(For previous posts in this series, see here.)

In a recent paper (Unconscious determinants of free decisions in the human brain, Chun Siong Soon, Marcel Brass, Hans-Jochen Heinze, & John-Dylan Haynes, Nature Neuroscience, vol. 11, no. 5, May 2008, 543-545), researchers used the more sophisticated modern technique of fMRI (functional magnetic resonance imaging) to measure brain activity. The paper is not available online without a subscription but you can read a news report on the results of their paper here.

This experiment was designed to meet two key concerns about the Libet studies: that the time interval between act and the precursor unconscious brain activity prior to act was too small to definitively rule out measurement errors, and that Libet’s team had not shown that the early brain activity was a predictor of a specific decision.

The fMRI studies find that our decisions as to what actions we will take originate in our unconscious neural activity and only later informs our conscious mind of it, thus providing strong evidence against the existence of free will. The paper describes what the researchers asked their test subjects to do while they were hooked up to fMRI measuring devices.

The subjects were asked to relax while fixating on the center of the screen where a stream of letters was presented. At some point, when they felt the urge to do so, they were to freely decide between one of two buttons, operated by the left and right index fingers, and press it immediately. In parallel, they should remember the letter presented when their motor decision was consciously made. After subjects pressed their freely chosen response button, a ‘response mapping’ screen with four choices appeared. The subjects indicated when they had made their motor decision by selecting the corresponding letter with a second button press. After a delay, the letter stream started again and a new trial began.

Each letter was shown on the screen for 500 ms before switching to a new one and this was the time marker used by the researchers to determine when the decision to push a button was made. Note that in this experiment there are two decisions involved: when to push a button and which button to push. The fMRI data enabled the researchers to use sophisticated decoding computer programs to detect predictive signal patterns in brain activity, even in the absence of an overall increased signal strength, the latter being what the earlier experiments had depended upon. This enabled the detection of far more subtle effects.

In the trials it turned out that patients pushed both left and right buttons equally often and they were conscious of the decision to press within a time interval of one second before actually pressing a button.

But it is the other results of the experiment that are dramatic. Soon et. al. found that there was precursor activity in regions of the brain other than the SMA regions probed by Libet, and that this activity occurred much earlier than the SMA activity. Furthermore, this activity also predicted which button was going to be pushed.

[T]wo specific regions in the frontal and parietal cortex of the human brain had considerable information that predicted the outcome of a motor decision the subject had not yet consciously made. This suggests that when the subject’s decision reached awareness it had been influenced by unconscious brain activity for up to 10 s. (my italics)
…
Notably, the lead times are too long to be explained by any timing inaccuracies in reporting the onset of awareness, which was a major criticism of previous studies. The temporal ordering of information suggests a tentative causal model of information flow, where the earliest unconscious precursors of the motor decision originated in frontopolar cortex, from where they influenced the buildup of decision-related information in the precuneus and later in SMA, where it remained unconscious for up to a few seconds.

This figure shows the time sequence of events that the study revealed.

The earliest precursors of an action lie in regions of the brain other than the SMA, which is the region that caused the electrophysiological effects that Libet was measuring, which explains why Libet (and Grey Walter before him) got just a half-second lead time while now it is a whopping 10 seconds. The new fMRI studies also enabled the researchers to determine that the leading brain activity selectively predicted the outcome of the subject’s choice of which button to push, and was not simply indicative of some nonspecific preparatory processes, which was the criticism made by Trevenna and Miller of the Libet team’s experiments.

They also found that the decision to push a button could be predicted up to five seconds before the act, and this information was present in the SMA and pre-SMA regions of the brain.

In this video clip, Marcus Du Sautoy records his experience of participating in this same experiment. (Incidentally, Du Sautoy, Professor of Mathematics at the University of Oxford, is the successor to Richard Dawkins as holder of the Simonyi Professorship for the Public Understanding of Science.)

Note that the predictions of which button to push were not perfect, with only around 60% accuracy. The absence of 100% accuracy is probably due to the lack of precision of the detecting apparatus and inadequacies of the pattern-recognition software, both of which are bound to get more sophisticated with time, thus increasing the accuracy of predictions. But the fact that the result is better than chance means that, as lead author Haynes says, “there’s not very much space for operation of free will” because “[t]he outcome of a decision is shaped very strongly by brain activity much earlier than the point in time when you feel to be making a decision.” Other researchers concur.

Dick Passingham, a cognitive neuroscientist at the University of Oxford in the U.K., says the paper clears up one of the major concerns about the original Libet experiment. “This activity that occurs earlier is … not just general preparation, it really is a proper decision,” he says.

Neurologist Mark Hallett of the U.S. National Institute of Neurological Disorders and Stroke in Bethesda, Maryland, says the study confirms his understanding of free will as a perception rather than a driving force.

This seems to pretty much kill the idea of free will as traditionally understood. As biologist Anthony Cashmore says, “The reality is, not only do we have no more free will than a fly or a bacterium, in actuality we have no more free will than a bowl of sugar. The laws of nature are uniform throughout, and these laws do not accommodate the concept of free will.”

The Ghost in the Machine seems to be well and truly exorcised.

Next: What about quick decisions?