The proportion valid effect in covert orienting: Strategic control or implicit learning?

https://doi.org/10.1016/j.concog.2009.07.013Get rights and content

Abstract

It is well known that the difference in performance between valid and invalid trials in the covert orienting paradigm (i.e., the cueing effect) increases as the proportion of valid trials increases. This proportion valid effect is widely assumed to reflect “strategic” control over the distribution of attention. In the present experiments we determine if this effect results from an explicit strategy or implicit learning by probing participant’s awareness of the proportion of valid trials. Results support the idea that the proportion valid effect in the covert orienting paradigm reflects implicit learning not an explicit strategy.

Introduction

In the covert orienting paradigm (Posner, Nissen, & Ogden, 1978), a cue (e.g., a peripheral onset) is used to direct the participant’s visual attention to a particular location before a target is presented. When attention is directed to the upcoming target location (i.e., a valid trial) responses are faster than when attention is directed to a non-target location (i.e., an invalid trial). One of the most robust findings in the covert orienting literature is that the cueing effect (i.e., the difference in response times between valid and invalid trials) increases in magnitude as the proportion of valid trials increases (e.g., Enns and Brodeur, 1989, Eriksen and Yeh, 1985, Johnson and Yantis, 1995, Jonides, 1980). This proportion valid effect is believed to reflect the endogenous control of visual attention. Specifically, individuals are held to strategically allocate attention to the cued location as a function of the cue’s spatial utility (i.e., the degree to which the cue location predicts the target location). Alternatively, the proportion valid effect may reflect a form of implicit learning (Seger, 1994) wherein associations between the cue and target location, developed outside the participants’ awareness, determine how attention is distributed. In the present investigation we test between these two accounts by determining the relation between the proportion of valid trials, participant’s estimate of the proportion of valid trials, and the magnitude of the cueing effect.

Attention shifts are often classified as either exogenous or endogenous (e.g., Yantis, 1998). Exogenous shifts of attention are fast and stimulus driven whereas endogenous shifts of attention are slow, voluntary and intention driven. The proportion valid effect in the covert orienting paradigm is taken to reflect this latter form of attention shift. Specifically, as the proportion of valid trials increases, participants are held to voluntarily allocate increasing amounts of attention to the cued location or to allocate attention to the cued location on a greater portion of trials. As a result, responses become faster on valid trials and slower on invalid trials producing a larger cueing effect. This strategic control account of the proportion valid effect reflects, in large part, the received view in the attention literature. Indeed, researchers have gone so far as to use the presence/absence of a proportion valid effect to make inferences about an individual’s ability to shift attention in a strategic or endogenous manner (Bartolomeo et al., 2001, Danckert et al., 1998, Enns and Brodeur, 1989, Maruff et al., 1998).

Rather than strategic control, the proportion valid effect may reflect a form of implicit learning (Seger, 1994). Numerous demonstrations exist that attest to the ability of the cognitive system to guide behavior based on information acquired implicitly (e.g., Cleeremans, Destrebecqz, & Boyer, 1998). For example, in a sequence learning task participants are faster to respond if there is an inherent structure built into the stimuli than when stimuli are random, independent of the participants’ awareness of that structure (Nissen & Bullemer, 1987).

Researchers have also demonstrated that visual attention can be guided based on information participants learn implicitly (Chun, 2000, Chun and Jiang, 1998, Chun and Jiang, 1999, Lambert, 2004, Lambert et al., 1999, Lambert and Sumich, 1996). For example, Chun and colleagues (Chun, 2000, Chun and Jiang, 1998, Chun and Jiang, 1999) have demonstrated that participants can localize a target faster when it is presented in a target–distractor spatial configuration that is repeated throughout the experiment. Critically, this effect occurs independently of whether participants are aware of the repeated target–distractor spatial configuration (Chun, 2000, Chun and Jiang, 1998, Chun and Jiang, 1999). According to Chun and colleagues an implicitly learned target–distractor relation is used, outside of awareness, to guide visual attention. Lambert and colleagues (Lambert, 2004, Lambert and Sumich, 1996, Lambert et al., 1999) have demonstrated similar results using the covert orienting paradigm. Lambert et al. (1999) presented participants with letters to the left and right of a fixation before a target was presented to either the left or right. The identity of the letters predicted the upcoming target’s location and participants demonstrated cueing effects consistent with that underlying structure (i.e., faster responses when the target appeared at the expected location). This occurred despite the fact that participants were unaware of the relation between the letters and the target location. Thus, research suggests that visual attention can be guided through implicit learning. According to an implicit learning account of the proportion valid effect, the spatial utility of the cue is learned implicitly and the increase in the amount of attention or the increase in the probability that the cue is attended occurs outside the participants’ awareness.

