The Neuroscience of the Self

What makes up our sense of self has been an important focus of discussion for centuries. It has vexed philosophers and thinkers who have debated what the self is—and is not—with passion. A few years ago, neuroscientists also entered the fray. Some, such as the neuroscientist and philosopher Georg Nothoff, even attempted to locate the self within the brain. Early efforts were motivated by an intuitive question: if we can localize vision, language, and motor control, why not the self?

With the rise of functional brain imaging in the 1990s, neuroscientists began designing experiments that contrasted self-referential mental states with non-self states. In the scanner, people might be asked to judge whether adjectives applied well to themselves or others—questions such as, “Am I honest?” These studies repeatedly identified activity along the cortical midline from front to back. Because these regions were more active when people thought about themselves than when they performed externally focused tasks, some researchers proposed that they formed a neural core of the self.

This idea gained further traction with the discovery of the so-called “default mode network” of brain regions—a set of interconnected brain areas that become active when the mind is at rest, engaged in introspection, retrieving memories from one’s past, or imagining one’s future. Because these mental activities feel intimately tied to selfhood, this collection of brain areas was sometimes described loosely as the brain’s “self network.” However, problems with this interpretation soon became apparent.

First, the same regions are active during many cognitive tasks that are not necessarily about the self. Second, different kinds of “self” tasks activated overlapping but not identical patterns of brain regions within the network. These findings led many researchers to conclude that what was being localized was not the self, but processes related to self-reference: self-evaluation, autobiographical recall, perspective-taking, and narrative construction. The consensus is that the self has not been localized by these sorts of brain scanning studies.

This comes as no surprise to many people. Daniel Dennett, the philosopher, famously argued that searching for the self in the brain was a category mistake, akin to searching for the center of gravity of an object. The self has no physical properties. It is not a real thing but a theorist’s fiction. “No one has ever seen or ever will see a center of gravity. As David Hume noted, no one has ever seen a self, either,” Dennett proclaimed. The self, according to Dennett, is a useful abstraction: a narrative construct that we create to explain who we are, not a biological object waiting to be found. It emerges from the stories we tell— to ourselves and to others—about our actions, beliefs, intentions, and experiences. It is illusory.

Dennett was comfortable to consider the self to be useful fiction. “We are all virtuoso novelists,” he quipped. “We try to make all of our material cohere into a single good story. And that story is our autobiography. The chief fictional character at the center of that autobiography is one’s self.” His theory rejected the idea of a single, unified “theatre” in the brain where experiences are somehow presented to an inner observer – the self. Instead, Dennett proposed that what we call conscious experience of the self is actually the outcome of many parallel, distributed, and competing neural processes—multiple drafts”—none of which is intrinsically privileged as the final version.

Dennett’s views align closely with those of the MIT AI pioneer Marvin Minsky, who argued that the self is simply the emergent result of many interacting brain processes. For Minsky, this “Society of Mind” is a collection of semi-independent agents—simple mechanisms that each do one small thing well, such as an aspect of perception, memory, or language. These agents or modules are not intelligent in themselves. They don’t understand, decide, or reflect. Nor is there any form of central controller. Control is distributed. Each module performs a limited function. Intelligent behavior arises from their coordination. Minsky was deliberately opposing the intuitive idea that somewhere in the brain there must be a unitary thinker—a captain of the ship. There is no such master agent, according to him.

These ideas have had great influence but also attracted fire. One important criticism of the view of the self as a fictional narrative and an emergent property of multiple, independent brain processes is that this does not explain why the self feels singular, not fragmented. As the NYU philosopher David Chalmers has argued, these models don’t explain the experience of being a unified self – the first-person subjective experience of being who we are. They deal with the “easy problem” of how cognitive functions operate in the brain, but not the “hard problem” of conscious experience. Why and how is any of this brain processing accompanied by subjective experience, such as what it feels like to see red, feel pain, or even be “oneself”? No amount of reductionist explanation, Chalmers argues, logically entails the existence of experience. You could, in principle, explain everything the brain does and still be left with this question about the self unanswered.

One important perspective on these issues comes from another side of neuroscience: the study of people who develop brain disorders. What happens if one part of the “Society of Mind” that normally underpins the self becomes dysfunctional? What is the first-person experience then? Researchers have been observing what happens to people who have suffered a stroke or developed a neurodegenerative condition such as Alzheimer’s disease. Some of these individuals can have highly focal impairment, limited to one cognitive process or module. They may, for example, have difficulty with visual perception, paying attention, retrieving information from memory, comprehending language, understanding concepts, and being motivated, and so on. 

Michael, for example, is a patient who came to see me in my clinic because he was worried about his increasing difficulty in finding the right word in conversation. To begin with, he seemed to be highly articulate, and it was difficult to appreciate what he was concerned about. As we talked more though, it began to be evident that he didn’t understand the meaning of some words that he really should know. When we got onto the topic of sports, for example, he told me he used to play rugby, but he couldn’t understand what I meant when I asked him which position he used to play in the team.

Through careful assessment, it became clear that Michael was suffering from a highly limited impairment in semantics—understanding the meaning of words and the concepts that are attached to them. His diagnosis was semantic dementia, an unusual neurodegenerative condition. Slowly over time, Michael’s difficulty with semantics began to have a greater impact. He didn’t understand jokes and became less inclined to talk. His friends thought that he’d lost his sense of humor and came to visit him less. He became more isolated from them and subsequently members of his family, even when they were aware of his diagnosis.

What do stories such as Michael’s reveal about our selves? At one level, they demonstrate that the self can be altered. People behave differently if their semantics and understanding of concepts are eroded. Similarly, for individuals whose perception might be impaired, start to experience visual hallucinations, have difficulty recalling information, or become disinhibited in their behavior. All these types of dysfunction in a limited cognitive module can lead to profound changes in a person’s identity. They can also alter how they appear to others and have a huge impact on their social identity and how they fit in their social circle. Crucially, though, none of these people lose their entire sense of self through the loss of any one cognitive module. They still have a first-person perspective on the fact that they are a person with a sense of self.

In this way, these neurological observations support the view that the self is not housed in one particular brain region, but rather emerges from the activities of a distributed constellation of cognitive modules. Focal brain disruption to one of these does not eliminate the self. It selectively alters it. Very basic brain functions clearly play a key role in determining who we are—our personal identity. They are integral parts of the “Society of the Mind” that creates our self. But they are also crucial to keeping us within society through our social identity. First-person experiences can continue after brain damage, although those experiences are different. However, although we can explain what is lost or disrupted (perception, memory, language, concepts, and so on), we don’t have a definitive answer to what makes up the self. 

And we don’t know whether philosophy, neuroscience, or both will be the discipline that helps us answer this question once and for all.