Benjamin Libet's classic 1983 experiment purported to show that preparatory brain activity precedes our conscious decision to move - a controversial finding interpreted by some as evidence that free will is illusory.
In Libet's study, participants reported the time on a clock at the instant they had decided to move a finger. This is less straightforward than it sounds. Visual processing is sluggish whereas participants were presumably instantly aware of when they'd made a conscious decision to move. This would have led them to report a decision time that was too early (i.e. at the instant of their decision, the participants' brains would only just have been getting round to processing an earlier time on the clock).
Libet's team realised this, so in a separate control condition they also asked participants to report the timing of an electrical stimulus applied to their hand - the error in this time estimation was then used to apply a correction to participants' estimates of when they'd made a movement decision.
But in a new study, Adam Danquah and colleagues point out that our different sensory modalities operate at different speeds. They copied the control condition of Libet's experimental set-up, but they asked participants to report not just the timing of a mild electric shock, but also of a flash in the centre of the clock, and the sound of a click (delivered through headphones).
The researchers found that the participants' estimates were less accurate (i.e. even earlier) for the visual flash and auditory click than for the electric shock. In other words, Libet would have arrived at a different estimate of when participants had made a decision to move if he'd used a visual or auditory control task to make his adjustment.
"The degree of variability in bias across modalities and studies means that it is very difficult to know what correctional standard, if any, can be applied to awareness times of endogenous events [e.g. decisions]," the researchers wrote.
However, defenders of free-will shouldn't take comfort in these new results. Danquah and his colleagues added an important note about the implications of their work: "the magnitude of the biases reported here suggests that they [Libet's team] underestimated the degree to which... [preparatory brain activity] preceded the intention to move!"
In a second experiment, Danquah and his colleagues also identified another problem with the Libet paradigm. The clock used by Libet featured a dot that circled the clock-face, rather like a second-hand. Danquah's team showed that the speed at which the dot circled the clock face also affected participants' time estimations - the faster the dot, the more accurate participants' estimates became.
"The results reported here have implications for the whole tradition of having participants locate temporally subjective events using the clock paradigm," the researchers concluded.
A DANQUAH, M FARRELL, D OBOYLE (2008). Biases in the subjective timing of perceptual events: Libet et al. (1983) revisited Consciousness and Cognition, 17 (3), 616-627 DOI: 10.1016/j.concog.2007.09.005
Disclaimer: I was a participant in this study several years ago, during my student days in Manchester.