Daniel Wolpert: The real reason for brains

Why We Have Brains

In this section, the speaker explains why humans and other animals have brains. The main reason for having a brain is to produce adaptable and complex movements.

The Purpose of Having a Brain

• Humans and other animals have brains to produce adaptable and complex movements.

• Movement is the only way we can affect the world around us.

• Sensory, memory, and cognitive processes are important only to drive or suppress future movements.

Evidence Supporting the Purpose of Having a Brain

• The sea squirt is an animal that has a nervous system but digests its own brain once it no longer needs to move.

• Trees and grass on our planet do not have brains.

Understanding How the Brain Controls Movement

In this section, the speaker discusses how well we understand how the brain controls movement. He compares human performance in tasks such as chess with that of robots.

Comparing Human Performance with Robots

• While computers can beat humans at chess, they cannot match human dexterity in manipulating objects.

• It is easier to determine an algorithm for playing chess than it is for being dexterous.

• Robotics technology for manipulation is still in its early stages compared to human performance.

Reverse Engineering Human Movement Control

• The speaker's group aims to reverse engineer how humans control movement by studying sensory feedback from vision, skin, muscles, etc.

• Controlling movement is difficult because sensory feedback can be extremely noisy.

Noise and Variability in Sensory Feedback

In this section, the speaker discusses how noise and variability in sensory feedback can affect our ability to localize objects and perform movements accurately.

Noise in Sensory Feedback

• When trying to localize an object with our hands, we can be off by several centimeters due to noise in sensory feedback.

• Movement output is also noisy, making it difficult to hit a target accurately.

• The outside world is both ambiguous and variable, adding to the overall noise in sensory movement tasks.

Bayesian Decision Theory

In this section, the speaker introduces Bayesian decision theory as a framework for understanding how the brain deals with uncertainty.

Beliefs and Probabilities

• Beliefs about the world are represented as probabilities between 0 and 1.

• These probabilities represent different levels of certainty or uncertainty about a belief.

Sources of Information

• There are two sources of information used for making inferences: data (sensory input) and prior knowledge (accumulated through experience).

• Bayes' rule provides a way to optimally combine these sources of information to generate new beliefs.

Example: Learning Tennis

• When learning tennis, one must make predictions about where the ball will bounce based on both sensory evidence (visual/auditory input) and prior knowledge (distribution of bounces during a match).

• Bayes' rule allows us to combine these sources of information optimally to generate beliefs about where the ball will land.

Predictive Processing

In this section, the speaker discusses how the brain makes predictions about future sensory feedback using predictive processing.

Predicting Sensory Feedback

• The brain uses predictive processing to make predictions about future sensory feedback.

• These predictions are based on a neural simulator of the physics of our body and senses.

Changing Perceptions

• Predictive processing can profoundly change our perceptions based on what we do.

• The transformation from movement command to sensory feedback is governed by the physics of our body and senses, but inside the brain, there is a neural predictor that uses predictive processing to make predictions about future sensory feedback.

The Brain's Predictive Capabilities

In this section, the speaker explains how the brain makes predictions and subtracts them from sensations.

Predictions and Sensory Information

• The brain combines external sensory information with internal sensory information to create one source of information.

• Distinguishing between external events and internal events is important because external events are more behaviorally relevant.

• To distinguish between external and internal events, the brain compares its prediction based on movement commands with reality. Any discrepancy should be external.

Tickling Study

• A sensation generated by oneself feels different than if generated by another person. This is evident in tickling studies where people can't tickle themselves as well as others can.

• The brain cancels out self-generated sensations by making precise predictions and subtracting them from sensations.

Testing Self-generated Sensations

• Children tend to get into fights that escalate in terms of force because they generate the movement command when hitting someone else, which leads them to think they hit less hard than they actually did.

• A study was conducted using two adults playing a game where they take turns applying force back and forth using a force transfuser while being briefed about the rules of the game separately so that they don't know what rules the other person is playing by.

The Brain's Control of Movement

In this section, the speaker discusses how the brain controls movement and the challenges it faces in doing so.

Bayes' Rule and Optimal Actions

• The brain generates actions based on beliefs, which should be optimal according to Bayes' rule. However, there is a gap between symbolic tasks and the movement system that must contract 600 muscles.

Stereotypical Movements

• Humans are extremely stereotypical in their movements due to dedicated neural circuitry in the brain that decodes this stereotyping. Biological motion can convey a huge amount of information through movement.

Planning Movements to Minimize Negative Consequences of Noise

• Movements get better over time through learning and evolution. Good movements are those that minimize negative consequences of noise, which increases with force. Therefore, planning movements to avoid big forces is important. This principle can explain a lot of data about how people plan their movements in daily life.

Importance for Disease, Rehabilitation, and Robotics

• Understanding how the brain controls movement is relevant for disease and rehabilitation since many diseases affect movement. Additionally, understanding this process can help improve robotic technology as well.

The Role of Movement in Brain Function

In this section, the speaker discusses how movement plays a crucial role in driving other aspects of brain function.

Movement as a Driving Force for Other Brain Functions

• Movement is not just an isolated function but rather plays an important role in driving other aspects of brain function such as memory and sensation. Studying these functions without considering their relationship to movement may lead to incomplete understanding or mistakes.