Elsevier

Consciousness and Cognition

Volume 17, Issue 3, September 2008, Pages 602-615
Consciousness and Cognition

How do we know what we are doing? Time, intention and awareness of action

https://doi.org/10.1016/j.concog.2007.03.007Get rights and content

Abstract

Time is a fundamental dimension of consciousness. Many studies of the “sense of agency” have investigated whether we attribute actions to ourselves based on a conscious experience of intention occurring prior to action, or based on a reconstruction after the action itself has occurred. Here, we ask the same question about a lower level aspect of action experience, namely awareness of the detailed spatial form of a simple movement. Subjects reached for a target, which unpredictably jumped to the side on some trials. Participants (1) expressed their expectancy of a target shift during the upcoming movement, (2) pointed at the target as quickly and accurately as possible before returning to the start posiment to the target shift if required and (3) reproduced the spatial path of the movement they had just made, as accurately as possible, to give an indication of their awareness of the pointing movement. We analysed the spatial disparity between the initial and the reproduced movements on those with a target shift. A negative disparity value, or undershoot, suggests that motor awareness merely reflects a sluggish record of coordinated motor performance, while a positive value, or overshoot, suggests that participants’ intention to point to the shifting target contributes more to their awareness of action than their actual pointing movement. Undershoot and overshoot thus measure the reconstructive (motoric) and the preconstuctive (intentional) aspects of action awareness, respectively. We found that trials on which subjects strongly expected a target shift showed greater overshoot and less undershoot than trials with lower expectancy. Conscious expectancy therefore strongly influences the experience of the detailed motor parameters of our actions. Further, a delay inserted either between the expectancy judgement and the pointing movement, or between the pointing movement and the reproduction of the movement, had no effect on visuomotor adjustment but strongly influenced action awareness. Delays during either interval boosted undershoots, suggesting increased reliance on a time-limited sensory memory for action. The experience of action is thus strongly influenced by prior thoughts and expectations, but only over a short time period. Thus, awareness of our actions is a dynamic and relatively flexible mixture of what we intend to do, and what our motor system actually does.

Introduction

Action awareness involves knowing what we are doing. We often have conscious experience of thinking about actions, and of controlling them, even in advance of actually moving. That is, there is a direct relationship between our intention to act and our conscious experience of acting. However, conscious intention may inform us about our actions only late in their temporal development (Libet, Wright, & Gleason, 1983), and perhaps not at all in the case of ‘automatic’ actions.

Several studies have shown that conscious perceptual experience and motor control can be dissociated. In the double-step reaching task (e.g., Castiello, Paulignan, & Jeannerod, 1991; Johnson et al., 2002a, Johnson et al., 2002b), the target of a reaching or grasping movement is unpredictably shifted during the course of movement. This produces corrections of the trajectory within 100–150 ms (Carlton, 1981, Day and Lyon, 2000, Paulignan et al., 1990, Paulignan et al., 1991, Prablanc and Martin, 1992, Soechting and Lacquantini, 1983, Zelaznik et al., 1983). Interestingly, these adjustments may occur independently of the conscious visual experience of perceiving the target moving. For example, Goodale, Pelisson, and Prablanc (1986) observed successful visuomotor adjustment in a pointing experiment in which the target occasionally jumped several degrees while the jump remained itself remained unnoticed by the participants. Pélisson, Prablanc, Goodale, and Jeannerod (1986) improved on this design by triggering the target shift during a voluntary saccade, thus ensuring that it was not consciously detected. Successful visuomotor adjustments were nevertheless observed. Fourneret and Jeannerod (1998) found that subjects adjusted their reaching movements in response to spatially distorted visual feedback about hand-movement trajectory. Their participants remained unaware that the direction of their movement differed from what they saw, for angular distortions up to 15°. Other experiments compared the latency of visuomotor adjustments with the latency of conscious detection of target shifts. Castiello et al. (1991) found that awareness of an unexpected target jump occurred some 300 ms after the motor system had initiated an appropriate movement correction. They suggested that the neural pathways underlying visuomotor adjustment and conscious awareness were dissociable, and had different time constants.