The strategic control and implicit learning accounts of the proportion valid effect differ with respect to the need for participants to be aware of the relation between the cue and target. According to the strategic control account the participant becomes aware of the spatial utility of the cue and this awareness causes the adoption of a strategy that leads to an endogenous shift of attention in response to the cue. Thus, the change in the distribution of attention reflects a deliberate, intentional, conscious act on behalf of the participant. In contrast, according to the implicit learning account, the relation between the cue and target is learned implicitly and attention is distributed accordingly. Thus, the change in the distribution of attention occurs without the explicit knowledge of the participant and is a consequence of learning that has occurred implicitly. These accounts are not mutually exclusive.

In a recent series of experiments, Bartolomeo et al., 2007, Bartolomeo et al., 2008 set out to determine if the proportion valid effect was dependent on the participant being aware of the relation between the cue and target. In three experiments Bartolomeo et al. (2007) manipulated proportion valid in the covert orienting paradigm. They used the inhibition of return (IOR) effect, a form of the cueing effect in which the participant is slower to respond on valid trials than invalid trials (Klein, 2000), as an index of attentional orienting. Inhibition of return occurs for peripheral abrupt onset cues at long cue–target intervals (e.g., 600–1000 ms; Klein, 2000). The proportion valid effect in the case of IOR consists of a smaller IOR effect as proportion valid increases (Bartolomeo et al., 2007). After each experiment, participants were asked to report (1) whether a relation existed between the cue and target and (2) whether cues predicted the target or non-target location. On the basis of their answers to these questions participants were divided into two groups, those that could identify the cue–target relation and those that could not. If the proportion valid effect in IOR required awareness of the cue–target relation, then only the former group should demonstrate a proportion valid effect. Across these three experiments, Bartolomeo et al. (2007) found no evidence that such awareness was necessary to observe an effect of proportion valid on IOR. Thus, even those participants who failed to notice any cue–target relation or indicated the incorrect cue–target relation demonstrated a proportion valid effect. These results are inconsistent with a strategic control account but consistent with the implicit learning account.

Bartolomeo et al. (2007) also reported an experiment in which they manipulated proportion valid using central arrow cues. Central cues were long considered the ideal cue type for indexing purely endogenous shifts of attention (Jonides, 1980), though this idea has come under attack recently (Ristic, Friesen, & Kingstone, 2002). Central cues do not produce inhibition of return (Klein, 2000), thus in this experiment the facilitatory, rather than the inhibitory, component of covert orienting was measured. Here, responses on valid trials are faster than on invalid trials and this effect increases in magnitude as proportion valid is increased. Using the same design as the IOR experiments, Bartolomeo et al.’s (2007) results were somewhat ambiguous. Whereas there was no interaction between group (i.e., those aware vs. not aware of the cue–target relation) and the proportion valid effect, the proportion valid effect was not significant in the group of participants who could not correctly report the cue–target relation. Bartolomeo et al. (2007) took this result as evidence that participants need to be able to report the cue–target relation in order to demonstrate a proportion valid effect with a central arrow cue. These results are consistent with a strategic control account. Overall, Bartolomeo et al.’s (2007) results suggest that strategic control might provide the best account of proportion valid effects with central arrow cues but implicit learning might provide the best account of proportion valid effects with peripheral cues.

In the present investigation we further examine the role of awareness of the cue–target relation in observing proportion valid effects by extending Bartolomeo et al.’s (2007) study in a number of directions. First, we introduce the use of a quantitative-continuous estimate of the participants’ knowledge of the cue–target relation. In the Bartolomeo et al. (2007) study awareness of the cue–target relation was qualitative-dichotomous; the cue–target relation was either “verbalizable” or not. Here, following each experiment we asked participants to estimate the percentage of valid trials. This provides a quantitative-continuous estimate of participants’ awareness of the proportion of valid trials. Thus, we can determine the degree of relation, if any, between participants’ awareness of the proportion valid and the magnitude of their cueing effect. If participants’ estimates of the proportion of valid trials correlates with the magnitude of their cueing effect this would be consistent with a strategic orienting account whereas no relation between estimates and performance would be consistent with an implicit learning account (see Dienes, 2008 for similar logic).

