by Bor, Daniel
This analogy illustrates the flexibility in a network of neurons. A neuron will only fire if it receives a certain number of firing outputs from other neurons, but this number can be increased or decreased in the short term, depending on what signaling chemicals are used in the gaps between neurons (the tiredness of night versus the caffeine fueled morning). These chemicals normally flood large collections of neurons indiscriminately. And if two neurons co-fire frequently, this in itself will physically strengthen the connections between the neurons and make it more likely that they will fire together in the future (the regular e-mails between the narrator and N. Uron). This leads to the famous neuroscience dictum that “neurons that fire together, wire together” (this is known as Hebb’s law, after the pioneer of the computational study of networks of neurons). This easing of information transmission between similarly behaving connected neurons is thought to be the main microscopic mechanism for learning and memory.
If we learn, say, that the word “square” corresponds to the visual input of a square, the neurons that visually represent an object with four equal sides at right angles to each other are becoming e-mail friends with those that hear the sound “square”—they prioritize their e-mails between each other, or, in other words, the visual square population of neurons collectively strengthen their connections with the sound “square” neurons. If there is any location to this learning, it is in the enhanced sensitivity between these groups of neurons, enabling them to activate each other more easily.
When I was a first-year undergraduate, my intuition was that there may be one or a few neurons that recognized my grandmother’s face, another local set that recognized a hammer, and so on. I naively saw the brain rather like a gigantic group of filing cabinets, with familiar buildings, say, neatly placed in one filing cabinet and the plots of novels I’d read alphabetized in another.
When I was told that in fact information is distributed throughout a network of neurons, that it is encoded partly in the strengths of the connections between neurons in huge networks spanning millions of neurons, that it isn’t really localized at all, but is a pattern of activity, I found this account shocking. It is in some ways the single most difficult neuroscientific concept to accept, but because of the way evolution works, it can’t be any other way. This system of neurons learning by building links with other neurons in a distributed fashion is a framework that can start small, but scale up exponentially, while some filing-cabinet version of the brain could only be made with a prospective plan by some god, not by evolution. It may be that a nematode worm, with its minuscule network of 302 neurons, can only learn a handful of things in its few weeks of existence, while I, with my 85 billion neurons, can learn many thousands of facts in my set of decades—but we nevertheless closely share the same underlying neuronal mechanisms for information processing (these worms even use many of the same neurotransmitter chemicals that we do).
This apparently simple neural system allows for incredible flexibility, especially in its larger forms. Just as DNA is written in a language understood by all life on earth, this ubiquitous neuronal binary language is potentially understood by the whole brain, which, for instance, allows for the easy exchange of information between regions. This helps explain why ferret brains can be rewired so that the auditory cortex can start to “see” if given input from the eyes, and why the visual cortex in blind people can easily adapt to process Braille.
ATTENTION AS A BRUTAL NEURONAL WAR
The semi-chaotic activity of our 85 billion neurons undergoes a kind of temporary natural selection every moment of our waking lives, as attention shapes the contents of consciousness. Rival coalitions of neurons compete with one another to be heard the loudest. Those with the most powerful voice recruit others to their cause, and suppress any dissenters, until the strongest thought is carried by millions of neurons, all with one voice—for instance, to look for the black hair of your lover as she approaches down the street. Every time you have a new thought, that idea has become the dominant clan of your internal world, following a violent battle between jostling, screaming tribes.
To illustrate just how attention shapes consciousness at the neural level, let me return to the e-mail analogy: Now I mentioned before that I’m a manager in a large company. In point of fact, I happen to work within the security department, and it’s my job to see that certain unwanted types keep well clear of the office building. Each security executive specializes in a single color (there is a visual color-processing region in the human brain, known as V4, with individual neurons firing most strongly for specific colors). My color happens to be a particular shade of yellow. If the junior boys at the back of the building (primary visual cortex, the first cortical station from the eyes) send me an e-mail that they’ve spotted my own lovely shade of yellow on the street, then a few e-mails from them will get me jumping up and down in my seat. I want to tell everyone I know that my color is around, so I fire off my e-mails quick as a flash, with lots of exclamation marks at the end of each sentence. The guy in the cubicle next to me is responsible for the color black. Now if we both start sending e-mails madly that our colors are around, then the security guy a level above me, down the corridor, just needs one or two e-mails from us both together for all hell to break loose. You see, if he learns that both black and yellow are on the jackets of those outside from even a few e-mails, he sends high-priority e-mails to everyone he knows—and he knows some pretty senior people. (Further into the visual stream, for instance in the inferotemporal cortex in the temporal lobes, there are neurons that are responsive to combinations of features, or even specific objects, by the connections they’ve made with other neurons. In the neural equivalent of the current example, a wasp has been detected, and the person is allergic to them.)
