THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8, 529-566 Printed in the United States of America Unconscious cerebral initiative and the role of conscious will in voluntary action Benjamin Libet Department of Physiology, School of Medicine, University of California, San Francisco, Calif. 94143 Abstract: Voluntary acts are preceded by electrophysiological "readiness potentials" (RPs). \Vith spontaneous acts involving no preplanning, the main negative RP shift begins at about -550 ms. Such RPs were used to indicate the minimum onset times for the cerebral activity that precedes a fully endogenous voluntary act. The time of conscious intention to act was obtained from the subject's recall of the spatial clock position of a revolving spot at the time of his initial awareness of intending or wanting to move (\V). \V occurred at about -200 ms. Control experiments, in which a skin stimulus was timed (S), helped evaluate each subject's error in reporting the clock times for awareness of any perceived event. For spontaneous voluntary acts, RP onset preceded the uncorrected Ws by about 350 ms and the \Vs corrected for S by about 400 ms. The direction of this difference was consistent and significant throughout, regardless of which of several measures ofRP onset or W were used. It was concluded that cerebral initiation of a spontaneous voluntary act begins unconsciously. However, it was found that the final decision to act could still be consciously controlled during the 150 ms or so remaining after the specific conscious intention appears. Subjects can in fact "veto" motor performance during a lOO-200-ms period before a prearranged time to act. The role of conscious will would be not to initiate a specific voluntary act but rather to select and control volitional outcome. It is proposed that conscious will can function in a permissive fashion, either to permit or to prevent the motor implementation of the intention to act that arises unconsciously. Alternatively, there may be the need for a conscious activation or triggering, without which the final motor output would not follow the unconscious cerebral initiating and preparatory processes. Keywords: conscious volition; event-related chronometry; free will; mental timing; motor organization; readiness potentials; unconscious processes; voluntary action One of the mysteries in the mind-brain relationship is expressed in the question: How does a voluntary act arise in relation to the cerebral processes that mediate it? The discovery of the ureadiness potential" (RP) opened up possibilities for experimentally addressing a crucial feature of this question. The RP is a scalp-recorded slow negative shift in electrical potential generated by the brain and beginning up to a second or more before a self-paced, apparently voluntary motor act (Deecke, Grozinger & Kornhuber 1976; Gilden, Vaughan & Costa 1966; Kornhuber & Deecke 1965). The long time interval (averaging about 800 ms) by which RP onset preceded a self-paced act raises the crucial question whether the conscious awareness of the voluntary urge to act likewise appears so far in advance. If a conscious intention or decision to act actually initiates a voluntary event, then the subjective experience of this intention should precede or at least coincide with the onset of the specific cerebral processes that mediate the act. This issue has recently been subjected to experimental tests and analyses, which I shall review briefly (Libet, Gleason, \Vright & Pearl 1983; Libet, \Nright & Gleason 1982; 1983). The experimental findings led us to the conclusion that voluntary acts can be initiated by unconscious cerebral processes before conscious intention appears but that conscious control over the actual motor performance of the acts remains possible. I shall discuss these conclusions and their implications for concepts of "the unconscious" and of conscious voluntary action. I propose the thesis that conscious volitional control may operate not to initiate the volitional process but to select and control it, either by permitting or triggering the final motor outcome of the unconsciously initiated process or by vetoing the progression to actual motor activation. (The reader is referred to our original cited research papers for the full details of the experimental techniques and observations together with their evaluation, etc.) o140-525XIB5/040529-3B/$06.00 529 © 1985 Cambridge University Press 1. Definitions of voluntary action and will Since the meanings assigned to the terms "voluntary action" and "will" can be quite complicated and are often related to one's philosophical biases, I shall attempt to clarify their usage here. In this experimental investigation and its analysis an act is regarded as voluntary and a function of the subject's will when (a) it arises endogenously, not in direct response to an external stimulus or cue; (b) there are no externally imposed restrictions or compulsions that directly or immediately control subjects' initiation and performance of the act; and (c) most important, subjects feel introspectively that they are WrEn, Libet: Cerebral processes and volition performing the act on their own initiative and that they act as wish. The are free to start or not to start significance of point (c) is sharply Illustrated m the case of stimulating the motor cortex (precentral gyrus) in awake human subjects. As described by Penfield (1958) and noted by others, under these conditions each subject regarded the motor action resulting from cortical stimulation as something done to him by some external force; every subject felt that, in contrast to his normal voluntary activities, "he," as a self-conscious entity, had not initiated or controlled the cortically stimulated act. The technical requirements of experiments do impose limits on the kinds of voluntary choices and settings available to the subject. The nature of the acts must be prescribed by the experimenter. In the studies to be discussed here the acts were to consist uniformly of a quick flexion of the fingers or wrist of the right hand; this yielded a sharply rising electromyogram (EMG) in the appropriate muscle to serve as a trigger for O-reference time. The subjects were free, however, to choose to perform this act at any time the desire, urge, decision, and will should arise in them. (They were also free not to act out any given urge or initial decision to act; and each subject indeed reported frequent instances of such aborted intentions.) The freedom of the subject to act at the time of his chOOSing actually provides the crucial element in this study. The objective was in fact to compare the time of onset of the conscious intention to act and the time of onset of associated cerebral processes. The specific choice of what act to perform was not material to the question being asked. Volitional processes may operate at various levels of organi7.ation and timing relative to the voluntary act. These may include consciously deliberating alternative choices as to what to do and when, whether or not to act, whether or not to comply with external orders or instructions to act, and so on. If any of these processes are to result in the motor perfonnance of a voluntary act, they must somehow work their way into a "final common motor activation pathway" in the brain. \Vithout an overt motor perfonnance any volitional deliberation, chOOSing, or planning may be interesting for its mental or psychological content, but it does not constitute voluntary action. It is specifically this overt performance of the act that was experimentally studied by us. In the present experimental paradigm subjects agree to comply with a variety of instructions from the experimenter. One of these is an expectation that the subject is to perfonn the prescribed motor act at some time after the start of each trial; another is that he should pay close introspective attention to the instant of the onset of the urge, desire, or decision to perfi)nn each such act and to the correlated spatial position of a revolving spot on a clock face "dock time"). The suhject is also instructed to allO\v each such act to 'lrise .. spon taneously,·' \\;thout deliberately planning or paying attention to the "prospect" of acting in advance. The subjects did indeed report that the inclination for each act appeared spontaneously Cout of nowhere"), that they were consciouslyaware of their urge or decision to act hefore each act that thev felt in conscious control of whether or not to act: and that they felt no external or psychological pressures that affected the time when they decided to act (Libet et a1. 1982; Libet, Gleason, \Vright & Pearl 1983). 530 THE BEHAVIORAL AND BRAIN SCIENCES {1985} 8:4 Thus in spite of the experimental requirements, the basic' conditions set out above li>r a voluntary act were met. Conditions for the subject's decision as to when t? act were designated to represent those one could aSS?Clate with a conscious, endogenously willed motor so that one could study the cerebral processes involved m such an act without confusing them with deliberative or preparatory features that do not necessarily result in action. Finally, one should note that the voluntary studied was defined operationally, including appropnate and reliable reports of introspective experiences. The definition is not committed to or dependent upon any specific philosophical view of the mind-brain relato tionship. However, some implications that are mind-brain theories will be drawn from the findmgs. 2. Cerebral processes precede conscious intention Two experimental issues have to be resolved in order to obtain a relevant answer to the questions about the relative timing of conscious intentions and cerebral processes in the performance of voluntary acts: (1) Is .the RP a valid indicator of cerebral processes that medIate voluntary acts? (2) How can one meaningfully measure the onset of the conscious intention, urge, or will -to perform a specific voluntary motor act? 2.1. RPs in voluntary acts Self-paced acts were used in the discovery ofRPs et al. 1966; Kornhuber & Deecke 1965) and in subsequent RP studies (e.g., Deecke et al. 1976; Shibasaki, Barrett, Halliday & Halliday 1980; Vaughan, Costa & Ritter 1968). Such acts have features that may compromise the exercise offree volition or confuse its interpretation: (a) Recording an RP requires averaging many events. When these self-paced acts are repeated in a continuous series, with irregular intervening intervals of 3-6 sec as selected by the subject, they become boring and may come to be performed in a stereotyped and almost automatic way, with no assurance that conscious control could be exercised in each trial. (b) Since subjects were asked to act within an allotted time interval, they may be under pressure consciously or unconsciously to plan to act within the time limit; that is, the subject's voluntary choice of when to act may be compromised by an external requirement. (c) Subjects are required not to blink until just after each act. The need to blink may impel the subject to act, thus serving as an external controlling factor. In a study of what we tenned "self-initiated" acts, these external forces were minimized or eliminated (Libet et aI. 1982). Each trial in an averaging series of 40 trials was initiated as a separate independent event after a flexible delay detennined by each subject's own readiness to proceed; there was no limit on the time in which subjects were to act; they were given the option to blink if necessary. For each trial, subjects were asked to perform a simple quick flexion of the wrist or fingers at any time they felt the "urge" or desire to do so; timing was to be entirely "ad lib," that is, spontaneous and fully endoge- Libet: Cerebral processes and volition RPI s RPll I m ,;J" .JiI f• .. r.. .. ········r s.s. I I RP n .. G.L. Cc S.B. I I \., . /jl A RP I ........ f · . ..... . • • ••••• , . • • • • • • - I •• I •••••••••••• ... I , ... • • • •,......... .. • r_ ""-', !\ r· " I :: ': uV 1000 ms + ! •••••• :-:. .... ,- '- - . -- • I"t .v .. : .. , . . --. ......... ...;,r...... CZ Cc t IV .\ AI! ... i .. MN _ C z •. .. Cc B.D. I ..... .. -.. - ........ •••• -j V +300 t Figure 1. Readiness potentials (RP) preceding self-initiated voluntary acts. Each horizontal row is the computer-averaged potential for 40 trials, recorded by a DC system with an active electrode on the scalp, either at the midline-vertex (C) or on the left side (contralateral to the performing right hand) approximately over the motor/premotor cortical area that controls the hand (CJ \Vhen every self-initiated quick flexion of the right hand (fingers or wrist) in the series of 40 trials was (reported as having been) subjectively experienced to originate spontaneously and with no preplanning by the subject, RPs labeled type II were found in association. (Arrowheads labeled N indicate onset of the" main negative" phase of the vertex recorded type II RPs in this figure; see Libet et al. 1982. Onsets were also measured for 90% of the total area of RP; see Table 1B). \Vhen an awareness of a general intention or preplanning to act some time within the next second or so was reported to have occurred before some of the 40 acts in the series, type I RPs were recorded (Libet et al. 1982). In the last column, labeled S, a near-threshold skin stimulus was applied in each of the 40 trials at a randomized time unknown to the suhject, with no motor act perfonned; the subject was asked to recall and report the time when he became aware of each stimulus in the same way he reported the time of awareness of wanting to move in the case of self-initiated motor acts. The solid vertical line through each column represents 0 time, at which the electromyogram of the activated muscle begins in the case of RP series, or at which the stimulus was actually delivered in the case of S series. The dashed horizontal line represents the DC baseline drift. For subject S.S., the first RP (type I) was recorded hefore the instruction "to let the urge come on its own. spontaneously" was introduced; the second RP (type II) was obtained after giving this instruction in the same session as the first. For suhjects C.L., S. B., and B. D., this instruction was given at the start ofall sessions. Nevertheless, each of these subjects reported some experiences of loose preplanning in some of the 40-trial series; those series exhibited type I RPs rather than type II. Note that a slow negative of self-initiated acts (RP) does not precede the skin stimulus in S series. However, shift in scalp potential that precedes evoked potentials following the stimulus are seen regularly to exhibit a large positive component with a peak close to +300 ms (arrow indicates this time); this P300 event-related potential had been shown by others to be associated with decisions about uncertain events (in this case, the time of the randomly delivered stimulus), and it also indicates that the subject is attending well to the experimental conditions. (Modified from Libet et al. 1982.) THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 531 Libet: Cerebral processes and volition nous. (For full technical details see Libet et al. 1982; Libet, Gleason, Wright & Pearl 1983.) Subjects reported that they were aware of the urge or intention to move before every act in the series; that is, the acts were not automatic or involuntary "tics." The absence of any larger meaning in this act appears to exclude external logical or other factors as controlling agents. Acts of thIs kind may thus be taken as paradigmatic examples of unrestricted volition, at least in regard to choosing when to act. The basic initiating process for these simpler volitional acts may be the same as that for the actual motor expression of other, more complex forms of voluntary action, since the latter are manifested behaviorally only when final decisions to move have been made. These self-initiated, endogenous acts were indeed found to be preceded by RPs (Libet et al. 1982). \Vhen all 40 self-initiated acts in an averaging series were performed with this spontaneous ad lib timing, with no reports of specific preplanning to act, the recordable averaged RP generally had an onset for its main negative rise at about 550 (±150) ms before the motor act began; these were called "type If' RPs (see Figure 1). (As is customary, the beginning of the muscle activity is signaled by the onset of the electromyogram, EMG, recorded at an appropriate muscle. This provides the "0time" trigger for averaging the preceding scalp potential at the vertex and for other timing features.) In some trials, subjects did report experiencing some general preplanning or preparation to act in the near future a few seconds before the act, despite the encouragement to be completely spontaneous. These occurrences were reported during the "debriefing" conducted at the end of each series of 40 trials. In those series that included even a small Humber of such reported experiences, a ramplike RP with onset at about -1050 ms (±175) was typically recorded (the "type I" RPs, Figure 1); these RPs were called type I because they resembled those RPs previously described for self-paced acts (e.g., Deecke et al. 1976). However, subjects all insisted that the more specific urge or intention to perform the actual movement was still experienced just before each act in a type I series, just as in the type II series; and they clearly distinguished this urge or intention from any advance feelings of preplanning to move within the next few seconds. In other experiments that required deliberate preplanning by instructing the subject to act at a preset time, there appeared a large ramplike RP that resembled the type I RP of our self-initiated acts. \Ve concluded, therefore, that the RP component that starts at about -550 ms, the one that predominates in type II RPs recorded when all acts in a 40-trial series are spontaneous, is the one uniquely associated with an exclusively endogenolls volitional process. The latter process is distinguished from a looser preintentionality or general preparation-to-act-soon that is not necessarily endogenous (Libet et al. 1982). 2.2. Timing the conscious intention to act It presented a difficult challenge to devise the operational criteria for determining the time at which the subjects become aware of wanting or deciding to act. One begins with the premise that this subjective event is only accessible introspectively to the subject himself; some k;nd of 532 THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 report of this by the subject is therefore a requirement (Libet 1966; 1973; 1981b). Conscious subjective experience in this case an awareness of the endogenous urge or to move, is a primary phenomenon; it cannot be defined in an a priori way by recourse to any externally observable physical event, including any. action not directly representative of the subject s mtrospective report (Beloff 1962; Creutzfeldt & Rager 1978; Eccles 1980; Libet 1965; 1966; 1981a; 1981b; Nagel 1979: Popper & Eccles 1977; Thorpe 1974). The report, whether a verbal one or some other motor indication (e.g., pressing an answer key), cannot be an immediate one made as soon as the conscious experience has occurred: (a) Cerebral preparations for the motor action of reporting might introduce some confusing RPs of their own. (b) There could be a substantial delay for neurally organizing and achieving the motor actions required to make the report. (c) \Vhen a premium is put on the speediness of a response, as in measuring reaction time to a stimulus, there is no assurance that the motor response directly indicates when an actual subjective experience has occurred. The fast response to a stimulus can represent an unconscious mental process; but when the subject becomes consciously aware of the stimulus some hundreds of ms later (Libet 1965; 1966; 1973), the experience can be subjectively referred backward in time to an early neural signal (Libet 1981a; 1982; Libet, Wright, Feinstein & Pearl 1979). For present purposes the experience of the time of the first awareness of wanting to move ("W") was related by the subject to his observation of the "clock position" of a spot of light revolving in a circle on the face of a cathode ray oscilloscope (CRO); the subject subsequently recalled and reported this position of the spot. (For technical details see Libet, Gleason, Wright & Pearl 1983.) Thus, the timing of this experience was converted to a reportable, visually related spatial image, analogous to reading and later recalling the clock time for any experience. This indicator of the time of first awareness of the intention to move could then be compared to (a) the actual time of the voluntary motor act, as indicated by the EMG recorded from the appropriate muscle, and (b) the time of appearance of the simultaneously recorded RP that is generated by the brain in advance of each act. For aU selfinitiated acts studied, the actual mean \Vs for each series of 40 acts averaged about -200 ms (Table 1); that is, subjects reported becoming consciously aware of the urge to move 200 ms before the activation of the muscle (EMG) (Libet, Gleason, \Vright & Pearl 1983). 2.3. Difference between RP onset and reported time of conscious Intention, W The RP onset time was found to be consistently in advance of\V, the time of initial awareness of wanting to move (Table 1). For all of the series in which all 40 acts were experienced as fully spontaneous and unplanned, the average RP onset of (type II, described above) was about -5.35 ms relative to the initiation of muscle action (as indicated by the EMG). Reported times of conscious intention to act (\V) in these same series \\lith type II RPs averaged about -190 ms. The average onset of these RPs precedes average \V by about 345 ms. {For the sIgmficance of the even larger discrepancy in series ex- .... Libet: Cerebral processes and volition Table 1. Average times (ms) of reported awareness and recorded readiness potentials (RP) for all experimental series on 5 subjects, in 6 or more separate sessions for each subject. Each series consisted of 40 trials in which subjects reported only \V or M or S times in that entire series. (Modified from Libet, Gleason, Wright & Pearl 1983.) A. Reported awareness times (ms) relative to recorded muscle activation (EMG). Wa nd Subject S.B. G.L. B.D. S.S. C.M. Grand averages Me (\V-S)b X n X n (M-S) X X n 8 8 7 7 8 -125 -282 -152 -246 -227 5 5 4 4 4 -123 -136 -249 -145 -165 4 4 4 4 4 -59 -202 +51 -118 -103 4 4 4 4 4 -19 -60 -32 -7 -20 38 -207 22 -160 20 -86 20 -28 B. reported time of conscious intention (\V) related to recorded RP onset, separated for type I and II (see text). Reported awareness times Onset of RP (in \V series) (Onset RP) minus (W) using onset of: (Onset RP) minus (W-S), using onset Type of RP, for W series nd \V RPKfN RP90% RP MN RP90% n RP MN RP90% II I 20 12 -192 -233 -535 -1025 -527 -784 -343 -825 -333 -522 14 6 -366 -950 -323 -585 a\V = time of first awareness of wanting to move (see text). bS was based on reported time of awareness of the sensation elicited by a near-threshold electrical stimulus pulse to the hand, delivered at a. randomly irregular time in each trial. The attentive and other conditions (subject's observing and recalling "clock time" for each S) closely paralleled those for the \V and M series, except that the event was an externally induced sensory one instead of a selfinitiated motor one. The difference (S) between reported and actual stimulus times may be regarded as a measure of the subject's error or "bias" when observing and reporting under the experimental conditions employed (see text and Libet et al. 1982; Libet, Gleason, \Vright & Pearl 1983). Almost all subjects exhibited a negative net bias for S (except for B.D.). For (W-S) values, the S bias exhibited by each subject is subtracted from the \V values available in the same sessions. eM was time reported for subjects' awareness that they were actually moving, instead of wanting to move as for W. The consistently negative though smaller values for M suggest that it reflects the time of initiation of the final motor cortical output, i.e., the 'endogenous "command to move" et al. 1983), rather than the awareness of proprioceptive sensory impulses evoked after onset of the movement (see text). d n = number of series, each of 40 trials. Each average or X value for n series is the mean of the mean \Vs (or mean Ms), each of which was detennined for each series of 40 trials (see Libet, Gleason, Wright & Pearl 1983). eOnsets ofRP, relative to KMG (electromyogram indicating that the activation of the muscle has started), are given for both the "main negative shift" as estimated by eye, and for the time at which the last 90% of the total area under the RP tracing begins. hibiting type I RPs, those recorded when some acts were preplanned, see Libet, Gleason, \Vright & Pearl 1983.) This timing relationship, with the "physical" (cerebral process) preceding the "mental" (conscious intention), held not just for average values of all series but for each individual series of 40 self-initiated acts in which RP and \V were recorded simultaneously. Although RPs of 40 events were averaged to produce the recorded RP, statistical and mathematical evaluation of the experimental data strongly supported the view that each individual RP precedes each conscious urge (see Libet, Gleason, Wright & Pearl 1983). The timing relationship also held regardless of which of the available parameters was used either to measure the onset of the RP (for the onset of its main negative component or for 90% of its area), or for \V (using either the "actual" or the "order" mode of recall of the clock position of the revolving spot at the time of conscious intention; see section 2.4.3). Confidence in the significance of the difference between RP onset and \V is further raised by the fact that it was almost invariably THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 533 Libet: Cerebral processes and volition large in all the individual series when compared to the standard error of the mean value for W in each respective series. In addition, the individual W time reported for each act in a series of 40 trials was almost never negative to (timed in advance of) the onset of the averaged RP recorded for that series. In view of the foregoing considerations (and additional methodological checks listed in Libet, Gleason, Wright & Pearl 1983), the substantial interval by which RP onset precedes W appears sufficiently reliable. Questions about the validity and meaning of the values must still be considered. 2.4. Validity of criteria for the time of a conscious Intention to act Because subjective experiences are not directly accessible to an external observer, it may be logically impossible for the external observer to detennine directly any feahIre of the experience (Creutzfeld & Rager 1978; Libet et al. 1979; Nagel 1979). This restriction applies also to the actual time of a subjective experience (Hamad, unpublished; Libet et al. 1979). We do not normally apply the criterion of logical impossibility to the validity of introspective reports by the people around us in everyday life although we do attempt to evaluate the accuracy of these reports. I do not know of any serious believer in Berkeleyan solipsism, even though that position may be logically unassailable. (On the other hand, the descriptions even of externally observable physical events cannot be regarded as haVing an absolute validity; they have. been appropriately viewed as mental representations or constmcts elicited by or developed from the available sensory experiences, e.g., Margenau 1984.) One is always faced, then, with the unacceptable alternative of not attempting to study a primary phenomenological aspect of our human existence in relation to brain function because of the logical impossibility of direct verification by an external observer. Or one can attempt to evaluate the accuracy of the introspective report and gain confidence in its validity by applying indirect controls, tests and converging operations. In the present study we rely on the subject's ability to associate his introspective awareness (of the urge or decision to move) with the (later reported) position of a visually observed revolving spot, the" clock time." The crucial experimental question thus becomes: Is there any convincing way of estimating what might be the discrepancy between actual and reported times (for the subject's introspective experience of the urge to move)'? The several independent types of control evidence discussed below proVide confidence that the accuracy of the reported clock times is sufficient for present purposes (i.e., for detennining the significance of the difference between RP onset and time of conscious intention). 2.4.1. Comparisons of simultaneous events. Our method requires that the subject observe simultaneously, for later report, the conscious urge or intention to move and a visual experience of "clock position" for the revolVing spot on the eRO. Subjective timing comparisons of simultaneous but disparate events are known to be subject to potential errors (see Boring 1957; Efron 1973; Sternberg & Knoll 1973). However, we introduced a control series in each experimental session to help measure such an error. For this. a skin stimulus was delivered 534 THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 at an irregular, randomized time after the start of each trial and the subject reported the time of his awareness of that stimulus. All procedures were otherwise the same as in series of self-initiated acts (except that awareness of the stimulus replaced awareness of the urge to move). The actual time of the stimulus in the control series was later known to the investigator, and the discrepancy between the subjecfs reported timing and the actual stimulus time could be objectively determined. To the extent that simultaneous observation of visual clock time and awareness of skin sensation shares similar processes and difficulties with simultaneous observation of clock time and awareness of urge to move, one may regard any measured "error" in reports of stimulus time as an estimate of the potential error in reports of W (time of awareness of wanting to move). Skin sensations were commonly reported to occur somewhat in advance of (negative to) the actual delivery time, reminiscent of the prior entry effect (e.g., Allan; 1978; Boring 1957). However, the amount of the error found in the stimulus series did not qualitatively alter the difference between onset of RP and W; in fact, it generally enlarged the difference (Table 1). 2.4.2. Judging onset time of an endogenous mental event. It might be proposed that subjects do not judge the onset of an endogenous mental event such as conscious intention the same way they judge the onset of an experience induced externally by a skin stimulus. In relation to such a suggestion we note: a. Each subject was instructed to "watch for" and report the earliest appearance of the awareness in question, and subjects did not raise any difficulties about doing this. b. The onset time even of an intracerebrally generated event of some complexity, although admittedly induced by an applied stimulus, can be reported with no significant delays. In earlier work (Libet et al. 1979), onset time of a vaguely perceived near-threshold sensation elicited by a. stimulu.s to a cerebral somatosensory structure (medIal lemmscus) was judged subjectively to differ by of ms from the sharper sensation elicited only a by a skin stImulus. In addition, both the medial lemniscus and the sensory cortex required repetition of stimulus 20 per sec) for at least 200 ms, to elicit any subjective experience at all in those experiments. Yet the subjects could consistently report a different onset time for each; they reported that the medial lemsensation began with no significant delay relative to the sensation elicited by a single pulse stimulus to the skin, whereas onset of the cortical sensation was delayed by the amount of the required stimulus duration (Libet et aL 1979). c. For two different though related endogenous mental voluntary act, the subjects even.ts related to the dIfferent onset times with an appropnate of dIfference. Under the identical expericondItIons for studying the self-initiated acts, the subjects were asked to report the clock time for their of actually moving (M) instead of for awareness of to move M values were, unexpectedly, to EMG-O time and slightly but conSistently reported times for awareness of skin stimulus (S) III whICh no movement was involved (see Table IA). J Libet: Cerebral processes and volition • Because M times were slightly before actual movement, this suggested that M may reflect awareness associated with the immediate initiation of cerebral motor outflow (Libet, Gleason, \Vright & Pearl 1983). This would be in accord with the findings by McCloskey, Colebatch, Potter & Burke (1983) that subjective timing of one's own "command to move" preceded the EMG by up to 100 ms; a sensation of having already moved, elicited by input from peripheral sensory sources, was found to be separately reportable with an appropriately delayed time. M thus appears to be an endogenous mental event, different from but related to W. Nevertheless, the subjects did not confuse their reports of onset times for M with those of\V; reports of \V times (for awareness of wanting to move) were consistently negative to (in advance of) M times (for awareness of actually committing the movement), by about 120 ms on the average. 2.4.3. Modes of reporting. One way to test and improve confidence in the validity of the reported timings lies in using different and independent but converging modes of observing and reporting. Two quite different modes were used for reporting the "clock positions" of the CRO spot at the time of awareness: (a) absolute readings and (b) order relative to final stopping positions of the CRO spot, varied randomly (see Libet, Gleason, Wright & Pearl 1983). Yet both modes produced values for W that were essentially indistinguishable. (\Vhen reporting in the "order" mode, subjects had to recall the position of the moving spot [at the time ofinitial awareness of the urge to act] only with respect to a final resting position ofthe spot that was varied randomly in different trials. Subjects needed to make judgments about whether the CRO spot came to rest at a clock position that was "earlier" or "later" than the recalled position of the revolving spot when they were aware of the urge; they did not have to specify an absolute clock position of the moving spot associated with \V [Libet, Gleason, Wright & Pearl 1983]. See also McCloskey et al. [1983] for an analogous order method for timing judgments.) 