Elsevier

Brain and Cognition

Volume 68, Issue 1, October 2008, Pages 42-48
Brain and Cognition

The role of the magnocellular and parvocellular pathways in the attentional blink

https://doi.org/10.1016/j.bandc.2008.02.119Get rights and content

Abstract

The attentional blink refers to the transient impairment in perceiving the 2nd of two targets presented in close temporal proximity in a rapid serial visual presentation (RSVP) stream. The purpose of this study was to examine the effect on human attentional-blink performance of disrupting the function of the magnocellular pathway—a major visual-processing pathway specialized in temporal segregation. The study was motivated by recent theories that relate the attentional blink to the limited temporal resolution of attentional responses, and by a number of poorly understood empirical findings, including the effects on the attentional blink of luminance adaptation and distraction. The attentional blink was assessed for stimuli on a red background (Experiment 1), stimuli on an equiluminant background (Experiment 2), and following flicker or motion adaptation (Experiment 3), three psychophysical manipulations known to disrupt magnocellular function. Contrary to our expectations, the attentional blink was not affected by these manipulations, suggesting no specific relationship between the attentional blink and magnocellular and/or parvocellular processing.

Introduction

An important question addressed by cognitive psychologists is how much time our cognitive system needs to turn relevant perceptual information into a representation that can be remembered or acted upon, before the system is available again for the next piece of information. Much of the research addressing this question has used the attentional-blink paradigm. In the standard version of this paradigm, participants are asked to report two targets that are embedded in a rapid serial visual presentation (RSVP) stream of distractor stimuli. All items are presented in the same location at a rate of about 10 items per second. Participants usually have no difficulty with reporting the first target (T1). However, if the second of the two targets (T2) is presented at temporal positions within about 500 ms of T1, report of T2 is considerably impaired—a phenomenon that is referred to as the attentional blink (Raymond, Shapiro, & Arnell, 1992). The attentional blink suggests that under the perceptually demanding conditions imposed by an RSVP stream, our cognitive system is rather limited in the rate at which durable representations of distinct perceptual stimuli can be formed.

One class of theories that have attempted to explain the mechanism underlying the attentional blink has emphasized the limited temporal resolution of our attentional system (Bowman and Wyble, 2007, Nieuwenstein et al., 2005, Olivers et al., 2007, Raymond et al., 1992). For example, Olivers (2007) has proposed that the attentional responses to RSVP stimuli are sluggish, generally lagging behind the stimuli that elicit them. As a result, they bias the processing not (just) of the eliciting stimulus, but also of the subsequent stimulus: Stimuli following targets receive a high attentional weight, and stimuli following distractors receive a low attentional weight. If T1 is followed by a distractor, the processing of the distractor is potentiated, which may in turn lead to a strong suppressive response that, if sufficiently long-lasting, may cause an attentional blink for subsequent stimuli (Raymond et al., 1992). This type of account can also explain various other results, including the finding that T2 performance is typically spared if T2 immediately follows T1 (Nieuwenhuis et al., 2005, Raymond et al., 1992), or if T2 is immediately preceded by a third target or a distractor sharing the target-defining property (Di Lollo et al., 2005, Nieuwenstein et al., 2005, Olivers et al., 2007).

The goal of the present research was to investigate whether the limited temporal resolution of the attentional system, as expressed in the attentional blink, might be mediated by some of the main characteristics of the visual system. Physiological and anatomical studies have revealed that the primate visual system consists of several parallel information-processing channels. The two major channels, the parvocellular and magnocellular pathways, originate in the retina, have distinct projections to the lateral geniculate nucleus, and remain in part segregated in cortical visual areas. The cells in the two pathways have very different physiological and functional properties (for reviews see Livingstone and Hubel, 1988, Schiller and Logothetis, 1990): Parvo cells respond in a relatively slow and sustained manner to visual stimulation, have a small receptive field, and are specialized in analyzing the color, shape, and other static surface properties of objects. Therefore, facilitation of the parvocellular pathway leads to increased spatial segregation (due to the small receptive field size) and increased temporal integration (due to the sluggish response profile). In contrast, magno cells respond much faster and more transiently, have larger receptive fields, and are specialized in analyzing movement and low-frequency information. As a result, facilitation of the magnocellular pathway promotes spatial integration and temporal segregation, an influence that opposes and complements that of the parvocellular pathway. In line with this physiological interaction, psychophysical tests have demonstrated that experimental manipulations that induce focused spatial attention cause a concurrent decrement in temporal resolution (Yeshurun & Levy, 2003).

