Free to Choose-Your-Own-Adventure?

Bandersnatch (2018), the new choose-your-own-adventure film on Netflix, has entertained millions of viewers worldwide since its release at the end of the last year. Besides enabling the viewers to choose the act of the protagonist to affect the plot, Bandersnatch has mesmerized the viewers even more with a compelling insight on free will ⎯ or the absence of free will. In the climax of the show, Stefan, the protagonist of the story, starts to doubt if he is choosing what to do based on his own free will (because the viewers are doing it for him!). Stefan goes as far as doubting his psychotherapist Dr. Haynes, who has indeed been exploiting him as the subject for a life monitoring experiment in one of the plots. By the way, do you know in real life there also is a Dr. Haynes who conducted studies that boosted a discourse on free will skepticism? In this post, we’ll explore his works and how they became neurophysiological evidences that make us question free will.  (Side note: THIS Dr. Haynes did not exploit any subject.)

The Libet Experiment

Before going through works of Haynes and his colleagues, we need to check out Libet’s classic free will experiment first. In the 1980s, Benjamin Libet, an American neurophysiologist designed an EEG experiment to determine the timing of when we make the decision to make a motor movement in our minds. In the experiment, subjects were asked to make a simple motor action such as flexing the wrist. Subjects were to make the movement freely whenever they wanted to; they just had to remember and report the time ‘when’ s/he consciously felt the urge or intention to do it referring to an oscilloscope timer in front of them. Meanwhile, electromyography (EMG) was recording the actual timing of the movement from electrodes attached to the subject’s right arm, and electroencephalogram (EEG) was recording brain activity of the subject.

After the experiment, something intuitively strange was found. Whereas the readiness potential (RP) was initiated 550ms before the action on average, subjects were conscious of their own decision only about 200ms before the action. In other words, the brain knew that the person will move the arm about 350m before that intention actually entered the person’s awareness. If what you are going to do is already determined without your awareness, how can you be sure that any of your action derives from your own free will?

Providing the first neurophysiological evidence that might prove the absence of free will, Libet’s experiment catalyzed countless free will debates and inspired thousands of researchers. One of them was, of course, Haynes.

10 seconds ago, you were meant to read this paragraph

 In 2008, Soon (a PhD student by then) and Haynes with their colleagues refurbished Libet’s freely paced motor-decision task with improved technology. First off, fMRI was used instead of EEG for more extensive imaging of the brain. While the Supplementary Motor Area (SMA), where RP was found in previous studies, only encodes information for late stages of motor planning, fMRI allowed to investigate if regions responsible for high-level planning stages are also involved in the unconscious phase of action planning. Secondly, subjects were to press either the button with their left index finger or with the right as voluntary motor movement. Giving the participants more than one behavioral option to choose from eliminated the argument that some leading brain activity might just reflect unspecific preparatory activation. Lastly, instead of a timer, subjects were presented with a stream of alphabet letters presented one after another for a half second each on the screen inside the scanner to mark the conscious decision time. Subjects only needed to remember which letter they saw at the time they decided to press the button. After the movement, three letters that came before the movement appeared on the screen and the subjects reported the letter they remembered with the buttons.

 To see if certain brain regions indeed encode decision-specific information before the motor decision pattern analysis was conducted on each fMRI image. The result was even more astonishing than the Libet’s study. While the conscious decision for the movement was made within 1 second prior to the execution, fMRI signals from a specific region of frontopolar cortex encoded predictive information about the movement 7 seconds prior to the conscious decision (although the prediction was statistically above chance but not perfect). Considering the sluggishness of BOLD responses, this means that the information about which button (left or right) to press was already in the brain as early as 10 seconds before the conscious decision..Also, the timing information was encoded in pre-SMA and SMA up to 5 seconds prior to the conscious decision, which is even earlier than RP that was observed in the Libet experiment.