Although people rarely listen to me, when the security manager responsible for yellow and black combinations sends e-mails about what he’s detected, they always listen to him. Everyone in the entire building passes on his information without question as soon as they get it. Thanks to a flurry of focused e-mails on the topic, all the security cameras swing to the sight of the spy, everyone discusses what to do with this enemy and what it might mean, and any guards outside get ready to move the interloper along, if need be. Almost immediately, absolutely the whole building knows about the spy, with his yellow and black clothes. The guys in my office who are responsible for any colors that aren’t yellow or black basically go for a break, as they know no one is going to listen to their e-mails for a while. I, for a change, become important, along with the guy sitting next to me who is responsible for black. Our e-mails are given priority—they are read first and acted on more often than not—as we report on the latest whereabouts of this probable spy to the security guy in the next corridor.
Although in the real brain, the guys responsible for yellow or black would be represented by the amalgam of thousands or even millions of similar neurons, this analogy illustrates many aspects of how attention can be generated in the brain. One important feature is that as information flows through the cortex, it increasingly gets filtered and combined to reveal its hidden, richer meaning.
Initially, our senses are constantly performing the first stage of filtering. As soon as visual information enters our cortex, any data concerning edges are already preferentially filtered for us. Then each simple detail activates a set of neurons tuned to only one or two basic features, but these neurons pass on that relatively raw information to later, more sophisticated neurons that are designed to represent ideas not just about single features, but combinations, or—even later in the stream—actual objects.
All of these separate neuronal populations, each activated by the various sensory inputs received at that moment, are trying to broadcast their own information, and each group is competing to shout with the loudest voice. But if certain combinations of basic features are spotted in the world that correspond, say, to a well-learned danger, then these set off a chain of actions in the brain. Those neurons t
hat have detected the threat are given priority over any others.
This winning signal now recruits all related areas. If you’ve experienced a fear of wasps, and one is buzzing around, anything remotely wasp-like suddenly is mistaken for a wasp—you initially think that horsefly over there is one too, and perhaps even, for a moment, mistake the fridge sound for a wasp’s noise. This is because all the neurons relevant to the details of a wasp are so active that these primed neurons are latching onto any tiny hint of it—and many false positive reports can occur. But if a true wasp is spotted, then all relevant neuronal regions will activate faster than usual and collaborate on recognizing the threat and then avoiding it. In the meantime, any neurons not currently coding for this danger will have their activity suppressed so as not to get in the way.
In other words, there is a constant competition occurring in the brain between different factions of neurons representing different chunks of information. Those groups that have the highest current biological relevance are given a leg up in activity, a head start, so that their influence rapidly spreads to more and more brain regions, until many areas of the brain will be dealing with this piece of information in their own way—and this amplification for this object or feature will be bouncing back and forth in the brain, constantly reinforced. At that point, the competition has effectively been won by this source of information, and any neurons representing some competing information will not only fail to broadcast their data widely, but have their activity inhibited by the neurons that are dealing with the relevant information. The winners in these battles are indeed oppressive victors, squashing any potential dissent, but in this way attention remains focused and we can respond efficiently to any danger, without being distracted.
This ideas-based, winner-takes-all system sounds simple, but emerging out of it is an amazingly flexible mechanism for much of the cortex in concert to shape itself according to some current purpose, whether it be a biological threat, such as a wasp sting, or a more complex task we’ve consciously set, such as composing a piece of writing. Large swaths of the cortex can recreate its own mountainous landscape of combinations of activated or inhibited neuronal coalitions in a highly pointed way, reflecting all sensory, memory, cognitive, and motor features of the goal of the moment.
ATTENTIONAL VICTORIES EMERGING INTO CONSCIOUSNESS
At what point does this attentional battle turn into consciousness? In line with Libet’s free-will experiments and Nikolov’s neuronal decision modeling data, at the first ramping up of decision activity, and the onset of the attentional war, there is no sign of awareness. Instead, neuronal activity probably becomes conscious when the battle is clearly won, when all features of a goal are significantly bound together in frenetic activity throughout the cortex, and all nonrelevant details are simultaneously shut down.
Consequently, when we spot that wasp, we don’t see lines of yellow independent from lines of black, both distinct from a couple of wings and all quite separate from a vague buzzing sound. Instead, the collaborative endpoint of the processing bursts through into awareness. We immediately know we are near a wasp, can hear and see it very much as a single yet compound object, orient our heads toward it, and already are thinking where to move away from it or what means we can use to push it out the window—and little else aside from that wasp at that moment occupies our consciousness.