2.4.4. Nonrecallable initial awareness of conscious intention?It might be argued that a non recallable phase of a • conscious urge exists, so that the reported time would apply only to a later, recallable phase of awareness. However, one should note that to report \V time, the subject need recall only the clock position of the revolving spot at the time he first becomes aware of the urge or intention to move and not necessarily the initial awareness itself. In any case, there is no evidence for a nonrecallable initial awareness. But, like some other conceivable hypothetical uncertainties in timing an endogenous mental event, such a hypothesis cannot be excluded since it is presently not experimentally testable. 2.5. RP as indicator of cerebral initiation For the experimental question about the initiation of a voluntary act, one must also consider whether the onset of recorded RP is a valid indicator of the time when cerebral processes begin to produce the act. The precise role of the cerebral activity represented by the RP in the initiation of the voluntary process is yet to be determined. It appears likely that the component of the RP associated with volitional preparation to act is generated in the supplementary motor area, a portion of the cerebral cortex located on the mesial surface of each hemisphere facing the midline (Deecke & Kornhuber 1978; Eccles 1982a; Libet et al. 1982). RPs associated with spontaneous self-initiated acts (type II) are indeed distinctly maximal at the vertex of the head (Libet et al. 1982), a scalp site that is above and adjacent to the supplementary motor areas. It has been proposed that the initial neuronal events in all voluntary movements arise in the supplementary motor areas (Eccles 1982). However, for present purposes it is not necessary that the full role of the supplementary motor area of the RP processes be established. It is only necessary to accept the RP as a valid indicator of minimum onset times for cerebral processes that initiate the voluntary act, even if these processes should be initiated elsewhere in the brain. It might be proposed that the RP does not indicate directly or indirectly the specific initiation of the voluntary act. Rather, the RP might represent preprogramming processes that develop periodically without signifying a volitional function. The actual initiation of a given voluntary act would then depend on conscious activation or triggering of one of these preparatory sequences so as to generate an actual motor discharge. Such a proposal would seem to be an ad hoc speculation not supported by the experimental evidence. (a) The proposal would predict that endogenous RPs appear repeatedly without any associated subjective awareness developing and with no actual voluntary movements occurring. This has not been experimentally demonstrated and would seem to be untestable with present techniques. The RP that precedes an individual voluntary act is not clearly discernible from the background rhythmic activity; averaging of the preperiods (1.4 sec) for 40 acts gave us a usable though still noisy RP shift at the vertex. However, one should note that individual spontaneous negative and positive slow potential (SP) shifts have been successfully recorded during 5-sec periods preceding a choice reaction test and found to be related to proficiency of performance (Born, \Vhipple & Stamm 1982). These interesting spontaneous SPs were apparently ma.ximal at frontal rather than vertex sites and they were either negative or positive in polarity; they presumably reflect processes different from those of the negative RP that is ma.ximalat the vertex and obtained in a different mental context. (b) The recorded RPs in selfinitiated acts do not exhibit any special electrophysiological event that might signal introduction of an activating process at the reported time of about -200 msec for the conscious urge (Libet et al. 1982; Libet, Gleason, \Vright & Pearl 1983). (For RPs in self-paced acts see also Deecke et al. 1976; Shibasaki et al. 1980.) (c) The available evidence suggests that an RP precedes every voluntary act as well as the conscious awareness of the urge to perform each act (Libet et al. 1982; Libet, Gleason, \Vright & Pearl 1983). Consequently, the proposal against RP initiation of the act would at best result in a two-stage mediation; "preparatory" cerebral processes would still unconsciously initiate the volitional sequence but consummation of the actual motor action would depend on a conscious control function. This sort of role for the conscious function is compatible with the thesis being advocated in this paper. Is it possible that the subject's introspective observaTHE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 535 Libet: Cerebral processes and volition tion of his conscious intention for each act would itself introduce a cerebral process that affects the recorded RP (a question raised by an anonymous editorial reviewer)? In a small number of experiments RPs were recorded for series of 40 self-initiated movements in which no reports of awareness time were requested from or made by the subjects. The RPs of these "no-report" in form and onset times to RPs of the report senes (Libet et a1. 1982; Libet, Gleason, Wright & Pearl 1983). Furthermore, reporting the time of awareness of a sensory stimulus delivered at a randomly irregular time ("S" series) required the same kind of attention and introspection by the subjects as did the reporting in self-initiated acts; yet there were no significant pre-event potentials at all in association with the stimulation experiments (e.g., Figure 1; Libet et al. 1982; Libet, Wright & Gleason 1983). One may conclude that the "introspective process" did not affect the RPs in any manner significant to the conclusions in the study, and that if there were any electrophysiological correlates of introspective observation or of the attentive state required for it, they are not manifested in the scalp recordings of RPs at the vertex. 3. Unconscious initiation of voluntary acts Onsets ofRPs regularly begin at least several hundred ms before reported times for awareness of any intention to act in the case of acts performed ad lib. It would appear, therefore, that some neuronal activity associated with the eventual performance of the act has started well before any (recallable) conscious initiation or intervention is possible. This leads to the conclusion that cerebral initiation even of a spontaneous voluntary act of the kind studied here can and usually does begin unconsciously. (The term "unconscious" refers here simply to all processes that are not expressed as a conscious experience; this may include and does not distinguish among preconscious, subconscious, or other possible nonreportable unconscious processes.) Put another way, the brain "decides" to initiate or, at least, to prepare to initiate the act before there is any reportable subjective awareness that such a decision has taken place. It might be argued that unconscious initiation applies to the kind of spontaneous but perhaps impulsive voluntarv act studied here, but not to acts involving slower deliberation of choices of action. The possible role of unconscious cerebral activities in conscious deliberation is itself a difficult and open question. In any case, after a deliberate course of action has been consciously selected, the specific voluntary execution of that action, i.e .• the cerebral activation and implementation of the actual motor deed, mav well be related to that for the ad lib kind of act we have ;tudied. Even when a more loosely defined conscious preplanning has appeared a few seconds before a self-initiated act, the usual specific conscious intention to perform the act was consistently reported as having been experienced separately just prior to each act by all subjects (Libet et al. 1982; Libet, Gleason, 'Vright & Pear11983). This leads me to propose that the performance of every conscious voluntary act is preceded by special unconscious cerebral processes that begin about 500 ms or so before the act. 5.36 THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 3.1. Cerebral basis of unconscious mental functions A role for "the unconscious" in modifying and controlling volitional decisions and actions was advocated long ago (e.g., Freud 1955; 'Vhyte 1960). This role was from analyses of strong but indirect evidence. The present experimental findmgs proVIde direct evidence that unconscious processes can and do initiate voluntary action and point to a definable cerebral basis for this unconscious function. In addition, these findings are in accord with a previous th.e general hypothesis that dealt with the q uesti?n subjective conscious experience of each mdlvIdual IS related to his cerebral processes and what this from unconscious processes. That hypotheSIS proposed that some substantial time period of appropriate cerebral activity lasting hundreds of ms may be required for eliciting many forms of specific conscious experiences (Libet 1965). The hypothesis developed out of experimental findings that cortical activities must sist for up to 500 ms or more before 1 sec before the preset time (Figure 2, "M-veto''), even though no actual muscle activation occurred (Libet, Wright & Gleason 1983). This resembles the RP of selfinitiated acts when preplanning is present (Libet et al. 1982, type I RP). The form of the "veto" RP differed (in most but not an cases) from those "preset" RPs that were followed by actual movements; the main negative potential tended to alter in direction (flattening or reversing) at about 150-250 ms before the preset time (Libet, \Vright & Gleason 1983). This difference suggests that the conscious veto interfered with the final development of RP processes leading to action. (\Vhether the above-mentioned ,MP or PMP components of RP are specifically eliminated bv such a conscious veto remains to be analyzed.) In an"y case, the preparatory cerebral processes associated with an RP can and do develop even when intended motor action is vetoed at approximately the time that conscious intention would normally appear before a voluntary act. The veto findings suggest that preparatory cerebral processes can be blocked consciously just prior to their consummation in actual motor outflow. As an alternative study, we might have randomly presented an external signal at which the subject would veto the prearranged or preset act. (External signaling to veto an act after a given self-initiated RP has begun is not technically feasihle, since the individual RPs arc not sufficiently discernible from the background EEG activity.) However, an externally signaled veto would not be an endogenous conscious process; as a quick reaction to a sensory signal it could even be generated unconSciously. It would of course be even more desirable to study the uninstructed veto of a spontaneous, self-initiated act, but, as mentioned, this is not presently possible technically because an objective trigger time for averaging RPs would not be available. 