User:Hakwan Lau/Proposed/Awareness of intention

• Dr. Hakwan Lau, Columbia University, New York, NY

• Dr. Brian Maniscalco, Columbia University, New York, NY

• Dr. Erin Christine Shaver, Columbia University, New York, NY

• Ms. Tashina Graves, Research Assistant, Psychology Department, Columbia University, NY, USA

In everyday life it seems as if we have straightforward access to, and control of, our intentions and associated behaviors. However, research in the fields of psychology and cognitive neuroscience has suggested that things are not as they seem. In particular, the causal role of conscious intention in controlling behavior, and the means by which we attribute agency for actions, have been brought under scrutiny, with some surprising results.

Discovery of the readiness potential (RP)

Figure 1: Readiness potential (negative plotted up)

One of the most surprising scientific discoveries related to the awareness of intention is the readiness potential (RP). Deecke and Kornhuber (1965) were interested in analyzing brain activity preceding voluntary movements. They asked subjects to perform self-paced finger movements while recording the subjects' electroencephologram (EEG). By averaging many segments of EEG activity in the time window preceding each finger movement, they discovered a distinct electrophysiological potential recorded at the vertex of the scalp reliably preceding self-paced movements. They named this event-related potential the bereitschaftspotential (German), also known subsequently as the readiness potential, and BP or RP for short. The RP (as mentioned above) is a slow, progressively negative electrophysiological potential whose onset occurs roughly one second before a spontaneously initiated action. The exact onset time of the RP depends on the nature of the task being performed, but it is always earlier than one would otherwise intuitively expect: one seems to perform spontaneous self-paced movements without much awareness of their preparations, but the brain somehow seems to be preparing these for a much longer time. Deecke and Kornhuber's demonstration of the RP was the first evidence that endogenously generated actions are preceded and perhaps initiated by neural activity.

Characteristics of the RP

Libet et al. (1982) recorded EEG while asking subjects to perform two different kinds of voluntary tasks. In one condition, subjects observed a clock hand approach a preset time on the clock face. When the preset time was reached, the subjects were instructed to press a button. In another condition, subjects were encouraged to make entirely spontaneous movements, without any pre-planning. Libet et al. found that an RP preceded both pre-set and spontaneous finger movements (albeit with different onset times and morphologies). Thus, the RP does not correspond to spontaneous movement per se, but is rather associated with movements that can be prepared in advance of any signal to move.

Libet et al. (1983) showed that the RP is also generated when subjects intend to press a button at a pre-set time, but "vetoes" this intention at the last instant. This finding demonstrates that the RP is not sufficient for action generation, but rather reflects processing related to action preparation.

Keller and Heckhausen (1990) further dissociated the RP from conscious intention. They recorded EEG and EMG while subjects performed a demanding cognitive task. Whenever a muscle movement registered on the EMG, the subject was interrupted from performing the cognitive task and asked about the nature of the movement: Was it spontaneous or preplanned? If spontaneous, was the subject aware of an urge to move before moving? In this way, it was possible to collect data on unconscious as well as consciously intended spontaneous movements in a somewhat naturalistic manner. Keller and Heckhausen found that preceding both unconscious and consciously intended spontaneous movements, there exists some form of RP. Thus, strictly speaking, the existence of an RP does not necessarily reflect conscious intention. However, Keller and Heckhausen also found that the distributions and intensities of the two forms of RP's differ, in that RP's preceding consciously initiated actions are more focused along the midline.

Taken together, these experiments suggest that the RP indexes a process of preparation for endogenously generated action that may occur even in the absence of conscious intention. However, conscious intention may modulate or intensify the RP, especially the medial component. There is evidence that one source of the RP that is likely to be related to conscious intention is located medially (see section 5). It is possible that the intention signal in that area needs to reach a certain threshold for us to become aware of the intention.

Libet's clock paradigm: RP precedes intention

Libet et al. (1983) reported seminal data that bear on the suggestion the RP may need to reach a certain intensity for us to become aware of our intentions. They devised a paradigm to ascertain the relative onsets of the RP and experienced intention relative to a subsequent movement. In the experimental condition, subjects viewed a dot revolving around a clock face as their EEG was recorded. They were instructed to wait for one full revolution of the dot (2.56 seconds) and then perform a well-defined finger movement whenever the urge to move spontaneously arose. Afterwards, subjects reported either the earliest time at which they felt an urge to move, or the time at which they actually did move. They made these time judgments by reference to the position of the dot on the clock face when the relevant event (urge or movement) occurred. In a control condition, subjects used the rotating dot to estimate the onset of a tactile stimulus applied to the hand at a random point in time.