These studies distinguish between motor performance and visual awareness. In contrast, Johnson et al., 2002a, Johnson et al., 2002b investigated the relation between the ability to make visuomotor adjustments, and the conscious experience of the adjusted movement itself. Participants made rapid pointing movements with blocked instructions to follow the target (pointing condition), or to move in the opposite direction (anti-pointing condition) if it jumped. After each movement, participants reproduced the spatial path of the movement just made, this time without any time constraint. The gap between the spatial path of the original pointing movement and the spatial path of the reproduced movement was used as a measure of motor awareness. In the pointing condition, participants showed reduced and delayed motor awareness: they reproduced the curved movement path evoked by the target shift both later and with lower amplitude than the adjustment executed in the original pointing movement. In the anti-pointing condition, however, participants’ corrections did not exhibit this dissociation between performance and motor awareness. Instead, the reproduced movements indicated that participants overestimated the speed and strength of the anti-point responses in the original pointing movements. Their motor awareness thus appeared to be influenced by the anti-pointing response that they should have made, or perhaps expected to make, and not by the weaker and slower response that they had actually made.

In many cases, such as the anti-pointing task described above, action awareness may depend on what we expect to occur, rather than on the physical movement of our body. This idea finds supporting arguments in studies dedicated to the dissociation between conscious expectancy and conditioning (Perruchet, 1985, Perruchet et al., 2006). When two events, E1 and E2, appear repeatedly in succession, the presentation of E1 tends to improve or modify the behavioural response to E2. Classical conditioning is perhaps the best-known example of such a priming effect. Crucially, this has been interpreted in two contrasting ways. First, the occurrence of E1 may generate a conscious expectancy of E2 (e.g., Bolles, 1972, Lovibond and Shanks, 2002, Tolman, 1932), which is in turn assumed to facilitate responding to the occurrence of E2. Second, classical conditioning has also been explained in terms of automatic activation: E1 facilitates the response to E2 as a mandatory consequence of their having been repeatedly associated in the past. Automatic activation, in this context, is therefore assumed to reflect previous experience with the association, independently of the agent’s conscious expectancy for E2.

The two interpretations differ radically in terms of their implications for conscious experience. The expectancy view takes conditioning to be dependent on conscious thought, while the priming view claims that conscious awareness is not necessary for conditioning to occur (e.g., Bush and Mosteller, 1951, Clark et al., 2001, Clark and Squire, 1998, Hull, 1943). These two hypotheses have been difficult to dissociate because most experimental settings involve learning schedules that modulate both automatic priming and conscious expectancy in parallel: under typical experimental conditions, identical effects are predicted by each view, for repeated associations between two events could result in either or both expectancy-meditated and activation-based facilitation of the processing of E2. As a solution to this problem, Perruchet (1985) proposed a methodology in which automatic priming and conscious expectancy can be opposed in simple conditioning. Here, we adapted this methodology to explore the relationships between expectancy, action and awareness of action. In the following, we briefly summarize Perruchet’s method and the main findings obtained through its application.

In Perruchet’s methodology, the US is presented in 50% of the trials. On each trial, before the CS is displayed, participants have to rate, on a graded scale, their expectancy that the US will be presented on the impending trial. Crucially, the sequence of trials is not random, but structured as shown in Table 1. This design results in the production of both E1-alone sequences and of sequences of E1–E2 pairings of various lengths. Thus, each trial can be described in terms of the nature and the length of the preceding sequence of identical trials. Both the probability of the conditioned response (CR) and subjective expectancy are recorded for each sequential context. In his 1985 work, Perruchet obtained two main results with this method. The first was the observation that recent experience of the occurrence of E2 influenced conscious expectancy of its repetition. Specifically, participants’ expectancy for E2 was the highest after a long sequence of E1-alone events and the lowest after a long sequence of E1–E2 pairings. Furthermore, this difference in degree of expectation was a function of the length of the preceding sequence and decreased linearly between these two extreme sequences—a well-known phenomenon called the Gambler’s Fallacy (Anderson, 1960, Burns and Corpus, 2004, Jarvik, 1951, Keren and Lewis, 1994), according to which people mistakenly believe that the probability of occurrence of successive independent random events depends on the recent history of their occurrence (i.e., mistakenly believing that “heads” is made more likely than “tails” after the observation that the past five tosses had each produced “tails”).