538 THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 4.2. Conscious "trigger" versus '\teto" An alternative mode of conscious control might lie in a requirement that a conscious "trigger" finally the unconsciously initiated cerebral processes to achIeve the an actual motor act. Conscious control would then active role in completing or consummating the volItIOnal process; the absence of a positive conscious trigger would mean no actual motor act occurs. If one grants the 1" availability of the veto process, then an active trigger becomes a redundant and unnecessary means of achIeving conscious control. On the other hand, it is conceivable that both modes of control, active trigger and veto age, are available. Whether by active positive triggenng or by vetoing the completion of the volitional the conscious function may be thought of as selectmg among the possible acts developed by the unconSCIOUS initiating processes. Would the appearance of a conscious trigger or also require its own period of prior neuronal activity, as IS postulated for the development of the conscious urge or intention to act and for a conscious sensory experience? of Such a requirement would imply that conscious the volitional outcome, whether by veto or by an activatof ing trigger, is itself initiated unconsciously. F or the volitional process to be exerted as a consctOuS initiative, it would indeed seem necessary to postulate that conscious control functions can appear without prior initiation by unconscious cerebral processes, in a context in which conscious awareness of intention to act has already developed. Such a postulate can be in accord either with a monist view, in which a conscious control function could be an ongoing feature of an already emergent conscious awareness (Margenau 1984; Sperry 1980), or with a dualist interactionist view (Popper & Eccles 1977). 5. Free will and individual responsibility This is not the place to debate the issue of free will versus determinism in connection with an apparently endogenous voluntary action that one experiences subjectively as freely willed and self-controllable (see Eccles 1980; Hook 1960; Nagel 1979; Popper & Eccles 1977). However, it is important to emphasize that the present experimental findings and analysis do not exclude the potential for "philosophically real" individual responsibility and free will. Although the volitional process may be initiated by unconscious cerebral activities, conscious control of the actual motor performance of voluntary acts definitely remains possible. The findings should therefore be taken not as being antagonistic to free will but rather as affecting the view of how free will might operate. Processes associated with individual responsibility and free will would "operate" not to initiate a voluntary act but to select and control volitional outcomes. (Voluntary action and responsibility operating behaViorally within a deterministic view would, of course, be subject to analogous restrictions. ) Some may view responsibility and free will as operative only when voluntary acts follow slower conscious deliberation of alternative choices of action. But, as already Commentary/Libet: Cerebral processes and volition noted above, any volitional choice does not become a voluntary action until the person moves. In the present study, the subjects reported that the same conscious urge or decision to move that they experienced just before each voluntary act was present and that it was similar whether or not any additional experience of general preplanning had already been going on. Indeed, the reported times for awareness of wanting to move were essentially the same for fully spontaneous acts and those with some preplanning (Libet, Gleason, Wright & Pearl 1983). One might therefore speculate that the actual motor execution even of a deliberately preselected voluntary act may well involve processes similar to those for the spontaneously voluntary acts studied by us. The urge or intention actually to perform the voluntary act would then still be initiated unconsciously, regardless of the preceding kinds of deliberative processes. The concept of conscious veto or blockade of the motor performance of specific intentions to act is in general accord with certain religious and humanistic views of ethical behavior and individual responsibility. "Selfcontrol" of the acting out of one's intentions is commonly advocated; in the present terms this would operate by conscious selection or control of whether the unconsciously initiated final volitional process will be implemented in action. Many ethical strictures, such as most of the Ten Commandments, are injunctions not to act in certain ways. On the other hand, if the final intention to act arises unconsciously, the mere appearance of an intention could not consciously be prevented, even though its consummation in a motor act could be controlled consciously. It would not be surprising, therefore, if religious and philosophical systems were to create insurmoyntable moral and psychological difficulties when-fliey castigate individuals for simply having a mental intention or impulse to do something unacceptable, even when this is not acted out (e.g., Kaufmann 1961). ACKNO\VLEDGMENTS This paper is based on a presentation at a conference, "Cerebral Events in Voluntary Movement," held at Castle Ringberg in \Vest Germany November 14-19, 1983, organized by J. C. Eccles, 0. D. Creutzfeldt, and M. Wiesendanger, under the auspices of the Max Planck Society (abstract in Experimental Brain Research, 1985). I thank Moreen Libet for helpful comments on an earlier draft of the paper. Open Peer Commentary Commentaries submitted by the qualified professional readership of this journal reill be considered for publication in a later issue as Continuing Commentanj on this arlicle. lntegratice ocercielcs and syntheses are especially encouraged. Problems with the psychophysics of intention Bruno G. Breitmeyer Department of Psychology, University of Houston, Houston, Texas. 77004 Several methodological and conceptual problems come to mind after a reading of Libet's article. For one, the timing of all consciously apprehended events under investigation was measured relative to the "clock position" of a dot revolving in a circle. Similar timing methods plagued by several problems have been used for over 100 years. Using a revolving dial, Wundt (1904) noted that the perceived time of a sensory event relative to the simultaneously visually perceived position of the rotating dial depended crucially on the angular rate of the dial's rotation and the other sense being stimulated. Libet's work is based on a single angular dot velocity; hence, despite acceptance of his particular implementation of the procedure by refereed journals, there is a very strong possibility that his measures are idiosyncratic. Moreover, the timing ofS, the awareness of a tactile stimulus, does not serve as a clear control that allows one to regard any timing "error" here as an indication of the potential error found in timing W, the awareness of the intent to act. First, judgments of intermodal sensory simultaneity depend on the particular senses investigated and the stimuli used. Besides the prior entry effect noted by Libet, intrinsic latency and processing rate differences among senses as well as latency differences introduced extrinsically by use of a near-threshold tactile stimulus relative to a clearly suprathreshold visual dot stimulus (Libet, Gleason, \Vright & Pearl 1983) render use of anyone estimate of timing error arbitrary and suspect. Second, attending to \V may not be eqUivalent to attending to S, as Libet assumes. Indeed, one can voluntarily allocate attention to endogenously produced cognitive/mental processes as well as to mental processes produced exogenously by sensory stimuli. However, in the latter case a compulsory, stimulus-evoked allocation of attention is typically also engaged, as illustrated by Remington's (1980) and Jonides's 1981) studies of attention to briefsuprathreshold visual stimuli. Insofar as Libet's near-threshold tactile stimuli were above threshold, their presentation would also evoke such an obligatory or nonvoluntary attention. Even if one were to pass over these pertinent methodological problems, several concerns of a more conceptual nature need addressing. First, in what sense can the voluntary acts as operationally defined by Libet be paradigmatic of volitional action generally, particularly when he draws certain weighty religio-ethical implications from his findings? As Libet admits, his experimentally reduced acts of finger/wrist flexion occur in the absence of any larger meaning. Hence they are as limited in application to our understanding of volitional action as use of nonsense syllables is to our understanding of memory. By what rules do we proceed from these experimental findings to human volitional action (or memory) occurring inextricably within a rich, varied, and meaningful context? \ViIIiam James (1950) held that a strictly voluntary act must be guided throughout its whole course not only by volition but also by idea and perception. Moreover, he observed that consciousness, besides being primarily a selective, intentional process, is more or less intense depending on action's being more or less significant and hesitant (nonhabitual), that is, where indecision is present to a greater or lesser degree. Consequently, one might at least require that subjects choose freely among several actions, each of which carries some practiC'al consequence (cost and benefit) rather than merely choosing to act or not in some stereotyped and inconsequential way. To counter the requirement that a strictly voluntary act be characterized by slow conscious deliberation and existential alternatives of action, Libet notes that no volitional choice becomes voluntary action until the person moves. The implication is that Libet's paradigmatic acts tap this final, effective conscious intent, which invariably appears approximately 3.50 ms after an RP is generated but 200 ms before one actually moves. It should be noted that the actions investigated by Libet have been performed (by myself and several of my colleagues) without awareness of intent to act. By requiring subjects to attend to a'wareness ofintent, Libet may have imposed intention artificially and in a way that is not comparable with more THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 539 ,- Commentary/Libet: Cerebral processes and volition The acts arc a step removed from the instructions, and the i.ssue of the source of timing for the irrcgularly rcpcated acts IS an important one, but the behaviors should not bc con.fuse.d with instances of free will. It is evcn possible that free Will, lIke the mind-body problem, will disappear as our undcrstanding of the physiology of experiencc increases. . In a sense the subjccts in the Libct expenments are asked to behave as though they had free will, whether such a thing really exists or not. Under these circumstances it is not clear whether we are seeing some fundamental property of the human nervous system or merely the program that the subject has set in.to To give another example of this process, consider a ma psychophysical experiment who is asked to draw boxes on of paper. The psychologist could study the box-drawmg machine as though it were designed only for this task, and the dynamics of the behavior, its physiological concomitants, and so on could be studied in detail. Box-drawing