Different methods of estimating the onset of the RP, and different methods of gathering subjects' estimates of the onsets of intention and action, provided converging results. The RP occurred about 550 ms prior to movement (where movement onset was measured by an EMG affixed to the moving hand). Subjects reported intention onset to occur at about 200 ms prior to movement, and surprisingly, movement was judged to occur about 85 ms prior to actual movement. In the control condition, tactile stimulus onset was reported to occur about 45 ms prior to actual stimulus onset. So, even when using the tactile stimulation condition as a control for latency judgments in this "Libet clock" paradigm, subjects' reported onset of intention occurred well after the onset of the RP.

Libet interpreted these results to mean that conscious intention is not causally responsible for initiating movements. If it were, then it should occur prior to, or simultaneously with, the cerebral precursors of movement. However, Libet argued that the data were still consistent with conscious intention having a causal role. Conscious processes associated with intention could not initiate a movement, but they could determine whether the unconsciously initiated movement would ultimately find behavioral expression. In other words, a "conscious veto" could exert inhibitory conscious control over unconsciously initiated behavior.

Later research by Haggard and Eimer (1999) refined Libet's results. They replicated Libet's main findings and then compared these to a "free selection" condition in which subjects could freely choose to spontaneously move either their left or right fingers. They reasoned that if intention onset occurs prior to action selection, and if action selection is a time consuming step in a serial process, then intention onset should occur earlier (relative to subsequent movement) when subjects are allowed to choose between possible actions. However, they found that allowing subjects to choose which finger to move did not change average reported onset of intention. Thus they concluded that intention occurs after action selection and so is associated with preparing for a specific movement. In addition to analyzing the RP, they also analyzed a related ERP component called the lateralized readiness potential (LRP) which provides a more specific index to motor preparation (Eimer 1998). They found that the LRP exhibits temporal correlation with intention onset whereas the RP does not. This finding suggests that the LRP is more closely related to the generation of intention than is the RP, giving further evidence that intention onset is associated with specific motor plans.

Doubt about the conscious veto

Although Libet's findings seemed to rule out a causal role for intention in initiating behavior, he argued they were consistent with intention playing the causal role of deciding which unconsciously initiated movements ultimately found behavioral expression. In the 200 ms between intention onset and subsequent action, one could consciously choose to suppress the action. However, later research has shown that the window between intention onset and action may be even smaller than 200 ms, potentially marginalizing any causal control that conscious intention could hold over action.

Lau et al. (2006) used functional magnetic resonance imaging (fMRI) to image the brains of subjects as they performed the classical Libet task. They found that the cingulate motor area (CMA) was activated when judging time of motor onset. But CMA activation was also found to be correlated with more distorted perceptions of motor onset. The more a subject's CMA was activated while attending to motor onset, the earlier in time the subject reported having made the movement, relative to the actual time of movement. Subjects who showed no CMA activation also showed no motor timing bias.

In previous data, Lau et al. (2004a) found that the pre-supplementary motor area (pre-SMA) was activated during the task of estimating intention onset in the Libet clock paradigm, just as CMA was activated while timing movement. This activation was correlated with reports of intention onset similarly to the way CMA activation was correlated with motor timing, in that greater pre-SMA activation was associated with earlier estimated onsets of intention. Based on these similarities, it seems natural to suggest that pre-SMA activation may bias intention onset judgments just as CMA activation biases motor onset judgments. If so, then reports that intention onset occurred 228 ms prior to movement in (Lau et al 2004a) may have been systematically biased by pre-SMA activation while attending to intention, and veridical intention onset may have occurred as late as 120 ms prior to action.

More extreme examples of late onset of intention are reported in Sirigu et al (2004). They found that patients with a lesion to an area in the parietal cortex reported intention onset as late as 50 ms prior to action. If the 200 ms between intention onset and action execution is important for suppression of undesirable urges, we should expect these patients to have major problems controlling their behavior. But this was not what has been reported.