In addition to a quantitative-continuous subjective estimate of proportion valid, we also assess the proportion valid effect in the context of facilitory cueing effects for both a peripheral and a central cue. Bartolomeo et al. (2007) assessed inhibitory effects of peripheral cues and facilitory effects of central cues thus leaving open the possibility that the differences found between the cue types reflect differences in the type of cueing effect indexed (i.e., facilitory vs. inhibitory). Here, short cue–target intervals were used with both cue types. Proportion valid effects therefore consist of an increase in the magnitude of the cueing effects as proportion valid increases. The last extension we introduce is the use of a discrimination task rather than a detection task. The discrimination task allows us to analyze error rates in addition to response times.

Participants performed a simple two choice discrimination task. Targets appeared in one of two locations and were preceded by a cue indicating either the target or the non-target location. When the cue indicated the target location this was considered a valid trial and when the cue indicated the non-target location this was considered an invalid trial. Two types of cues were used; an abrupt onset peripheral cue (Experiment 1) and a central arrow cue (Experiment 2). These cue types were chosen because they have been used frequently in studies of the proportion valid effect. Proportion valid was manipulated such that for one group of participants 50% of the trials were valid and for the other group 75% of the trials were valid. In addition, after the experiment, participants were asked to estimate the percentage of valid trials.

According to the strategic control account of the proportion valid effect, participants in the 75% valid condition should provide a higher estimate of the proportion valid than in the 50% valid condition. Furthermore, participants’ estimates of the proportion of valid trials at the individual level should correlate with the magnitudes of their cueing effects.

Section snippets

Participants

Sixty undergraduate students (30 in each proportion valid group) were paid $4 each for their participation. All participants reported normal or corrected-to-normal vision.

Design

The experimental design consisted of a 2 (Proportion Valid: 50% vs. 75%) × 2 (Cue Condition: Valid vs. Invalid) mixed design. Proportion valid was manipulated between subjects.

Apparatus

Micro Experimental Lab (MEL 2.0; Schneider, 1988) software controlled timing and presentation of stimuli and logged responses and RTs. Stimuli were

Experiment 2: central cue

In Experiment 2 a central cue (e.g., arrow) was used in place of the peripheral abrupt onset cue used in Experiment 1. Central cues also demonstrate a Proportion Valid by Cue Condition interaction. Thus, similar to Experiment 1, we can assess the degree to which awareness of the cue–target relation determines the presence/absence of the proportion valid effect. In the Bartolomeo et al. (2007) experiment, participants who did not demonstrate awareness of the cue–target relation did not show a

General discussion

The present experiments investigated the mechanism responsible for the proportion valid effect in the covert orienting paradigm by determining the relation between participants’ awareness of the proportion of valid trials and the magnitude of the cueing effect. Two theories of the proportion valid effect were contrasted. According to the control account, awareness of the proportion of valid trials plays a central role, whereas according to the implicit learning account, awareness of the

Conclusion

In the present investigation we demonstrated that proportion valid effects can occur independently of participants’ knowledge of the proportion of valid trials. Thus, we argue that proportion valid effects are better viewed as an example of implicit learning than strategic control. We have attempted to incorporate proportion valid effects with other instances involving implicit learning in visual attention in an attempt to highlight the interaction between attention and memory.

Acknowledgment

This work was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) Canada Graduate Scholarship to EFR and an NSERC operating Grant (018395) to JAS.

References (36)

  • J. Danckert et al.

    Inhibitory processes in covert orienting in patients with Alzheimer’s disease

    Neuropsychology

    (1998)
  • Z. Dienes

    Subjective measures of unconscious knowledge

    Progress in Brain Research

    (2008)
  • C.W. Eriksen et al.

    Allocation of attention in the visual field

    Journal of Experimental Psychology: Human Perception and Performance

    (1985)
  • C.L. Folk et al.

    Involuntary covert orienting is contingent on attentional control settings

    Journal of Experimental Psychology: Human Perception and Performance

    (1992)
  • C.K. Friesen et al.

    The eyes have it! Reflexive orienting is triggered by nonpredictive gaze

    Psychonomic Bulletin and Review

    (1998)
  • B. Hommel et al.

    Symbolic control of visual attention

    Psychological Science

    (2001)
  • B. Hommel et al.

    A feature-integration account of sequential effects in the Simon task

    Psychological Research

    (2004)
  • L. Jimenez et al.

    Which attention is needed for implicit sequence learning?

    Journal of Experimental Psychology: Learning, Memory & Cognition

    (1999)
  • Cited by (0)

    View full text