centers could be found in the brain, box-detecting circuits could be described in the visual system, and the prebox potentials could be analyzed. The artificiality of the task, though, would not be apparent no matter how detailed the analysis; in fact, the more detailed the analysis the less likely it is that the results will bc interpreted as specialized operations of a more general-purpose machine. The subject has programmed himself to behave as if he were a box drawer and nothing else. Similarly, Libel's RPs may have characteristics uniquc to the rather specialized and tasks required of his subjects. This is not to say that Libet s paradigms are invalid but only that they should be interpreted with caution. The temptation to overgeneralize a specific task with its unique demand characteristics may also be related to the generalization of the veto principle at the end of Libel's article. The Bible's injunction not to commit adultery, we may expect, will be handled very differently from Libel's injunction not to move the fingers on a given trial. The confusion of levels is an error that I have called "Uttalism"after Uttal's (1971) injunction that properties of single-cell receptive fields cannot automatically be applied to behaviors of the whole organism. This problem has arisen in visual where neurophysiologically based models, whether computer simulations (Bridgeman 1971; 1978; Weisstein 1972) or qualitativc theories (Breitmeyer & Ganz 1976), rely on mechanisms too limited to reflect the subleties of real human behavior. No amount of tinkering with these theories will deal with practice and attention effects, for example, nor will they explain strong effects of rather small differences in stimulus patterns on masking. Similarly, the Libet data, important as they are, should not be confused with physiological studies of self-control in human behavior. The finding that consciousness enters after the beginning of Free will and the functions of consciousness an identifiable set of neurological events can be viewed in the context of consciousness as a neurological system like any other, Bruce Bridgeman with specific jobs that help the organism to function effectively. Zentrum fUr Interdisziplinare Forschung, UniversiUit Bielefeld, 48 Bielefeld Its jobs include handling situations that are difficult, dangerous, 1. Federal Republic of Germany or novel (Norman & Shallice 1980), and it serves among other Libet attempts nothing Jess than a bcginning of the physiology of things to establish action schemata, order their priorities, and free will. an area where philosophical work preViously has monitor their progress. Thus consciousness must be involved enjoyed a total lack of cmpiric-al rcstraint. Thc philosophical when a behavior is about to be executed, if that behavior might issues won't go away yet, hO\vever, and they remain important interfere with other ongoing schemata. In Libel's special case to interpreting the experiments. Two problems deservc special the only ongoing task is to sit still. Here, that stage of organizing comment: the demand characteristics ofthc cxperiment and the a behavior that first requires access to consciousness can occur from millisecond-level operations to long-tenn only a few hundred milliseconds before the behavior begins. \Ve behavioral planning. do not yet k"lloW what happens in the more general case, when A careful analysis of the experimental conditions reveals that other action programs are being executed at the same time. the subjects' wills were not as free as the Libet article implies, for the small, sharp movements that they were instructed to make were not freely \....iIled but were requested by the experiConsciousness and motor control menter. Thc will of a subject was no more free in this design than Arthur C. Danta in reaction-time experiments; the only difference between this Department of Philosophy. Columbia University. New York. N. Y. 10027 experiment and the latter paradigms is tbat the instruction and the movement arc decouplcd in time. \Vhile performing the It is a !ruth universally acknowledged that a physiologist in task. the subjects do nothing morc than obey the instmctions. posseSSIOn of a metaphysical prejudice must be in want of ecologically and existentially valid voluntary and intentional acts. To illustrate, up to this point I was not consciously aware of intending to write down these thoughts. Yet a prior intention to write a critique occurred days ago. In fact, however, I could have chosen to intentionally write out my critique word by word that is, with clear awareness of each intent to write each work just prior to writing it. Yet this or Libel's "hyperintention" brought about by self- or by experimental instruction in no way represents my voluntary actions in general. At best the hyperawareness of intention functions as a monitor retrospecting on my much earlier plan, decision, or intention to write rather than as an instigator, motivator, or modulator of writing activity. In this view, the awareness of intent, though it falls just after the onset of RP and just before the onset of movement, poses neither a scientific nor a philosophical problem and has little if any bearing on issues of free will and responsibility. Finally, even if one admits the legitimacy of Libel's procedure and interpretation, Libet hedges on and skirts around an important issue. Libet would have it that one can discuss the operational possibilities of conscious control of action on purely phenomenological grounds without commitment to specific philosophical alternatives such as detenninism versus free will or epiphenomenalism versus mental intervention. Such a phenomenological bracketing is well-nigh impossible since it asks one to suspend any thesis of reality including the metaphysical assumptions hidden behind the very scientific enterprise being undertaken by Libet. In the context of his work, how can one talk of possibilities of conscious control, and not tum this talk into idle chatter, without taking a stand in particular on epiphenomenalism versus mental (conscious) intervention? On the one hand, if the conscious pennissive "trigger" or rcstrictive "veto" is preceded by causally efficacious yet unconscious neural activity just as in the case of the consciously experienced intent to move (Hamad 1982), then that consciousness is mere afterthought, a reflection on events outside its causal control and, therefore, epiphenomenal. On the other hand. consciousness is a fact to each of us. Insofar as its existence is undeniable, it is a troublesome and abiding enigma, particularly to any accepted version of natural evolution. For to have evolved it must be as causallv efficacious as is the hand that writes these words. Hence including any conscious or "veto," calls for some form of mental intervention. As scientists, we cannot stand on the sidelines and suspend or bracket the thesis of natural evolution. To do so would further mystify consciousness to a degree warranting silence. 540 THE BEHAVIORAL AND BRAIN SCIENCES (198S) 8:4 • Commentary/Libet: Cerebral processes and volition philosophical help. It is inconceivable save with reference to some such prejudice that Libet would find it necessary at the end of his paper, to postulate functions whose existence would be incompatible with everything he had up to that point been at pains to show. These are "conscious control functions," which "can appear without prior initiation by unconscious cerebral processes." But everything up to then would have disposed us to believe that motor acts are the consequence of exactly such initiating processes, revealed to the consciousness of the agent about 350 ms after onset, with the motor act itself taking place about 150 ms thereafter, barring endogenous intervention. But then, in that last fateful interval, abruptly and without experimental motivation, between the intention and the act falls the shadow of alien ideas. These are the "conscious control functions" that "trigger" or "veto" the act and that spring, cerebrally unsummoned, into being. Freud famously said the hysterical symptom seems to have no knowledge of anatomy. When a physiologist relaxes his laboratory scruples in favor of what must be physiologically mysterious, he is to be diagnosed as in the grip of a kind of metaphysical hysteria. Surely conscious control functions have some physiological substance if they have physiological effects. And surely it should be an empirical matter whether or not their occurrence be cerebrally initiated through that kind of neuronal activity which precedes the occurrence of subjectively experienced intentions or '\vantings to act." So why should it seem necessary to postulate them as thus unpreceded unless one believes precedent unconscious activity must queer some theory held dear by the writer - perhaps a position on the free will question? If Libet is right that "the present experimental findings do not exclude the potential for 'philosophically rear individual responsibility and free will," why should he act as though they did exclude that by postulating what he feels must be in place in order that responsibility and freedom have application? Philosophy must learn to live with scientific truth. It seems to me that the existence of free will does not have as close a connection with "conscious deliberation of alternative choices of action" as Libet supposes. Choosings between alternative courses of action, in the preponderance of motor acts we perform, occur as the outcome of deliberations of which we are barely conscious, if at all. A slow-motion film of Matisse shows the artist making countless decisions with his fingers that at normal speed looks like a single confident chalk stroke defining the edge of a leaf. He mayor may not have been conscious of each decision, but I suspect that he was conscious only of drawing a leaf. Consciousness, in moral theory, plays its role only in connection with premeditation, for which there is neither time nor occasion in the sort of spontaneous choosings we do in life and in the sort oflaboratory Libet's work presumes. Happily, we are so wired that deliberation may occur without the mediation of consciousness at all. Consciousness is evolution's gift to us for rather special deliberative employment having to do, as responsibility and free will have to do, with courses of action - with projects - rather than the basic sorts of acts involving the simple flexion of a muscle or the moving of a hand to no further purpose. Suppose one were to designate as intentions the entire cerebral processes that eventuate in motor acts, rather than restricting the intention to that fragment of the cerebral process which becomes conscious? The concept of intention was framed well before there was knowledge of cerebral process, but once it is accepted that much of deliberative action transpires without becoming conscious to the agent - because its being conscious would reduce our efficiency as agents - the concept might easily be extended to cover more than would have been necessary in periods when the mental and the conscious were closely identified. \Ve might indeed think, in those cases in which some segment of the intention becomes conscious, of the preceding segment as preconscious intention. Then, in the standard ease, this is what happens: The intentional is formed; some milliseconds later the agent becomes conscious of his intention; some milliseconds later the motor act occurs as