In the 1980s, neuroscientist Benjamin Libet found that the brain starts preparing a simple movement before people feel they have decided to make it.1PubMed. Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential). The unconscious initiation of a freely voluntary act That result is often cited as proof that free will is an illusion. What it actually shows is narrower. It concerns split-second, meaningless choices, and what the early brain activity means is still debated. The bigger surprise may be that conscious awareness seems to keep a late chance to stop an action rather than to start it.2PubMed Central. The point of no return in vetoing self-initiated movements
What Libet Actually Measured in 1983
Libet did not discover the brain signal at the center of his experiment. Since the 1960s, researchers had known that a slow electrical change, recorded from the scalp, builds up as much as a second or more before a self-paced movement.3Neurophysiology of Consciousness. Readiness-Potentials Preceding Unrestricted’ spontaneous’ vs. Pre-Planned Voluntary Acts Scientists named it the readiness potential and took it as a sign that the brain was gearing up to act.4Neurophysiology of Consciousness. Readiness-Potentials Preceding Unrestricted’ spontaneous’ vs. Pre-Planned Voluntary Acts Its early part comes from areas on the inner surface of the frontal lobe that help plan movements.5World Neurology. What the Libet Experiment Reveals About Free Will
What Libet added was a way to time the conscious urge. Volunteers sat still, watched a dot sweep around a clock face, and waited for a spontaneous urge to flex the wrist. There was no reason to move at any particular moment, which turns out to matter later. After moving, they reported where the dot was when they first felt the urge, while electrodes on the scalp recorded the brain the whole time.6PubMed. Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential). The unconscious initiation of a freely voluntary act
The clock was deliberately fast, with the dot circling in a couple of seconds. That sped-up dial let volunteers point to a position that could be converted into a time within a fraction of a second. Researchers translated the reported position into a time, called W, for the conscious intention.
Each volunteer repeated the task dozens of times, and the recordings were averaged, because a single trial’s electrical trace is too noisy to show the slow build-up. The averaged trace was lined up on the moment the muscle actually moved, and the reported urge was placed on that timeline. The readiness signal started at least several hundred milliseconds before W.7PubMed. Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential). The unconscious initiation of a freely voluntary act
Libet concluded that the brain can begin a voluntary act before the person is aware of deciding.8PubMed. Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential). The unconscious initiation of a freely voluntary act The gap between brain signal and reported urge has been found again many times.9PubMed. (Non)-experiencing the intention to move: On the comparisons between the Readiness Potential onset and Libet’s W-time Popular accounts turned “starts before awareness” into “the brain chose for you,” but Libet himself did not conclude that free will was gone, as later sections show.
The finding is about timing, not about who or what did the choosing. The signal was only visible after the movement had happened, in averaged trials, so it could not predict a move in real time. Nothing in the recording shows it holding a decision.
Later Studies Pushed the Gap Further
Libet’s averaged scalp signal could only be seen looking back from the movement. Later studies tried to predict a choice before it happened, which is a harder and more convincing test. A computer is trained on many earlier trials and then guesses new ones, and how often it guesses right is the accuracy figure these studies quote.
One study recorded from 256 individual neurons in a movement-planning area. It predicted an upcoming decision to move with over 80% accuracy 700 milliseconds before people were aware of it, and it could estimate when the decision would happen to within a few hundred milliseconds.10PubMed Central. Internally generated preactivation of single neurons in human medial frontal cortex predicts volition That requires electrodes inside the brain, so it was done only in patients already undergoing surgery. Even good prediction leaves room for one miss in five, and predicting a choice is not the same as showing that nothing conscious mattered.
