Imagine sitting in a laboratory with one instruction:
Move your finger whenever you feel like it.
Do not plan the moment.
Just watch a rapidly moving clock and remember when you first became aware of the urge to move.
Electrodes record your brain activity.
Your finger moves.
Then the researchers compare three moments:
When did brain activity begin to change?
When did you report becoming consciously aware of wanting to move?
When did the movement happen?
In the 1980s, Benjamin Libet and colleagues reported a result that became one of the most famous findings in the science of volition: a slow electrical signal associated with movement preparation—the readiness potential—began before participants reported conscious intention.
A dramatic interpretation followed.
Your brain decides before you do. Therefore free will is an illusion.
It is a powerful headline.
It is also much stronger than the experiment itself.
What Libet actually found
Libet's study compared EEG activity with participants' reports of when they first felt the wish or intention to act.
The readiness potential began several hundred milliseconds before the reported conscious intention.
At minimum, the result challenged a simple picture in which a fully formed conscious command appears first and the brain merely executes it afterward.
That is scientifically important.
But it does not automatically tell us what the readiness potential is.
Does it mark a completed decision?
General motor preparation?
A buildup toward movement?
Spontaneous neural fluctuation that becomes visible when many trials are averaged?
And what exactly does the reported moment of conscious intention measure?
Those questions became central to the decades of research that followed.
The hidden assumption in “your brain decided first”
The phrase sounds decisive because it quietly treats “your brain” and “you” as two rival agents.
But neural activity is not evidence that someone other than the person made the decision.
Every serious theory of human choice expects decisions to have physical implementation in the nervous system.
The relevant question is narrower:
What kind of neural process occurred before conscious awareness, and what role did conscious deliberation play in the action?
That cannot be answered merely by placing one timestamp before another.
The readiness potential may not be a hidden decision
A major alternative interpretation came from computational work by Aaron Schurger and colleagues.
They modeled spontaneous self-initiated movement as an accumulation-to-threshold process. In a task where participants are told to move at some unspecified moment, ongoing neural fluctuations can drift until activity crosses a threshold and movement occurs.
When researchers align many trials backward from the moment of movement and average them, those fluctuations can produce the familiar ramp-like readiness potential.
Under that model, the early buildup does not have to mean that the brain secretly committed to one decision long before consciousness arrived.
It can partly reflect how spontaneous neural activity and threshold crossing look after the data are time-locked to the action.
A 2021 review of the readiness potential concluded that modern computational models substantially complicate its classic interpretation and its use in arguments against free will.
This does not erase the signal.
It changes what the signal may mean.
Even the famous timing result is less secure than the legend
A 2021 meta-analysis reviewed Libet-style experiments.
The broad temporal pattern often resembled Libet's original finding, but the authors highlighted something surprising: for the most philosophically important comparison—the time difference between unconscious brain activity and reported conscious intention—only a small number of studies contributed suitable evidence, and uncertainty was substantial.
The point is not that “Libet was disproven.”
It is that decades of cultural certainty grew around a result whose measurement and interpretation are more fragile than the slogan suggests.
Scientific findings can be real while their philosophical reputation outruns them.
What about predictions ten seconds before a choice?
In 2008, Chun Siong Soon and colleagues used fMRI pattern classification while participants freely chose between left- and right-hand button presses.
Patterns in prefrontal and parietal regions contained information about the eventual choice several seconds before participants reported conscious awareness of deciding.
The result is fascinating.
But “predictable” does not mean “already fixed with certainty.”
Reviews of this work note that early classification performance was only modestly above chance—roughly in the 55–60% range in the relevant analyses—not 100% deterministic prediction.
That distinction is enormous.
A weak statistical bias before awareness could reflect developing preferences, background neural states, sequential effects or other precursors without showing that an irrevocable decision had already been completed.
Prediction is evidence of information.
It is not automatically evidence of metaphysical inevitability.
The biggest problem: what kind of choice are we studying?
Most famous free-will experiments use deliberately meaningless choices.
Press left or right.
Move whenever you feel like it.
Choose now, but do not use a reason.
That design is useful because it strips away complexity.
It may also strip away much of what people actually care about when they say “free will.”
Real decisions often involve reasons, values, consequences, memory, conflict and deliberation.
Should I tell the truth?
Leave the relationship?
Take the job?
Donate the money?
Change my mind?