intended. Why do we need an extra "trigger" since there is no empirical basis for its existence but only a "necessary postulation"? It would be like requiring a trigger in mechanics in order to explain the fact that a body, moving in a straight line with uniform velocity, continues to move in a straight line with uniform velocity, when in fact all we need is an explanation of acceleration, or change in direction and velocity. Why should not the intention be enough to trigger the movement? I surmise that Libet thinks that simply allowing to take place what is already in process is too passive a role for conscious intention if freedom is to be robust enough for our moral vision of ourselves. In my view, all we need to explain is changes in intention. But these can be well under way before we are conscious of the change, with the entire cerebral process, including the fragment of it that is conscious, as the veto of the previous intention. There is plenty of time to abort the action if the intention arises before consciousness of veto. In brief, instead of the conscious control functions playing the special on-off role of metaphysical switches, we have the play of cerebral processes, in which consciousness informs us of what we have decided to do. Whether these decisions themselves are free belongs to a different topic, but my claim is that freedom and consciousness have less to do with each other, and certainly so in the execution of simple behaviors, than Libet supposes. Once he realizes that it is only because he believes that they have much more to do with each other than the data he presents justifies, he may drop from the inventory these curious operations that owe their existence in his article to an insufficiently self-conscious agenda. Knowing what we are embarked upon need not be a causally inert fact about ourselves when in fact we are embarked upon projects with horizons wider than the circumscribed boundaries of the laboratory. In these straitened confines, the projects to which responsibility and freedom have application scarcely can flourish. Commonly we do not simply move our hands; we do so with larger purposes in mind-to wave away a canape, to signal the death of a gladiator, to stifle by gesture the cackle of subordinates, to set up perturbations for the distraction of a wasp, or to express some agitation or other througll the language of the body. Our minds bent upon these, consciousness simply assures us we are in contact with ourselves. The time course of conscious processing: Vetoes by the uninformed? Robert W. Doty Center for Brain Research, University of Rochester Medical Center, Rochester, N. Y. 14642 Perhaps the most important feature of this latest in the series of ingenious experiments by Libet and his colleagues is the demonstration it provides that the neurophysiological basis of conscious awareness can be subjected to meaningful analysis. This has profound philosophical import, the more so since it adds further evidence for the probable uniqueness of the neural processes accessible to or directly producing conscious experience. It has long been apparent that many, indeed probably most, neural transactions are utterly devoid of or incapable of an element of consciousness-for example, autonomic regulation, hormonal release, adaptations in visuomotor control, cerebellar activity. and all neuronal discharge during most of a night's sleep (see Doty 1975). A particularly dramatic example is the loss of visual sensation despite demonstrably continuing retinal input when one is viewing a Ganzfeld (Bolanowski & Doty 1982) or absolutely fixated image (Rozhkova, Nickolayev & Shchadrin 1982); the same is probably true for the disappearance of stimuli rotating about a fixed locus in the peripheral visual field (Hunzelmann & Spillmann 1984). On the other hand, in these instances the absence of a direct conscious concomitant to the THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 541 Commentary/Libet: Cerebral processes and volition neuronal activity in the forebrain clearly does not mean that such activity is inaccessible to consciousness. Rather, these phenomena of visual loss are probably an extreme example of the workings of that still mysterious tool of consciousness, selective attention. Thus, it is apparent that there are neural processes that lie forever outside the domain of conscious experience and that there are others for which a conscious concomitant is elective. In still other instances it seems that information garnered from sensorial processes lacking an experiential component can nevertheless be incorporated into the guidance of movements consciously controlled. These issues have previously been well discussed in these pages in relation to the phenomena ofafferen t discharge from muscle spindles (Roland 1978) or blindsight (Campion, Latto & Smith 1983). However, the fact that unconsciolls neuronal activity is constantly in play dUring movement seems well recognized, as in the common inability to perform properly a habitual, rapid movement while endeavoring to exert conscious control over all its components. (Try intellectuaUy constructing and planning the motions of your fingers in tying your shoes!) Now, .perhaps Libet's experiments are detecting this, the unconscIOus components of an organized movement. There is a voluntary initiation of these components, just as there can apparently be a voluntary cancellation (veto) of them. The actual decision to release the movement occurs only against the background of readiness, the point at which the subconscious set of the neuronal program, possibly being arranged in striatedcerebellar circuitry, is acceptably complete. The unconscious part, just as i.n tying one's shoe, proceeds pari passu with, and ahead of, the overt and consciously released movement; but thIS does not mean that the unconscious components proceed or arise independently of conscious control. After neurons for each are all embedded and intertwined WIthm the same and one does not know yet whether the nellr:o nal resulting in conscious perception are a f mamfestahon of a special type of neuron or a 'al fc a r'ty 'th' • speCI orm 0 c IV) \VJ m groups of neurons of diverse form and chemistry r It t?at !his is a much more satisfactory 0 I et s observations, that an aura of unconSCIOUS for movement perpetual Iv surrounds th focus of consciousness, and that 'the aptly rea mcss.potential" (Komhuber& Deecke 1965) which L'b t pnor to the "decision" to actually th ) e ment, IS a manifestation of this process The alte f e he to favor, is that "the brain" co.ntrol to prepare movements, which can then\ conSCIOusly allOWed or consciously "vet d" Th e flaw m this interpretation is that ifth oe. e great wholly outside conscious controi ho: :ey;atory n:ovement is thcn "know" what·U 'if"u a conscIous process l? WI ensue It fails to veto the b . , In this scheme, consciousness is rele ated rain s b .. mtUltive process of guessing what it may be that lip to and being ever on the alert that th d ram IS Ufi(.'onscious do not set in motion some . e of the conscious plan. While such views f bact. mappropnate tv the permissihle in the poetic fantasy of may be '. psychology, thev. art' not neurophvsioloO"icallv • e." j convlncmg. summation: The timing of vOlun intentIons by cortical activity tary John C. Eccles Max-PIanck-lnstitut fiir Biophysische Chemie Gatti of Germany , ngen. Federal Republic My commentary starts with an acceptance of th .' findings reported by Libet. With great ingen .e e;:raordlllary Ulty able to tmin subjects to report retrospecti I. th .. has been ve)l e timmgoftheir 542 THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 voluntary intention to make a simple sharp movement. I am not concerned with the suhtle distinctions he makes hetween types of conscious endogenously willed motor actions, for example, whetheror not the suhject was cognizant of planning in advance. For me the decisive discovery is that the subjectively experienced onset of intention to move is about 200 ms before the muscle activation and about 350 ms after the onset of the readiness potential (RP), which provides some integrated signal of the cortical activity preceding the movement. To simplify my hypothesis, I will assume that the voluntary intention to move acts on the supplementary motor area (SMA) [see Goldberg; "Supplementary .Motor Area Structure and Function," this issue} and thence through the various pathways to the motor cortex and so by the pyramidal tract to bring about the movement (cf. Eccles 1982b). It is very tempting to follow Libet in interpreting these findings as establishing that cortical activity (of the for example) initiates not only the voluntary movement but also, after some hundreds of milliseconds, the introspective experience of having initiated the movement, which thus becomes an illusory experience. I shall consider later Libet's veto hypothesis, by which he attempts to preserve the responsibility of the conscious sclfby means of its power to veto the ongoing cortical activities that would otherwise lead to the movement. I now present a hypothesis that accepts all of Libel's experimental observations but that nevertheless preserves fully the role of conscious intention in initiating the movement. The hypothesis has several components. (1) It is proposed that there is a fluctuating background of activity in the cerebral cortex and in the S"'IA that can in part be generated by the reticular activating system and .. was proposed by Oshima (1983), possibly to involve a set for movements. (2) As discovered by Libet, the mental intentions reported by subjects begin about 200 ms before the movement. The hypothesis is that these intentions tend to be timed unconsciously by the subjects so as to take advantage of the spontaneous in the cortical activity «(1) above). Since the RP as observed formed by the averaging of a large number (fifty to hundreds) 0 recordings of scalp potentials with zero time given by the onset of the electromyogram, it is a mistake to assume tacitly that.the averaging eliminates the random fluctuations. If there IS a tendency for the initiation of the movements to occur during excitatory phases of the random spontaneous activity, the earher phase of the RP may be no more than the averaging of the premonitory spontaneous activity. If that is so, the RP does signify that cortical activity initiates the movement. Instead, hypothesis is that the spontaneous fluctuations of corti . mereIyadjust the phase of the conscious Illl . 't'Iat'10n to the ac ti Vlty intention some 200 ms before the movement. . (3). It is further postulated that this timing th.e intention relatIon to the phases of cortical actiVIty IS a learn phenomenon haVing the advantage that it secures opportunistically the most effective occasions for initiating voluntary actions. The lower right comer of Figure 1 illustrates the e hypothesis. It is to be noted that the activities of the reciprocally related to the mental intentions, the arrows directed both ways across the frontier between mind and brrun. (4) In the further development of the hypothesis we have to consider how the mental event of an intention can cause changes the neuronal responses of the SMA. Let us first focus attention on a single synaptic bouton, which may be, for example';:i a pyramidal cell of As shown for very diverse cen d synapses by Jack, Redman and "Vong (1981) and by Kom an Faver (1985), a impulse evokes the fcl: the bouton of a single synaptic vesicle probabIllsticalIy, .' • probability factor being usuallv less than 1 in 2. This can be increased or with consequent changes synaptic effectiveness. As described by Akert, Peper, an Sandri (1975), each bouton has a single paracrystalline stru: ture, the presynaptic vesicular grid that holds about 50 synaptic If ;d :rr Id Commentary/Libet: Cerebral processes and volition INTERACTION OUTER SENSE World 2 light Color Sound Smell INNER SENSE Thoughts Feelings Memories Dreams Figure 1. (Eccles). Information flow diagram for brain-mind interaction in human brain. The three components of World 2outer sense, inner sense, and psyche or self - are diagrammed with their communication shown by arrows. Also shown are the lines of communication across the interface between \Vorld 1 and \Vorld 2 - that is, from the liaison brain to and from these World. 