Categories
科技報導

How does the brain align our experiences with expectations?



The Book of Days (1864) by Scottish writer Robert Chambers tells a strange legal case: In 1457, in a small school named Lavegny In town, a sow and her piglet are accused of killing a child and eating part of it. After careful consideration, the court sentenced the sow to death, but acquitted the innocent piglets because they were too young to realize the seriousness of their crimes.

How does the brain align our experiences with expectations? 1

Our brain predicts the results of our actions and shapes reality into what we expect. This is why we believe that “seeing is believing”.

To modern people, it seems counterintuitive to subject pigs to criminal trials, because many people believe that it is human beings who have a clear understanding of behavior and results that distinguish us from other animals. A free-range pig may not know what it is chewing, but humans can certainly control their actions and be vigilant about the imminent consequences. However, although we have established our own identity and social relationships on the basis of this insight, psychology and neuroscience are constantly revealing that for the human brain, even the simplest interaction between us and the material world and society is monitored. , Will also face many difficulties. In the face of these obstacles, the brain needs to rely on predictive mechanisms to align our experience with expectations. Under normal circumstances, such consistency is very useful, but it can also deviate our experience from objective reality, thereby reducing our clear insight.

When monitoring our behavior, one of the challenges the brain faces is the inherent ambiguity of the information they receive. In fact, we “indirectly” experience the world outside the head through the sensory system: peripheral organs and nerve tissues receive and process different physical signals, such as light shining on the eyes or skin. Although these neural circuits are very complex, the brain’s sensory system has many common flaws in biological systems: imperfect wiring, leakage in transmission, and the system is plagued by noise—just like the crackling of a poorly tuned radio masks the real Transmission signal.

Noise is not the only obstacle. Even if these neural circuits can be transmitted with perfect fidelity, our perception experience is still incomplete. This is because the sense organs can only capture the “shadows” of objects in the external world. To illustrate this point, imagine how our visual system works. When observing the world around us, we sample the spatial patterns of light reflected by different objects. These lights fall on the plane of the eyes to form a two-dimensional image, which is stored in the initial part of the visual cortex and forms what we see The foundation of things. This process is very strange, but it also brings challenges, that is, how to reconstruct the real three-dimensional world from the two-dimensional shadows projected on the sensory surface.

Let us recall our own visual experience. This challenge does not seem to be difficult to solve. The world that most people see is three-dimensional. For example, when you look at your hand, a special two-dimensional sensory shadow will be cast on your eyes, and your brain will successfully construct a three-dimensional image of the hand that includes skin, muscles and bones. However, reconstructing a three-dimensional object from a two-dimensional shadow is actually what engineers call an “ill-posed problem”—it is basically impossible to solve it with only sampled data. This is because countless different objects can cast the same shadows as real hands. How does your brain pick the correct three-dimensional image from all possible options?

The brain faces a second challenge in effectively monitoring our actions, and that is speed. Our sensory system must depict a rapid and continuous flow of input information. Even for the simplest actions, it is important to quickly perceive these dynamic changes: when drinking coffee, if we cannot accurately predict when the coffee will reach the lips, the whole cup of coffee will most likely be spilled on us. However, the biological mechanisms we use to detect and transmit sensory signals are still not perfect, which makes it difficult for the brain to quickly and accurately describe what we are doing. The cost of time can’t be ignored: Although the signal takes only a fraction of a second from the eye to the brain, and using this information to guide a continuous action also takes a fraction of a second, these fractions of a second may be a dry shirt. The difference with a wet shirt.

For a long time, psychologists and neuroscientists have been thinking about what strategies our brain might use to overcome the problems of ambiguity and speed? People are increasingly realizing that both of these challenges can be solved by prediction. The key idea here is that observers do not only rely on the information currently input to their sensory system, but combine this information with the “top-down” expectations of the world.

Start with the baby’s “do and see”

This is not a very new point of view. In the 19th century, the erudite German physicist Hermann von Helmholtz proposed that the ill-posed problem of producing reliable perception from fuzzy signals can be solved through the process of “unconscious inference” , That is, the observer has a predetermined understanding of how the world is constructed, and an accurate visual image can be derived from this. For decades, this view has penetrated into cognitive psychology, especially when the British psychologist Richard Gregory put forward the concept of “perceptions as hypotheses” in the 1970s. after that. Gregory likened the process of sensory perception to scientific methodology: Just like scientists interpret evidence through current theories, our perception system can also integrate the fuzzy evidence received from the senses into the situation based on its own environmental model. in.

