Why Are You Always Thinking About Yourself?

Elegant neuroscience research sheds light on an age-old question.

Posted Jan 19, 2018

“It doesn’t mean we can compute everything. The rules of the game by which we make the computation, the laws underneath everything that makes nature work, are simple. That doesn’t mean we can figure everything out.”  

— Richard Feynman, American Quantum Physicist

Do human beings have a built-in tendency to be "self-ish"? Not selfish in the usual sense, but rather does baseline brain activity default to self-related processing — egocentrism — when we are in repose? If so, what is the brain mechanism for this, and what are the implications?

It stands to reason that brains should be selfish. From a basic evolutionary sense, the individual's first order of business under ordinary circumstances is to preserve the individual. Of course, part of this self-preservation involves attending to others, a need which typically subserves individual survival. Imagine a child so preoccupied with her parents that she sets aside her own basic developmental needs. That's an evolutionary dead end. 

Freud discussed “primary narcissism,” an innate tendency to be self-centered, which is present from the earliest stages of life: "Loving oneself is the libidinal complement to the egoism of the instinct of self-preservation” (Freud, 1914). Later in life, the primary focus on oneself remains crucial to basic organismic survival, and social relations become more nuanced and interactive with self-referential processing. Freud goes on to suggest that in illness one's focus on others is turned back upon oneself in the form of “secondary narcissism,” resulting in megalomania and the inability to properly empathize and form relationships with other people.

When we are caregivers, typically as adults with other adults and children, but sometimes precociously, our personal altruistic needs lead us to consider the other's needs, but we can argue that even in making the ultimate sacrifice for others, we are still thinking first of ourselves, even if implicit or unconscious. In the end, whether deliberative or spontaneous, such decisions are generally made by oneself, even if in the context of powerful social and cultural factors.

Researchers are very interested in this question and are studying the default mode network (DMN), a brain network involved in generating experience when we are not doing anything else — sometimes referred to as the brain's "idle state," like an automobile left running while parked. (For a brief overview of three important brain networks and TMS therapy, including default mode network, salience network, and executive control network, go here.)

In their compelling work on this question, neuroscientists Meyer and Lieberman (2018) investigate the causal relationship between brain network anatomy and function, and whether there is a basic tendency for human beings to gravitate toward thinking about ourselves. Special thanks to the study authors for assistance in preparing this article.


In reviewing the existing research underpinning their work, Meyer and Lieberman note that self-related processing has been found to be associated with midline parts of the brain. If we pull apart the two hemispheres of the brain (like separating halves of an orange), the areas which make up the DMN are clustered in that region. It makes sense, because that area of the brain, so critical for basic function, is somewhat protected, nestled within the surrounding tissues. In the prior research reviewed, the authors note that participants answered questions requiring either self-referential or social information processing. For example, they would answer questions about another person’s motives for doing something, like donating money, or be asked to engage in tasks requiring self-reflection. During studies like this, participants would go back and forth between resting phases and active phases in order to see whether different brain areas at rest facilitated, or “primed,” different types of information processing.

The medial prefrontal cortex (MPFC), along with its sub-parts, has been shown to be one of the most important regions for self-referential processing, a key node in the DMN. In terms of naming, parts of the cortex are often designated “Brodmann’s Area,” shortened to “BA” with a corressponding number, from the work of the great German anatomist Korbinian Brodmann. Numerous studies have suggested, but not definitively demonstrated, that MPFC/BA10 is uniquely implicated in self-referential processing, active when we think about our own personalities, imagine our future selves, or consider that other people are paying close attention to us. Furthermore, MPFC/BA10 has increased metabolic activity when the brain is in idle or free from outside distraction. This area ramps up when the DMN kicks in during breaks in attention and with prolonged rest.

A nearby area, the dorsal medial prefrontal cortex/BA9 (DMPFC/BA9), becomes more active when thinking about others. DMPFC/BA9 appears to process social information, and is thought to provide the information to MPFC/BA10 required for thinking, and making decisions about, oneself. Increased resting activity in each respective area has been shown to prime task performance on their respective mental activities, either self-referential or social information processing. However, these studies have not been definitive, because it is possible that these brain areas are actually priming general information processing, or semantic processing, rather than the specific functions related to self or others, because of how those earlier studies were designed.

The current study

Meyer and Lieberman sought to determine whether activity in MPFC/BA10 is specific to self-referential processing by measuring participants' reactions under four different conditions — when considering their own qualities, when considering someone else’s qualities (e.g., a familiar figure like Barack Obama), when processing general semantic information (e.g., a well-known place like the Grand Canyon), and when not trying to perform a particular task (the brain at rest, activating the DMN). The Grand Canyon condition requires participants to make a judgment about the Grand Canyon (the authors use the example of whether or not it is dry), which requires semantic processing, independent of self or other processing.

They performed fMRI scans to check the brain activity in 18 participants (approximately half were women) and measured how quickly they were able to make judgments for each of the four conditions. Each participant completed 40 trials presented at random (see figure), alternating among rest, self-referential processing, social information processing, and general information processing. Brain activity levels were analyzed against reaction times in each series of trials to see whether activity in the MPFC/BA10 uniquely and specifically directs self-referential processing. If increased activity in the MPFC/BA10 at rest (when the DMN is active) predicted better performance as reflected by faster times on self-related judgments — but not on social or general semantic judgments — this would provide strong evidence that this brain area at rest is responsible for our tendency to think about ourselves at rest, all other factors being equal.