In 2019, Uri Maoz and colleagues directly compared arbitrary choices with deliberate decisions involving real charitable donations. They found the expected readiness potential before arbitrary decisions but not before the deliberate choices.
Their result does not prove free will.
It does undermine the assumption that one neural signature from arbitrary finger movements automatically generalizes to reasoned decisions that matter.
The feeling of agency is another question again
There is also a difference between free will and the sense of agency.
Sense of agency is the subjective feeling that “I am causing this action” or “I controlled that outcome.”
Researchers study it using explicit reports and indirect measures such as intentional binding, in which voluntary actions and their effects can be perceived as closer together in time.
Modern reviews describe agency as emerging from multiple cues and mechanisms: motor predictions, sensory feedback, expectations, contextual information and retrospective interpretation can all contribute.
That means your feeling of control is not an infallible detector of metaphysical freedom.
It can be strengthened, weakened or misattributed.
But the reverse mistake is just as important:
showing that the feeling of agency is constructed does not prove that agency itself is unreal.
Vision is constructed too. That does not mean there is nothing to see.
Neuroscience and philosophy are answering different parts of the problem
“Free will” is not one simple laboratory variable.
Philosophers disagree about what kind of control matters.
Some compatibilist theories argue that freedom can exist even in a causally determined world if actions flow from the agent's reasons, values and capacities in the right way.
Incompatibilist theories hold that genuine freedom requires something determinism cannot provide, often including a stronger ability to do otherwise or a special kind of sourcehood.
Skeptical positions doubt that humans possess the freedom required for ultimate moral responsibility.
A button-press experiment cannot decide among these theories unless its data connect to the specific condition the theory says freedom requires.
That is why the Stanford Encyclopedia of Philosophy describes disputes about free will as involving metaphysics and ethics as well as empirical science—and notes that the interpretation of Libet-style evidence remains controversial.
What the neuroscience does show
The strongest conclusion is more interesting than either extreme.
Human action is prepared by processes that are not fully available to introspection.
Conscious awareness is not a transparent control room watching every causal step from the beginning.
Choices can be biased by prior neural states.
The subjective moment “I decided now” is not necessarily the first event in the decision process.
Agency itself appears to be constructed from multiple predictive and retrospective signals.
Those are substantial findings.
They challenge naive introspection.
They do not yet amount to a laboratory proof that free will exists—or that it does not.
A better way to think about choice
Instead of asking whether a mysterious inner self appears from nowhere and overrides the brain, ask what capacities an agent actually has.
Can you represent alternatives?
Can you learn from consequences?
Can reasons change your behavior?
Can you inhibit an impulse?
Can you revise a plan when evidence changes?
Can you reflect on a desire without automatically acting on it?
Neuroscience can study many of those capacities.
Philosophy then asks which of them, if any, are sufficient for the kind of freedom that matters.
That moves the debate away from one dramatic timestamp and toward the architecture of control.
The deeper point
The most unsettling free-will experiments became famous because they seemed to reveal a hidden betrayal.
You thought you chose.
But the brain had already acted.
The deeper lesson is less theatrical.
The brain is not the stranger under the floorboards.
It is part of the system doing the choosing.
What remains unresolved is how unconscious preparation, conscious deliberation, reasons, prediction, inhibition and responsibility fit together—and what standard of freedom we expect that system to meet.
Neuroscience has made the old philosophical problem more precise.
It has not made it disappear.
Key Takeaways
- Libet found readiness-potential activity before participants reported conscious intention to make a simple movement.
- That temporal order does not by itself prove that the readiness potential represents a completed unconscious decision.
- Modern accumulator models provide an alternative explanation for the classic readiness-potential buildup.
- A meta-analysis found substantial uncertainty in the most philosophically important timing comparison from Libet-style studies.
- fMRI studies have predicted simple choices above chance before reported awareness, but early prediction was probabilistic rather than certain.
- Readiness potentials found for arbitrary choices may not generalize to deliberate, reason-based decisions.
- The sense of agency is a measurable subjective experience built from multiple cues; it is not identical to metaphysical free will.
- Neuroscience constrains theories of choice, but the existence and nature of free will remain philosophically disputed.
Evidence boundary
The article distinguishes measured neural timing and agency research from philosophical conclusions. It does not present neural precursors as proof that free will is impossible, and it does not present critiques of Libet-style experiments as proof that libertarian free will exists. Empirical claims are separated from metaphysical interpretation and moral-responsibility questions.