Relationship Between Awareness of Intention to Act and General Metacognitive Capability

The Relationship Between Awareness of Intention to Act and General Metacognitive Capability

Abstract

Theories of metacognition have explored metacognitive intention to act and its relationship to other metacognitive capability and to various unrelated tasks.  Prior theories have found that relationships do exist between metacognitive capability and unrelated variables such as agency and work performance.  This study investigated whether a linear negative correlation exists between metacognitive intention to act and metacognitive sensitivity variables.  Metacognitive intention to act and metacognitive sensitivity were tested on 119 Birkbeck College psychology students by replicating the intention to act and metacognitive sensitivity experiments of prior studies.  The results of a Pearson’s correlation and r² analyses showed a very minor positive but statistically non-significant correlation between these variables.  These findings reinforce earlier studies showing that measures of metacognitive sensitivity and metacognitive confidence do not appear to be dependent on each other.  However, the findings also run counter to other prior studies suggesting a relationship between metacognition and seemingly unrelated variables.  Procedural limitations of this experiment question its validity and wider investigation into the relationship among multiple metacognitive and other variables and their associated partial correlations would add to the examination of metacognitive theories.

Introduction

The study of metacognition, our awareness of intention to act as distinct from the action itself and the role and order of associated cerebral activity, is a fertile vein of psychological research.  Experimenters have sought to define the scientific nature of free will by studying the role of metacognition in voluntary muscle activity (Libet, Gleason, Wright & Pearl, 1983; Haggard, 2005).  Over the 35 years since the seminal study of Libet et al (1983) metacognition study has undertaken a variety of paths to better understand our awareness of intention to act.

In their 1983 study, Benjamin Libet et al examined the relationship in time among the urge to act, the readiness-potential (RP) associated with the act recorded by EEG, and the actions themselves.  Metacognition in that study was identified as a specific point in time when subjects became aware of their urge to act.  This was determined by subjects stopping a revolving point of light on a clock face when they first became aware of the intention to act.  The study’s findings showed that the RPs occurred in time a minimum of several hundred milliseconds before the recorded urge to act, which in turn were prior to the initiation of the motor activation for the act itself.  As we are not conscious of the brain activity represented by the RPs, these findings call into question our ability to exercise conscious instigation or control over our actions.

Further inquiry looked at the importance of metacognition in understanding our impact on the outside world, defined in the concept of ‘agency’ (Jeannerod, 2003; Haggard, 2005; Metcalfe & Greene, 2007).  Patrick Haggard’s review of studies in this area builds on Libet’s findings and concludes that conscious intention to act is “a direct consequence of pre-movement brain activity in the frontal and parietal motor areas” (Haggard, 2005, p.294).  In that review, Haggard also refers to correlations between the urge and effect aspects of intention, suggesting that we associate our actions with their effects on the environment, giving rise to a sense of agency. 

Jeannerod (2003) reinforced the importance of agency as a fundamental factor in our understanding of an independent self and tied this to the relationship between the neural signals associated with muscle movements and sensory cues arising from the movements themselves.  Metcalfe & Greene (2007) undertook experiments examining factors that may influence our metacognition of control over our actions.  After completing a number of tasks with varied conditions of influencing factors such as speed or turbulence, subjects were asked to rate their performance.  The results showed a correlation between the subjects’ judgement of performance in the task and their judgement of agency or feeling of control.  Metcalfe & Green (2007) concluded that we have metacognitive awareness of our agency that closely tracks our actual control of events due to our actions, reinforcing metacognition as a principal factor in our concepts of self in our environment.

In their 2010 study, Lysaker, Dimaggio, Carcione, Procacci, Buck, Davis & Nicolo recognised that schizophrenia is associated with impaired metacognitive capability and examined the relationship between impaired metacognitive capacity due to schizophrenia and prospective work function.  They found that subjects with higher metacognitive ability had higher levels of work function, suggesting a relationship between metacognition and seemingly unrelated variables.

In 2015 Sam J. Gilbert used perceptual discrimination tasks to examine the relationship between metacognitive awareness and confidence.  This ‘metacognitive sensitivity’ was established by looking at the relationship between confidence in performance and actual performance in the task (Gilbert, 2015). This was done in part by testing the correlation between subjects’ metacognitive judgements of their ability to perform well at the tasks and their use of intention offloading which in Gilbert’s study was the use of memory aides.  Confidence was measured by the mean confidence rating of subjects across trials.  Key elements of Gilbert’s experiments sought to build on earlier inquiries which showed no correlation between these two variables (Song, Kanai, Fleming, Weil, Schwarzkopf & Rees, 2011).  Importantly, Gilbert’s findings were consistent with Song et al showing that the measures of metacognitive sensitivity and metacognitive confidence do not appear to be dependent on each other.

Although Gilbert drew conclusions of independence between metacognitive sensitivity and metacognitive confidence, he did not specifically investigate the statistical correlation between the two but rather focused on the correlation of each across different perceptual determination tasks (Gilbert, 2015).  Likewise, Libet et al (1983) examined the role of metacognition and brain function in causing voluntary action but stopped short of investigating the relationship between metacognition of intention to act and the metacognitive sensitivity of perceptual judgements.