The process of combining more primitive pieces of information to create something more meaningful is a crucial aspect both of learning and of consciousness and is one of the defining features of human experience. Once we have reached adulthood, we have decades of intensive learning behind us, where the discovery of thousands of useful combinations of features, as well as combinations of combinations and so on, has collectively generated an amazingly rich, hierarchical model of the world. Inside us is also written a multitude of mini strategies about how to direct our attention in order to maximize further learning. We can allow our attention to roam anywhere around us and glean interesting new clues about any facet of our local environment, to compare with and potentially add to our extensive internal model.
The example of our attentional system being driven by some biologically important external object may be the normal form of attention in the animal kingdom, where most species have a considerably simpler mental life than we do. But if you have far greater processing capacity and a more elaborate internal model of the world, then you also have far more choice about what to attend to, as so much more than the latest obvious threat or sign of food could potentially aid your biological goals, if analyzed carefully.
Without any obvious external threat forcing itself on our attentional system, how is the choice made to attend to any of the seemingly infinite options? On the one hand, consciousness and complex thought can constrain the process in a number of ways. We can logically interrogate the reasoning behind each major option and heavily favor one side in the neuronal battle to follow. We are certainly not limited to the small set of more instinctive attentional filtering systems. Instead, we can consciously create almost any kind of neuronal filter, strongly boosting attention for one feature of the inner or outer world and suppressing others. In this way, we can seemingly choose what to attend to—in other words, what to be aware of.
On the other hand, despite our impressive conscious ability to bias what we attend to, the basic competitive neuronal mechanism of attention is just the same in these seemingly voluntary, internal choices of attention as it is for the immediate external drivers of our attentional system. Out of the many fighting voices in our minds, conscious control is but one choice, commonly pitted against a set of various bullying unconscious desires, where only a single voice can win out, to ruthlessly recruit to its cause all conceptually related neurons throughout the cortex and suppress any dissenters.
Here, with these neuronal wars, we can return to the question of emergentism, where advanced ideas on one level emerge out of the interactions of simpler, lower-level objects, with a clear proposal for the mechanism of emergentism in relation to awareness: Multiple factions of neurons competitively interact, with two kinds of feedback—a positive form that can rapidly boost neuronal activity, and a negative form that can rapidly inhibit it. The complex interplay between these two opposing feedback loops at the level of local neurons can dynamically tune and activate much of the cortex, giving rise to highly flexible, global, synergistic information processing and consciousness.18
OVERESTIMATING THE VALUE OF EMOTIONS
So far in this chapter, I’ve been describing how attention is a key component of consciousness. Attention is a filtering and boosting mechanism, taking the entirety of the sensory input we receive, including much that is irrelevant to us, and converting this into a far more finite, refined output containing only those items that are most germane to our current goals. It is this output that we are conscious of and that I will explore in more detail for the remainder of this chapter. But first, I turn to two particular forms of conscious content: emotions and self-consciousness.
It would be churlish to underestimate the role that emotions play in our conscious lives. Emotions continuously, profoundly shape our thoughts and behavior, and many would say it is the panoply of so many vivid sentiments that give true color and meaning to life. I don’t deny any of this. But some theorists have put emotions as both the core evolutionary driving force of consciousness and its main contents. There are good reasons to reject such a position.
For instance, there are many moments in our lives when we’re not conscious of any particular emotion, even though we’re nevertheless very much aware of something.
But what would happen if we lived our entire lives in such a grey state? Would we still be conscious? In the book Descartes’ Error, Antonio Damasio discussed a patient, known as Elliot, who in many ways is a modern version of Phineas Gage. Elliot, also following damage to his orbitofrontal cortex, also underwent a radically shifted mental world. His life is now almost devoid of any emotions whatsoever
. He goes through life continuously in a neutral emotional gear. But his level of consciousness seems hardly dented, even if his social decisions are abnormal, just as Gage’s were. In total contrast, a severe attentional deficit following brain damage, as I’ll describe in the next chapter, crushes awareness.
At the same time, intense emotions can seem to shrink consciousness in unhealthy ways. For instance, some people become sufficiently nervous when speaking in public that they stumble over words, or momentarily forget what they meant to say. Their minds go blank, and it almost feels as if their awareness of everything except the object of their nerves has disappeared. Ramp this up many notches, and terror, an emotion designed to save our lives, regularly kills because it profoundly diminishes consciousness and therefore intelligent control. For instance, it is tragically common in plane crashes for passengers to be trapped in their seats because they repeatedly press the safety belt as if it had a car’s release button instead of the plane’s metal lever. The most minuscule conscious analysis of the situation is unavailable to them because they are paralyzed by fear.