We hypothesized that the usual instruction to identify targets within an RSVP stream places the cognitive system in a processing mode that shifts the relative contribution of the parvocellular and magnocellular pathways toward the former. However, although parvo cells are presumably more sensitive to the detailed features of the stimuli typically used in attentional-blink tasks, their prolonged neural response periods render them unsuitable for segregating the individual stimuli. That is, when two stimuli are separated by a brief interval, the corresponding neural responses are likely to be integrated over time, resulting in decreased temporal resolution. In contrast, magno cells respond instantly and vigorously to the type of luminance flicker presented by an RSVP stream, leading to distinct neural representations for each of the individual stimuli. Together, this suggests that performance in the attentional-blink task might benefit from experimental manipulations that favour a more dominant contribution of the magnocellular pathway to the processing of the RSVP stream. In contrast, performance might be impaired by manipulations that disrupt magnocellular function.

There are several lines of (indirect) evidence for this hypothesis. Perhaps the most supportive evidence concerns the effect of luminance adaptation on the attentional blink. Giesbrecht and colleagues compared performance on an attentional-blink task after 40 min of dark adaptation (scotopic viewing condition) and after 40 min of adaptation to ambient light (photopic viewing condition; Giesbrecht, Bischof, & Kingstone, 2004). An attentional blink was observed only in the photopic viewing condition. Interestingly, physiological and psychophysical experiments have demonstrated that under scotopic viewing conditions, visual processing is dominated by the magnocellular pathway (Benedek et al., 2003, Purpura et al., 1988), presumably because there is strong rod input to the magnocellular pathway, but negligible rod input to the parvocellular pathway. Rods are retinal cells that form the primary source of information under scotopic viewing conditions. According to our hypothesis, the dominant contribution of the magnocellular pathway under scotopic viewing conditions is consistent with the absence of an attentional blink under such conditions.

The parvo/magno hypothesis would also offer an intriguing explanation of the counter-intuitive finding that the attentional blink is ameliorated by manipulations that promote divided visual attention. For example, Olivers and Nieuwenhuis (2006) found that the attentional blink is smaller following explicit instructions to participants to “concentrate a little less”, and to “pay a little less attention” to the RSVP stream. Similarly, the attentional-blink magnitude is much reduced if T2 is presented at an unattended location quite far away from T1 (Kristjánsson & Nakayama, 2002). These and other distraction manipulations (Arend, Johnston, & Shapiro, 2006) may be assumed to reduce the attentional focus on the RSVP stream. Importantly, as discussed above, there is evidence indicating that increases in focused spatial attention cause a concurrent decrement in temporal resolution, possibly as a result of the mutual trade-off between spatial and temporal sensitivity of the parvo- and magnocellular pathways (Yeshurun, 2004, Yeshurun and Levy, 2003). This raises the possibility that the reduction in attentional-blink magnitude under distracting conditions reflects the increased magnocellular involvement and accompanying increase in temporal resolution associated with a reduction in focused spatial attention.

There are various other striking similarities between the attentional blink and properties of the magnocellular system. For example, magno cells show maximal sensitivity to flicker at around 10 Hz (Lee, Martin, & Valberg, 1989), which is more or less the same frequency as the RSVP stream in most attentional-blink experiments. Furthermore, backward masking of the targets by subsequent RSVP items is crucial for the occurrence of an attentional blink (Brehaut, Enns, & Di Lollo, 1999). This is consistent with the observation of an increased effect of backward masking under task conditions that attenuate magnocellular activity (Okubo & Nicholls, 2005). Finally, there is some evidence that the magnitude of the attentional blink varies with stimulus size and with the requirement to identify either the global aspects or the local details of RSVP stimuli (Lawson et al., 2002). Although the exact influence of these factors, and in particular their interaction, requires further investigation, the broad pattern of results appears consistent with the well-documented role of the magnocellular pathway in processing low spatial frequencies and global aspects of a scene (Breitmeyer and Breier, 1994, Chikashi et al., 1999).

Here we report three experiments that were designed to test the possible influence of the relative contribution of parvo- and magnocellular activity on the attentional blink. Each of the experiments capitalized on the differential sensitivity of the magno- and parvocellular pathways, by using psychophysical manipulations shown to be successful in previous behavioral and neurophysiological research.

A subset of cells in the magnocellular pathway is inhibited by red diffuse light, due to an inhibitory surround that is selectively sensitive to long wavelengths. Therefore, the use of a red background disrupts performance on tasks requiring high temporal resolution (e.g., Breitmeyer and Williams, 1990, Wiesel and Hubel, 1966). In Experiment 1, we exploited this property to manipulate the relative involvement of the parvo- and magnocellular pathways in an attentional-blink task. On each trial, participants were required to identify the two digits that were embedded in an RSVP stream of letter distractors. In one condition the stimuli were presented against a red background, and in another condition they were presented against a green background. The conditions were identical in terms of the luminance contrast between stimuli and background, and in terms of general luminance. Our prediction was that the magnitude of the attentional blink would be larger with a red than with a green background, because the red light was assumed to weaken magnocellular involvement in processing the RSVP stream.