Brain scans produced the “seven seconds” headline. People choosing between two buttons showed hints of the choice about 7.5 seconds before they reported deciding.11PubMed Central. Tracking the Unconscious Generation Free Decisions Using UItra-High Field fMRI But the best accuracy was about 57%, only a little better than the 50% of a coin toss.12PubMed Central. Tracking the Unconscious Generation Free Decisions Using UItra-High Field fMRI
The two results are not interchangeable. The single-cell result is far more accurate, but it covers only whether and when to move, in a small group of patients. The scan result covers which of two buttons, in healthy volunteers, and is much weaker. Single-cell recordings and scans also measure the brain very differently, from individual neurons firing to blood-flow changes, so their results can’t simply be stacked into one finding.
A long lead time with barely better than chance accuracy says the early signal is weak, not that the decision was made seven seconds ago. For any such headline, the useful question has two parts: how early, and how often right.
Can You Still Stop an Action Your Brain Has Started
Libet suggested that even if an urge starts unconsciously, the conscious mind might still block it. This is sometimes called “free won’t.” In a later test, people could stop a self-started movement even after the readiness signal had begun.13PubMed Central. The point of no return in vetoing self-initiated movements Volunteers got ready to move at a set time and were then given a signal to stop. That tested whether they could cancel a movement already being prepared, rather than just reporting an urge.
Because stopping worked when the cue came early enough, the preparation signal can’t be a fixed commitment to move. But stopping only worked if the stop cue came more than about 200 milliseconds before the movement.14PubMed Central. The point of no return in vetoing self-initiated movements Past that point there is no turning back.
The veto gives conscious awareness a role even if urges start below awareness: it can act as a final check on what the brain has begun. Control may work more as editing urges than generating them, like pulling your hand back from a second biscuit. Some argue that the veto must itself be caused unconsciously, so it settles nothing. But the data show the action isn’t locked in when the signal starts. Whether stopping is “free” is a philosophical question the experiment can’t answer. It shows only that the process can be interrupted, which is a smaller claim and the one the data support.
Is the Readiness Signal Really a Decision
The readiness signal is seen only after averaging many trials lined up to the moment of movement. That detail matters for what it means. One model says that when there’s no reason to move at any particular time, random ups and downs in brain activity decide when a movement threshold is crossed.15PubMed Central. An accumulator model for spontaneous neural activity prior to self-initiated movement Lining up trials at the movement then makes those random fluctuations look like a steady rise.16PubMed Central. An accumulator model for spontaneous neural activity prior to self-initiated movement
In the model, brain activity drifts up and down at random, a bit like static, and a movement happens when it happens to cross a threshold. The timing of the crossing is set largely by chance rather than by a plan. Looking back from each movement, the stretch just before it is likely to include an upswing, since an upswing is what caused the crossing. Averaging many such stretches gives a smooth rise that looks like preparation, though each single trial was just noise that happened to cross.
So the signal may show the brain on its way to a decision, not a decision already made.17PubMed Central. What Is the Readiness Potential? The model also correctly predicted new patterns in people’s brain recordings and behaviour.18PubMed Central. Specific Relationship between the Shape of the Readiness Potential, Subjective Decision Time, and Waiting Time Predicted by an Accumulator Model with Temporally Autocorrelated Input Noise
If the signal is partly noise, “the brain decided first” is no longer the natural reading. On this account the sense of “I decided now” could arrive when the threshold is about to be crossed, and the gap with awareness may mean little about who chose. The model doesn’t say the activity is meaningless: it says the activity sets a readiness level and chance tips the final timing. Researchers still debate which interpretation is right, so the signal’s meaning is open.
How Reliable Is the Moment People Say They Decided
Changing the clock’s speed and number of markings changes when people say they felt the urge.19PubMed Central. Manipulations of Libet clock parameters affect intention timing awareness People also misjudge the timing of outside events like a beep or a flash, depending on the clock and sense involved.20PubMed. Biases in the subjective timing of perceptual events: Libet et al. (1983) revisited So the reported urge time partly reflects the apparatus rather than purely the person’s inner moment, and the error isn’t limited to inner feelings.
