The Experience Machine

Unboxing the Experience Machine

• The number of neuronal connections carrying signals backward in this way is estimated to exceed the number carry signals forward by a very substantial margin.
• The brain, at 2% of human body weight is estimated to account for around 20% of total bodily energy consumption.
• The bulk of what the brain does is learn and maintain a kind of model of body and world - a model that can then be used, moment by moment, to try to predict the sensory signal. These predictions help structure everything we see, hear, touch, and feel.
• As a brain encounters new sensory information its job is to determine if there is anything in that incoming signal that looks like important "news" - unpredicted sensory information that matters to whatever it is that we are trying to see or do.
• Hermann von Helmholtz - "given everything I know, how must the world be for me to be receiving the pattern of signals currently present?" This is the question that perceptual systems are built to resolve.
• If the signal is poor, the brain chruns out "good hallucinations" by filling in and fleshing out the missing signal according to what it expects to hear.
• The "top-down" flow of information
• "Controlled hallucination" - When inner guessing completely rules the roost, we are just hallucinating, full stop. But when it is appropriately sensitive to sensory stimulations - via prediction error signals - the guessing is controlled, and the world becomes known to the mind.
• Linear predictive coding (Cluade Shannon) - The trick is trading intelligence and foreknowledge against the costs of encoding and transmitting all the information. Transmit only whatever turns out to be different from the predicted patterns. Wherever there is detectable regularity of any kind, prediction (and hence this form of data compression) becomes possible.
• However complex or high-level the predictions, it is prediction errors that must then carry the news, signaling differences from the expected and thereby keeping us in touch with a changing and sometimes surprising world.
• In the predictive processing architecture of the brain, it is thought that different neuronal populations specialize in different things, so that each "higher" level can use its own specialized knowledge and resources to try to predict the states of the level immediately below it. A level that specializes in predicting whole words might use its knowledge to help predict states at a lower level whose specialty is recognizing letters. The word level might be predicted by a higher level that specializes in sentences.
• All that flows forward is news - deviations from what is expected. This is efficient.
• Our own actions and histories sculpt the onboard prediction machinery that in turn sculpts human awareness, right down to the level of what seem to us to be basic sensory experiences.
• Millions of years of evolution have determined the bedrock configuration of the machinery we command at birth: the early wiring of the brain, the structure of our sense organs, and the shape of our bodies. Courtesy of all that, we start our journey already armed with plenty of hard-won knowledge.
• Creatures with lungs are already structurally "expecting to breathe".
• Creatures like us specialize in learning about their worlds on the basis of repeated sensory encounters. And we drive learning by trying to predict our own sensory flows.
• The brain learns by looking for better and better ways to predict that unruly sensory barrage. Very young infants seem to spend most of their time doing just this, trying to find useful patterns in the sensory stream.
• One way to predict quite a lot about the most likely next word in a sentence is to implicitly know a lot about grammar. And a good way to learn grammar is to try, again and again, to predict the next words you are going to hear. As those attempts continue, your brain can slowly, unconsciously, discover the regularities that will enable you to do a better job.

