Book Report: Incognito

David Eagleman’s Incognito might be almost 10 years old, but the perspectives raised are still very relevant. Below is a summary of the key ideas in the book, with some reflection on top. Here are 8 key topics that stood out.

1. Team of Rivals: A Competition in your Brain

The author puts forth the model of the team of rivals, in which the brain is not a single entity, but a series of “sub-agents”, all competing to accomplish a subset of the tasks needed to sustain life. These functionalities are overlapping, and in competition with each other. There is a determination of fitness, and selection. As this cycles, the brain as a whole can continue improving.

In software engineering, people often take one good shot at a given problem, and once there is a satisfactory solution, they move on to solve something else. Nature is such a fantastic engineer because it experiments and reinvents itself at such a massive scale. There's always the possibility that a given solution is "optimal”, but of course, for perception or intelligence-based tasks, that is never really the case. We should always be experimenting, always be reinventing ourselves, as well as our creations to accomplish any type of goal.

2. The Creation of Narratives

The human brain is a pattern recognition machine. There is data, there is a prediction, and a prediction error. When a prediction error occurs, the model is updated. When the model is correct, then the brain continues to reinforce that pathway, and frees up the attention for other, more interesting tasks. For humans, these models often take the form of a narrative, which intuitively describes the data that has been observed so far.

Eagleman describes a card game in which participants of the study are given rewards depending on the cards that they choose to place on the table. Rewards are calculated in a way that is too complex for any participant to accurately determine on their own, but at the end of the game, every participant confidently reported an explanation of how the rewards were being computed. It was done with insufficient evidence, and in all cases were wrong, but they were convinced nonetheless.

We are always creating narratives, on top of observed data and decisions, as a way to make sense of the world. This extends from predicting the trajectory of a ball, to predicting the rules of a simple game, to pondering our very existence. I imagine this played a large role in the creation of religion as well.

3. Our Consciousness is the Tip of the Iceberg

Consciousness is simply the tip of the iceberg of our brains, and there is an enormous amount of computation happening behind the scenes. There is enough going on to overwhelm our conscious thought if it was surfaced. Further, there are some operations that are so important that we’re better off guarding them from our conscious thought (e.g. keeping our heart beating). Our consciousness is being presented newspaper headlines, and being spared the details. As an analogy, the president might be told that soybean production has increased over the last year, but they don’t need to know the actions of every worker in every field.

Also, our conscious decision making is heavily influenced by unconscious thought, in ways that are often not even apparent to our conscious mind. For example, men are more attracted to women at certain moments in their menstrual cycle, and it is unclear exactly why. The changes are subtle - perhaps increased ear and eye symmetry contributes, but the exact mechanism is unknown.

4. Our conscious experience is a poor reflection of reality

We are tempted to believe that our senses provide an accurate reproduction of the world around us. Our eyes are video cameras, and our ears provide a faithful representation of the sound waves going through the air. However, our reality is shaped by our sensory limitations, and by our interpretation on top of that data. Our eyes don't detect a significant portion of the electromagnetic spectrum, and our ears are also limited to a certain range of sounds. There are many senses that animals possess that we cannot even imagine, such as sensing electric fields, or perceiving ultraviolet light, or smelling with fidelity thousands of times greater than our own. We don't miss having these senses because we can't imagine what they would be like, but if you flip the argument and asked a bloodhound what the experience of a human might be like, they say that we have a very deprived existence.

Then, our vision is not simply a faithful capturing of all the photons hitting our eyes. We actually only have high-resolution in the centers of our eyes, and it deteriorate significantly towards the periphery. We have blind areas in the middle of our sight, but since we move our eyes so quickly our brain can create an image that appears complete. Then, there is our interpretation of the data. What we find attractive is hard to imagine changing, but just think that there are frogs that find other frogs extremely sexually attractive and humans that find other humans extremely sexually attractive, but these perceptions are arbitrary. Evolutionary forces and desires color are very existence.

5. Burning into Circuitry

Human brains can take a computationally intensive task, which requires quite a lot of thinking, and practice it until it becomes automatic. One example is learning to play tennis - at the start you are very slow and clumsy, until you progress in your proficiency and movement becomes efficient and refined. One neuroscientist friend described the hippocampus as the R&D part of the brain that figures out how to do something, and then later stamps out a more efficient version of the program elsewhere in the brain. We even do this when writing software, where the first iteration of a program is inefficient, then is improved with a refactor, and finally it could even be implemented in hardware to operate at maximum efficiency.