The current experiment replicates in part the Libet (1983) and Gilbert (2015) experiments.  It seeks to build on their work to examine whether there is a specific correlation between metacognition of intention to act and metacognitive sensitivity.  As the study seeks to establish a correlation between these separate variables, neither is a dependent or independent variable for the purposes of this examination.  The experiment tests whether the metacognitive sensitivity from Gilbert’s coloured grid tasks, expressed as d’, will negatively correlate with metacognitive intention measured by replication of Libet’s clock face task.  Findings of this nature would suggest that metacognitive intention may have common processes with other metacognitive tasks and help to answer the open question of the open question of metacognitive capability’s relationship with other unrelated tasks.

Method

Design

Subjects were asked to complete two separate tasks.  Task One was a perceptual determination task replicating the Gilbert coloured grid task (Gilbert, 2015) and Task Two was a metacognition of intention task replicating the Libet clock face task (Libet et al, 1983).  A within-subjects design was used.  All subjects were asked to complete 150 experimental trials in Task One to determine which of two grids contained more coloured squares.  In Task One subjects also were asked to rate their degree of confidence in their determination of which grid contained more of the coloured squares in each trial.  Task One was designed to test metacognitive sensitivity.  In Task Two all subjects were asked to complete 40 experimental trials to determine when they first felt the urge or intention to act and when they took the action itself.  Task Two was designed to test metacognitive intention.

Participants

The subjects were Birkbeck College psychology students participating in the experiment as a routine part of their study.  The sample was 119 adult students; 87 females and 32 males between 19 and 61 years of age (N = 119, mean age = 30.55, St. dev = 9.81).  Subjects were verbally informed of their right not to participate in the experiment or to stop at any time.  Formal consent was sought and provided on-line.

Materials

Subjects had access to a computer-based experiment preloaded onto computers provided by the experimenter.  Each had their own computer, well-spaced to avoid distraction.  The computer representation included a data collection screen covering age, gender and subject number.  The computer representation also included a series of 150 perceptual determination trials and 40 metacognitive intention trials.  Subjects were provided with oral and written instructions and randomly assigned subject numbers.

Procedure

For Task One, subjects were instructed to use computer keys to indicate which of two identical grids contained more coloured squares.  They were provided with practice blocks to familiarise them with the procedure of the experiment.  Subjects were instructed to address their attention to a point between the grids and to use their best judgement to determine which grid had more coloured squares.  They were also asked avoid guessing.

Once the experiment was initiated two identical grids of 400 squares each appeared on computer screens side by side for 250ms.   Each grid contained different numbers of coloured squares.  After 250ms the coloured grids disappeared and were replaced by identical grids that contained no coloured squares.  These uncoloured squares remained on the screen until subjects made a judgement as to which grid contained more coloured squares and indicated this by a computer keystroke.  Subjects were then asked to provide an indication of how confident they were in their judgement by sliding a cursor along a computer representation of a spectrum ranging from ‘Guessing” to ‘Confident’.  The subjects were presented with 150 trials.

For Task Two, subjects were presented with a clock face with seconds marked out in five second intervals.  They received instructions to make use the clock face in two ways: first to stop a clockwise rotating dot wherever they felt like doing so, trying not to stop the dot at the same time in each trial; and second to indicate the time on the clock face at which they first noticed the intention to act to stop the rotating dot.

On commencement of the task subjects were presented with an immobile clock face for 250ms.  In the first part of the task, subjects were asked to use a computer key to stop a clockwise rotating dot whenever they wanted to.  Subjects were instructed that the dot would rotate the clock face 3 times.  Each of the three rotations was programmed to occur over 2560ms.  In the second part of the task, subjects were presented with an immobile clock face and were asked to indicate where on the clock face the dot was when they first became aware of their intention to stop the dot in the first part of the task.  During this second part of the task the clock face remained on the computer screen until subjects gave a response.  Subjects were presented with 40 trials in Task Two. 

At the conclusion of the tasks, subjects were requested to complete an unrelated survey.

Results

The shape of the data for each of the variables, metacognitive sensitivity and metacognitive intention to act, showed widely different characteristics; metacognitive sensitivity (M = 0.621, St. dev. = 0.344) and metacognitive intention judgement (M = – 50.01, St. dev. = 224.17).

It is useful to examine the shape of the data by scatterplot to gain an initial view of positive or negative correlation between the variables.  This is presented Figure 1, below.

Figure 1. Raw Data and Fit Line for Metacognitive Sensitivity and Intention Variables

To confirm the apparent relationship between the variables initially identified in the scatterplot view of the data, a Pearson’s bivariate correlation test was performed and 95% confidence intervals were calculated.  The results of this analysis showed a minor positive correlation between metacognitive sensitivity and metacognitive intention but that the relationship is not statistically significant, r = .049, 95% BCa CI [-.159, .232], p = .298.