Psychiatry and Neurology - Chronic Pain, Autism, Schizophrenia, and PTSD

• Types of pain:
  • Nociceptive pain - indicates actual or threatened bodily damage
  • Neuropathic pain - is caused by damage or disease affecting the sensory systems that deliver the experience of pain, or the processing of pain information.
  • Nociplastic pain - arises form abnormal processing of pain signals without any clear evidence of either tissue damage or any other recognized systemic pathology.
• Human experience arises at the meeting point of predictions and sensory evidence and this meeting is flexibly determined by the brain's best estimate of their relative reliability and significance (precision).
• Only predictions that our brains estimate to be reliable get to exert a powerful influence on our sensations. If you don't trust an authority, then all bets are off.
• Placebo-induced chagnes have been shown to reach far down, altering responses even at the level of the spinal cord. Our active expectations of pain or relief are somehow impacting the whole web from which experience itself is constructed.
• False expectations, once they get a grip on us, become increasingly resistant to change.
• Functional neurological disorders are entirely genuine but appear not to be caused by any kind of anatomical or structural change or conventional disease process.
• Precision variations control which bits of what we know and what we sense will be most influential, moment by moment, in bringing about further processing and actions. Precision variation is what attention really is.
• Attention is the brain adjusting its precision-weightings as we go about our daily tasks, using knowledge and sensing to their best effect. By attending correctly, I become better able to spot and respond to whatever matters most for the task I am trying to perform. Precision estimation is thus the heart and soul of flexible, fluid intelligence.
• Strongly anticipating pain, numbness, weakness, or other symptoms alters patterns of attention (precision-weightings) in ways that can either amplify or entirely generate the experience - which then seems to confirm those very expectations.
• There is no such thing as raw or correct experience of a medical symptom.
• After several bouts of back pain, people start to process the world differently... their pain becomes embedded among the things they associate with themselves.
• We need to move away from thinking of pain as a simple sensation, a direct signal of damage or potential damage, to a view of pain as a perception.
• There are two ways for predictive processing to go wrong:
  • Underweight predictions and expectations (underuse the brain's predictive powers) - making it hard to detect faint but predictable patterns in a noisy or ambiguous environment.
  • Overweight expectations (ignoring environmental conditions) - in extreme cases, this leads to hallucinations.
• Evidence now favors the idea that a tendency to overweight the sensory evidence (enhanced sensory precision) is the core difference separating autism spectrum condition from the more neurotypical profile. Autism spectrum individuals constantly encounter an excess of highly detailed and apparently very salient sensory information, coming from both inside their own body and the outside world. This sensory excess impedes the moment-by-moment identification of the broader context or scenarios. The emphasis on every aspect of sensory detail effectively makes it impossible to spot the larger forest for the trees.
• Faced with an endless stream of rich and apparently attention-demanding sensory information, an individual with autism spectrum condition might start to self-select more predictable environments, becoming increasingly wary of complex social encounters. Repetitive and stereotyped behaviors such as rocking or hand-flapping might also emerge, as these would offer a clever way to ensure (by self-generating) a predictable streams of sensory input. Yet another way to reduce sensory surprises is to develop extreme expertise in a restricted domain.
• Schizophrenia - Falsely generated, highly weighted prediction error forces the brain to seek a new predictive model. The resulting hypotheses (shadowy organizations, telepathy, alien control, and strange beliefs involving the internet), may appear bizarre to the external observer. Yet from within they constitute the best explanation available.
• During the onset of psychosis, the world begins to look somehow different or strange, reflecting the presence of persistent, unresolved prediction errors. Those errors then slowly drive the system to form increasingly radical hypotheses in an effort to accommodate them.
• Predictive brains control action as well as perception, so the delusional person will actively seek out confirming evidence for their radical hypotheses. New information may itself be interpreted differently so as to appear to confirm or consolidate the radical beliefs. The cycle of error thus becomes viciously self-protecting. Such pernicious outcomes seem to be the Achilles heel of the predictive brain.
• Up to 30% of those exposed to a highly traumatic event will go on to develop some form of PTSD, which is well correlated with unusually large increases in precision-weighting on the prediction error signal in response to unexpectedly negative outcomes. In the most severely affected individuals, the response to failing to predict the showk was to radically overweight the missed cue and thereby become hypersensitive to its occurrence in the future.

Action as Self-Fulfilling Prediction

• The brain predicts how things would look and feel if the action were being successfully performed, and by reducing errors relative to that prediction, the action or movement is brought about. Predicting just how it would look and feel to hit that perfect drive, or make a killer serve, brings that longed-for result about.
• By making prediction the common root of both perception and action, predictive processing (action inference) reveals a hidden unity in the workings of the mind. Action and perception form a single whole, jointly orchestrated by the drive to eliminate errors in prediction.
• Actions come about because we mentally represent the completed effects of the action. Representing some desired end result, such as grasping a cup, automatically recruits (in the skilled agent) the set of motor commands needed to make that very thing happen.
• The brain learns, through training and experience, to predict what we would see and feel if - but only if - our bodies were moving in just the right ways so as to achieve our goals.
• By predicting the specific sensory effects of that motion, and then rapidly quashing the errors that resulted by actually moving my head and eyes in the right ways, I brought into view the scene outside my window.
• By launching a cascade of sensory predictions, and then rendering them true by means of action (thus eliminating the resulting prediction errors), the brain creates the desired movements. I strongly predict looking out the window and that prediction acts as a kind of self-fulfilling prophecy.
• Motor control works in much the same way as perception. In each case, the brain is seeking to achieve a fit between what is predicted and what the sensory evidence suggests. But in the case of action, the fit is achieved by altering the sensory evidence to bring it into line with the prediction.
• To move my arm, I must give high weighting to the desired future state (arm moving) rather than the undesired present state (arm not moving). That means actively disattending to the present state of the arm, hence dampening down that sensory information.
• For many processes, waiting for feedback cues (such as wilting plants) is a bad idea, as what is really needed is ongoing preemptive action. Systems that instead predict the future from their current state, current actions, and a model of how those actions will affect that state, are always one step ahead of the game.
• Active sensing - Sensing itself is an intelligent action, aimed at delivering just the right information, just in time for use. As embodied agents we are able to act on our worlds in ways that actively generate new patterns of sensory stimulation.
• Expert seeing
• Goal-directed behavior involves using predicted outcomes to help structure the actions that will best serve to make those outcomes real.
• If I plan to become a better surfer, my brain needs to make the realistic-yet-optimistic prediction that I will indeed later be such a surfer. With that goal (long-term prediction) active, I can use what I know about how things work in the world to identify important steps along the way, generating a policy that might - according to my current skill set and personal circumstances - include moving to the coast, taking classes, or vacationing in Tarifa.
• Realistic optimisim - We must at some level strongly predict that we will occupy the states that we can plausibly attain and that best realize our goals. We wil then act in ways designed to eliminate errors calculated relative to the optimistic-yet-realistic prediction that those goals are achieved.