The Brain's Real Purpose: It's All About Movement

Why do we have brains? It's a question that seems fundamental, yet the answer might surprise you. Neuroscientist Daniel Wolpert proposes a compelling idea: our brains primarily evolved not for thinking or feeling, but for controlling movement. This perspective shifts our understanding of cognitive function, suggesting that everything from sensory perception to memory serves the ultimate goal of influencing how we move.

The Primacy of Movement

Movement is our primary means of interacting with the world. Consider these points:

• Communication: Speech, gestures, writing, and sign language all rely on muscle contractions.
• Sensory Input: Sensory memory and cognitive processes are crucial, but only insofar as they drive or suppress future movements.
• Evolutionary Advantage: There's no evolutionary benefit to remembering childhood memories or perceiving colors if these don't impact future actions.

To illustrate this point, consider the sea squirt. This creature possesses a nervous system during its juvenile, free-swimming stage. However, once it settles on a rock, it digests its own brain and nervous system for food. This drastic act underscores the brain's core function: when movement is no longer necessary, the brain becomes expendable.

The Challenge of Understanding Movement

If movement is so central, how well do we understand its control? The answer, according to Wolpert, is "extremely poorly." While machines excel at tasks like playing chess, they struggle with the dexterity of a five-year-old. This disparity highlights the complexity of movement control.

Robotics vs. Human Dexterity

Consider these comparisons:

• Chess: Computers can beat humans, a problem considered solved.
• Dexterity: A child can easily outperform the most advanced robots in tasks requiring fine motor skills.

The difficulty lies in the unknown algorithm for dexterity. It requires perceiving and acting on the world, a feat that remains elusive for robots.

Cutting-Edge Robotics

Even the most advanced robots struggle with tasks that humans find simple. For instance, a robot trained to pour water into a glass requires extensive programming and lacks the adaptability of a human. There's no generalization from one task to another.

In contrast, consider Emily Fox, a world-record holder in cup stacking. Her speed and precision are remarkable, yet we have little understanding of the neural processes behind her performance. Reverse engineering human movement control is a key focus in neuroscience.

The Noisy Reality of Movement Control

Controlling movement seems straightforward: send a command, muscles contract, and the body moves. However, the signals involved are far from perfect. Sensory feedback is noisy, meaning random interference corrupts the signal. This noise affects both sensory input and motor output.

• Sensory Noise: Trying to locate your hand under a table can be off by several centimeters due to sensory noise.
• Motor Noise: Even when aiming for the same spot repeatedly, movement variability results in a wide spread.
• Ambiguity: The external world is ambiguous and variable. A teapot may be full or empty, and its state changes over time.

Society values those who can minimize the consequences of noise. Golfers, for example, are highly rewarded for their ability to consistently hit a small ball into a hole despite these challenges.

Bayesian Decision Theory: A Framework for Understanding

To cope with noise and variability, the brain employs strategies rooted in Bayesian decision theory. This framework, popular in statistics and machine learning, provides a way to understand how the brain deals with uncertainty.

The core idea is to make inferences and take actions based on beliefs about the world. These beliefs are represented as probabilities, ranging from 0 (complete disbelief) to 1 (absolute certainty). Bayesian inference combines two sources of information:

• Data: Sensory input provides real-time information.
• Prior Knowledge: Accumulated knowledge and memories shape expectations.

Bayesian decision theory offers a mathematical approach to optimally combine prior knowledge with sensory evidence to form new beliefs. This process is crucial for learning new movement skills.

An Example: Playing Tennis

Imagine learning to play tennis. To decide where the ball will bounce, you use:

• Sensory Evidence: Visual and auditory cues provide information about the ball's trajectory.
• Prior Knowledge: Experience tells you that the ball is more likely to land in certain areas of the court, depending on your opponent's skill.

Bayesian inference combines these sources to predict the ball's landing point, allowing you to react accordingly.

The Brain as a Neural Simulator

The brain predicts sensory feedback by creating a neural simulator of the body's physics and senses. As a movement command is sent, a copy is run through this simulator to anticipate the sensory consequences. This process allows the brain to distinguish between external and internal events.

Tickling: A Clear Example

The sensation of being tickled by someone else is far more intense than tickling yourself. This difference arises because the brain subtracts predicted sensations from actual sensations. When you tickle yourself, the predicted sensation cancels out the actual sensation, reducing the ticklishness.

Experiments using robots to control tickling further demonstrate this principle. By introducing time delays between the movement and the tickle, researchers can manipulate the perceived ticklishness.

Children Fighting: An Unexpected Insight

Consider children fighting in the back seat of a car. Often, each child claims the other hit harder. This inconsistency can be explained by the brain's predictive mechanism. When a child hits another, they predict the sensory consequences and subtract them off, leading them to underestimate the force they applied. The recipient, lacking this prediction, feels the full force of the blow.

Experiments involving adults applying force to each other's fingers confirm this phenomenon. Participants consistently escalate the force they apply, suggesting they underestimate the force they are producing.

From Beliefs to Actions: Minimizing the Consequences of Noise

Bayesian decision theory suggests that actions should be optimal given one's beliefs. However, there's a gap between symbolic tasks (e.g., drinking, dancing) and the movement system, which must coordinate hundreds of muscles.

Despite the infinite ways to perform a movement, humans are remarkably stereotypical. This consistency suggests dedicated neural circuitry for decoding these patterns.

The Importance of Minimizing Noise

Good movements minimize the negative consequences of noise. For example, when intercepting a ball, different paths require different forces. Since noise increases with force, the brain chooses the path that minimizes force and, therefore, variability.

By planning movements to minimize the negative consequences of noise, the brain optimizes performance.

Conclusion

The brain evolved to control movement. Understanding how it does so is an intellectual challenge with implications for disease, rehabilitation, and robotic technology. Even seemingly simple animal tasks involve dramatic complexity within the brain. By recognizing the primacy of movement, we gain a deeper understanding of the brain's true purpose.