It seems that the creation of the intention experience may even extend to the period following movement, as reported in Lau et al (2007). They found that in the Libet clock paradigm, judgments of intention onset could be altered as late as 200 ms after the subject's action by means of transcranial magnetic stimulation (TMS). This implies that at least some aspects of intention experience are constructed during or immediately after the corresponding action.

If the experience of intention has causal control over behavior, there should be a window of time between intention onset and action during which intention can exert its influence. This window is reported to be about 200 ms long by subjects, but in fact it may be as small as 120 ms. Even 50 ms durations between intention and action can exist without any discernible effects on behavior. And even 200 ms after an action is executed, the experience of intention is still being constructed in the brain. These observations provide converging evidence to suggest that intention experience may not subserve a "conscious veto" function.

Neural correlates of intention

Figure 2: Attention to intention

The pre-motor area located on the medial frontal wall is likely to be involved in the representation of intention for spontaneous actions. These include the pre-supplementary motor area (pre-SMA) and the supplementary motor area proper (SMA proper); sometimes the two areas are collectively called the SMA. Fried et al. (1991) found that electrically stimulating the SMA can cause human subjects to report the urge to make specific movements. Thaler et al. (1995) found that monkeys with lesioned SMA can't learn to make spontaneous movements, although they can learn to make cued movements.

Lau et al. (2004a) used fMRI to image subjects' brain activity as they performed the Libet task. They found that when subjects make judgments about intention onset, there is activation in the pre-SMA, right dorsal prefrontal cortex (DPFC), and left intraparietal sulcus (IPS). The pre-SMA and right DPFC also show enhanced functional connectivity during this task, in that their activity is better correlated when making intention onset judgments than when making movement onset judgments. The authors interpreted this network to reflect attention (reflected by a process carried out in the DPFC, and via its interaction with other motor) to intention (represented in pre-SMA) during the intention onset judgment task, although it was not clear whether the DPFC was driving SMA activation or merely monitoring it, or both. This finding is also consistent with previous results that the pre-SMA is involved in the generation of free choice (Lau et al 2004b).

Additionally, several EEG source modeling studies (Ball et al. 1999, Erdler et al. 2000) and high temporal resolution fMRI studies (Weilke et al. 2001, Cunnington et al. 2003) have produced results consistent with the SMA being a neural generator for the readiness potential.

Cued intention

Outside of the context of spontaneous action, in the literature the word 'intention' is sometimes used to refer to a form of motor preparation (e.g. Snyder et al 2000, Kalaska and Crammond, 1995, Toni et al 2001). In particular, such preparation should be specific to the movement type (i.e. not just general attention to the target of action), but not confounded with the actual execution of the movement. Typically, such 'intention' activity is related to the parietal cortex. This type of intention is different from what we have focused on in this article, as those actions are essentially cued and determined by an external stimulus, i.e. they are not spontaneously generated.

Intentional Binding

Others have studied the effects of conscious intentions on the perceived timing of subsequent events. Haggard et al. (2002) investigated how various representations of motor action (e.g. those involved in motor preparation, motor output, and sensory feedback from motor output) are bound together to produce one coherent experience of voluntary action. In a set of baseline conditions, they asked subjects to use a rotating clock hand to judge the time onset of: (i) voluntary movements; (ii) movements induced by TMS to the motor cortex; (iii) the clicking noise of a sham TMS directed at the parietal cortex; and (iv) a tone. In the experimental conditions, a tone was presented 250 ms after voluntary movements, TMS-induced movements, and sham TMS.

They found that for voluntary movements, the judged onset of movement was later than in the baseline condition, while the judged onset of the tone was earlier than baseline, as if the perception of the two events were "attracted" in time. For TMS-induced movements, the opposite occurred; the perception of the movement and subsequent tone were "repelled" in time, relative to baseline. For sham TMS trials, judgment of sham TMS and tone onset did not differ from baseline.

The authors interpreted these results to reveal an "intentional binding" effect. When one's behavior is experienced as causing a subsequent event, the two events are intentionally bound. That is, the behavior and the subsequent event are perceived to occur closer together in time. The opposite effect holds when behavior and subsequent event are not plausibly linked by causal intention. That is, they are perceived to occur farther apart in time, as if to signal explicitly that the behavior was not an intentional cause of the event.

