A major limitation in human information processing concerns the ability of the brain to process two temporally close meaningful stimuli. This limitation is illustrated by the attentional-blink paradigm: When the second of two target stimuli (T1 and T2) follows the first one in close temporal proximity (within 500 msec) in a rapid stream of distracter events, it often goes unnoticed (Ward, Duncan, & Shapiro, 1996
; Raymond, Shapiro, & Arnell, 1992
). Many theoretical accounts of this so-called attentional-blink deficit have relied on some form of limited attentional resource that is allocated to the leading target at the expense of the trailing target (Jolicoeur & Dell’Acqua, 1998
; Ward et al., 1996
; Chun & Potter, 1995
; Shapiro, Raymond, & Amell, 1994
). For instance, it has been postulated that when many resources are devoted to T1 processing, too few may be available for subsequent T2 processing, rendering its representation vulnerable to distracter interference (Shapiro et al., 1994
). In line with this idea, several recent event-related potential (ERP) studies (Kranczioch, Debener, Maye, & Engel, 2007
; Sessa, Luria, Verleger, & Dell’acqua, 2007
; Slagter et al., 2007
; Martens, Elmallah, London, & Johnson, 2006
; Martens, Munneke, Smid, & Johnson, 2006
; Shapiro, Schmitz, Martens, Hommel, & Schnitzler, 2006
; Sergent, Baillet, & Dehaene, 2005
) have shown that the ability to accurately identify T2 is related to the latency and/or amplitude of the T1-elicited P3b, a brain-potential index of resource allocation (Wickens, Kramer, Vanasse, & Donchin, 1983
). For example, a delayed or larger T1-evoked P3b has been observed in trials in which T2 was missed (i.e., blink trials) versus detected (i.e., no-blink trials) (Sergent et al., 2005
). These electrophysiological findings support the notion that the amount of resources invested in T1 processing markedly influences the processing capacity available for T2, and thus, whether T2 will be consciously perceived or not.
The attentional blink does not represent an absolute bottleneck in information processing. This is most clearly indicated by the fact that most participants are able to identify both targets on at least a portion of the trials (e.g., Martens, Munneke, et al., 2006
). At present, it is unclear why physically identical information sometimes does and sometimes does not reach awareness. It is notable in this respect that the attentional blink has been observed to get smaller when measures are taken that prevent an overinvestment of attentional resources in stimulus processing. For example, T2 detection has been found to benefit from a diffusion of attention (Arend, Johnston, & Shapiro, 2006
; Olivers & Nieuwenhuis, 2005
). Moreover, we recently found that 3 months of intensive training in a style of meditation, which allegedly reduces elaborate object processing (Lutz, Slagter, Dunne, & Davidson, 2008
), reduced brain resource allocation to T1, as indexed by a smaller T1-elicited P3b, and improved T2 detection, with no impairment in T1 detection (Slagter et al., 2007
). This common style of open monitoring (OM) meditation, also known as Vipassana meditation, consists in being attentive moment by moment to anything that occurs in experience, whether it be a sensation, thought, or feeling (Lutz, Slagter, et al., 2008
). Usually, one starts by focusing or stabilizing concentration on an object such as the breath. Then one broadens one’s focus, cultivating a nonreactive form of awareness. This form of awareness is nonreactive in the sense that, ideally, one does not become caught up in judgments and affective responses about sensory or mental stimuli. Although initially the practitioner frequently “clings” to objects in a way that takes up resources available to process information related to current experience, eventually, a trait is thought to emerge such that one can sustain the “non-clinging” state in which one is attentive to the content of experience from moment to moment. As participants were not engaged in formal meditation during task performance, our observations that intensive OM meditation reduced T1 capture and improved T2 detection are in line with the idea that one long-term effect of this style of meditation is a reduction in the propensity to “get stuck” on an object. Together with previous behavioral findings (Olivers & Nieuwenhuis, 2005
), these data support the idea that the attentional blink is due to an overinvestment of attentional resources in stimulus processing, and that this suboptimal processing mode can be counteracted by manipulations promoting a less object-focused state of attention.