An urge has no sharp edge; it builds rather than switching on. You must split attention between the clock and your own feeling, then report afterward, relying on memory of a fleeting moment. The W time could be off by enough to matter for a gap of a few hundred milliseconds, since that is about the size of the possible error.
Errors don’t erase the finding that brain activity starts first, but they make the exact size of the gap uncertain. The sensible reading is that the earlier signal is probably real but the exact lead is not known.
A cleaner design would not depend on a remembered clock position. Some later studies used other ways of marking the moment of decision or tested whether people could respond to a prompt. Whatever the method, it still relies on the person’s report of something inner, so some uncertainty about timing remains in every version.
Wrist Flicks Versus Decisions That Matter
These experiments study choices about acting right now, not plans like phoning a friend tomorrow.21Oxford Handbooks Online. Free Will and Science Planning ahead is outside what they test, though when people worry about free will they usually mean exactly those considered choices.
An arbitrary choice is one where nothing favours either option, such as which of two identical buttons to press. A deliberate choice has reasons and consequences that make one option better. Libet-style tasks remove reasons on purpose, because the aim was to catch a spontaneous urge. That design is also why the results may not carry over. When nothing hangs on the choice, random brain noise is free to tip it.
In one study where choices decided real donations to charities, the readiness signal showed up for arbitrary picks but was absent before deliberate ones.22eLife. Neural precursors of decisions that matter—an ERP study of deliberate and arbitrary choice That suggests the two kinds of choice may work differently. Another study found a stronger signal before deliberate decisions.23Neuroscience of Consciousness. A multi-measurement study of the relation between deliberation and volition The two used different tasks and ways of defining deliberate choices, so a difference in the signal may reflect design more than a real conflict in the brain. Whether the classic signal applies to decisions that matter is unsettled.
Weighing reasons happens over seconds or days, far beyond the experiment’s window. Deciding whether to take a job involves comparing options, talking to others and changing your mind, none of which fits in a few seconds in a lab. If noise tips tiny choices, that doesn’t imply noise also decides weighty ones. Still, the lab results are a fair reminder that some of what feels like a decision, such as when to reach for a glass, may start before you notice.
Does It Change Responsibility or How People Behave
Criminal law asks whether a person had an intention and understood its likely consequences.24PubMed Central. Criminal Responsibility and Neuroscience: No Revolution Yet Where that intention came from in the brain doesn’t change responsibility under current law.25PubMed Central. Criminal Responsibility and Neuroscience: No Revolution Yet Someone who plans and carries out a crime has reasons, knowledge and the chance to stop, and Libet’s meaningless wrist flicks are far from the planned acts criminal law deals with.
An early study found people cheated more after reading that free will is an illusion.26Psychological Science. The Value of Believing in Free Will: Encouraging a Belief in Determinism Increases Cheating It got wide attention because it suggested that learning about determinism could make people behave worse. But careful attempts to repeat it largely failed.27PubMed. Does encouraging a belief in determinism increase cheating? Reconsidering the value of believing in free will A review of 145 experiments found such messages lower belief in free will but showed no clear effect on behaviour.28PubMed. Manipulating Belief in Free Will and Its Downstream Consequences: A Meta-Analysis The worry that learning about Libet will make you less honest or less responsible isn’t supported by the evidence. A sensible takeaway is that your sense of choosing is partly a construction, but your reasons and values still shape what you do.
Real breakdowns of agency look different. In alien hand syndrome, brain damage makes a hand perform purposeful-looking acts the person didn’t intend.29PubMed Central. Alien hand syndrome: Pathophysiology, semiology and differential diagnosis with psychiatric disorders The hand may reach, grab or fumble with objects while the person says they did not mean it. In one case, scans showed the usual movement areas active but extra activity in a region tied to sense of self.30PubMed Central. Alien Hand Syndrome: Neural Correlates of Movements without Conscious Will The feeling of authorship is a brain function that can fail, unlike in Libet’s healthy volunteers, who had the normal sense of owning their actions.

