-frequency power, that is, the “event-related desynchronization” (Pfurtscheller and Aranibar, 1977), whose amplitude depends on the physical features and novelty of the stimulus (Aston-Jones et al., 1999). One might speculate that endogenous arousal and arousal ensuing exogenous stimulation modulate different regions (Sturm and Willmes, 2001), for example, depending on the activity of these regions. Accordingly, exogenous and endogenous arousal must be dissociated: the EEG low-frequency power corresponding to the event-related desynchronization was analyzed separately from the parts corresponding to the endogenous fluctuations of arousal during the blocks.