The second result obtained by Perruchet was the striking observation that the strength of the conditioned response followed a completely different course than that obtained for conscious expectancy. Specifically, eye blink strength increased linearly as a function of the number of recent previous occurrences of E1–E2 trials, confirming well-documented laws of conditioning and associative memory (e.g., Anderson, 2000). This result allowed Perruchet to dissociate between the two classes of interpretations described above. Repeated exposure to E1–E2 pairings strengthens behavioural responses to E2 while the very occurrence of such pairings simultaneously decreases conscious expectancy for E2! In other words, conditioning reflects a priming process that can be completely dissociated from conscious expectancy—a result that, when combined with other findings, suggests the involvement of independent processes in the determination of behavioural responses in this paradigm. In a subsequent study, Perruchet et al. replicated these findings with voluntary responses (keypresses in a simple reaction time task), suggesting that the dissociation method designed by Perruchet (1985) is robust and not limited to reflex responses.

Further research in the conditioning domain suggested that the relation between conscious expectancy and behaviour depends critically on the temporal features of the task. Clark et al. (2001) used Perruchet’s experimental approach to compare effects on behaviour of two training conditions: delay conditioning and trace conditioning. In delay conditioning, the CS co-terminates with the US ; in trace conditioning, a 1000 ms trace interval is introduced between CS offset and US onset. Clark et al. found that learned behaviour and conscious expectancy developed in parallel in trace conditioning, but not in delay conditioning. In trace conditioning, awareness of the CS–US relationship is mandatory for conditioning to take place, whereas this is not the case under delay conditioning. The relationship between conscious expectancy and behaviour thus appears to be modulated by temporal factors. One interpretation of Clark et al.’s findings is that awareness is necessary to bridge the temporal gap between the CS and the US; another is that it takes time for conscious expectancy to influence behaviour (Cleeremans, 2005, Cleeremans, 2006, Cleeremans and Sarrazin, 2007).

In this context, our study drew on both the visuomotor adjustment literature and on the associative priming literature, both reviewed above, to investigate the awareness of action. Thus, we explored whether retrospective awareness of a aiming movement is influenced by prior conscious expectations about the occurrence of a target shift, by the actual details of the movement that was just executed, or by both. To do so, we asked participants, on each trial, (1) to express their expectancy that a target shift would occur during the upcoming movement, (2) to point to a virtual target as quickly and accurately as possible (returning to the start position immediately thereafter), and (3) to reproduce the spatial path of the movement itself.

Thus, for each trial, we have three measures of different aspects of performance, each collected very closely to each other: (1)an expectancy judgment, which measures participant’s instantaneous conscious intention prior to movement onset, (2) path execution measures, which characterize the time course and the spatial extent of the actual movement itself, and (3) path reproduction measures, which characterize participant’s conscious awareness of the movement they just executed. The first and latter measures inform us about the preconstructive (intentional) and reconstructive (motoric) aspects of action awareness, respectively, and both can be compared to the actual movement itself.

To manipulate expectancy systematically, the material was organized in just the same way as was Perruchet’s material, that is, the sequence of successive trials participants were exposed to contained both long series of trials where a target shift occurs, and long series of trials where no shift occurs. We would thus expect the recent history of shifts to influence expectancy judgments in the same manner as the recent history of reinforcements influences expectancy judgments in the conditioning and simple RT protocols developed by Perruchet, that is, participants should fall for the Gambler’s fallacy in their conscious predictions about whether the next trial will contain a target shift or not.