 Simply put, it seems that when we are going to make a decision and what the outcome of the decision will be are determined in our brain a few seconds before we consciously make that decision. In fact, if that’s the case, it’s wrong to say we ‘consciously make a decision’, since we are just executing what was heralded in the brain 10 seconds ago. And 10 seconds is regarded as forever in timelines of the neuroscience world (Also, interestingly, the time limit the viewer has to choose the action for the character in Bandersnatch is 10 seconds). If our consciousness is not even there when our actions are shaped, do we really have a ‘free will’?

Arguments against arguments – Do the Libet paradigm reflect what we reckon a free will?

With this study in 2008, Haynes and his colleagues successfully supplemented and developed Libet’s study. However, there were more arguments against their study yet to be solved. One example of them was that the responses by the subjects might have been affected by the previous response choices. If sequential dependency was present throughout the trials, the whole analysis should have been done again with different measures. In 2013, Haynes and his colleagues analyzed the statistical properties of the responses of each individual-subject in their previous experiments. With this work, they could confirm that the subjects’ behavior was random enough with sufficiently high entropy rate. They also used a modelling approach to directly confirm that the information was related to upcoming choices, not to the previous choice. In the same year, Haynes and his colleagues dealt with a more philosophical question as well. Do simple motor-decisions represent every choice of our daily lives? Decisions we make every day are not confined to flexing the wrist or pressing a button. They are often more complex, engaging abstract thoughts and concepts. What if prognostic information in the brain can only predict simple motor movements and not abstract decisions? Haynes and colleagues set up a new experiment this time to respond to this  question. In this experiment, five numbers were added to the one-letter-screen usedin the previous task; four on each corner and one right above the letter at the center. Two of the four numners in the corner were always the sum and the subtraction of the central numbers on the previous two screens. Subjects were asked to freely conduct an addition or a subtraction at any point of time, remember what letter was on the screen when they decided to do the calculation, and report the letter after the calculation. The results showed that choice-predictive information was again encoded in a frontopolar region and a region spread from the precuneus to posterior cingulate just as in the previous study, although it preceded the conscious decision by 4 seconds instead of 7 this time. Still, a great counterpunch to those who expected Libet-style freely paced decision task would show results only for simple motor decisions.

However, is the thing observed in Libet-style experiment indeed ‘free will’? What is free will anyways? In 2017, Haynes and his colleagues tried to answer this more fundamental and philosophical question to prove external validity of their experiments. The problem of the concept of free will was that there was no clear and dominantly agreed definition for it. In this case, what was necessary was to figure out what common people mostly assumed as free will. In order to do this, a survey  consisted of 12 scenarios describing different actions that 668 respondents had to  evaluate to which degree each was free in terms of several factors. As a result, it was shown that lay people regard conscious intention to be an important factor for a free action, and consider spontaneous and immediate actions such as the responses made in the experiments of Libet and Haynes to derive from free will. In contrast, decisions based on deliberation and reasons were not regarded as free.

Your free choice to believe in free will

One might raise a tricky question here: Don’t these whole action-prediction studies actually contribute to fighting against determinism? Because you know, what can be predicted can also be prevented. Yes you can fight fate, but only within a time limit. Schultze-Kraft and Haynes conducted an experiment in which the participants were asked to act more unpredictably to beat the predictions of a real-time EEG classifier. It turns out that people can still veto a prepared action and choose not to make it, as late as 200ms before the movement onset. But then again, let alone this 200ms-long ‘point of no return’, being able to intervene with a brain process to make an action only means that yet another brain process to cancel the action is overtaking. One may keep investigating for a veto to veto and a veto to veto to veto and so on, but it seems that the idea of free choice will still remain in question.  As Jean-Paul Sartre once said, life is C(choice) between B(birth) and D(death). Surely, your choices do shape your life and who you are. Now since we’ve learned these choices might not be under our control, how can we explain our lives? Maybe you should shout out in the air like Stefan did in Bandersnatch to find the person who’s been making the choices for your life from the other dimension. But be careful, your show might be just turned off with a remote control.