The relationship between the variables was further examined by using an r² value which indicated that metacognitive sensitivity and metacognitive intention share only .24% of variability and thus show a negligible relationship, r² = .0024.  This result indicates that 99.76% of the variability in each of these variables is still to be accounted for by other variables not examined here.

Discussion

The results demonstrate that the metacognitive sensitivity and metacognitive intention variables are widely different as depicted in the means and standard deviations of the two variables.  The results further show that the variables have virtually no correlation between them as evidenced by the r and r² values and the statistically non-significant nature of their relationship.  These results reinforce the findings of Gilbert (2015) and Song et al (2011) showing that the measures of metacognitive sensitivity and metacognitive confidence do not appear to be dependent on each other.  Further, the results fail to show a possible reliance across metacognitive tasks on common or similar processes.

The current study sought to examine the hypothesis that there would be a negative correlation between metacognitive sensitivity (d’) and metacognitive intention to act.  Proof of this hypothesis would imply that earlier metacognition of intention to act would be associated with higher metacognitive sensitivity, suggesting that metacognition of intention to act somehow crosses over to other tasks.  However, the results of this examination do not support this hypothesis.  The results showed a very minor positive correlation between the variables tested here and that the relationship was not statistically significant.  Thus, the question of the relationship between metacognitive intention to act and other variables remains open.  In addition, the question of whether metacognitive tasks share common or similar processes remains open as well.

The inconclusive nature of the results here runs counter to the results in the work of Metcalfe & Greene (2007) and Haggard (2005) regarding the correlation between metacognition sensitivity and another separate variable, our sense of agency.  Given the importance of our sense of agency for placement of ourselves in our environment explained by Jeannerod (2003), Metcalfe & Greene (2007) and Haggard (2005), if the correlation and possible influence of metacognition sensitivity on unrelated variables such as our sense of agency remains open we still cannot with confidence anchor our actions as primary influencers on the world around us.  The results of the current experiment also run counter to the findings of Lysaker et al (2010) that showed a relationship between high metacognitive capability and other tasks, in that instance prospective work performance.

Notwithstanding the lack of support for the experimental hypothesis, this experiment had a series of limitations that question reliance on the results.  There were multiple issues with the procedure: subjects were in different rooms; computers failed; and there were repeated interruptions.  In addition, instructions were not fully documented and relied on verbal augmentation making replicability imprecise.  Another limitation was that this experiment only tested the correlation between two variables.  Including other variables through unrelated tasks and examining partial correlation among them in further research could add to the investigation of correlations among metacognitive variables and among metacognitive and other variables such as agency and work performance.  This wider exploration could advance the examination of the role of metacognition in voluntary action and the exploration of the processes associated with free will.

Conclusion

The experimental results found no statistically significant correlation between metacognitive intention to act and metacognitive sensitivity.  The apparent absence of a relationship between these two variables calls into question whether different metacognitive variables share common or similar processes.  It also calls into question prior theories indicating apparent relationships among these variables and among metacognitive variables and other variables.  However, flaws in the experimental procedure raise extraneous variable issues casting doubt on the findings.  Further investigation into the relationship among multiple variables and associated partial correlations and would add to the examination of metacognitive theories, our sense of an independent self and the nature of free will.

References

• Gilbert, S. J. (2015). Strategic use of reminders: Influence of both domain-general and task-specific metacognitive confidence, independent of objective memory ability. Consciousness and Cognition, 33, 245-260.
• Haggard, P. (2005). Conscious intention and motor cognition. Trends in cognitive sciences, 9(6), 290-295.
• Jeannerod, M. (2003). The mechanism of self-recognition in humans. Behavioural brain research, 142(1-2), 1-15.
• Libet, B., Gleason, C. A., Wright, E. W., & Pearl, D. K. (1983). Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential) the unconscious initiation of a freely voluntary act. Brain, 106(3), 623-642.
• Lysaker, P. H., Dimaggio, G., Carcione, A., Procacci, M., Buck, K. D., Davis, L. W., & Nicolò, G. (2010). Metacognition and schizophrenia: the capacity for self-reflectivity as a predictor for prospective assessments of work performance over six months. Schizophrenia research, 122(1-3), 124-130.
• Metcalfe, J., & Greene, M. J. (2007). Metacognition of agency. Journal of Experimental Psychology: General, 136(2), 184.
• Song, C., Kanai, R., Fleming, S. M., Weil, R. S., Schwarzkopf, D. S., & Rees, G. (2011). Relating inter-individual differences in metacognitive performance on different perceptual tasks. Consciousness and cognition, 20(4), 1787-1792.

Appendix

IRM Lab Experiment 14/03/2018 – Instruction Sheet for Participants

The experiment we going to conduct today involves you making some tapping movements with your right hand (index and middle finger on the DOWN and UP arrow keys, respectively) and making some judgments about some things you see on the screen. You will make these judgments with your left thumb (held over the 1 and 2 number keys). The image below shows how to arrange your hands.

Before you start up the experiment:

1. Turn off your phone
2. Make sure that your monitor is positioned at a comfortable height
3. Make sure that the UP/DOWN arrow keys line up with the centre of the monitor
4. Type runexperiment (all lower case) into the open window and press ‘Enter’