Evidence for predictive awareness


Is seems generally accepted in neuroscientific circles that the brain predicts the results of motor action, constructs a prediction of sensory signals, compares the prediction with the outcome and uses the error to correct motor action and perceptual processes. It is also accepted that the prediction is singular and global in nature. Many assume it is Bayesian, at least in spirit. Below is some representative evidence for this predictive process being linked to the awareness that we experience. What we are aware of is the prediction.

(This was written to appear elsewhere and so I have avoided the use of the word consciousness in order to bypass fruitless discussion. But most neuroscientists do not avoid the word and in fact, many consider it the object of their investigations. If you find the wording awkward, then just substitute consciousness for awareness.)

  1. The failure of simultaneity between moving events and stationary objects: Nijhawan’s experiment had an object move across the visual field and pass a flashbulb on the way. The flashbulb flashed at exactly the moment that the object passed it. Subjects reported perceiving the object pass the bulb before the bulb flashed. The brain ‘sees’ a split second into the future for moving objects but not individual stationary ones. Bai’s showed that there were more mistaken ‘out’ calls in tennis than ‘in’ calls because the officials perceive the ball is having moved further then it had.

  2. The systematic nature of some types of visual illusion: Changizi’s investigation of illusions mathematically predicts the extent that certain attributes of an object (smaller size, slower speeds, greater luminance contrast, farther distance, low eccentricity, greater proximity to the vanishing point, greater proximity to the focus of expansion) produce similar perceptual effects (increased perceived size, increased perceived speed, decreased perceived luminance contrast, deceased perceived distance). This shows that the visual system uses mechanisms for compensating for neural delay during forward motion in order to ‘perceive the present’.

  3. Problems with timing of normal sensory input and direct stimulation of the cortex: Libet’s experiment showed that stimulation of the skin reached awareness in a much shorter time then stimulation of the exposed surface of the cortex. This prompted him to propose a system of backward referrals of the timing of events. A much cleaner explanation is a predictive projection into the future occurs in normal awareness.

  4. Comprehension of language has a predictive nature: Berkum reported that event-related-potentials show the same pattern for unexpected words as for actual grammatical errors. This implies a ‘look ahead’ feature in language comprehension that produces surprise when predictions are wrong.

  5. There is an error registering system in the brain: Firth states that the brain is Bayesian, correcting its understanding on the basis of comparing prediction with current input. He states that the dopamine signal is a prediction error indicator. Dopamine neurons become more active if a good surprise happens, do not change their activity if there is no surprise and become less active if a negative surprise happens. Menon’s investigation of Go/NoGo experiments shows clear error-related activity (named ERN and ERP) in a wide network of cortical areas. Others have slightly different statements of the error registering system but all seem to agree that errors are identified and corrected. A prediction is needed in order to generate an error signal.

  6. Prediction is required for movement: Morsella has a theory that explains much about awareness. It postulates that it is used to meld together separate skeletomotor plans to avoid conflict. The conflict would only be apparent if our awareness predicted the course of the somewhat independent plans to see how they interacted with each other and the environment before the conflict actually happened. Llinas and Roy postulate that the main function of brain is a global one to implement intelligent motricity through prediction of the consequences of impending motion. They have outlined a thalamocortical system to do this that resembles the ‘neural correlates of consciousness’.

  7. Energy use by the sensory areas of the brain is higher for unexpected input: Alink’s investigation found that the response in V1 to unexpected signals was higher than for expected ones. This is thought to be the result of feedback from higher levels of predictive information.

  8. Reaching awareness takes time: Results of the bulk of experiments following the neural events leading up to awareness, it takes about 300msec from event to awareness of the event. There appears to be no evidence that we appear to live our lives a third of a second out of sync with the world. Predictive awareness eliminates this problem.

2 thoughts on “Evidence for predictive awareness

Leave a Reply

Your email address will not be published. Required fields are marked *