This is one of the key use cases for our conscious thought - to identify new challenging programs that we want to learn, and to ensure that we practice them until they become efficient and automatic. A good lesson here is to be as deliberate as possible with the programs that we are burning down into our circuitry. As my piano teacher once said, practice carefully, because it's hard to unlearn bad habits.

6. Free Will

It seems no neuroscience book can get around this particularly thorny problem.

Logically, there are two possible explanations of the decisions that we determine to be free will: they are either deterministic, or not deterministic. We resist the idea that they are deterministic, but if they’re not deterministic that implies they are random, and that’s not very satisfying either. People sometimes draw inspiration from chaos theory - that it’s simply too complex to possibly model. While this indicates that we won’t build a decision-predicting device anytime soon, it actually endorses the deterministic model as well.

There are fascinating experiments in which the brain activity associated with a decision occurs seconds before the person reports having made such a decision. This implies the brain is has pre-baked a decision, and serves it up for the conscious mind to take credit for when complete. This is like a government department doing a deep analysis and deciding on a new initiative, presenting the results and recommendations to the president, and having the president think that it was their decision from the beginning.

We all have a strong sensation of having free will, but actually digging into the circuitry and trying to pinpoint the exact location of the free will has been unsuccessful. Further, every part of the brain is highly interconnected, and dependent on, other areas of the brain, so it’s unclear if there could be a “prime mover” of the brain. Science will continue to try and reconcile the sensation of having free will with the facts that such a sensation doesn’t quite fit the physical realities of the system. For now I will assume it is only a sensation until evidence arises to describe otherwise.

The author has an interesting point when it comes to the blame-worthiness of individuals in a legal context. There are clear examples in which individuals are powerless to prevent themselves from committing crimes. He provides an example of someone with a brain tumor who ends up shooting a bunch of people, and another where an individual drives 14 miles to murder his mother-in-law in her sleep while (legitimately) sleep-walking. In the latter case the individual was found to be not guilty.

So, if you consider an individual who experienced feel alcohol syndrome, was exposed to lead paint and other neurotoxins, were abused for most of their lives, and had a strong genetic predisposition for violence and poor impulse control, then where do we draw the line between an individual's choice to commit the crime and it  being a deterministic outcome of their environment. Indeed, there are many people whom we are try today who would almost certainly be determined to be not guilty, if only we had the ability to characterize their genetic makeup, their brain physiology, and their particular life circumstances, with sufficient precision.

8. Strengthening the Brain

Once someone is sentenced, the question becomes, what should be done to fairly punish the individual? Eagleman raises the valid point that punishment is only humane if it is a viable path for people to correct their behavior in the future. Then, you might ask, what is the most effective path for people to correct their behavior in the future? It is actually possible for an individual to strengthen their critical decision making, and their pre-frontal cortex more generally, such that they can resist the temptation of making bad choices in the future.

He looks at Mel Gibson as an interesting example of someone who contains many competing selves. He infamously had a an anti-semitic tirade during a traffic stop while drunk, and when sober he was very apologetic and entirely condemned that particular line of thinking. There were many sides to this debate, some people saying that alcohol is a truth serum that reveals somebody’s true colors. Others were inclined to let him off the hook because people do all kinds of stupid things while drunk, and the real “you” is the one present while sober. Both of these miss the point - in reality we all contain multitudes, and the question is whether the short term or the long term versions of ourselves might “win” in a given circumstance. The problem is not necessarily that people think bad thoughts, if that is their reptilian, irrational brain thinking, because it is not possible (or desirable) to control people’s thoughts. However, what does represent a crime, and is truly reprehensible, is failing to constrain these harmful thoughts, and temper them with the reason and morality that are critical for someone that is a functioning member of society.

The key to improving recidivism is to find ways for impulse control to be trained. This gels well with my philosophy that everything is a muscle, and we can be trained with the right workout. He describes a tool which allows people to observe, and therefore improve, their self control via neuro-feedback. In this scheme, the patient looks at a screen, and a bar raises and lowers with more or less exertion of self-control. Over time, the skill improves, and (in theory) in the future when the urge to shoplift arises, for example, there is little risk of it being acted upon.

Closing Thoughts

Finding ways of measuring and systematically improving the brain will always be a keen area of interest for me, and I will follow up on the neuro-feedback techniques described. It's also fascinating to consider the "team of rivals” experimentation framework as an approach for more effectively achieving goals when writing software, running an organization, or even living your own life. We should always be reinventing ourselves, and looking for more effective approaches to reach our goals.

Show Comments