Finally, to explore the effects of temporal factors on our different measures (keeping in spirit with the delay vs. trace conditioning protocols described above), our design included conditions in which a 6-s delay was inserted, either between expectancy judgment and the initial pointing movement, or between the initial pointing movement and the reproduced path.

Section snippets

Participants

Seven student volunteers were tested on the basis of informed consent and with local ethical committee approval. They were paid 10 € an hour to participate. All were right-handed and they had normal or corrected-to-normal vision as per self-report.

Apparatus

The experiment was conducted in a small, dark testing booth. Participants sat at a desk with their right hand resting comfortably on a start button. A video marker was taped over the nail of the right index finger. A Hamamatsu video recorded movement

Results

Results will be presented in three parts. First, we examine the extent to which conscious expectancy judgments are sensitive to the recent occurrence of target shifts, that is, we ask whether people’s predictions about what will happen in the next trial follows the Gambler’s fallacy. Second, we examine how the initial pointing movement is influenced by the occurrence of a target shift. Finally, we present analyses concerning the “motor awareness gap”, that is, the impact of both expectancy and

Discussion

This study was aimed at (1) exploring the content of action awareness, viewed here as a “balancing act” between a sensorimotor rendering of an actual, just-executed movement and an intentional rendering of what is consciously expected to happen, and at (2) showing how temporal factors (i.e., the occurrence of delays between expectancy judgments, action execution, and action reproduction) modulate the content of action awareness. To address these issues, we contrasted actual motor performance

References (37)

  • U. Castiello et al.

    Temporal dissociation of motor responses and subjective awareness

    Brain

    (1991)
  • R.E. Clark et al.

    Trace and delay eyeblink conditioning: Contrasting phenomena of declarative and nondeclarative memory

    Psychological Science

    (2001)
  • R.E. Clark et al.

    Classical conditioning and brain systems: The role of awareness

    Science

    (1998)
  • A. Cleeremans

    Conscious and unconscious cognition: A graded, dynamic, perspective

  • Cleeremans, A., & Sarrazin, J. C. (2007). Time, Action, and Consciousness. Human Movement Science, 26,...
  • B.L. Day et al.

    Voluntary modification of automatic arm movements evoked by motion of visual target

    Experimental Brain Research

    (2000)
  • Dijkerman, H. C., & De Haan, E. H. F. (in press). Somatosensory processes subserving perception and action. Behavioral...
  • M.A. Goodale et al.

    Large adjustments in visually guided reaching do not depend on vision of the hand or perception of target displacement

    Nature

    (1986)
  • Cited by (18)

    • Delusions of control in schizophrenia: Resistant to the mind's best trick?

      2018, Schizophrenia Research
      Citation Excerpt :

      By introducing systematically delayed feedback regarding these movements, the authors were able to manipulate the perceived onset of motor intentions, leading them to conclude that conscious intentions are inferred on basis of the movements' effects rather than directly perceived during its execution. Under certain conditions, healthy individuals deny being the initiator of self-caused actions (Sarrazin et al., 2008; Wegner, 2002; Yoshie and Haggard, 2013), or conversely experience an intentional state for actions that were in fact controlled by others (Lynn et al., 2010; Wegner, 2002). Together, these findings suggest that intentions do not arise prior to the corresponding actions, but rather at the same time or even after the action has taken place.

    • The Neurology of Consciousness: An Overview

      2015, The Neurology of Consciousness: Cognitive Neuroscience and Neuropathology
    • A second-person neuroscience in interaction

      2013, Behavioral and Brain Sciences
    View all citing articles on Scopus

    Jean-Christophe Sarrazin is a Marie Curie Postdoctoral Fellow at the Cognitive Science Research Unit (ULB, Belgium), Axel Cleeremans is a Research Director with the National Fund for Scientific Research (FNRS, Belgium), and Patrick Haggard is a Professor at the Institute of Cognitive Neuroscience (UCL, England). This work was supported by a grant from the European Commission (Grant MEIF-CT-2005-515499). We thank two anonymous referees for their comments on a previous version of this article, as well as Thelma Coyle and Frank Buloup for their assistance in conducting the experiment.

    View full text