The Cognitive Science of Free Will

How to make the best decisions is one of the “Key Questions” the @lantis Learning Community is interested in.

This lesson is focused on how the brain constrains our decision making by fooling us about Free Will

One of the main questions I investigate is what Cognitive science can tell us about how we make decisions.  Here is a White Paper on “Agency, freedom, and responsibility.”

Why looking at the Cognitive Science of Free Will is important.

We need to understand the cognitive science of free will because we need to be able to adjust for our biases.  The Serenity Prayer asks us to change the things we can change, accept the things we cannot change, and learn to know the difference.  By studying how the brain works we can make better decisions by being able to adjust for those things we cannot change.

What is free will?

“Free Will” seems to be a familiar feature of our everyday lives — most of us believe that what we do is up to us. For example, I’m deciding what to write now.  No one is forcing me to write these particular words.  I acted freely when I decide to go for a run this afternoon.

Free will is not just that I move about in the world to achieve a goal, but that I exercise meaningful control over what I decide to do. My decisions and actions are up to me in the sense that they are mine — a product of my values, desires, beliefs, and intentions.

Correspondingly, it seems to me that at least at times I could have decided to and done something else than what I did. I decided to go for a run this afternoon, but no one made me and I wasn’t subject to any compulsion; I could have gone for a coffee instead, at least it seems to me.

I Googled “Free Will” and got 24,740,000,000 results.

Apparently there are two aspects to the term “Free Will.”  First, Free will is the ability to choose between different possible courses of action unimpeded.  And second Free will is closely linked to the concepts of moral responsibility, praise, guilt, sin, and other judgments that apply only to actions that are freely chosen.

I am not going to focus on the second definition.  I am not going to focus on how Free Will is closely linked to the concepts of moral responsibility, praise, guilt, sin, and other judgments that apply people may or may not ascribe to the term “free will.”

Rather I am focusing on the First definition, “Free will is the ability to choose between different possible courses of action unimpeded.”

Given this definition, my position is our actions are always constrained by any number of external or internal factors.  So, we have free will, but that free will is bounded by both known and unknown constraints.

How do neuroscientists study free will?

There are plenty of sensational claims about the brain science of free will out there and lots of back and forth about whether or not science disproves free will (e.g., “My brain made me do it”).

The current neuroscience of free will traces its lineage back to an influential experiment by Benjamin Libet and his colleagues. The majority of our actions begin with bodily movements, and most of us think that when we decide to move (e.g., decide to pick up my cup of tea), first I, the agent or person, decides and then I hand off control, so to speak, to the brain circuits for motor control to execute the action.

It was known since the 1960s from work by Kornhuber and Deecke that there is a slow buildup of negative brain activity in the supplementary motor area (SMA) and pre-SMA measurable by electroencephalography (EEG) just prior to voluntary (i.e., movement initiated by the participant) bodily movement. This brain activity, called the readiness potential (RP), was taken to be neural preparation to move for spontaneous movements and starts about a half-second before the time of the movement.

So Libet and his fellow researchers ask when does the agent appear in relation to the RP? The agent’s decision has to be something measurable in the lab, so Libet asked participants to make movements (of the finger or wrist) at a time of their choosing and then report after the fact when they were first aware of their decision or urge to move using a modified clock (termed ‘W time’).

Libet found, contra the commonsense expectation, that the average reported time of first awareness of the decision to move, W-time, occurred almost a third of a second after the start of the RP. So Libet (and select others since) concluded that the RP is the brain’s unconscious decision to move with the agent’s decision occurring later

Libet took this as evidence that the conscious agent or self doesn’t initiate, or kick off, preparation to act, the unconscious brain does. He argued that this result is representative of how all of our voluntary movements are produced, and, if so, then the agent’s conscious decision to act doesn’t initiate the process leading to movement. But if the agent doesn’t play this initiating role in acting, how can it be up to me how I act?

These results have worried a lot of folks and inspired a booming research enterprise in cognitive neuroscience and philosophy. One shouldn’t jump to the depressing conclusion, though, that we don’t act freely or don’t really deserve any of the moral reactions others have to our actions; there is a healthy discussion on how the original Libet results can be interpreted as consistent with that picture of us humans as self-governing and free and moral persons.

W-time is taken to indicate the moment of awareness of a decision. Can we capture “moments of conscious awareness” scientifically?

Since the initial publication of Libet and colleagues’ study, worries about whether we could measure the time of conscious awareness have been voiced. After all, we are talking here about the timeframe of milliseconds.

In these studies, all of the events measured prior to movement in the lab are happening within one second before the participant wiggles a finger or hand (now button presses are the preferred movement). Libet argued that W-time within a reasonable range was reliable since we can see how accurately participants in the lab estimate the time of other events, such as skin shocks. The reliability of W-time has recently been challenged yet again with a new study that concludes that depending on the order in which participants complete certain tasks in the experiment, W-time can be strikingly different (i.e., there is an order effect; see here).

