Lost in time...: The search for intentions and Readiness Potentials
Introduction
In 1983 Libet et al. measured the brain activity of six participants using EEG (electroencephalography) while they were performing voluntary movements with their right hand. While making these movements, the participants watched a clock and had to report the time at which they first felt their intention to act. Libet et al. used the Readiness Potential (RP) as an indicator of the neural preparation for a voluntary movement and found that the RP preceded the intention to act by 350 ms and the actual movement by 500 ms on average. This implied, according to Libet et al., that the experience of consciously willing an act arises after our brain has already started preparing that act.
A major reason for the substantial interest in Libet-style experiments derives from the possibility that our conscious intentions or decisions to act follow from brain processes of which we are not aware, and hence that our will or decisions are not ‘free’. For instance, Libet (2004) says:
Establishing the time of the conscious will relative to the onset of brain activity (the RP) was clearly important. If conscious will were to follow the onset of the RP, that would have a fundamental impact on how we view free will. (p. 125)
Spence (1996) says:
If the findings of Libet and co-workers are replicated, then we must conclude that ‘decisions’ to act arise prior to our conscious awareness of them. Thus our ‘decision’ or ‘freedom’ is illusory (if by these terms we mean conscious phenomena). (pp. 83–84).
Given these suggested implications, it is understandable that the empirical findings caused excitement in different branches of cognitive science, which resulted in a great number of reinterpretations (Dennett, 2003, Mele, 2010, Rosenthal, 2002) and replications (Haggard and Eimer, 1999, Keller and Heckhausen, 1990, Trevena and Miller, 2002). Even up to this day, many variants of the Libet experiment are being published (Bode et al., 2012, Fried et al., 2011, Matsuhashi and Hallett, 2008, Schurger et al., 2012, Soon et al., 2008).
The debate about RPs and the implications for the causal efficacy and freedom of conscious will is difficult in part because of the technical issues regarding the RP measurements, and the conceptual complexity when defining the nature of ‘an intention’. There is considerable variation in terminology as to what the participant is asked to do in the experiment: whereas Libet, Gleason, Wright, and Pearl (1983) asked for a report of the onset of the conscious awareness of wanting to perform a given self-initiated movement, Soon et al. (2008) asked to report the onset of the urge to act and Trevena and Miller (2002) asked for a report of the instant of the decision to “go now”. In our experiment, we asked our participants to report the onset of their intention to act (i.e., the moment in time at which they became aware that they wanted to perform an act). This makes it difficult to establish exactly what the implications of the acquired data are. This indicates that the conceptual interpretation of Libet-style experiments can be seen as quite variable or ‘soft’. As more often in cognitive neuroscience, the issue of ‘hard data and soft concepts’ presents a serious obstacle to reach a consensus in the field, in spite, or perhaps even because of a growing number of empirical studies. However, even speaking of ‘hard data’ might be saying too much in the case of Libet-type experiments.
Although there are a growing number of replications of Libet’s experiment, upon closer examination the exactness of the measurements is open to discussion, which is especially relevant when one takes into account the requirements for establishing a causal relationship between an RP and an act. In order to calculate the RP onset, several manipulations such as averaging and filtering are necessary. These preprocessing steps are commonly found in the literature on Libet-type experiments (Haggard and Eimer, 1999, Libet et al., 1983, Matsuhashi and Hallett, 2008, Trevena and Miller, 2002) and seem to be minor technical details in the analysis of EEG data. However, each of these manipulations has its implications for the data, e.g. resulting in a shift of the RP onset forward or backward in time (Trevena & Miller, 2002). Furthermore, getting a reliable estimate of the onset of the intention to act is quite a challenge due to the differences in the participant’s understanding of what an intention is and what they need to do and report, as indicated above. In Libet’s experiment (and in ours), we are not measuring the actual intention to act but the subjective report of this intention. All this makes the theoretical interpretation of the relation between the RP in the supplementary motor cortex and its interpretation in cognitive terminology (i.e., intention, urge, will, decision or even action preparation) quite ambiguous. We would therefore like to suggest that debates about Libet-style experiments are plagued by the problem of ‘semi-hard data and soft concepts’.
In this article, we present our own replication of Libet’s experiment and use it to illustrate how technical and conceptual issues can influence establishing whether the RP reliably and consistently precedes the intention to act. We will focus on two points that (to our knowledge) have not been addressed in the literature. First of all, by comparing the results of three different ways of measuring the onset of the RP, in combination with three different sets of electrodes used to measure the RP, one can notice large differences in RP–intention timings. For instance, RP timing differences within one user can be found to vary from −898 ms to as little as −47 ms. Although the general order found in Libet’s results (first RP then awareness) is replicated, the data are not quite as ‘hard’ as one may think. Differences in individual participants are obscured when looking at the grand average over all participants (as done in previous research). For instance, although the RP precedes the intention to act in the grand average, we found the opposite pattern of results in one participant (participant 5): the intention actually preceded the RP onset. We will discuss the consequences this may have for the interpretation of the data in terms of a causal relationship between the RP and an intended act. Secondly, although we indicate there are several reasons to believe that the ‘inverted results’ of participant 5 are not valid, we take it up as a thought experiment. Imagine one would find a participant whose intentions would reliably precede the RP, how is one to interpret such a case? Should one be forced to conclude that here we have a case of a causally efficacious consciousness, or even a ‘soul’ or ‘res cogitans’? By examining the Libet experiment with such a hypothetical case we aim to elucidate the debates on the freedom and causal efficacy of conscious will.