In more recent history, the explanation of these ideas is that the brain is assumed to be “Bayesian”, that is, that the brain’s perception is based on expectations. In 1763, two years after the death of British statistician and Presbyterian pastor Thomas Bayes, one of his theorems was finally published. Bayes’ theorem describes how to make rational inferences by combining observations and prior knowledge. For example, if you hear the ticking sound of water droplets on a hot summer day, it is likely that you left the sprinkler somewhere instead of raining. Proponents of the “Bayesian brain” hypothesis believe that when the brain evaluates “bottom-up” sensory signals based on “top-down” knowledge—about the probability that an event may occur— Probability inference will happen.

How does the brain align our experiences with expectations? 2

Perception is difficult because two different objects may cast the same “shadow” on your sensory system. Your brain can solve this problem by relying on the size and shape of known things, such as hands.

It turns out that the pattern of neural connections seen in the cerebral cortex—a large number of reverse connections from “higher” to “lower” areas—supports these views. University College London neuroscientist Karl Friston (Karl Friston) and his colleagues designed an influential brain function model-hierarchical predictive coding (hierarchical predictive coding). This theoretical model suggests that in any given brain region (such as the early visual cortex), one set of neurons encodes sensory evidence from the outside world, and another set of neurons represents the current “belief” about what the world contains. According to this theory, perception unfolds as input evidence adjusts our “belief”, and “belief” itself determines our experience. However, it is crucial that the large-scale connections between brain regions make it possible to use prior knowledge to prioritize certain “beliefs.” This allows observers to use top-down knowledge to increase the “volume” of the signals they expect, thereby giving these signals more weight as the perception unfolds.

Letting top-down predictions penetrate into perception helps us overcome speed issues. By pre-activating the parts of our brain related to sensation, the perception system can be effectively “one step ahead”. In fact, a recent study by neuroscientist Peter Kok and others found that when we expect an event to occur, the template for that event will appear in the brain’s visual activity before the actual event. This “first step” can provide a shortcut for quick and effective behavior.

Shaping perception into what we expect can also help us overcome ambiguity. As Helmholtz envisioned, if we lean towards the most likely explanation, then we can get reliable perceptions from fuzzy data. For example, when we look at our hands, because the brain has strong expectations about the types of objects it will encounter, it accepts “correct assumptions”-these are indeed hand-shaped objects, not countless One of other possibilities.

When it comes to our own actions, these expectations come from experience. In our lifetime, we gain a lot of experience by performing different actions and experiencing different results. This situation may have started in the babyhood of babbling. Those apparently random kicks, arms swings, and head rotation movements give children the opportunity to issue different action commands and observe different results. This experience of “doing and seeing” establishes an expected connection between movement and sensory representations, and between behavior and perception.

One reason to speculate that these connections are formed through learning is that there is evidence that these connections have extraordinary flexibility even in adulthood. Experimental psychologist Celia Heyes (Celia Heyes) and her team conducted related research at University College London, showing that even short-term studies can reconnect actions and perceptions. This way of establishing connections can sometimes be The natural anatomy of the human body conflicts.

The brain scan experiment illustrates this point well. When we see other people moving their hands or feet, the part of the brain that controls the relevant parts of our body will also become active. However, an interesting experiment led by Caroline Catmur, a psychologist at University College London, found that if subjects were to reverse the experience-clap their hands while watching others clap their feet, and vice versa -This mapping can be reversed. After this experience, when the subject sees the action of clapping the feet, the movement areas related to the hands become active. These findings, and others like them, provide convincing evidence that these connections are obtained by tracking probabilities. This probabilistic knowledge can shape perception and allow us to activate the template of expected action results in the sensory areas of the brain; in turn, this also helps us overcome perceptual ambiguities and quickly provide a “correct” perceptual interpretation.

Action experience will bias our expectations

In recent years, a group of neuroscientists have put forward another view, that the brain will selectively edit the expected results of our actions. Proponents of this view believe that it is more important for humans to perceive surprising and unpredictable parts of the world-such as a coffee cup accidentally slipping from our fingers. Filtering out the expected signal will mean that the sensory system contains only unexpected “errors”, so that our limited sensory loop bandwidth can only transmit the most relevant information.