Ownership of action

It seems intuitive that we have privileged access to the authorship of our own actions. That is, we have ready access to certain knowledge about which actions we have and have not caused. However, work by Wegner et al. (2003) investigating "facilitated communication" casts this proposition into doubt. Subjects (playing the role of "facilitators") were asked to place their fingers on two keys of a keyboard, while a confederate (playing the role of "communicator") placed his or her fingers on top of those of the subject. Subjects were given headphones with which they listened to questions of varying difficulty. Confederates were given headphones as well, and subjects were led to believe that the confederates would be hearing the same questions, although in fact the confederates heard nothing. Subjects were told to detect subtle, unconscious movements in the confederate's fingers following each question. When such movements were detected, the subject should press the corresponding key in order to answer on the confederate's behalf.

It was found that subjects answered easy questions well above chance levels. If they had performed the task strictly according to the instructions, however, they should have performed at chance. Therefore, subjects must have been directing their own key presses. Nonetheless, they attributed a significant causal role for the key presses to the confederate. The degree to which subjects answered easy questions correctly was not correlated with the degree to which they attributed causal responsibility to confederates, suggesting that the generation of action ("action production") and attribution of action to an agent ("action projection") are independent processes.

In a subsequent experiment, Wegner et al. manipulated subjects' beliefs regarding the efficacy of facilitated communication. One group of subjects was led to believe that facilitated communication was a verified technique while another group was led to believe that it was a discredited technique. The group led to believe that facilitated communication works well attributed more causal responsibility for the key presses to confederates, although even the skeptical group attributed some causal agency to the confederates. This action projection effect held even when subjects had no direct contact with confederates, but were told to empathize with them at a distance to try to discern their responses to the questions.

From these results Wegner et al. concluded that the process by which we attribute causal responsibility for our actions is independent from the means by which we actually produce actions. The two seem to run in parallel rather than being inextricably linked. Furthermore, they concluded that action attributions are not foolproof but malleable and vulnerable to error. When there are plausible candidates for the causes of one's actions other than one's own self, confusion over authorship may ensue. The plausibility of other agents having caused one's own actions, and hence the potential for authorship confusion, is a function of one's prior beliefs.

These results compliment what has been discovered in neuroscience regarding the function and awareness of intention (see section 4), in that they similarly support the notion that one's sense of agency is not being determined 'online'. It depends on many other factors, many taking place before or after the action. This is why one's sense of agency is manipulable in many different ways.

Many patients with a diagnosis of schizophrenia report that their actions are not caused by themselves, but by alien forces, a symptom known as a delusion of control. This symptom no longer appears so bizarre, given the demonstrations above that everybody is vulnerable to errors of action attribution. Such symptoms may result from a combination of deviant sensory experiences with abnormal prior beliefs (Frith, 2005).

References

• Kalaska, J. F., and Crammond, D. J. 1995. Deciding not to GO: Neuronal correlates of response selection in a GO/NOGO task in primate premotor and parietal cortex. Cerebral Cortex. 5: 410–428.

• Kornhuber HH, Deecke L. Hirnpotentialänderungen bei Willkurbewegungen und passiven Bewegungen des Menschen: Bereitschaftspotential und reafferente Potentiale. Pflugers Archive. 1965;284:1–17

• Libet B, Wright EW Jr, Gleason CA. Readiness-potentials preceding unrestricted 'spontaneous' vs. pre-planned voluntary acts. Electroencephalogram Clinical Neurophysiology. 1982 Sep;54(3):322-35.

• Libet B, Wright EW Jr, Gleason CA. Preparation- or intention-to-act, in relation to pre-event potentials recorded at the vertex. Electroencephalogram Clinical Neurophysiology. 1983 Oct;56(4):367-72.

• Keller I, Heckhausen H. Readiness potentials preceding spontaneous motor acts: voluntary vs. involuntary control. Electroencephalogram Clinical Neurophysiology. 1990 Oct;76(4):351-61.

• Libet B, Gleason CA, Wright EW, Pearl DK. Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential). The unconscious initiation of a freely voluntary act. Brain. 1983 Sep;106 (Pt 3):623-42.