2 The liaison brain is the columnar arrangement mdlcated by the vertical broken lines. vesicles, and somehow it controls the probability of their emission. The hypothesis is that the immaterial mental event of intention acts analogously to a probability field of quantum mechanics, as proposed by Margenau (1984), and modifies the of emission of a synaptic vesicle by a presynaptic Impulse. Thus an intention is effective only insofar as there is an adequate quota of presynaptic impulses; hence the necessity for the learned timing of intentions in relation to the fluctuating waves of SMA background activity. (5) Any effect of a mental intention in altering probabilities of quantal emission from a bouton is orders of magnitude too small to cause the sequence of neuronal actions leading to an effective discharge of motor pyramidal cells. It is conjectured that there has to be an immense collusive action of the mental intention on the multitude of boutons on one neuron and on a large assemblage of similarly acting neurons. This is in accord with the findings of Brinkman and Porter (1979) that, when a monkey is carrying out a voluntary act, there is excitation of many similarly acting neurons in the supplementary motor area 100 to 200 ms before the onset of the electromyogram. (6) Furthermore, according to the hypothesis there is also a reverse flow of information (Figure 1), the SMA activity being subconsciously "sensed" when a mental intention is being initiated. This is the most obscure component of the hypothesis. Yet it is generally recognized that in the perceptual areas of the cortex much activity can occur subconsciously, as in the refined experiments of Libet (1973) on somatosensory perception, where weak repetitive stimulation of the somatosensory cortex may have to continuc for 0.5 sec before the cortical activitv reaches the threshold for conscious perception. . The veto experiments of Libet are very ingenious and offer further evidence of mental control of cortical activity with the late flattening of the RP. . In conclusion, the hypothesis here presented offers a general explanation of the findings of Libet while preserving the essential character of dualist interactionisms. The early phase of the RP may be no more than an artifact arising from the technique of averaging. There is no scientific basis for the belief that the introspective experience of initiating a voluntary action is illusory. NOTE Commentator's mailing address: CH 6611 Contra (TI). Switzerland Brain mechanisms of conscious experience and voluntary action Herbert H. Jasper University of Montreal and the Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada H3Z 1E7 For many years Libet has heen carrying out carefullv controlled becrucial electrophysiological experiments on the tween electrical stimulation and responses in sensory cortex and pathways in the conscious human brain and vcrbal reports of conscious awareness with the surprising result that it seems to require considerable time (about 500 ms) for activity in sensory systems to reach the threshold of conscious awareness. Thc precise neuronal mechanisms involved in this delay have not been specified. It has long been known from experiments carried out under light barhiturate anaesthesia or natural sleep that evoked potentials and unitary responses from single cells in sensory cortex (somatic, visual, or auditory) are preserved, even including the complex information processing involved in feature detection in visual cortex as studied by Hubel and Wiesel, in states that probably preclude conscious awareness (light barbiturate anaesthesia). Libct now uscs the "rcadincss potential" (RP) to time unspecified cortical events that precede an ad libitum voluntary motor act as compared to the timing of the subject's conscious awareness of intention to move, with thc surprising conclusion that willed voluntary movements arise out of brain mechanisms that preccde conscious awareness of the intcntion to move and must therefore be subconscious. Controls on the reliability of subjcctive reports of the timing of conscious awareness of intention to move depend on the accuracy of memory, introducing another important factor that in my opinion has not given adequate considcration. Is it not possible that brain mechanisms underlying awareness may occur without those which make possiblc the recall of this awareness in memory afterward? Patients with cpilcptic automatisms, for example, may carry out many apparently intcntional complcx motor acts, often remarkably appropriate ones (such as driving in traffic), without being able to recall having donc so afterward. A similar state of apparcntly "automatic" behavior may occur with certain drugs such as scopolamine. I realize that it may be impossible to dissociate mechanisms of awareness from those of memorY recall under thc conditions of these experiments, but there is problem here that should be given serious consideration. Concerning the more philosophical implications of thcse studies, Libet should be commended for his ingenious and precise experiments, which have clarified, if not solved, the ageold problem of mind-brain relationships. I agree that mental events can be considered scientific data even though they are difficult to measure, and that they may well playa most important role in the direction of behavior and consequently of the brain mechanisms underlying this behavior, while at the samc time mental events must depcnd upon highly integrative brain functions (i.e., interactionism rather than dualism). It may well bc that there are specialized neuronal systems extending throughout cortical and suhcortical structures but scparate from spccific afferent and efferent pathways to cerebral cortex, which mediate mechanisms of conscious awareness, analogous to the outworn hypothesis of the reticular system or the "ccntrencephalic system" of Penfield. Libet has providcd us with important temporal constraints on two aspects of this problem: the temporal summation required for conscious awareness and the delay in awareness of conscious intention of voluntarv movement. I would suggest that he now direct more of his attention to brain circuits separate from the primary sensory or motor pathways in the search for mechanisms more closely related to mechanisms of consciousness, as originally suggested by Hughlings Jackson in his search for brain mechanisms of "highest level seizures." THE BEHAVIORAL AND BRAIN SCIENCES {1985} 8:4 543 Commentary/Libet: Cerebral processes and volition Voluntary intention and conscious selection in complex learned action A Richard Jung Department of Neurophysiology, University of Freiburg, 0-7800 Freiburg, Federal Republic of Germany Libel's experiments are limited to the recording of readiness potentials (Kornhuber & Deecke 1965), which precede the decision to make or veto brief finger flexions. These simple movements are made voluntary, but the will acts here only as a trigger. Willed intention is more important in goal-directed and complex movements such as writing. These also contain many unconscious mechanisms and become partly automatized by learning. Slow brain potentials recorded during action may give additional information complementing the analysis of readiness potentials that appear before movement. I agree with Libet that the conscious will mainly selects and controls ollr action and that unconscious preparatory cerebral mechanisms are important. I doubt Libel's assertion, however, that the subject's will does not consciously initiate specific voluntary acts. It is true in complex and learned movement too that several more or less unconscious motivations contribute to the action. In man, however, even emotional or instinctive actions and skilled movements can be voluntarily initiated, directed, and set for their duration, as they can be inhibited and blocked by will. \Villed intention is normally related to consciousness. Cerebral correlates of intention. The interaction of instinctive, willed, and learned factors in human decisions to act can be demonstrated by skilled movements and mental activity such as language and calculation tasks. Cerebral correlates of these conscious acts have been recorded in man Gung 1984). The electrophysiological correlates of goal-directed and writing movements are large surface negative potentials that appear as an increase of the readiness potentials at the precentral and parietal cortex (Griinewald-Zuberbier et al. 1978; Jung et al. 1982). The aiming potentials terminate in a positive shift when the goa} is reached (Figure lA). The preparatory body posture and balance accompanying the consciously steered goal-directed movement become unconscious after the primarily will-controlled movement is trained Gung 1981; 1982). The aiming potentials are probably related to the willed performance of goal-directed movements and to their programming. Normally, our consciousness is concerned only with the goal and not with the automatic and learned mechanisms of action involved in its pursuit. Owing to the limits of conscious information content, conscious intention is only a small part of the whole action program. Umited capacity of consciousness. In conscious perception and voluntary action the information flow of the human nervous system is extremely reduced from the input of 107 to about 2050 bits/sec (Kiipfmiiller 1971). This narrow range of consciousness necessitates selective processing and automatized programs for all voluntary skilled movements Gung 1981). Such unconscious motor programs are acquired by learning. Let me explain the selective and restricted role of the consciolls contribution to complex action by the experience of goal-directed movements and other tasks. As a subject in the experiment shown in Figure lA, I was consciously aware of my aiming intention during the action and of two other intentions that were in the background and less salient. The first intention, to direct the object to the goal, began with the readiness potential of 1 sec duration and continued fi)r 3 sec. The second, to fLxate the target and not to look to my hand, was less conscious, and the third, to suppress blinking by staring, was sometimes interrupted by involuntary blinks. Of course, special activation of arm muscles, needed during the task, was not conscious. Hence, the voluntary conscious intention to reach the target was combined with a negative veto to avoid eye movements and associated with automatized hand movements. 544 THE BEHAVIORAL AND BRAIN SCIENCES (1985) 8:4 o z 6 Readiness a 8 sec writing potenti...
Purchase answer to see full attachment