The basis of this “delete expectations” hypothesis is that studies have shown that when we experience predictable results of actions, the activity in the sensory areas of the brain decreases. If we feel the skin being touched or see a hand moving, the different areas of the brain related to somatosensory or vision become more active. However, early research found that when we touch our skin—by ​​touching our hands or palms—the activity of these brain regions is relatively reduced compared to external sensations. When we observe a hand that matches our own actions, we also find similar activity inhibition in the visual area of ​​our brain.

Daniel Yon, a cognitive neuroscientist and experimental psychologist at Goldsmiths College, University of London, and his colleagues studied how these predictive mechanisms work. In a recent study, he and Clare Press of Birbeck College, University of London, and cognitive neuroscientists Sam Gilbert and Floris de Lange ( Floris de Lange) collaborated on a study. They asked volunteers to undergo an MRI scan and recorded their brain activity while performing a simple task. These participants were asked to move their fingers and observe the movement of the avatar’s hand on the screen. Whenever they perform an action, the hands on the screen either perform an expected synchronized action (moving the same finger) or an unexpected action (moving a different finger). By observing the patterns of brain activity in these two situations, researchers can figure out how to anticipate changes in the process of perception processing.

In short, the researchers’ analysis showed that when participants were consistent with their behavior, the results they saw contained more information. Careful observation revealed that these “sharp signals” in the visual area of ​​the brain were accompanied by some suppressed activities-but only in the parts sensitive to unexpected events. In other words, the predictions made during the action seem to edit out unexpected signals, produce clearer representations in the sensory areas of the brain, and are more inclined to our expectations.

These findings indicate that it is our expectations that shape neural activity and enable our brains to show the behavioral results we expect to unfold. This is consistent with the results of more and more psychological literature, indicating that our behavioral experience is biased toward our expectations.

In 2007, Japanese psychologist Kazushi Maruya (Kazushi Maruya) led a study that used a technique called “binocular rivalry”. In the binocular competition experiment, the observer is in a device that can present different images in the left and right eyes. When these images are very different, the observer’s perception experience is usually dominated by one of the images, and occasionally there will be fluctuations between competing options. Maruko and his colleagues created a visually competitive experience, allowing the observer to see a moving sphere in his right eye when he sees a flashing black and white pattern in his left eye. The researchers found that what is interesting is that when they associate the moving sphere with the observer’s ongoing activity, the image is more likely to “win out” the competition. The conscious experience here is dominated by predictable action results.

Experiments such as visual competition and many other experiments have shown that our perceptual experience is affected by behavior. For example, when a pianist presses a key, the pitch of the series of notes that they hear will rise or fall. More importantly, how our experience of the passage of time can also be manipulated by behavior-when our movement speed slows down, other events seem to last longer-and we tend to see the same direction as our own behavior Blur motion. Since our expectations usually become reality, shaping our perceptions according to our beliefs allows us to understand more clearly how we affect our surroundings.

Mental illness and expected ability

The most fascinating possibility is that this predictive mechanism may play an important role in helping us interact with society. After all, humans seem to abide by a set of rules when it comes to greeting each other, talking in turns, and responding to each other. One of the most common and structured ways of responding to each other is through imitation. Cognitive scientists define “imitation” as a situation where the observer will imitate the other’s body movements-the gestures used, the gait of walking, etc.; if you see the person you are talking to rubbing or shaking their feet , You will most likely do the same.

A series of studies have shown that the imitated experience has an important impact on the quality of our interactions. When we allow our actions to be imitated by others, we can increase our trust in these people, making the relationship harmonious and a sense of belonging. In 2003, psychologist Rick van Baaren of the University of Nijmegen in the Netherlands led a study that found that waiters who were instructed to imitate diners received more tips.

However, although the imitated experience can serve as a powerful social lubricant, we can only perceive the behavior of others towards us through an imperfect sensory system. If the brain cannot notice the occurrence of these reactions, the pro-social potential of the imitated behavior will not be fully realized. But if we can predict the social consequences of our actions, just like we predict the physical consequences of our actions, and use this anticipation mechanism to make others more easily perceivable by us, this problem can be solved.