• Libet, B. (1985). Unconscious cerebral initiative and the role of conscious will in voluntary action. Behavioral and Brain Sciences, 8, 529-566.

• Haggard P, Eimer M. On the relation between brain potentials and the awareness of voluntary movements. Exp Brain Res. 1999 May;126(1):128-33.

• Eimer M (1998) The lateralized readiness potential as an on-line measure of selective response activation. Behavioral Research Methods Instrum Comput 30:146–156

• Haggard P, Clark S, Kalogeras J. Voluntary action and conscious awareness. Nature Neuroscience. 2002 Apr;5(4):382-5.

• Lau HC, Rogers RD, Passingham RE. On measuring the perceived onsets of spontaneous actions. J Neuroscience. 2006 Jul 5;26(27):7265-71.

• Lau HC, Rogers RD, Passingham RE. Manipulating the experienced onset of intention after action execution. J Cognitive Neuroscience. 2007 Jan;19(1):81-90.

• Lau HC, Rogers RD, Ramnani N, Passingham RE. Willed action and attention to the selection of action. Neuroimage. 2004b Apr;21(4):1407-15.

• Lau HC, Rogers RD, Haggard P, Passingham RE. Attention to intention. Science. 2004a Feb 20;303(5661):1208-10.

• Fried I, Katz A, McCarthy G, Sass KJ, Williamson P, Spencer SS, Spencer DD. Functional organization of human supplementary motor cortex studied by electrical stimulation. J Neuroscience. 1991 Nov;11(11):3656-66.

• Thaler, D., Chen, Y.C., Nixon, P.D., Stern, C.E., Passingham, R.E., 1995. The functions of the medial premotor cortex. I. Simple learned movements. Exp. Brain Res. 102, 445– 460.

• Ball, T., Schreiber, A., Feige, B., Wagner, M., Lu¨cking, C.H., Kristeva-Feige, R., 1999. The role of higher-order motor areas in voluntary movement as revealed by high-resolution EEG and fMRI. NeuroImage 10, 682–694.

• Erdler, M., Beisteiner, R., Mayer, D., Kaindl, T., Edward, V., Windischberger, C., Lindinger, G., Deecke, L., 2000. Supplementary motor area activation preceding voluntary movement is detectable with a wholescalp magnetoencephalographic system. NeuroImage 11, 697–707.

• Weilke, F., Spiegel, S., Boecker, H., Gra¨fin von Einsiedel, H., Conrad, B., Schwaiger, M., Erhard, P., 2001. Time-resolved fMRI of activation patterns in M1 and SMA during complex voluntary movement. J. Neurophysiology. 85, 1858–1863.

• Cunnington R, Windischberger C, Deecke L, Moser E. The preparation and readiness for voluntary movement: a high-field event-related fMRI study of the Bereitschafts-BOLD response. Neuroimage. 2003 Sep;20(1):404-12.

• Haggard P, Clark S, Kalogeras J. Voluntary action and conscious awareness. Nature Neuroscience. 2002 Apr;5(4):382-5.

• Wegner DM, Fuller VA, Sparrow B. Clever hands: uncontrolled intelligence in facilitated communication. J Personality Social Psychology. 2003 Jul;85(1):5-19.

• Wegner DM, Sparrow B, Winerman L. Vicarious agency: experiencing control over the movements of others. J Personality and Social Psychology. 2004 Jun;86(6):838-48.

• Sirigu A, Daprati E, Ciancia S, Giraux P, Nighoghossian N, Posada A, Haggard P. Altered awareness of voluntary action after damage to the parietal cortex. Nature Neuroscience. 2004 Jan;7(1):80-4.

• Snyder LH, Batista AP, Andersen RA. Intention-related activity in the posterior parietal cortex: a review. Vision Res. 2000;40(10-12):1433-41.

• Toni I, Thoenissen D, Zilles K. Movement preparation and motor intention. Neuroimage. 2001 Jul;14(1 Pt 2):S110-7.

• Frith C D. The self in action: lessons from delusions of control. Conscious Cogn. 2005 Dec;14(4):752-70.

See also

Action selection, Binding by Synchrony, Electroencephalogram, Functional Imaging, Functional Magnetic Resonance Imaging, Intentionality, Models of Consciousness