The study has appeared on a number of blogs in a number of different contexts but always making the same point – that imaging can detect activity which suggests the decision to press a left or right button is made a number of milliseconds BEFORE we are consciously aware of having made the decision. The accuracy of the prediction based on imaging has been reported at approximately 60% if I recall correctly.

Whilst this accuracy is greater than chance, meaning the prediction has some validity, I think that the conclusion that humans only have an illusion of free will is a very poor one. The decision to press a left or right button is essentially a “random” decision, as random as choosing A or B can ever be from within your own mind, rather than flipping a coin. My point is that, in a “random” decision such as this, is it really surprising that micropatterns in neural activity, which could easily be described as neural “noise” give the person a subconcious inclination to chose one button over another? As the person may feel they are choosing the button “at random” using their own “free will” this is not surprising at all.

Now, if we consider a “reasoned” as opposed to “random” use of free will, the pattern might be very different. Consider quitting smoking cigarettes, which is a hard act of free will, requiring “willpower”. In this case, although it is not yet possible to detect this, we could hypothesize that there would be micropatterns in neural activity relating to dopamine activation of synapses following nicotine inhalation, that give the person a subconcious craving to continue smoking. If an MRI scanner could detect this activity, we might be able to predict if a person would find it easier or harder to quit smoking, with say, 60% accuracy.

Whilst it is arguable that a person might not be able to quit cigarettes if the addiction and craving exceeds their “willpower” it is ridiculous to argue that the person in the study above would not have been able to change their mind and press a different button, should they so wish. This ability, to inhibit our actions and responses, is tested by neuropsychological tests such as the Stroop Test. I would argue that the ability to inhibit our actions and/or change our minds, examples of Executive Function, are an example of a limited capacity for “Free Will” or “Willpower”. The fact that we might chose button A randomly over button B and that this random decision is influenced by neural noise of which we are not conscious does nothing to disprove the concept of Free Will.

I meant to put the following passage (from your post) in quotation marks:

In this case, ‘your consciousness is only aware of some of the things your brain is doing’ seems to me to be the crux of the phenomenology-neurology tangle.

Not to get into the hair-splitting game too much, but is it accurate to say that we are aware of *anything* our brains are doing? I might infer that certain things are happening, say, in my occipital cortex when I see something at a certain orientation, but this isn’t awareness of what my brain is doing.
My gut feeling is that the crux of the phenomenology/neurology divide isn’t so much a lack of phenomenal access to brain events as it is the absence of a mediating step between our phenomenological descriptions and our scientific descriptions, which tend to be based on different ontological (perhaps folk-ontological?) commitments.

I think you actually get at this in your closing remarks on this post.

But I’m also glad to know that my account doesn’t seem like overly-precious semantic hair-splitting, at least to some readers.