Meyer & Lieberman, 2018
Source: Meyer & Lieberman, 2018


First, they found that MPFC/BA10 was indeed more active in advance of self-referential tasks, and the higher the activity, the better the performance. This area did not prime performance on social or general semantic tasks, confirming the particular role of this area in generating a default egocentric bias to our mental activity. Furthermore, higher MPFC/BA10 activity actually slowed down performance on the Grand Canyon task, providing additional evidence that this area is not only not involved with general information procession, but that self-referential processing can get in the way of thinking about other things.

Meyer & Lieberman, 2018
Source: Meyer & Lieberman, 2018

In contrast, DMPFC/BA9 activity predicted faster performance for self-referential and social cognition, supporting the hypothesis that this area may be directly involved in thinking about others, as well as providing the necessary social information required for making judgments about oneself. Meyer and Lieberman (2018, references redacted for clarity) highlight the complex relationship between self-referential and social cognition over time as reflected in the interaction between these two key areas of the PFC, noting:

Why would DMPFC/BA9 prime multiple forms of social cognition while MPFC/BA10 preferentially prime self-referential processing? On the one hand, some of the same mechanisms we use to understand others may be helpful for understanding ourselves, and thus, DMPFC/BA9 may commonly support self and other processing. However, over the course of development, we also gain crystallized knowledge about our own identities and ways of experiencing the world. Crystallized self-knowledge comes from a combination of social sources (e.g., “he thinks I am nervous”) and visceral sources (e.g., “I feel like I am nervous”). As a hub of the default network, MPFC/BA10 has strong functional connections both to the DMPFC [dorsomedial prefrontal cortex] subsystem that supports generic social reflection well as sensory and limbic regions, which convey information about internal visceral states. Thus, through repeated coupling with other sources of information, self-reflection may become routinized in MPFC/BA10 and function at rest to preferentially facilitate self-referential processing.

Further considerations

This research provides the first direct evidence that we are egocentric because of higher baseline activity in a specific brain area (the MPFC/BA10, central to the DMN). When the brain is at rest, the DMN is more active, and this biases us to shift to thinking about ourselves over others. This built-in tendency must vary from individual to individual as a function of many different potential factors, arguably the result of evolution. As a very basic illustration, consider that babies who are more concerned about their mother's discomfort during feeding than their own needs would do poorly compared to babies who take in sustenance without hesitation. While it makes sense that the fundamental egocentric tendency is necessary for survival and would, therefore, be selected by evolutionary processes, survival also requires thinking about other people and the environment in general, particularly as we get older, and the role of reciprocity becomes more important to survival in an increasingly social world. Self-referential processing, in this view, is the cornerstone of the individual’s resting mind, but is shaped by experience over the course of development.

It is important to repeat studies like this with different groups to determine the effects of environmental and developmental influences, such as culture (more communal versus more individualistic cultures), life experience (people who grew up in more optimal environments, where they could focus on sophisticated developmental needs, versus those who had to attend to others in order to ensure basic survival), profession (professions which require a higher level of social cognition, such as therapists, versus professions which do not, such as computer programmers), personality (more narcissistic versus less narcissistic, as a function of the “Big Five” personality traits), and different pathological states (PTSD, depression, anxiety, etc.). For example, default mode activity has been shown to be disrupted in patients with PTSD, as increased attention to external threats can interfere with self-referential processing and working memory function (e.g. Lanius et al., 2015).

Along clinical lines, measuring activity in brain areas such as these may be useful for tracking and guiding therapeutic progress. For example, does effective therapeutic work shift DMN activity in measurable ways, and are there consistent neural signatures marking clinical efficacy for people seeking help with different issues? What is the impact of different therapeutic approaches on brain function, and how does this translate into behavior and well-being? How do psychiatric medications affect brain activity when it comes to self and other processing? Can neuromodulation approaches, like TMS, tDCS, neurofeedback and others, usefully target these and related brain regions? What impact do interventions like mindfulness and compassion-based practice have on self-referential processing for people who have either excessive and ruminative thoughts about themselves, or inadequate or erroneous social information processing (both common issues for people with various psychiatric conditions, including depression, social anxiety, and developmental trauma)?

In the meantime, consider that when you settle down and are not engaged in any particular activity, and your default mode network kicks in, the "reason" you tend to think about yourself is because of activity in a small part of your brain. Moreover, and perhaps amusingly, when you contemplate this very thought, your thinking is likely being primed by the very brain area you are imagining. Of course, rather than being "neuroreductionistic," it's important to understand specific findings like these in context. This research opens up avenues for a deeper understanding of who we are, as well as potential translation to clinical application.


Freud, S. (1914). On Narcissism. The Standard Edition of the Complete Psychological Works of Sigmund Freud, Volume XIV (1914-1916): On the History of the Psycho-Analytic Movement, Papers on Metapsychology and Other Works, 67-102.

Lanius RA, Frewen PA, Tursich M, Jetly R, & McKinnon MC. (2015). Restoring large-scale brain networks in PTSD and related disorders: a proposal for neuroscientifically-informed treatment interventions. European Journal of Psychotraumatology, Published online 2015 Mar 31. doi:  10.3402/ejpt.v6.27313

Meyer ML & Lieberman MD. (2018). Why People Are Always Thinking about Themselves: Medial Prefrontal Cortex Activity during Rest Primes Self-referential Processing. Journal of Cognitive Neuroscience, 2018 Jan 8:1-9. doi: 10.1162/jocn_a_01232. [Epub ahead of print]