Recent electroencephalography (EEG) work has shown that the attentional state of the observer is reflected in ongoing or baseline neural activity prior to stimulus presentation, and can predict subsequent visual perception performance (e.g., Hanslmayr et al., 2007
; Thut, Nietzel, Brandt, & Pascual-Leone, 2006
; Linkenkaer-Hansen, Nikulin, Palva, Ilmoniemi, & Palva, 2004
). To our knowledge, only one previous attentional-blink study has examined the relationship between baseline neural activity and T2 perception (Martens, Munneke, et al., 2006
). In this study, mean EEG activity in a 1024-msec fixation period before task onset was compared between a group of participants with a relatively large attentional blink (“blinkers”) and a group of participants with a relatively small attentional blink (“nonblinkers”). No differences between groups in mean baseline EEG activity were observed, but the nonblinkers showed an earlier T1-elicited P3b than the blinkers. These findings might be taken as evidence that the state of the system right before task onset does not strongly determine the ability of the system to process T1 efficiently, thus arguing against the idea that the observer’s mental state influences conscious target perception. However, an across-subjects comparison, as was used in the Martens, Munneke, et al. (2006)
study, may have obscured small differences in baseline mental state related to successful T2 detection. Furthermore, Martens et al. only looked at the amplitude (or power), not phase, of oscillatory activity, whereas phase synchrony might provide a more sensitive measure of attentional state (see, for example, Hanslmayr et al., 2007
). This possibility is substantiated by findings from other EEG studies, which have reported within-subject differences in synchronous EEG activity between blink and no-blink trials in the alpha (Kranczioch et al., 2007
), beta (Gross et al., 2004
), and gamma (Nakatani, Ito, Nikolaev, Gong, & van Leeuwen, 2005
) frequency ranges, not only after T1 presentation but also right before and during distracter presentation. This variation of findings showing differential processing of the distracter stimuli points to differences in attentional state or the way attention is directed prior to the presentation of T1. It also highlights an important role for phase synchrony in the attentional blink, in line with recent work indicating that phase synchrony is essential in the formation of transient neuronal assemblies that constitute the meta-representations required by sensory awareness (Singer, 1999
) and in the communication therein (Fries, 2005
). Of particular importance, some of this work has associated sensory awareness with increased consistency across trials in the phase of oscillatory EEG activity in the delta (e.g., Lakatos, Karmos, Mehta, Ulbert, & Schroeder, 2008
) and alpha (e.g., Palva, Linkenkaer-Hansen, Naäätänen, & Palva, 2005
) frequency bands, as indexed by the phase-locking factor (PLF; Tallon-Baudry & Bertrand, 1999
). Phase consistency may thus provide a particularly useful measure in the study of conscious target perception.
The aim of the current study was to further examine the impact of intensive OM training on brain function and behavior, and investigate why physically identical information sometimes does and sometimes does not reach awareness. To this end, we evaluated the time course and correlation of neuronal oscillations associated with T2 perception by analyzing unreported aspects of EEG data from our previous study (Slagter et al., 2007
). Specifically, the current study tested two hypotheses. The first hypothesis concerned the effects of the previously observed reduction in brain resource allocation to T1 on the overall readiness of the system to be disturbed by a new target stimulus (i.e., T2). The second hypothesis was based on the idea that attentional processing of T1 can be predicted by the baseline mental state of the observer. These two hypotheses and corresponding predictions are described in detail below.
As mentioned above, training in OM meditation is thought to leave the system more open to, or ready to process, whatever thoughts, feelings, or sensations may arise, without grasping to an explicit object. Based on this description, we hypothesized that not only will such training reduce elaborate T1 processing, as we previously observed (Slagter et al., 2007
), but it will also render the system more susceptible to T2, as resources should be more consistently available to process new information. As mentioned above, it is possible to reveal whether the recorded EEG signals at a given latency show a consistent or nonrandom phase relationship to a presented stimulus across trials using the PLF (Tallon-Baudry & Bertrand, 1999
). Using this measure of stimulus-locked phase variability, we examined whether intensive training in OM meditation decreased trial-to-trial variability in the recruitment of processes related to the conscious perception of T2. We also directly investigated the relationship between the previously observed reduction in brain-resource allocation to T1 (as indexed by T1-elicited P3b amplitude) and the ability of the system to be perturbated by a new target stimulus (i.e., T2). Our specific prediction was that those individuals who showed the greatest reduction in brain-resource allocation to T1 would show the greatest increase in T2-locked phase consistency. In addition to effects of intensive mental training on phase variability, we also explored the possibility that such training affected the amplitude of neuronal oscillations.
To address the second hypothesis, the current study examined whether the observed reduction in brain-resource allocation to T1 was associated with any changes in baseline mental state. As mentioned above, the mental state of the observer has been found to influence the attentional blink, with a less object-focused state of attention promoting a smaller attentional blink, presumably by reducing the propensity to overinvest resources in T1 processing (Olivers & Nieuwenhuis, 2005
). Our second hypothesis therefore was that intensive mental training would be associated with changes in baseline mental state, as reflected by the state of oscillatory brain activity before task onset. We predicted that such changes would be predictive of target perception. In contrast to the Martens, Munneke, et al. (2006)
study, the relationship between baseline mental state and conscious perception was interrogated within the same
individual, and effects of intensive mental training on the amplitude as well as phase of baseline oscillatory activity were examined.
Data were collected from 17 participants at the beginning and end of a 3-month meditation retreat during which they meditated for 10 to 12 hr/day (practitioner group). Control data were collected from 23 matched participants interested in learning about meditation (novice group), who received a 1-hr meditation class and were asked to meditate for 20 min daily for 1 week prior to each session. In each session, participants performed an attentional-blink task in which they had to identify two targets (both numbers) embedded in a rapid stream of distracter letters () while their EEG was recorded. T2 could follow T1 either within (334 msec; short-interval trial) or outside (672 msec; long-interval trial) the time window of the attentional blink. Participants were not engaged in formal meditation during task performance.
Figure 1 Attentional-blink task. On every trial, between 15 and 19 items were presented at the center of the screen, preceded by a 1780-msec fixation cross. Most of the items were letters, presented for 50 msec each and followed by a 34-msec blank. On T2-present (more ...)