Was Libet wrong? The question has often been asked since the famous experiments which found that a detectable “Readiness Potential” (RP) showed that our decisions were made half a second before they entered our consciousness. There are plenty of counter arguments but the findings themselves have been reproduced and seem unassailable.
However, Christian Jarrett reports two new pieces of research which shed fresh light on the issue (Jarrett’s book Great Myths of the Brain is very sensible, btw, and should be read by all science journalists).
The first piece of research shows that although an action may be prepared by the brain before we know we’ve decided to act, we can still change our minds. Subjects were asked to press a button at a moment of their choice after a green light came on. However, if a red light appeared before they pressed, they were asked to refrain. The experimenters then detected the RP which showed the subjects were about to press the button, and used the red light to try to cancel the intention. They found that although there was a ‘point of no return’, so long as the red light appeared in time the subjects were able to hold off and not press the button after all.
That means that a significant qualification has to be added to Libet’s initial findings – but it’s one Libet himself had already come up with. He was aware that the emergent action could be suppressed, and memorably said it showed that while we might not have free will, we could still have ‘free won’t’. I don’t know much use that is. In the experiment the subjects simply responded to a red light, but if the veto is to make consciousness effective again we seem to need the veto decision to happen instantly and overtake the action decision somehow. That seems problematic if not paradoxical. I also get quite confused trying to imagine what it would be like to veto mentally a decision you’re not yet aware of having made. Still, I suppose we must be grateful for whatever residual degree of freedom the experiments allow us.
The second piece of research calls into question the nature of the RP itself. Libet’s research more or less took it for granted that the RP was a reliable sign that an action was on the way, but the new findings suggest that it is really just part of the background ebb and flow of neural noise. Actions do arise when the activity crosses a certain threshold, but the brain is much quicker about getting to that level when the background activity is already high.
That certainly adds some complexity to the picture, but I don’t think it really dispels the puzzle of Libet’s results. The RP may be fuzzier than we thought and it may not have as rigid a link to action as we thought – but it’s still possible to predict the act before we’re aware of the decision. What if we redesigned the first experiment? We tell the subject to click on a target at any time after it appears; but we detect the RP and whip the target away a moment before the click every time. The subject can never succeed. The fact that that is perfectly possible surely remains more than a little unsettling.