Claire Press and Daniel Yong explored this view in a 2018 study. They analyzed whether the expected mechanism in the behavior process affects our perception of imitation by others. They did find a sign of anticipation: the perception of the expected action result is visually stronger. This sign will continue for a few seconds after our own actions, indicating that our anticipation mechanism is very suitable for predicting the imitation response of others.

This enhancement mechanism may be particularly important in demanding sensory environments: for example, we are more likely to find friends waving at us in a crowded room. Another relatively underexplored possibility is that this expectation mechanism may cause us to feel that some people are more similar to ourselves. If slow movement makes people feel that other parts of the world also look slower, then when we feel sad, the slow, sluggish movements we make may also make us think that others have become slower and more pessimistic, which causes A kind of deviation.

Of course, sometimes using expectations to construct our experience is a double-edged sword. Obviously, when you realize that your expectations are sometimes not fulfilled, you may encounter two-sided results. If you forcefully lift an empty teapot that you think is full of water, the empty container will accelerate much faster than you expected. Likewise, if you tell an out-of-date joke, you may see a blank face instead of the laughter you expected.

In this case, we don’t seem to be very comfortable with reshaping our perceptual experience—for example, misperceiving the movement of an empty teapot to make it slower than it actually is; or editing the expressions of social partners to make them look happier— To make these events look more in line with our expectations. The reason is very simple, because these are our expectations deduced in advance. This perceptual error is more likely to occur when the sensory world provides insufficient information and is expected to be given more weight. It takes a process to produce a reliable experience, and accidental misunderstandings may be the price we pay for this process.

Nevertheless, it is easy for us to see what kind of harm such perceptual errors can cause. Many of human society, culture, and legal systems rely on the notion that under normal circumstances, we all know what we are doing, so we can be responsible for what we do. In most cases, humans will misperceive their own behavior. This statement is a bit ironic in science, but in some cases, perceptual errors caused by expectations may lead to serious consequences. For example, an experienced doctor may have strong expectations about how the patient will respond to a simple operation (such as a lumbar puncture), which may cause them to develop cognitive biases when unusual reactions occur. If doctors make deviations and misperceive patients’ symptoms, will their moral or legal status change? If we think that any perceptual errors may be caused by the professional knowledge of doctors, will this problem become more complicated?

Although anticipation has its dark side, think about it and see how difficult the world will become without anticipation. This idea has long troubled psychiatrists, who believe that some unusual experiences in mental illness may reflect the disturbance of expected abilities.

A particularly strange case is the “control delusion” that appears in some schizophrenic patients. Patients with this delusion report a painful experience, and they feel that their behavior is driven by some external force. A patient described his abnormal experience to the British psychiatrist CS Mellor: “My hands and arms are moving, and my fingers pick up pens, but I can’t control them. What they do has nothing to do with me. .”

Echoing these vivid case reports, many experimental studies have revealed that people with schizophrenia sometimes find it difficult to recognize their behavior. For example, in an experiment led by Nicolas Franck, a psychiatrist at the University Hospital of Lyon in France, in 2001, researchers showed schizophrenia patients and healthy control volunteers video feedback about their behavior; These videos have been processed in various ways, such as spatially distorting the footage or adding time delay. Researchers found that people with schizophrenia performed poorly when they discovered these mismatches, which suggests that they have relatively poor perception of their behavior.

This lack of behavior monitoring and the attendant delusions may arise because these patients have experienced the breakdown of a mechanism that allows them to predict the consequences of their actions. Patients may therefore lead to changes in the way they experience behavioral outcomes, which in turn makes them more inclined to form strange beliefs. In particular, when losing the “” benefit of top-down prediction, it may make people’s experience of their own actions relatively vague, and it also makes it difficult for people to determine what their actions lead to. This ambiguity can be very painful in itself, and being tortured by this unusual experience for a long time may cause a person to have “delusion emotions”-feeling that strange things are happening and need something strange, almost delusional. Explanation.

Finally, there seems to be a new perspective from psychology and neuroscience that hope plays a key role in shaping how we experience behavior and its results. Although in a world of perception that is inherently vague, combining expectations and perceptions may be a powerful way to monitor our behavior, but when expectations are not fulfilled, this process may lead us to misdescribe our behavior. as a result of. This fictitious experience may weaken our clear perception of our own behavior, and it is this perception that distinguishes us from naive animals. When it comes to behavior, we may only see what we are willing to believe, and sometimes we may not know what we are doing.