In Section 2, we will first go into more detail on the background of our project. Sections 3 Experimental methods, 4 Data analysis, 5 Results will describe our replication study and Section 6 illustrates how seemingly minor technical details can greatly influence the reliability and consistency of the time relation between the RP and intention. Section 7 will aim to elucidate the debate on the existence of free will and the causal efficacy of conscious will by discussing the consequences of finding a participant which intention onset precedes their RP onset (as found in our replication study). We will point out that Libet-style results are not as ‘hard’ as one might have thought and cannot yet be taken as proof for the type of conclusions that are often formulated regarding the non-existence of free will. What possible type of relation the RP has with the act, will be discussed in Section 8.
Section snippets
Background
Research in cognitive neuroscience (CNS) often encounters difficulties that could be characterized by the phrase ‘hard data, soft concepts’. Despite a plethora of increasingly detailed and sophisticated measurements of brain activity (‘hard data’), their relationship with the cognitive processes of interest often is a continuous matter of debate, due to the potential variation in meaning of the cognitive terminology (‘soft concepts’). This issue is especially prominent in research concerning
Experimental methods
In order to replicate the results found by Libet et al. (1983), we designed a similar experiment, but used a slightly different task. In our experiment, similar to that of Soon et al., 2008, Haggard and Eimer, 1999 and Trevena and Miller (2002), participants have to press either a left or a right button with their left or right thumb respectively, instead of making quick movements with only their right hand. This altered task makes it possible to investigate not only the when-component of an
Preliminary EEG analysis
In order to analyze the individual trials in the recorded EEG, the raw data was sliced in trials of 7 s around each button press (from 4 s before each button press until 3 s after). Because we recorded both left and right hand responses, the data of each trial was labelled either as a left or a right button response. Furthermore, the raw EEG data was down sampled from 2048 Hz to 128 Hz.
Only the data recorded in the testing phase was analysed. Furthermore, we split the data in sound trials (10% of
Results
Table 1 contains the following results for each participant, the average over these individual participant results and the grand average:
- •
The number of left (L) and right (R) hand responses.
- •
The three sets of electrodes used to calculate the RP onset.
- •
The mean onset of the intention to act.
- •
The mean RP onset measured by eye over the three different sets of electrodes.
- •
The mean RP onset measured using the 90% area method over the three different sets of electrodes.
- •
The mean RP onset measured using a t
The problem of semi hard data and soft concepts
We did find that the RP precedes the intention to act when looking at the average over all participants, but do these findings accurately reflect the patterns in the data of the individual participants? Do they reflect what occurs in individual trials? In this section we will describe that although we were able to successfully replicate the results of Libet et al. (1983), the relation of this semi hard data (the measured RP onset) to the underlying soft concepts (the onset of the intention to
A small thought experiment
The results for participant 5 are likely due to a misunderstanding of the task at hand by the participant. However, what would be the consequences if these results were in fact correct? We would like to reflect briefly on this scenario. It would mean that we found an opposite pattern of results compared to that of Libet et al.; for this participant, the average onset of the intention to act was reported prior to the average onset of the RP. The conclusion that is commonly drawn from Libet-type
Conclusion
We were able to replicate the results found by Libet et al. (1983): the RP we measured preceded the intention to act by 544 ms and the button press by 700 ms on average over all participants measured with the 90% area measure over Cz. Using the best set of electrodes instead of Cz resulted in similar onsets: an average RP onset of 688 ms before a button press. Measuring the RP over Cz, C3 and C4 resulted in somewhat later onsets: an average RP onset of 535 ms before a button press. Overall, the eye
References (46)
- et al.
The neural encoding of guesses in the human brain
Neuroimage
(2012) - et al.
Internally generated preactivation of single neurons in human medial frontal cortex predicts volition
Neuron
(2011) The timing of conscious states
Consciousness and Cognition
(2002)- et al.
What is the bereitschaftspotential?
Clinical Neurophysiology
(2006) - et al.
What is folk psychology?
Cognition
(1994) - et al.
Cortical movement preparation before and after a conscious decision to move
Consciousness and Cognition
(2002) - et al.
Brain preparation before a voluntary action: Evidence against unconscious movement initiation
Consciousness and Cognition
(2010) - et al.
Why we may not find intentions in the brain
Neuropsychologia
(2014) - et al.
The Berlin Brain–Computer Interface: EEG-based communication without subject training
IEEE Transactions on Neural Systems and Rehabilitation Engineering
(2006) On a confusion about a function of consciousness