In Experiment 2, we made use of another property of the magno- and parvocellular pathways: Stimuli with low luminance contrast and equal color to the background preferentially activate the magnocellular pathway, whereas stimuli that are equiluminant but of a different color than the background preferentially activate the parvocellular pathway (Kaplan and Shapley, 1986, Schiller and Logothetis, 1990, Steinman et al., 1997). Accordingly, we compared the attentional blink under conditions of (low) luminance contrast vs. color contrast. On the basis of pilot work, we selected a set of colors such that in a stationary setting, the stimuli were perceived as slightly better visible in the color-contrast condition than in the luminance-contrast condition (cf. Omtzigt, Hendriks, & Kolk, 2002). Therefore, if the attentional blink would be more pronounced in the color-contrast condition, this finding could be unambiguously ascribed to a weaker involvement of the magnocellular pathway.

Finally, in Experiment 3, we used extended periods of flicker adaptation or motion adaptation as a way of fatiguing the magno cells and disrupting the function of the magnocellular pathway (Clifford and Wenderoth, 1999, Green, 1981, Pantle, 1971). Each block of RSVP trials was preceded by a 2-min period of adaptation to flicker or motion, using parameters shown to be successful in previous research (Green, 1981). The control condition began with 2 min of adaptation to a stationary gray field. In all three task conditions, a re-adaptation period of 2 s was presented between the RSVP trials (Chapman, Hoag, & Giaschi, 2004). Our prediction was that the attentional blink would increase in size following flicker and motion adaptation, due to disruption of magnocellular function.

To foreshadow the results, in none of the three experiments our predictions were confirmed: The attentional blink was not systematically affected by manipulations that changed the relative involvement of the magno- and parvocellular pathways.

Section snippets

Participants

Sixteen students (12 female, age 18–27 years) from Leiden University participated in the experiment in return for €6, or course credit.

Stimuli, design, and procedure

Each trial started with a 0.4 × 0.4° fixation cross, presented for 1000 ms in the center of the display. Subsequently, the fixation cross was replaced by an RSVP stream of 21 uppercase letters, each measuring approximately 0.5 × 0.5°. Each letter was randomly drawn (without replacement) from the alphabet and presented for 74 ms, followed by a 26-ms blank interval.

Participants

Seventeen students (15 female, ages 18–24 years) from Leiden University participated in the experiment in return for €6, or course credit.

Stimuli, design, and procedure

All details were as in Experiment 1, except for the following. The RSVP items were presented for 74 ms and were separated by a blank interval of 46 ms, resulting in a 120-ms item onset asynchrony. All stimuli were presented against a bright yellow background (RGB = 199, 199, 0; CIE x-, y-coordinates = .401, .519; luminance = 74.2 cd/m2). In the luminance-contrast

Participants

Twelve students (nine female, ages 18–24 years) from Leiden University participated in the experiment in return for €10, or course credit.

Stimuli, design, and procedure

All details were as in Experiment 1, except for the following. The stimuli were presented in black against a gray background (RGB = 128, 128, 128; CIE x-, y-coordinates = .279, .304; luminance = 34.2 cd/m2). To make the task more difficult, the blank interval between RSVP items was shortened to 19 ms, resulting in a 93-ms item onset asynchrony, and no trial-to-trial

General discussion

We tested the prediction that experimental manipulations that disrupt magnocellular processing and bias parvocellular processing should affect attentional-blink magnitude. This prediction was based on the notion that the neural response profile of magno cells (i.e., short-latency, transient) is more suited for resolving the challenges posed by an RSVP stream than the response profile of parvo cells (relatively slow and sustained). We suggested that our hypothesis is consistent with, and

Acknowledgments

This research was supported by VIDI grants from the Netherlands Organization for Scientific Research (S.N. and C.O.). We thank Durk Talsma for the idea that inspired the current research, and Heleen Slagter for fruitful discussions.

References (37)

  • B.A. Steinman et al.

    Transient visual attention is dominated by the magnocellular stream

    Vision Research

    (1997)
  • Y. Yeshurun

    Isoluminant stimuli and red background attenuate the effects of transient spatial attention on temporal resolution

    Vision Research

    (2004)
  • I. Arend et al.

    Task-irrelevant visual motion and flicker attenuate the attentional blink

    Psychonomic Bulletin & Review

    (2006)
  • G. Benedek et al.

    Human scotopic spatiotemporal sensitivity: A comparison of psychophysical and electrophysiological data

    Documenta Ophthalmologica

    (2003)
  • H. Bowman et al.

    The simultaneous type, serial token model of temporal attention and working memory

    Psychological Review

    (2007)
  • J.C. Brehaut et al.

    Visual masking plays two roles in the attentional blink

    Perception & Psychophysics

    (1999)
  • M. Chikashi et al.

    Effects of background color on the global and local processing of hierarchically organized stimuli

    Journal of Cognitive Neuroscience

    (1999)
  • C.W. Clifford et al.

    Adaptation to temporal modulation can enhance differential speed sensitivity

    Vision Research

    (1999)
  • Cited by (0)

    View full text