Other researchers are currently exploring alternative ways to measure a decision to move in the lab, including work by Parés-Pujolràs and co-authors, who have been using an online (i.e., pre-movement) measure of the agent’s awareness of a decision to move (here).

In these studies, participants watch a continuous stream of letters on a computer while spontaneously pressing a button. Every now and then, though, the letters change color. When this happens participants are told to press the button just then if they were already aware of their preparing to press the button soon. These kinds of online measures of awareness may yet prove to be more reliable ways of getting at whether people have conscious intentions to act in the lab.

What’s the latest work on neuroscience of free will?

Two of the hottest topics seem to be, first, what exactly the RP, that negative build-up of brain activity pre-movement, really signifies and, second, how we can make our voluntary actions in the lab more ecologically valid.

As to the first, the past decade has seen researchers investigating if we have evidence that the RP really does stand for a decision to move or, alternatively, if the RP just is the brain’s being biased to move in some way (say, left, instead of right) without the commitment to do so.

Others test the possibility that the RP isn’t really movement specific activity at all (e.g., general cognitive preparation to perform a task voluntarily). Others, such as Schurger and colleagues, have argued via empirical studies that the RP is the neural signature that we pick up when are actions are generated by neural noise crossing some threshold (here). That possibility would be alarming as then our actions, which we take to be undertaken by me for reasons, may really just be the passive result of fluctuating brain activity.

As to the second hot issue, researchers are now attempting to design tasks in the lab that are closer to the kind of decisions and action that we engage in daily. Libet argued that a simple movement like a wrist flex or button press could stand-in for the more complex actions, as the RP has been shown to occur prior to more complex movements in the lab. Hence we could give a unified explanation of the timing of events involving practical decisions and bodily movements.

But many have voiced concern that when to press a button or whether to press a left or right button, just isn’t the right kind of action to stake a claim that we as agents don’t initiate our actions via our conscious intentions to act. Hence, some of the ongoing work involves making the choice of which button to press or when to press it meaningful via rewards or penalties for skipping ahead or value-laden options, such as charity donations.*

And, of course, there are plenty of neuroimaging tools at the disposal of cognitive neuroscientists. Some of the most interesting replications and extensions of the Libet findings have been done using single-cell recording and fMRI among other technologies (see here and  here, respectively).

In fact, the neuroscience of free will has been and currently is the focus of some major research grants, such as the Big Questions in Free Will project (2010-2014, Principal Investigator Dr. Alfred Mele) and the Consciousness and Free Will project (2019-, a collaboration across 17 PIs), each of which involves philosophers and numerous neuroscientific labs worldwide. From these grants, we should expect further clarity on what’s going on under the hood, so to speak, when we decide what to do and act voluntarily.

Are there any other results in neuroscience that tell us something intriguing about our agential control?

Yes, one of the aspects of our lives that seems the most undeniable is that we really do experience ourselves as in control of our movements and their effects in the world. There is a large body of work in cognitive neuroscience that focuses on this sense of agency via research on what’s been termed intentional binding (for a recent academic review see here).

Basically, if you ask participants in clever experimental set-ups to judge whether some event (e.g., icon moving on a computer screen) was the outcome of their agency or someone else’s (i.e., “I did that” judgments), participants tend to misjudge an outcome to be a result of their own agency if it is a positive one and misjudge an outcome to be the result of another’s agency if it is a negative one. That is, there is a self-serving bias to the explicit sense of agency judgments (For interesting results in this regard see Wegner and Wheatley’s 1999 paper here and other earlier work in psychology on attribution theory).

Cognitive neuroscientists have found a methodology to study our sense that we are in control of our actions and actional outcomes without surveying participants’ explicit “I did that” judgments. Instead, experimenters ask participants to judge the time of various events, including their movements  (e.g., a button press) and the sensory outcomes of those movements (e.g., a beep following the button press). What researchers have found is that if you voluntarily press a button and hear a tone as a consequence, you are going to judge that the time of the movement and the time of the tone are much closer together in perceived space than if you are caused to move (via neural stimulation) and hear a tone as a consequence.

In other words, the perceived time of the action and the tone “bind together” in perceptual space when you act voluntarily as opposed to when you are caused to move or simply judge the time of events without acting (here). What’s intriguing about this research on agency, then, is that our perceptual judgments about the world seem to distinguish when we act from when something is done to us. Research work on intentional binding has tackled more ecologically valid issues of sense of agency when acting under emotional distress, due to coercion, and in the face of options.*

* Neuroscientists working on more representative kinds of decisions and/or sense of agency in more ecologically valid contexts include researchers in the UCL Action and Body Lab at University College London and The Brain Institute at Chapman University, among others.